Pollution Control in Fertilizer Production

title: Environmental Science and Pollution

Control ; 10
author: Hodge, C. A.
publisher: CRC Press
isbn10 | asin: 0824791886
print isbn13: 9780824791889
ebook isbn13: 9780585362298
language: English
Fertilizer industry--Waste disposal,
subject
Fertilizer industry--Environmental aspects.
publication date: 1994
lcc: TD799.F47P65 1994eb
ddc: 668/.62
Fertilizer industry--Waste disposal,
subject:
Fertilizer industry--Environmental aspects.
Pollution Control in Fertilizer Production
 Environmental Science and Pollution Control Series
Series Advisors
Dr. Suzanne Lesage
Environment Canada
Burlington, Ontario, Canada
Dr. Willy J. Masschelein
Consultant
Brussels, Belgium
Dr. Hans van Bergen
International Flavors and Fragrances
Hilversum, The Netherlands
Dr. Constantine Yapijakis
The Cooper Union
New York, New York, U.S.A.

1. Toxic Metal Chemistry in Marine Environments, Muhammad Sadiq

2. Handbook of Polymer Degradation, edited by S. Halim Hamid,
Mohamed B.
Amin, and Ali G. Maadhah
3. Unit Processes in Drinking Water Treatment, Willy J. Masschelein
4. Groundwater Contamination and Analysis at Hazardous Waste
Sites, edited
by Suzanne Lesage and Richard E. Jackson
5. Plastics Waste Management: Disposal, Recycling, and Reuse,
edited by Nabil
Mustafa
6. Hazardous Waste Site Soil Remediation: Theory and Application of
 Inno-
vative Technologies, edited by David J. Wilson and Ann N. Clarke
7. Process Engineering for Pollution Control and Waste Minimization,
edited by
Donald L. Wise and Debra J. Trantolo
8. Remediation of Hazardous Waste Contaminated Soils, edited by
Donald L.
Wise and Debra J. Trantolo
9. Water Contamination and Health: Integration of Exposure
Assessment,
Toxicology, and Risk Assessment, edited by Rhoda G. M. Wang
10. Pollution Control in Fertilizer Production, edited by Charles A.
Hodge and
Neculai N. Popovici

Additional Volumes in Preparation

Groundwater Contamination and Control, edited by Uri Zoller
Handbook of Toxic Properties of Pesticides, Nicholas P.
Cheremisinoff and
John A. King
Combustion and Incineration Processes: Applications in
Environmental
Engineering. Second Edition, Revised and Expanded, Walter R.
Niessen
Hazardous Chemicals in the Polymer Industry, Nicholas P.
Cheremisinoff
 Page i

Pollution Control in Fertilizer Production

edited by
Charles A. Hodge
Tennessee Valley Authority
Muscle Shoals, Alabama

Neculai N. Popovici
Chemenco Ltd.
Bucharest, Romania
Library of Congress Cataloging-in-Publication Data
Pollution control in fertilizer production / edited by Charles A. Hodge, Neculai N.
Popovici.
p. cm. -- (Environmental science and pollution control series; 10)
Includes bibliographical references and index.
ISBN 0-8247-9188-6 (alk. paper)
1. Fertilizer industryWaste disposal. 2. Fertilizer industryEnvironmental
aspects. I. Hodge, C. A. II. Popovici, Neculai N. III. Series: Environmental
science and pollution control; 10.
TD799.F47P65 1994
668'.62dc20 94-4798
CIP
The publisher offers discounts on this book when ordered in bulk quantities. For
more information, write to Special Sales/Professional Marketing at the address
below.
This book is printed on acid-free paper.
Copyright © 1994 by Marcel Dekker, Inc. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in any form or
by any means, electronic or mechanical, including photocopying, microfilming,
and recording, or by any information storage and retrieval system, without
permission in writing from the publisher.
Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 10016
Current printing (last digit):
10 9 8 7 6 5 4 3 2 1
PRINTED IN THE UNITED STATES OF AMERICA
 Page iii

Foreword
I am honored to introduce Pollution Control in Fertilizer Production
edited by Drs. Charles A. Hodge and Neculai N. Popovici. The editors
are to be commended for undertaking such an awesome but important
and necessary work. Never has such a book been written from an
environmental perspective. For the first time, nearly every aspect of
pollution control and abatement for the mining, manufacturing, and
marketing of chemical fertilizers is addressed under one cover.
The significance of this book is even greater when one considers that
industrial pollution as an environmental concern is an issue in
virtually every corner of the world. This book should be a comfort to
the environmentalist while serving as a wealth of knowledge to the
world fertilizer industry. The editors have tapped some of the best
scientific and engineering minds working on pollution prevention and
control in global fertilizer production.
The reader will be surprised, I think, to learn of the many existing and
emerging technologies used to prevent and control effluents in the
fertilizer production process. But the book goes beyond the
technological advancements in pollution control in fertilizer
production, discussing remediation technology and its application and
potentialan excellent addition to an already quality publication.
I commend this book to anyone associated with the mining,
production, and marketing of fertilizerfrom the CEO to the plant
operator to the environmentalist. All will greatly benefit from its
insights.
JOHN T. SHIELDS
SENIOR VICE PRESIDENT
NATIONAL FERTILIZER AND
ENVIRONMENTAL RESEARCH CENTER
TENNESSEE VALLEY AUTHORITY
MUSCLE SHOALS, ALABAMA
 Page v

Preface
Pollution Control in Fertilizer Production is a harmonious and
comprehensive discussion of the aspects related to fertilizer
production pollution control. The fertilizer industry is approached on
the basis of major end-products and by-products from the point of
view of environmental pollutants to the atmosphere, water, and soil.
The book presents solutions capable of eliminating or diminishing
pollution by this industry and covers methods of converting effluents
and by-products to useful products where possible.
The degradation of some parts of the environment over the years has
caused increased public concern and a strong desire to overcome this
problem. The adverse aspects of fertilizers on the environment have
been given much media attention. Over the past decade, the number of
environmental protection laws and regulations affecting chemical
production and the fertilizer industry have steadily increased on the
local and national levels, and most recently, on the regional and
international levels. This comprehensive book on methods of reducing
environmental pollution and the positive actions taken by the fertilizer
industry fills a gap in the literature. To date, no other book has
covered the pollution control technology required for the production,
handling, and distribution of fertilizers for environmental legislation
compliance.
Pollution Control in Fertilizer Production evaluates the
environmental polluting potential of fertilizer production processes. It
presents solutions capable of eliminating or diminishing pollution and
covers methods of converting effluents and by-products to useful
products. The latest environmental protection technology is presented
by 33 recognized experts in a manner that can be used in the planning,
renovation, or operation of facilities, and in the remediation of
discontinued facilities.
The book addresses specialists and designers, constructors, production
supervisors, and administrators involved with fertilizer technology
and environmental regulation
 Page vi
compliance. The chapters covering the transport, storage, containment
facilities, and use of fertilizers will provide useful technical
information for fertilizer vendors and distributors. Both academic and
industrial personnel involved with fertilizer technology and
environmental protection will find the book a useful reference source.
CHARLES A. HODGE
NECULAI N. POPOVICI
 Page vii

Contents
Foreword iii
Preface v
Contributors xi
I. Introduction
1. Growth of the World Fertilizer Industry 1
Charles A. Hodge
2. Pollutants, Wastes, and By-Products of the Fertilizer 9
Industry
Neculai N. Popovici
3. Regulation and Pollution Control in the Fertilizer 23
Industry
Charles A. Hodge
II. Nitrogen Fertilizer Industry
4. Ammonia Production 35
Mircea Turtureanu
5. Ammonia Pollution Control Measures 51
Thomas A. Czuppon, Hassan M. Gomaa, and Stan A. Knez
6. Ammonia Storage, Loading, and Transport 73
Robert H. Roberts and John G. Baguley
7. Nitric Acid Production 81
Eugen-Petre Cristescu
 Page viii

8. Impact of NOx Abatement on the Fertilizer Industry 93

Jumpei Ando
9. Ammonium Nitrate and Nitrochalk 99
Robert E. Nitzschmann and Johannes G. A. Reuvers
10. Urea Production 109
Hidetoshi Uchino
11. Control of Water Pollution and Water Treatment in the 119
Nitrogen Fertilizer Industry
Lucian Negulescu
III. Phosphate Fertilizer Industry
12. Introduction to the Phosphate Fertilizer Industry 145
Manny Sierra
13. Phosphate Rock Storage, Transportation, and 151
Processing
Manny Sierra
14. Sulfuric Acid Production 161
Teodor Ciobanu
15. Sulfuric Acid Storage and Transport 187
Robin W. Strickland
16. Impact of SO2 Abatement On the Fertilizer Industry 189
Jumpei Ando
17. Phosphoric Acid by Wet Process: Recovery of 197
Fluorine-Containing Gases
W. R. Parish
18. Phosphoric Acid by Wet Process: Phosphogypsum, 209
Transport, Storage, and Utilization
Jumpei Ando and John E. Cameron
19. Phosphoric Acid by Wet Process: Pond Water 225
Management
John E. Cameron
20. Phosphoric Acid by Wet Process: Radioactive 237
Components in Phosphoric Acid and Their Removal
Florin T. Bunus
21. Phosphoric Acid by Wet Process: Heavy Metals in 253
Phosphoric Acid and Their Removal
Marinela Man
22. Granular Triple Superphosphate and Single 269
Superphosphate
Henryk Jan Górecki
23. Utilization of Fluorine from Phosphate Fertilizer 299
Plants
Henryk Jan Górecki
 Page ix

IV. Potassium Fertilizer Industry

24. Introduction to the Potassium Fertilizer Industry 337
Maurice O. Klein, Kenneth W. Reid, and Elmar L.
Goldsmith
25. Environmental Aspects in Mining and Production of 343
KCl
Kenneth W. Reid and Maurice O. Klein
26. Long-Term Solid Waste Management and KCl 361
Facilities Decommissioning
Kenneth W. Reid and Maurice O. Klein
27. Air Emissions at KCl Facilities 369
Maurice O. Klein, Kenneth W. Reid, and Elmar L.
Goldsmith
28. KCl Handling, Storage, and Transport 387
Elmar L. Goldsmith
V. Mixed Fertilizer Industry
29. MAP, DAP, and NPK Fertilizers: Pollution Control 393
Practices in Japan
Jumpei Ando and Ichiki Hamada
30. MAP, DAP, and NPK Fertilizers: Pollution Control 405
Practices in the United States and Europe
Frank P. Achorn
31. Nitrophosphate-Based NPK Fertilizer: Air Pollution 419
Control
Robert E. Nitzschmann and Johannes G. A. Reuvers
32. Nitrophosphate-Based NPK Fertilizer: Wastewater 433
Treatment
Paula-Gabriela Petreanu
33. Nitrophosphate-Based NPK Fertilizer: Radioactive 449
Components and Their Removal
Florin T. Bunus
34. Storage and Transport of Solid Fertilizers 461
Robert E. Robinson
VI. Liquid Fertilizer Industry
35. Ammonium Phosphate Solutions and Suspensions 467
Mark T. Holt
36. Hot and Cold Blend Liquid Fertilizer Plants 473
Derek A. Palgrave
37. Containment Systems for Liquid Fertilizer Plants 477
Horace C. Mann, Jr.
VII. Plant Site Remediation
38. Remediation Measures for Existing Plant Facilities 485
and Decommissioned Facilities
Susanna Tomlinson
39. Remediation Measures for Distribution and Dealer 491
Facilities
Horace C. Mann, Jr.
Index 493
 Page xi

Contributors
Frank P. Achorn SE-ME, Inc. Chemical Process Consultants,
Florence, Alabama
Jumpei Ando Faculty of Science and Engineering, Chuo University,
Tokyo, Japan
John G. Baguley Process Engineering Department, The M. W.
Kellogg Company, Houston, Texas
Florin T. Bunus Institute of Atomic Physics, Bucharest, Romania
John E. Cameron New Wales Plant, IMCAgrico Company, Mulberry,
Florida
Teodor Ciobanu* Inorganic and Fertilizer Department, Ipran-
Iprochim, Bucharest, Romania
Eugen-Petre Cristescu Eurochim S.R.L., Bucharest, Romania
Thomas A. Czuppon Product Technology Development Department,
The M. W. Kellogg Company, Houston, Texas
Elmar L. Goldsmith Research Department, Kalium Canada, Ltd.,
Regina, Saskatchewan, Canada
Hassan M. Gomaa Technology Management, The M. W. Kellogg
Company, Houston, Texas
*Retired
 Page xii
Henryk Jan Górecki Institute of Inorganic Technology and Mineral
Fertilizers, Technical University, Wroclaw

*, Poland

Ichiki Hamada Coop Chemical Company, Tokyo, Japan

Charles A. Hodge National Fertilizer and Environmental Research
Center, Tennessee Valley Authority, Muscle Shoals, Alabama
Mark T. Holt Chemical Development Department, Tennessee Valley
Authority, Muscle Shoals, Alabama
Maurice O. Klein Outreach Department, International Fertilizer
Development Center, Muscle Shoals, Alabama
Stan A. Knez Process Design Department, The M. W. Kellogg
Company, Houston, Texas
Marinela Man* Institute of Chemical Research, Bucharest, Romania
Horace C. Mann, Jr.* Field Programs Department, Tennessee Valley
Authority, Muscle Shoals, Alabama
Lucian Negulescu Faculty of Energetics, Polytechnical University,
Bucharest, Romania
Robert E. Nitzschmann Fertilizer Production Division, BASF
Aktiengesellschaft, Ludwigshafen, Germany
Derek A. Palgrave Technical Consultant, Bury St. Edmunds, Suffolk,
England
W. R. Parish Westinghouse Savannah River Company, Aiken, South
Carolina
Paula-Gabriela Petreanu Department of Phosphorus and Complex
Fertilizer, Ipran-Iprochim, Bucharest, Romania
Neculai N. Popovici Department of Engineering and Consulting,
Chemenco Ltd., Bucharest, Romania
Kenneth W. Reid Environmental Affairs, Potash Corporation of
Saskatchewan Inc., Saskatoon, Saskatchewan, Canada
Johannes G. A. Reuvers Fertilizer Production Division, BASF
Aktiengesellschaft, Ludwigshafen, Germany
Robert H. Roberts Office of Chief Engineer, The M. W. Kellogg
Company, Houston, Texas
Robert E. Robinsont** Robert E. Robinson and Associates, Beaver,
Pennsylvania
*Retired
**Formerly of Renneburg Division, Heyl & Patterson, Inc., Pittsburgh,
Pennsylvania
 Page xiii
Manny Sierra* Consultant, Clearwater, Florida
Robin W. Strickland Crescent Technology Inc., New Orleans,
Louisiana
Susanna Tomlinson* National Fertilizer and Environmental Research
Center, Tennessee Valley Authority, Muscle Shoals, Alabama
Mircea Turtureanu Research and Development Division, Department
of Chemical and Petrochemical Industry, Bucharest, Romania
Hidetoshi Uchino Licensing and Patents Department, Toyo
Engineering Corporation, Chiba, Japan
*Retired
 Page 1

1
Growth of the World Fertilizer Industry
Charles A. Hodge
Tennessee Valley Authority
Muscle Shoals, Alabama
The positive role of fertilizer use in agricultural production and
productivity is well known. The economic benefit of fertilizer use has
contributed to increased agricultural productivity and higher living
standards. With the increased influence of fertilizer use on total crop
production, the public, environmentalists, and legislators are more
keenly aware of fertilizer production, use, and its fate. In the past few
decades, fertilizer production facilities have increased in individual
capacity, and the fertilizer complexes have become more concentrated
in specific geographic locations. Thus there has been a greater need to
decrease the percentage or total loss of undesirable by-products and
effluents to the environment [1].
The growth of the world fertilizer industry will be affected
increasingly by environmental concerns and pollution control
compliance of production activities. Environmental concerns are
becoming more closely linked with economic decisions as to the
location of new facilities, production costs, and international trade
agreements. There exist great disparities in the environmental
regulations and their enforcement from one country to another.
Present efforts to harmonize environmental regulations throughout the
world will affect the foregoing issues and have an influence on the
growth rate and geographical location of future fertilizer production
facilities.
The consumption of fertilizer in the developed countries has increased
to rates that are approaching that which yields the highest economic
return. Fertilizer use and application rates in the developing countries
lag those of the developed countries but are increasing at rates higher
than in the developed countries. This trend is expected to continue for
several decades. Similarly, the trend of increasing pollution control
compliance and environmental regulation experienced by the fertilizer
industry in the developed countries is expected to continue. These
same issues will also need to be addressed and dealt with in the
developing countries.
 Page 2
Fertilizer use was one of the most important inputs for increasing
world grain yields from 1950 to 1980. The trend in world fertilizer use
per capita is shown in Fig. 1. The future growth rate of fertilizer
production is expected to parallel the demand for food and other
agricultural products, which will have to keep pace with the world
population growth rate. Progress has been made in decreasing the
world population growth rate. World population was growing at a
2.1% rate in 1970 and at a 1.7% rate in 1990. Projections of world
population are estimated by the United Nations [3,4] to increase from
4.8 billion in 1985 to 6.1 billion in the year 2000 and to about 8.4
billion in the year 2025 (Fig. 2). Europe's population growth rate is
the lowest in the world and is expected to continue at about the same
0.2% increase per year. However, the current population growth rates
are much higher: 3.0% in Africa, 1.9% in South America, and 1.8% in
Asia. These growth rates will cause Africa's population to more than
double to 1.6 billion and Asia's population to reach nearly 5 billion
people by the year 2025. The estimate of the population growth in the
developing countries for the period 19852000 is about 31%. The UN
estimates the population growth in the same time period in the
developed countries to be only 9% (Fig. 3) [3].
For many decades scientists have expressed concerns that the world's
food production capability would not keep up with population growth.
However, during the past 30 years, food production has nearly tripled,
with increased fertilizer input and improved plant varieties responsible
for most of this increase. With this increased use of chemical
fertilizers and the increased concerns of controlling chemical pollution
from all sources in our environment, more emphasis is being placed
on controlling the major sources of these chemicals going into the
environment [6]. Many of the major air pollution chemicals from
sources such as power plants, transportation vehicles, and industrial
emissions can also be emitted from fertilizer production processes.
Two of the most serious air pollutants are the oxides of sulfur and
nitrogen.
World fertilizer use of 26 kg per capita has been relatively constant
during the decade

Figure 1
World fertilizer use per capita. (From Ref. 2.)
 Page 3

Figure 2
World population growth. (From Ref. 5.)
of the 1980s (Fig. 1). Based on this conservative correlation to
population growth, the future total world fertilizer use is expected to
grow from 130 million metric tons in 1985 to 165 million metric tons
in the year 2000 and 227 million metric tons in the year 2020 (Fig. 4)
[7,8]. Many factors have been affecting the geographical location of
new fertilizer production facilities, continued operation of existing
facilities, and the decommissioning of fertilizer facilities. These
factors include the costs of production, raw materials, and energy,
market conditions, and other factors. One of these other factors has
increasingly become the environmental compliance factor. The
fertilizer industry has been challenged on both the national and
international levels to meet the current and increasingly more
restrictive rules and regulations affecting fertilizer production
pollution control and
 Figure 3
Estimated population growth rate, 19852000. (From Ref. 3.)
 Page 4

Figure 4
World fertilizer use. (From Refs. 7 and 8.)
environmental compliance. The foregoing factors have influenced the
growth rate of the fertilizer production industry such that in the past
decade these growth rates have been much higher in the developing
countries.
In Western Europe fertilizer production has decreased during the last
decade. Western Europe lost 2 million tons of nitrogen production
capacity from 1985 to 1990 and 0.7 million tons in the preceding 5
years. Estimates by the European Fertilizer Manufacturers Association
(EFMA) for the European Community (EC) member countries
indicate a decline in fertilizer consumption during the decade
19902000 (Table 1). North American fertilizer production has stayed
relatively constant during the last decade (Fig. 5). Consumption of
fertilizer has also changed little during the past 10 years. Exports of
phosphate rock, phosphate fertilizers, and potash fertilizers from
North America are a significant part of the total world trade in these
products [11]. The low-cost production of these products in North
America portends that this situation can be expected to continue
during the next decade. Nitrogen fertilizer production has been
affected very significantly by the cost of natural gas feedstock.
Nitrogen fertilizer production in Canada and Mexico have increased
based on their economical natural gas availability. In the United
States,
Table 1 Estimates of Future Mineral
Fertilizer Consumption for N and P2O5 for
EC/EFMA Member Countries
Year Million tons N Million tons P2O5
19901991 10.25 4.44
19951996 9.36 3.96
20002001 8.95 3.81
13% reduction 14% reduction
Source: Ref. 10.
 Page 5

Figure 5
North American fertilizer production. (From Ref. 9.)
the cost of natural gas has been more favorable than in Western
Europe but not as low as in other areas of the world with large natural
gas reserves and production. Industry analysts project that nitrogen
production in the United States will remain relatively constant in the
near future, with few, if any, new production facilities built. Imports
of nitrogen fertilizer are expected to increase their share of the U.S.
market long term.
The developing world will have the greatest increase in fertilizer
consumption in the coming decades. This increase is based on the
lower current use base compared with the developed world and based
on the greater need for agricultural production due to large population
increases [12]. The United Nations estimates that about 90% of the
population increase will occur in the developing countries. The
demand for agricultural products by this increasing population in the
developing world will thus affect fertilizer use rates to a much greater
extent in the developing countries. With the need for agricultural
production to increase the most in the developing countries, there is a
corresponding need for crop production inputs, including fertilizer.
Economic considerations will have the major influence on the location
of these new fertilizer production facilities. Location of new fertilizer
production facilities will preferentially take place in the developing
world, where the increased production capacity is needed the most.
Among the factors influencing the economics will be the location of
the product market, transportation costs, and projected long-term
market demand. The developing countries will face the greatest
challenges in producing fertilizers and importing fertilizers required
for their increased food production. The increased fertilizer
production will affect the local environment. This environmental
impact will depend on local environmental compliance laws and
regulation, technology used, operation practices, maintenance, and
local conditions. Other aspects related to production facilities that
affect the environment depend on the handling, treatment, use, and
disposal of by-products and wastes produced in obtaining raw
materials and in production of the final fertilizer products.
 Page 6
Nitrogen production is expected to increase most significantly in
regions of the world where there are both large reserves of low-cost
natural gas and proximity to the product market. Ammonia is the raw
material for most nitrogen fertilizers. The preferred feedstock for
ammonia is natural gas because it is usually the most economical, the
process is simpler, and there is less pollution than with the use of
other feedstocks, such as naphtha or coal. The trend in ammonia
production has been away from regions that import naphtha and
regions with high-cost or imported natural gas. Table 2 shows the
proven natural gas reserves for the major worldwide natural
gasproducing countries. Evaluation of the information presented in
this table reinforces the conclusion that the most economical increase
in nitrogen fertilizer production will occur in the less developed
countries, particularly in the countries of the Middle East. The
percentage of world nitrogen fertilizer exports from the Middle East
will continue to increase. Such countries as China and India will be
pressed to develop further their domestic nitrogen fertilizer production
industry. Indonesia and Malaysia are well positioned for above-
average growth in nitrogen fertilizer production.
The high price of natural gas in Western Europe compared to the cost
in the Middle East and the United States will continue to have an
adverse impact on the European nitrogen fertilizer production industry
[18]. In Central Europe, Eastern Europe, and the Commonwealth of
Independent States (CIS; the former Soviet Union), the cost of
fertilizer production and sales price were subsidized in the past.
Fertilizer, particularly nitrogen fertilizer, has been exported for foreign
exchange while food was imported. With a transition from a centrally
planned economy to a market-based economy, the fertilizer industry
of these countries will change significantly.
The phosphate industry is concentrated in a small number of countries
with large-scale intensive mining and chemical processing facilities.
Production and proven reserves of phosphate rock for the leading 10
producers are shown on Table 3. Approximately 75%
Table 2 International Natural Gas
Reserves
1990 Proven
natural gas
reserves
(trillion cubic
Country meters)
CIS 52.0
Iran 17.0
Canada 7.6
Argentina 7.2
United Arab 5.5
Emirates
Saudi Arabia 5.1
U.S. 4.9
Qatar 4.6
Iraq 3.1
Venezuela 3.0
Indonesia 2.4
Norway 2.3
Mexico 2.1
Malaysia 1.5
India 1.1
China 1.0
Source: Ref. 13.
 Page 7
Table 3 Major World Phosphate Reserves and
Production
Proven reserves 1991 Production
Country (million tons) (million tons)
U.S. 1200 47
CIS 1300 37
Morocco 5900 22
China 210 18
Tunisia 270 7
Jordan 90 6
Israel 10 3
South 2500 3
Africa
Others ~800 16
Source: Ref. 15.

of the current phosphate rock production comes from just four

sources: the United States, the CIS, Morocco, and China.
There are usually economic efficiencies associated with converting
phosphate rock to phosphate fertilizers near the mine site compared
with shipping the rock to distant locations for conversion. With
transportation a significant part of the price of fertilizer, the cost of
shipping the low-analysis rock compared with the product phosphate
fertilizer must be considered. Large quantities of by-product gypsum
are produced in the production of wet-process phosphoric acid. This
by-product gypsum needs to be disposed of in an environmentally
acceptable manner. This environmental consideration is becoming a
larger factor in the continuing operation of existing phosphate
fertilizer facilities, the expansion of existing facilities, and the location
of new facilities.
The world reserves of potash are extremely large and widely
distributed. Potash supplies that can be competitively developed are
located in a more limited number of countries. The major producers of
potash fertilizers include Canada, the United States, the CIS,
Germany, Israel, and Jordan [9]. The potash industry has encountered
decades of overcapacity and low profits. With this economic
environment, future production increases are expected to come from
expansion of the most competitive existing production facilities. Little
of the current potash production is located in the developing countries.
Much of the increase in potash consumption in the Far East and Asia
is expected to be supplied by imports. The environmental aspects of
potash mining, production, and by-product disposal are dealt with
extensively in other chapters of the book.
References
1. Agricultural and Environmental Policies: Opportunities for
Integration, OECD, Paris, 1989.
2. Current World Fertilizer Situation and Outlook, 1985/861991/92,
FAO, Rome, 1987.
3. Population Studies, No. 98, U.N. Department of International
Economic and Social Affairs, New York, 1986.
4. 1990 World Population Data Sheet, Population Reference Bureau,
Inc., Washington, D.C., 1991.
5. Population Studies, No. 112, U.N. Prospects of World
Urbanization, Department of International Economic and Social
Affairs, New York, 1989.
 Page 8
6. World Commission, World Commission on Environment and
Development: Our Common Future, Oxford University Press, Oxford,
1987.
7. Yearbook of Fertilizer Statistics (various years), FAO, Rome, 1991.
8. Yearbook of Food and Agricultural Statistics (various years), FAO,
Rome, 1991.
9. Yearbook of Production Statistics (various years), FAO, Rome,
1991.
10. IFA Proceedings of the 1992 Production and International Trade
Committee Meeting, Paris, 1992.
11. Yearbook of Trade Statistics (various years), FAO, Rome, 1990.
12. L. R. Brown, State of the World 1992, W.W. Norton, New York,
1992.
13. E. Moore and E. Crousillat, Prospects for gas-fueled combined
cycle power generation in the developing countries, Yearbook of
Energy Statistics, United Nations, New York, 1991.
14. TFI Proceedings of the 1992 World Fertilizer Conference,
Washington, D.C., 1992.
15. Mineral Commodity Summaries 1992, Bureau of Mines, U.S.
Department of the Interior, Washington, D.C., 1992.
 Page 9

2
Pollutants, Wastes, and By-Products of the Fertilizer
Industry
Neculai N. Popovici
Chemenco Ltd.
Bucharest, Romania
I
Introduction
The increasing demand for food globally has been an incentive for the
development of various means to raise the agricultural productivity of
land. In striving for higher and higher crop yields, fertilizers have
made a significant contribution, and the world production of fertilizers
has experienced continuous growth. At an early stage, fertilizer
production grew rapidly in developed countries. Later, the fertilizer
industry gradually developed in countries possessing important
resources of raw materials, as well as in countries with large
populations where food had a high priority.
As a consequence of criticism from a variety of groups in the
developed countries, fertilizer consumption has declined in many of
these countries. This decrease has also been due to more scientific and
more efficient methods of applicationthat is why it has not resulted in
decreased soil productivity. However, fertilizer consumption in
developing countries had not been below the optimum level
recommended by modern agricultural techniques. Therefore, in these
geographical areas, where the demographic growth rate is also very
high, the critical problem of providing food for human communities
continues.
A continuous increase in fertilizer consumption has stimulated
activities that have resulted in greater efficiencies, continuous price
reductions, and an increase in agricultural production. Nevertheless,
with all the technical progress achieved on a worldwide scale, the use
of fertilizers remained a potential source of environmental pollution.
Thus more efficient technological processes were developed and
investigated by researchers. These were refined and new engineering
solutions forwarded that resulted in many improvements. One
proposal was the creation of production units of very large capacity.
This led to a concentration of production within huge fertilizer
complexes (combines) that could produce up to 2 to 3 million tons a
year.
Concentration of production within large industrial complexes
increases the potential
 Page 10
for possible pollution of the local environment. Of course, the
problem of environmental protection from fertilizers or intermediary
products resulting from this industry is expanding into other areas,
such as trade, transport, distribution, and use, each with specific
characteristics. Considerable differences have arisen in the degree of
environmental protection in developed countries and that in
developing countries. The concern in developed countries over the
activities involved in producing and using fertilizers is very intense. It
is clear that greater attention should be paid to pollution in the
developing countries, especially because many of them continue to
increase the production and utilization of fertilizers. The domain of
fertilizer manufacturing, including the preparation of such raw
materials as sulfur, phosphate rock, and potassium salts, raises
numerous problems of protection of the environment which require
solution during the construction and operation of industrial
complexes.
II
Pollutants, Wastes, and By-Products
The manufacture of fertilizers generates many chemical compounds
which if released to the environment could detrimentally affect quite
large areas, including fauna and flora, water, and soil. By accounting
for such pollutants it is possible to get a picture of the noxious
substances that occur in the fertilizer industry (Table 1). The specific
noxious effect differs from one chemical compound to another and is
determined by the concentration, combined or accumulated effects,
and economic and social implications to the area affected by pollution.
In evaluating the effects of pollution, one must consider the
immediate impact as well as indirect influence of the pollutants. The
pollutants generated by the fertilizer
Table 1 Noxious Substances in the Fertilizer Industry
 Pollutants in
Atmosphere Water Wastes By-products
NH3 NH4+ Wastes from Pyrite ashes from
phosphate sulfuric
mines acid manufacturing from
pyrite
CO2 NO3 Wastes and Phosphogypsum from
F PO43 soluble salts wet-
NOx SO42 at potassium process phosphoric acid
mines manufacture
SOx KCl Acid sludges in H2SiF6 solution from
sulfuric fluorine
SiO2 powder NaCl acid gas washing in
Pyrite ashes manufacture phosphate
powders from and phosphoric acid
pyrites and fertilizer
gases from
nonferrous
metallurgy
Phosphate K2SO4
Powders SiO2
Fertilizer F Spent catalysts CaCo3 from
powders Sludges and nitrophosphate
polluted manufacturing by Odda-
water type process
Noise From plant Liquid effluents CO2 gas from ammonia
washings manufacturing
Radiation from Liquid
phosphogyp- effluents
sum storage
 Page 11
industry present different levels of impact on the environment. Thus
the atmosphere is severely affected by large quantities of SOx and
NOx released by fertilizer plants which are added to similar emissions
resulting from other industrial activities, such as the electric power
[1], steel, and nonferrous metallurgy industries, as well as pollution
caused by internal combustion motors (emissions of NOx). Those
emissions are the cause of ''acid rains" that fall over large areas,
including those occupied by forests and agricultural surfaces [2].
Similarly, emissions of gases with fluorine from units processing
phosphates add to the quantities of other volatile fluorine compounds
used at present that are lost to the atmosphere which contribute to the
reduction of the ozone layer that protects the earth from the dangerous
effects of ultraviolet rays [3].
The most noxious pollutants of waters are compounds containing
nitrogen, which are soluble in water (especially nitrates) and heavy
metals (especially cadmium). The role of nitrate is evident in the
incidence of methemoglobinemia ("blue baby disease") as well as in
gastric cancer. That is why the nitrate content of drinking water is
limited to 50 mg NO3 per liter in EEC directives [47].
Pollutants resulting from the processing of phosphates in wastewaters
include PO4 3, SO4 2, and Fions. In addition to the fact that their
degree of harm is somewhat lower, with the exception of fluorine
compounds, they have the advantage that they form compounds
having low solubility when in contact with Ca2+ ions. This ion is used
in a controlled neutralizing process and is found in numerous types of
soils. Figure 1 presents the relation between the content of these three
ions in a sample of residual water resulting from a fertilizer unit after
being neutralized with Ca(OH)2.
Natural factors can diminish or accentuate the effects of pollution, but
at present it is not possible to rely on natural dispersion, dilution, or
on natural purification of the environment with time. In the past,
fertilizer plants producing wet-process phosphoric acid that were
situated near a seashore disposed of their residual phosphogypsum
directly in the sea or had it transported by barge and discharged in the
sea. This practice was based on the relatively low solubility of the
residual product in seawater. Where such practices still exist,
authorities are taking more and more severe measures to restrict them
[3]. Similarly, in factories where sulfuric acid and nitric acid are
manufactured, large quantities of residual gas (in the case of old
technologies) are dispersed into the air through stacks. Nowadays
such disposal is not tolerated since there exist new technologies that
effectively reduce the residual gas emissions [8,9].
Based on the foregoing, one might ask: Are the mineral fertilizers
harmful or beneficial? It is our intent in this book to provide an
answer so that the utilization of mineral fertilizers can continue to
benefit humankind.
III
Recent Pollution Reduction by the Industry
Diminishing the negative environmental effects associated with
fertilizer manufacture has been a serious concern of the industry. As a
result of specialists' efforts, the potential for pollution generated by
fertilizer producers can be reduced considerably (Table 2) [1,6,814]. It
has been possible to achieve pollution abatement results due to efforts
in various areas, such as the following:
1. Developing new, less polluting technologies. For example:
a. Double catalysis and double absorption (DC/DA) in sulfuric acid
units.
b. Pressure absorption in nitric acid units [8,9,15].
 Page 12

Figure 1
Parameter variation during pH neutralization.
2. Adopting elaborate special technologies designed to destroy
polluting substances in the units generating them. For example:
a. Methods of SOx absorption from effluent gases discharged from
sulfuric acid units.
b. Technologies for destroying NOx effluents discharged from nitric
acid plants [9,15].
3. Diminishing the utilization of raw materials that have high pollutant
contents. For example:
a. Coal for ammonia production.
b. Pyrites for sulfuric acid production.
4. Employing engineering solutions to reduce pollution risks and
decrease pollution effects as follows:
 Page 13
Table 2 Reduction of Pollution Effluents in the
Manufacture of Fertilizers During the Past 20
Years
From: To:
Nitrogen fertilizers (NH3,
HNO3)
CO2 40%
reduction
N2O (ppm) 1500 300
NOx (ppm) 1000 80
Sulfuric acid
% S discharge 2 0.3
SO2 (in exhaust) (%) 0.2 0.03
NPK granulation
Liquid discharge (%) 1 0
Gaseous discharge, dust 300 50
(mg/m3)
Phosphoric acid
Liquid discharges
Gypsum slurry (% P2O5 6 1
losses)
Concentrated P2O5 (%) 1.3 0.2
Heavy metals: Cd (ppm) 2 0.5
Gaseous discharge (mg/m3) 15 3
Source: Ref. 4.

a. Choosing sites more suitable from an environmental protection

viewpoint.
b. Using an air-or water-cooling system in a closed circuit.
c. Selecting more resistant construction materials and efficient
leakproof systems aimed at reducing the risk of uncontrolled
leakage.
 d. Using more efficient equipment for some technological
operations, such as scrubbers with superior washing efficiency.
5. Recirculating contaminated effluents. For example:
a. Contaminated waters and purged liquids from nitrophosphates
units.
b. Diluted washing acids from sulfuric acid units based on pyrites.
c. Ammonium-contaminated waters following stripping operations
in ammonia and urea plants.
6. Using modern, nonpolluting processes for chemical treatment of
some of the effluents from fertilizer plants, such as:
a. Biochemical or ion-exchanger treatment of residual waters
containing NH4 + [6].
b. Reduction of radioactivity and heavy metals content from
phosphoric acid through solvent extraction.
Of course, all these means should be considered during the design and
construction of new plants, but many can also be used within existing
plants when upgrades are considered for economic reasons.
Another direction of research in the fertilizer industry during the last
few decades has been the discovery and implementation of the best
technical solutions for the industrial utilization of wastes and by-
products, among them phosphogypsum, pyrite ashes, fluorine
compounds, CaCO3, and CO2 resulting from units producing,
respectively, phosphoric acid, sulfuric acid, phosphate processing,
nitrophosphates, and ammonia, as well as wastes from potassium ore
mining. Phosphogypsum utilization has been a major problem for
 Page 14
decades because each ton of P2O5 produces 4 to 5.5 tons of
phosphogypsum. Many hundreds of millions of tons of
phosphogypsum have accumulated near phosphoric acid plants and
continue to do so. Research has been carried out worldwide to solve
the problems connected with this particular waste [1619].
Figure 2 presents various routes that can be followed by researchers in
finding and proposing an ideal solution for the recirculation of sulfur.
Using the MullerKuhne method, sulfuric acid and cement are obtained
simultaneously from phosphogypsum. This industrial-scale process is
being used in a fertilizer factory in Phala Boarwa, South Africa, built
in 1972, that has a daily production capacity of 350 tons of H2SO4.
This attractive solution has become uneconomical, causing the
cessation sulfuric acid production from anhydrite (the original
MullerKuhne process), a process developed in Germany as a "crisis"
or "wartime" technology.
An intensive study has also been made of the utilization of fluorine as
a solution to the problem of H2SiF6 resulting from phosphorus
fertilizer manufacturing plants. Figure 3 illustrates the technologies
and utilization that have been studied for fluorine processing. Many of
those mentioned could not be applied practically for economical
reasons or because the final products are of low quality [1214,20,21].
However, a residual product that results in large quantities from
nitrophosphate manufacturing (by the Odda process and its
derivatives)CaCO3has found an excellent use as a corrective factor for
acidic soils. It also contains 2 to 3% NH4NO3, which is beneficial to
plant growth.
Producers of sulfuric acid from pyrites have to deal with a
troublesome by-product, ashes resulting from the roasting of pyrites,
large quantities of which are produced per ton of H2SO4 (0.7 to 0.75
ton of by-product). There have also been numerous studies of and
processes developed for separation of nonferrous metals from residue,
the last yielding good iron ore. But there are cases in which such
processes are not used for economic reasons.
An effluent produced in great quantities from ammonia plants is CO2
gas. It is generally known that this emission, primarily from the
burning of fossil fuels, is responsible for climatic warming, and onset
of the "greenhouse effect." The manufacture of ammonia generates
large quantities of CO2, for which numerous uses have been found, as
shown in Fig. 4. These uses depend on the profile of the fertilizer
plants and the capacity of the market to absorb the products
manufactured from the CO2. Nevertheless, there exists a large excess
of CO2 from ammonia plants that is purged to the atmosphere. There
are liquid effluents from fertilizer plants containing dissolved
chemical fertilizers that can be used in irrigation water (diluted). They
can also be used for the production of liquid fertilizers on a seasonal
basis to reduce costs.
As far as wastes and solid by-products are concerned, if there is no
possibility of using them economically, proper storage should be
considered to avoid environmental pollution. For this purpose the
characteristics of the soil and the groundwater table should be known.
Geological data should be found for the most adequate solutions and
for the most suitable measures of waterproofing. In Fig. 5 an example
is given for the hydraulic storage of phosphogypsum.
The pollutants, wastes, and by-products mentioned above constitute
the main potential environmental effects of fertilizer plants. With this
awareness of potential problems the industry has emphasized efforts
aimed at minimizing to the greatest practical extent the negative
effects on the environment. Thus to the question: "Mineral
fertilizerharmful or beneficial?" one may conclude that the balance is
weighted toward fertilizer utilization.
In the construction of new fertilizer plants one must investigate all
new and updated
 Page 15

Figure 2
Possible uses of phosphogypsum. (From Ref. 16)
 Page 16

Figure 3
Fluorine sources, fluorine products, and their primary applications. (From Ref. 21)
 Page 17

Figure 4
Commercial applications of CO2 coming as waste from ammonia plants.
 Page 18

Figure 5
Gypsum pond water seepage control. (From Ref. 16.)
 Page 19
technologies for reducing noxious emissions to the lowest possible
levels. Plants should be provided with adequate
cleaning/purification/treatment facilities to eliminate harmful
substances. Selection of the most appropriate site is also a very
important factor in building a new industrial unit. In addition, the
climatic characteristics as well as social aspects and economic factors
must be considered, to determine their influence, positive or negative,
on site selection from an environmental viewpoint.
Environmental protection problems are becoming more difficult for
many existing fertilizer plants. At the time of their construction,
pollution standards were not as severe as in today's milieu. Also, the
pollution-abatement technologies were less well developed. In
addition, the physical wear of existing installations can be an
additional source of pollution [4]. Nevertheless, even for such cases it
is possible to find and apply technical solutions that can help to
ameliorate the situation and lessen the danger of pollution.
When the technological process conditions are similar, the volume of
pollutants is usually related to the plant capacity. If an increase in
production capacity is dictated by economics, the need for additional
pollution protection acts in a countervailing manner. Generally
speaking, one must consider environmental protection problems,
especially in the fertilizer industry, in relation to fertilizers' net effect
on food prices, which are a vital sector of the economy. A solution is
needed that balances economic concerns and environmental
preservation. Applicable technical solutions should solve the problem
of specific pollutants entirely, not transfer the pollutant from one
medium to another, such as from the atmosphere into the water. That
is why the antipollution solutions chosen should ensure a total
solution of the problem.
IV
Environmental Awareness of the Fertilizer Industry
To ensure adequate nutrition for humankind now and in the future,
fertilizer use is requisite. Fertilizers help to raise the level of
agricultural production and help satisfy the need for food on a global
basis.
It is possible that many public officials concerned with social
problems and protection of the environment consider fertilizers to be a
"necessary evil." Accepting this "verdict" does not mean that fertilizer
production should be seen as an evil that must be left without
limitation in causing harmful effects to the environment. On the
contrary, there is a need to use all means to diminish these effects. In
this respect, the public should be better educated to understand the
problem in all its complexity. This activity should involve the wide
variety of people associated with the industry.
Another part of the environmental protection framework involves
factors outside the specialized domain of production, such as
environmental regulations and taxes related to environmental
protection measures. Involved are regulations at the governmental
and/or local level, an organized and specialized network for auditing
and control, adequate technical means for testing and intervening, and
so on. Participating in these activities are organizations and private
individuals who are strongly in favor of environmental protection and
who sometimes act with remarkable tenacity and efficiency, especially
in the most developed countries [4].
Elaborating on the laws and norms that refer to environmental
protection, especially in the case of fertilizer plants, the technical level
that can be realized as well as the possible economic level that this
industry can afford should be taken into account. That
 Page 20
is why the norms elaborated should be a reasonable compromise
between the objective envisaged and the technical and economical
levels attainable. The norms imposed for effluents should not lead to
unreasonable restraint and curtailment of fertilizer production
[4,7,22]. It is necessary to ensure that local administrative authorities,
mass media, and institutions and organizations active in the social,
medical, and educational fields all play an active role in protection of
the environment. That is why the education of the public must be
extended over a wider group, to involve them in this activity either
directly or indirectly.
Coexistence between the fertilizer industry and the social media
should be based on the ability to communicate continuously, to avoid
psychological pressures that might degenerate into social tensions,
followed as a rule by drastic decisions having negative effects [23].
Therefore, good communication is needed between the fertilizer
industry and the public to ensure good relations and to avoid
exaggerated interpretations generated by ignorance.
Continuous monitoring systems are advisable, a notable example
being recent publicity concerning the presence of cadmium as a very
harmful element. Cadmium has been detected in chemical fertilizers,
causing strong opposition to their use. However, workers at a U.K.
research station have determined an annual addition to the soil of only
32 g of Cd, from which 16 g came from farmyard manure, 14 g from
the atmosphere, and 2 g from single-superphosphate application [4].
Good relationships should be maintained with public authorities and
the mass media. There should be organized educational actions for
people living in neighborhoods surrounding fertilizer production
facilities. Similarly, any steps that are taken for technological
improvements to the fertilizer units and means to improve the
environment should be made known to the public. This information
would help to increase the confidence of the public in the concern of
manufacturers to protect the environment [23].
This book presents the technical level of worldwide progress in the
field of fertilizer manufacturing considered from the point of view of
potential harmfulness to the environment. Also noted are existing
technical solutions and possibilities for solving various problems in
this field. Shown also are the limits reached at the current stage of
technological development. This information is meant to help
manufacturers and specialized institutions in their care for the
environment and to give others the opportunity to enrich their
knowledge of this field.
References
1. H. Kuroda et al., Recent technological development in SOx and
NOx removal systems for flue gases from thermal power plants,
Hitachi Review, Vol. 36, No. 6, 1987, pp. 329338.
2. B. W. Innes, It's raining nitrates and sulphates, ChemTech July
1984, pp. 440447.
3. N. N. Popovici, Fertilizer industry: environment pollution sources,
UNIDO Expert Group Meeting on Minimizing Pollution from
Fertilizer Plants, Helsinki, 1974.
4. L. M. Maene, IFA and the environment, British Sulphur 7th
European Fertilizer Forum, Ghent, November 2224, 1992.
5. British Sulphur Corp. Ltd., Nitrogen, No. 182, NovemberDecember
1989, pp. 1317.
6. British Sulphur Corp. Ltd., Nitrogen, No. 168, JulyAugust 1987,
pp. 3844.
7. British Sulphur Corp. Ltd., Nitrogen, No. 167, MayJune 1987, pp.
1417.
8. British Sulphur Corp. Ltd., Nitrogen, No. 174, JulyAugust 1988,
pp. 3435.
9. British Sulphur Corp. Ltd., Nitrogen, No. 171, JanuaryFebruary
1988, pp. 2534.
 Page 21
10. R. Maddury and P. Naidu, Modeling of a Venturi Scrubber for the
Control of Gaseous Pollutants, American Chemical Society,
Washington, D.C., 1985.
11. British Sulphur Corp. Ltd., Phosphorus and Potassium, No. 110,
NovemberDecember 1980, pp. 2730.
12. British Sulphur Corp. Ltd., Phosphorus and Potassium, No. 145,
SeptemberOctober 1986, pp. 3436.
13. British Sulphur Corp. Ltd., Phosphorus and Potassium, No. 152,
NovemberDecember 1987, pp. 3640.
14. P. K. Ramasubbu and N. Natarajan, Improving performance of
complex fertilizer plant, Fertilizer News, December 1986, pp. 5056.
15. J. Roiron, How to solve the problems of NOx pollution from nitric
acid plants, Nitrogen '86: The World Nitrogen Fertilizer Conference,
Amsterdam April 2023, 1986.
16. N. N. Popovici, Utilization of Phosphogypsum (Phosphogypsum
Still a Problem), Engineering Company for Inorganic Chemistry and
Fertilizer Industry (IPRAN), 1982.
17. F. Kilz et al., Production of sulphuric acid and cement from the
phosphogypsum, Journal of Technology and Development (issued by
the Arab Federation of Chemical Fertilizer Producers), Vol. 12, No. 4,
1987, pp. 2131.
18. N. Singh et al., Rendering phosphogypsum suitable for plaster
manufacture, Indian Journal of Technology, Vol. 22, January 1984,
pp. 2832.
19. British Sulphur Corp. Ltd., Phosphorus and Potassium, No. 172,
MarchApril 1991, pp. 2832.
20. A. W. Frazier et al., Chemical behaviour of fluorine in production
of wet-process phosphoric acid, Environmental Science and
Technology, Vol. 11, October 1977, pp. 10071014.
21. N. N. Popovici et al., Phosphate Rocks Used in Fertilizer
Industry: A Valuable Source of Fluorine Products, Engineering
Company for Inorganic Chemistry and Fertilizer Industry (IPRAN),
1986.
22. British Sulphur Corp. Ltd., Phosphorus and Potassium, No. 175,
SeptemberOctober 1991, p. 27.
23. F. Koppatschek, Training for right to know, Farm Chemicals,
Summer 1991, pp. 6263.
 Page 23

3
Regulation and Pollution Control in the Fertilizer
Industry
Charles A. Hodge
Tennessee Valley Authority
Muscle Shoals, Alabama
I
Regulations Affecting the Fertilizer Industry
The coming decades will undoubtedly see increased industrialization.
As in the developed nations, the production of industrial by-products
and wastes will increase in the less industrialized countries. Based on
the experience of the past few decades, increased emphasis will be
placed on preventing and limiting environmental pollution. Having
encountered the effects of environmental pollution earlier, the
developed countries have been in the forefront in establishing
pollution control laws and regulations covering proposed facilities,
existing facilities, and decommissioned facilities [1].
We live in a new environmental era. The number and complexity of
environmental regulations have grown dramatically in the last two
decades and will continue through the turn of the century. In the
United States, the first environmental regulation affecting the fertilizer
industry came with the Federal Air Pollution Control Act introduced
in 1970 (the Clean Air Act). The U.S. Environmental Protection
Agency is responsible for implementing the statutes of this law. The
first air pollutants covered by this law of concern to the fertilizer
industry were SOx and NOx. Since the passage of the Clean Air Act, a
series of additional U.S. federal environmental laws affecting the
fertilizer industry have been passed [2]:
1. Federal Air Pollution Control Act (Clean Air Act), 1970
2. Federal Water Pollution Control Act (Clean Water Act), 1972
3. Safe Drinking Water Act, 1974
4. Toxic Substances Control Act, 1976
5. Resource Conservation and Recovery Act, 1977
6. Comprehensive Environmental Response, Compensation, and
Liability Act, 1980
7. Pollution Prevention Act, 1990
 Page 24
All of these laws are technology based except the Safe Drinking Water
Act, which is health based. The technology-based environmental laws
set the limits for pollutant discharges on the basis of the best
practicable control technology available at the time of enactment of
the law. The health-based law sets limits on the concentration of
specific chemicals in the water based on the potential health hazard
for which the water is used. Summaries of the major U.S.
environmental laws affecting the fertilizer industry are given below
[3,4].
A
Federal Air Pollution Control Act
The Federal Air Pollution Control Act is also known as the Clean Air
Act. This law, which was derived from the original Clean Air Act of
1955 and the Air Quality Act of 1967, was enacted in 1970 with
amendments in 1977 and 1990. With this act, the U.S. Environmental
Protection Agency (EPA) was required to establish national ambient
air quality standards and was empowered to control hazardous air
pollution emissions. Amendments to the act put into place a
standardized basis for rules relating to new source performance
standards and hazardous air pollution standards. The Clean Air Act
regulates emissions of hazardous air pollutants. Title III of the Clean
Air Act lists 189 pollutants, among which the fertilizer industry must
be concerned with particulate matter, NOx, HNO3, HF, SO2, H2SO4,
and radionuclides, including radon. Ammonia was initially included
on the list of extremely hazardous substances but was removed from
the list by Congress. However, individual states in the United States
must implement this act using the federal standards as minimums.
Individual states can set their own emission standards for ammonia
[5].
The Clean Air Act and its amendments contain over 800 pages, and
additional detailed regulations will continue to be written and
implemented for individual chemicals. Similarly, specific rules and
regulations for individual states will evolve with time using the Clean
Air Act as the minimum standard. In this act, toxins are considered to
be at the ''major source" threshold (subject to regulation) when the
emission rate is 10 tons/y or higher for one of the listed pollutants or
at 25 tons/y or higher for any combination of pollutants. Regulation of
the emissions will occur automatically for new facilities through the
state permit systems. Existing facilities will have 3 years to reach
compliance after the date that relevant air toxic regulations become
effective. To avoid the requirements for obtaining operating permits,
fertilizer producers that are not major sources must document that
they are "non-major sources" of air pollution. Existing facilities can
be prepared for the new permitting process by verifying compliance
with existing permits on a periodic basis, collecting the data and
information needed for the new permits, or reducing emissions to
levels below the major source threshold.
B
Federal Water Pollution Control Act
The Federal Water Pollution Control Act of 1972 and its amendments,
the Clean Water Act of 1977 and the Water Quality Act of 1987,
focused on improving water quality standards in the United States.
These requirements are based on health standards intended for various
water use categories, which are:
1. Water for full-contact recreational use
2. Water for support of fish and wildlife
3. Public water supply
4. Water for agricultural and industrial use
 Page 25
Both direct and indirect discharges of water are covered by this act.
Point sources regulated in the fertilizer production industry include
water treatment plants, process waste streams, liquid discharges from
cooling towers, and sewage discharges. Discharge of any pollutant
into U.S. waters from point sources is prohibited by this act unless the
discharge is authorized in a permit from the National Pollution
Discharge Elimination System (NPDES). The NPDES permit
specifies required monitoring, control and reporting, and allowable
concentrations of pollutants in the discharge [6].
The Clean Water Act allows individual states to operate the NPDES
permit system instead of the federal EPA. Again the individual state
requirements must meet the federal requirements as a minimum. At
the present time about 80% of the states issue NPDES permits, and
the EPA issues permits for the remaining states [7]. In addition to the
industrial water discharges covered by the Clean Water Act, storm
water discharges associated with industrial activity are also covered
by the law. This includes storm water discharges from industrial areas
where material handling equipment or activities, raw materials,
intermediate products, final products, waste materials, by-products, or
industrial machinery are exposed to storm water.
Effluent guidelines and standards exist in EPA's NPDES Guidance
Manual for a large number of industries. Fertilizer producers can find
specific guidelines and standards under one or more of the following
industry categories:
1. Fertilizer manufacturing point source
2. Phosphate manufacturing point source
3. Mineral mining and processing point source
4. Inorganic chemicals manufacturing point source
A permit is required under the Clean Water Act for the discharge of
process wastewater from new sources and from new dischargers. In
the permit application toxic pollutants and hazardous substances must
be identified that are present in the process wastewater discharge.
Examples of chemicals that may be found in fertilizer production
water discharges that must be identified and quantified include
nitrogen dioxide, phosphoric acid, sodium bifluoride, sodium fluoride,
and sulfuric acid. To meet the standards for a permit, information is
required on the quantities, treatment facilities, and processing for the
chemicals listed above and any other listed chemicals discharged in
process wastewater.
C
Safe Drinking Water Act
The Safe Drinking Water Act (SDWA) of 1974 had health-based
standards for setting the maximum contaminant levels in water used
for the public water supply. The act contains requirements for
monitoring, reporting, and public notification. The standards in this
act are also used as reference points for the protection and remediation
of water resources under several other EPA programs, such as the
Clean Water Act, and as standards for state programs [8]. Although
the Safe Drinking Water Act is oriented primarily toward public water
supply systems, industry must take proper precautions to prevent
contamination of groundwater and surface waters that serve as a
source of public drinking water [9].
Potential liability can result from not having an effective
environmental management system in the process industries. Severe
penalties and very high liability can result from many different
contaminant sources. A summary of some of the major sources of
potential water contamination from the process industries follows.
 Page 26
1. Continuous or periodic emissions, leaks, and spills
2. Wastes and by-products deposited in landfills
3. Wastes in surface containment facilities, such as ponds, pits, and
lagoons
4. Large accidental releases of intermediates, product, or waste
products
5. Open storage of product or waste products
6. Improper waste management by another party of off-site waste
material from a production facility
Later amendments to the Safe Drinking Water Act listed additional
contaminants with standards on the maximum contaminant level in
drinking water. The chemicals listed as contaminants include organics,
inorganics, metals, herbicides, and pesticides. Some of the chemicals
encountered in the fertilizer industry on the SDWA contaminant list
are nitrates, fluorides, sulfate, zinc, cadmium, radon, and materials
causing turbidity in water [4].
D
Toxic Substances Control Act
The Toxic Substances Control Act (TSCA) of 1976 covers organic
and inorganic substances, including combinations that have been
determined to be a risk to human health or toxic in the environment.
An initial inventory of chemical substances was made by requiring
industry to report all chemical substances produced or imported in
1977. This act also controls newly manufactured chemicals. The act
requires a premanufacture notification of any new chemical or the
reporting of any proposed significant new use of the chemical. Import
certifications are required for chemicals and mixtures under this act.
The inventory of chemical substances was updated in 1986 and 1990
and will continue being updated at 4-year intervals. TSCA requires
special reporting and record-keeping requirements for specific
chemical compounds. Chemical compounds that require special
reporting and record keeping are listed by the EPA in the publications
identified as 40-CFR-704, GPO#055-000-00254-1, and GPO#055-
000-00361-1.
E
Resource Conservation and Recovery Act
The Resource Conservation and Recovery Act (RCRA) of 1976 was
enacted to establish a federal waste management system for hazardous
waste. It also authorized a program for regulation of hazardous waste
that included inspection and enforcement authority. RCRA has since
been amended several times. Wastes covered by RCRA now include
waste types categorized as hazardous wastes, municipal solid wastes,
special large-volume wastes, and nonhazardous industrial solid
wastes. Recent emphasis of RCRA waste management philosophy has
been to favor pollution prevention, recycling, and waste treatment
[10]. Criminal penalties exist for noncompliance with RCRA.
Individual states are required to compile and maintain an inventory of
active and inactive waste sites within the states. RCRA is one of the
major congressional acts that has resulted in much detailed regulation,
penalty provisions, and high liability provisions for contaminated
sites.
F
Comprehensive Environmental Response, Compensation, and
Liability Act
The Comprehensive Environmental Response, Compensation, and
Liability Act (CERCLA) enacted in 1980 had as its goal responding to
and cleaning up sites from past hazardous
 Page 27
waste activities and responding to current waste releases. CERCLA is
also known as the "Superfund" act since part of this act includes
regulations, penalty provisions, and liability for contaminated-site
cleanup [11]. CERCLA specifies that responsible parties may be liable
for government response costs or any other necessary costs incurred in
the event of a hazardous substance release or threatened release from
a facility. The definition of a hazardous substance regulated by
CERCLA is listed under the CAA, CWA, TSCA, or RCRA acts
described earlier. In case of an accident or emergency, release of a
hazardous substance above the Reportable Quantities level requires
reporting the release to the National Response Center [12]. Reportable
Quantities are specified in 40-CFR Part 302 of CERCLA. For
example, the accidental collapse of a gypsum pond wall could release
fluids to a river. The chemical composition and quantity of fluid
released would be covered under the Reportable Quantities
provisions. In reporting an emergency release, the following
information is required:
1. Location of release
2. Description of material released
3. Quantity of material released
4. Time and date of release
As part of CERCLA, the Superfund Amendments and Reauthorization
Act (SARA) of 1987, companies were required to report on the fate of
more than 300 chemicals [4]. This information is reported publicly
and the numbers are totaled for individual chemicals on a nationwide
basis. This Toxics Release Inventory (TRI) has been used as a basis
for measuring progress in reducing pollution. The EPA has helped to
ensure the accuracy of TRI reporting through thousands of plant site
visits and fines totaling millions of dollars. Since 1987, total releases
of these 300+ chemicals have been reported. In the 25 chemicals of
highest tonnage release, chemicals encountered in the fertilizer
production industry include ammonia, ammonium nitrate, ammonium
sulfate, nitric acid, phosphoric acid, and sulfuric acid. Based on data
available for 1990, a significant reduction of 44% in the release rate of
the top 25 chemicals was achieved in the period 19871990. This
reduction is even more significant when considering the 11% increase
in total production of these chemicals during the same period. One
effect of the 1987 toxic chemical release requirements has been a
strong focus by industry on preventing pollution by various methods
of reducing wastes produced in chemical facilities. These methods of
pollution prevention include process changes, operational changes,
equipment changes, chemical substitutions, and product changes [13].
G
Pollution Prevention Act
The Pollution Prevention Act (PPA) of 1990 was enacted to orient
pollution control more in the direction of pollution prevention and in
reducing pollution at the source compared with previous legislation,
which focused on waste that had already been produced. The U.S.
Environmental Protection Agency delineated a multilevel hierarchical
approach to waste management in the PPA. The objective of this
hierarchy is to give priority first, to source reduction; second, to
recycling and treatment; and last, to waste disposal. This waste
management hierarchy is shown in Fig. 1.
The pollution philosophy has been promoted by the EPA by starting
an Office of Pollution Prevention. This office has now been merged
with EPA's Toxic Substances Office. In this philosophy the best waste
management system is believed to be that which
 Page 28

Figure 1
Waste management hierarchy.
avoids the generation of waste initially rather than one that reacts to
waste production through recycling, waste treatment, and disposal. In
many cases the prevention of pollution can be more profitable
economically than other methods of waste management. Economic
incentives for pollution prevention include higher product production
efficiency, better resource conservation, less legal liability, lower cost
of waste product management, and less stringent regulatory levels.
Pollution prevention results will become more and more subject to
measurement on both the national scale and on an individual company
basis with the more recent waste reporting requirements of EPA's
Toxic Release Inventory and the Clean Air Act.
II
Environmental Compliance Costs
For the fertilizer industry, environmental compliance costs are
increasing at a higher rate than for any other segment of its cost
structure. These environmental compliance costs affect the design and
construction of new facilities. They are an increasingly higher
percentage of operating costs. These environmental costs include
compliance costs for past, present, and future environmental
management. Subsequent chapters cover various environmental
processes encountered in plant site location, permitting, design,
operation, decommissioning of facilities, and plant site remediation.
 Page 29
The expense associated with environmental compliance in the
fertilizer industry has a substantial influence on the cost of doing
business. Data recently supplied by 16 North American integrated
fertilizer producers demonstrate the upward escalation of
environmental compliance cost [14]. Environmental capital
investment for these facilities increased from $33 million in 1987 to
$106 million in 1991. When total operating cost is included, the total
cost increased from $123 million to $231 million (see Fig. 2). The
same 16 fertilizer producers projected that the total cost for
environmental compliance would reach nearly $300 million by 1992.
In addition to capital costs there are increased operating costs, such as
continuous control and monitoring of gas, liquid, and solid effluents.
Long-term protection and monitoring costs for groundwater are also
incurred. Environmental compliance also affects what must be done in
decommissioning a plant facility, remediation of its industrial site,
restoration of mined land, and the long-term care of by-products and
wastes on land associated with a fertilizer production facility [15].
The U.S. Environmental Protection Agency estimated that federally
mandated environmental compliance and cleanup cost $115 billion in
the United States in 1990 [13]. For the United States the EPA
estimates that these environmental costs will amount to 2.7% of the
gross national product (GNP) in the year 2000 [161. Also, the EPA
projected that these costs would grow at double-digit annual rates
during the 1990s. The EPA's cost estimate to meet compliance for the
Resource Conservation and Recovery Act (RCRA) alone in the
United States illustrates this double-digit cost increase through the
year 2000 (Table 1). These cost estimates for RCRA compliance have
escalated from $174 million in 1983 to $1.7 billion in 1987 and are
projected to cost $11.7 billion in 1995.
Although the cost of environmental compliance is high, the cost of
being in violation of the federal environmental laws and regulations
can also be quite high, both monetarily

Figure 2
Cost of environmental control (16 North American
integrated basic fertilizer producers). (From Ref. 14.)
 Page 30
Table 1 RCRA Compliance Cost
Estimates
Pollution control and cleanup
costs
Year (billions of dollars)
1983 0.17
1987 1.7
1995 11.7
2000 13.2
Source: Ref. 17.

and in terms of public image. During the four years 19891992, the
enforcement of these laws and regulations in the United States
through monetary fines has increased dramatically, with over $5
billion collected from industry and individuals. This trend in fines for
not complying with the environmental regulations is shown in Fig. 3.
The technology needed for control and prevention of environmental
pollution is generally available. The choices involve primarily
selection of the most appropriate and economical pollution control
technology to meet the required compliance standards. New control
technology and new pollution control standards continually affect the
decision process. Most large modern fertilizer production facilities
employ technology that is very efficient in controlling pollution.
However, the degree to which this technology is used
 Figure 3
Cost of environmental fines in the
United States. (From Refs. 18 and 19.)
 Page 31
from one country to another varies widely based on the environmental
standards mandated in each country. These various levels of
environmental standards enforcement result in different levels of
production cost competitiveness on an international basis.
The environmental benefits accrue primarily to the geographical
regions where the strictest environmental standards are implemented.
However, the higher compliance costs affect both a company's
competitiveness and its profitability. Because of the disparities on an
international basis of the environmental standards and their economic
effects, there is a growing awareness of the need for international
agreements to share environmental compliance costs more equitably.
To date this subject has been dealt with only at the discussion stage.
Methods of implementing international cooperation have been
proposed, including internationally agreed-upon minimum industry
standards and international trade agreements specifying that products
be produced using agreed-upon minimum environmental standards.
Without an industry-wide agreement or more uniform international
standards, the environmental costs may impede fertilizer industry
expansion in those countries that have the most stringent
environmental standards.
III
European Environmental Laws and Regulations
The regulatory culture in Europe is quite different from that in the
United States. Environmental legislation in the United States is
characterized by extensive adversarial procedure and litigation.
Regulatory decisions in Europe tend to be developed from informal
consultation processes, with litigation being rare. The number of
environmental regulations in Western Europe has been growing at an
exponential rate during the past 20 years. Early legislation in the
European Community (EC) included the Treaty of Paris in 1951 and
the Treaty of Rome in 1957. By 1992, over 300 environmental
legislation items had been adopted by the EC since the 1957 Treaty of
Rome. It is beyond the scope of this book to cover environmental
legislation in detail. Therefore, only the majors laws and concepts are
covered.
EC law can have supremacy over national law and can be enforced in
the member states. The European Community legislation forms are
defined in Article 189 of the Treaty of Rome. The Treaty of Rome
created the European Economic Community (EEC). The 1987 Single
European Act (SEA) amended the treaty, setting December 31, 1992,
as the deadline for European integration. The provisions of this act
form the basis of the EC environmental protection law.
Environmental concerns are addressed in Title VII and Article 100a in
the SEA. Article 100a is the basis for most environmental legislation.
The SEA's Title VII, the Environmental Title, contains Article 130r,
which states the EEC's environmental objective: "to preserve, protect
and improve the quality of the environment, to contribute toward
protecting human health and to ensure product and rational utilization
of natural resources." The principle "the polluter pays" is also derived
from Article 130r. The "polluter pays" principle means that
environmental damage should be paid for by the party responsible for
the pollution. Article 130t states that member states can enact their
own environmental protection legislation, provided that the member
states' legislation has more stringent standards than those of the
corresponding EC regulations.
Federal process legislation in the EC is less extensive than legislation
in the United States. However, transboundary pollution requires
central environmental legislation.
 Page 32
Enforcement of EC central environmental legislation is not carried out
by direct authority of the EC but relies on enforcement by the member
states. Enforcement is encouraged by infringement actions in the EC's
Court of Justice. One problem with this system is the large variation
in the effectiveness of implementation and enforcement of
environmental directives from one member state to another.
The European Commission is an executive body that proposes
legislation to the Council of Ministers. The European Commission
staff, located primarily in Brussels, implements the rules and ensures
that EC law is applied correctly. The Council of Ministers adopts laws
following proposals from the European Commission. The
Environmental Ministers adopt environmental laws on a unanimous
basis of the members. The Environmental Ministers act on air, water,
chemical regulation, and waste laws. Legal judgments of
environmental cases take place in the Court of Justice, Europe's
highest court. National courts can be overruled by this court in
subjects in the field of EC law. To date, most environmental cases
have been brought to the European Commission to implement
directives not implemented by individual members.
Legal instruments of the EC are listed below:
1. Regulations. These are directly applicable to each member nation.
They can be compared to national laws. Both the Council of Ministers
and the European Commission are empowered to issue regulations.
2. Directives. These must be implemented by the member nations.
Results to be achieved are binding. The main difference between a
regulation and a directive is that regulations are directly applicable,
and for directives, the member nation chooses the form and methods
of integrating them into national legislation.
3. Decisions. These are binding on those to whom they are addressed,
such as a government or individual company. It has been held by the
European Court of Justice that decisions are capable of direct effect.
This means that certain provisions of the directive in question must be
precise and unconditional. The national regulatory authorities and
courts are obligated to give legal force to decisions that have direct
effect.
4. Recommendations. These are not binding.
To date, most environmental laws have been written in the form of
directives. Considerable variation in environmental laws exists in the
EC because of different interpretations of the directives by the
member nations. In 1991, an environmental directive came into force
dealing with air quality limit and guide values for sulfur dioxide and
airborne particulate matter. This Directive, 89/427/EEC, was an
amendment to the earlier Directive 80/779/EEC and related to
approved methods used for sampling and analysis. Directive
85/203/EEC relates to air quality standards for nitrogen dioxide. A
draft directive has been written for nitrogen pollution from
agricultural sources. This directive is aimed at protecting from nitrate
pollution underground aquifers used for drinking water. Directive
85/337/EEC requires that the member state regulatory authority take
into account the environmental impact of major construction projects.
Several proposals for environmental directives have been made for the
future. One such proposal is on integrated pollution prevention and
control (IPPC). The principles of this proposed directive are similar to
those of the British system, including integrated pollution control and
best available techniques. A permitting procedure specified in the
directive would require permits for air, water, and wastes or a single
permit for all three. Requirements of the permits would relate to
construction of a facility, 5-year operation
 Page 33
periods, and a commitment to remediate the site. Emission limits
would apply to specified substances to ensure that the appropriate
environmental quality objectives are met.
Current and future EC environmental legislation for industry relates to
best available technology (BAT). The best available technology
techniques not entailing excessive costs (BATNEEC) and best
practical environmental options (BPEO) are being incorporated into
this new legislation. The European Fertilizer Manufacturers
Association (EFMA) has been providing BAT information for various
types of fertilizer production.
Another proposed EC strategy for control of environmental pollution
is for an environmental auditing program for all types of chemical
production facilities. This is commonly referred to as an "eco audit."
Action programs suggest proposals for environmental policy from the
European Commission. The Fifth Action Program targeted several
sectors, including industry and agriculture. Initiatives and
recommended time goals from the EC's Fifth Environmental Action
Program 19922000 are given below.

ContinuingStricter implementation directives on nitrate

pollution of groundwater
19921993 Collection and updating of groundwater data
1992 on Additional specific emission standards to protect
groundwater
1993 Integrated water management and protection
1993 Protection and rehabilitation of aquifers
By 1995 Monitoring and control methods on groundwater
By 1995 Amendment proposals for existing legislation on
 SOx, NOx, and particulates
1995 Phosphate reduction directive for groundwater
By 1997 Identification of existing or potential problems
from cadmium
By 1999 Identification of existing or potential problems
from deposition of sulfur and nitrogen compounds

References
1. BNA Editorial Staff, U.S. Environmental Laws, Bureau of National
Affairs, Inc., Washington, D.C., 1986.
2. Environmental Compliance Manual: A Guide to Pollution Control
Regulations, J. Keller & Associates, Washington D.C., 1991.
3. R. Hall, Environmental management, the viewpoint of the regulator
at the national level, Phosphorus in the Environment Workshop,
Tampa, Fla., March 2527, 1992.
4. GII Environmental Statutes (various editions 19821987),
Government Institutes, Inc., Rockville, Md., 1987.
5. D. Selmi and K. Manaster, State Environmental Law, Clark
Boardman, Callaghan, New York, 1991.
6. D. Rona, Environmental Permits: United States, Van Nostrand
Reinhold, New York, 1988.
7. D. Jessup, Guide to State Environmental Programs, The Bureau of
National Affairs, Washington, D.C., 1988.
8. NAP Ground Water Quality Protection; State and Local Strategies,
National Academy Press, Washington, D.C., 1986.
9. L. Canter and R. Know, Ground Water Pollution Control, Lewis
Publishers, Chelsea, Mich. 1986.
10. J. Wright, Managing Hazardous Wastes: A Programmatic
Approach, The Council of State Governments, Lexington, Ky., 1986.
 Page 34
11. S. Deutsch and A. Tarlock, eds., Land Use and Environmental
Law Reviewed 1991, Clark Boardman, Callaghan, New York, 1991.
12. G. Lowry and R. Lowry, Lowrys' Handbook of Right-to-Know and
Emergency Planning, Lewis Publishers, Chelsea, Mich., 1988.
13. A. Thayer, Pollution reduction, Chemical & Engineering News,
Vol. 7, No. 46, 1992.
14. TVA, TFI-Fertilizer Facts, Washington, D.C., 1992.
15. J. Shields, U.S. fertilizer industry and the environment: from mine
to farm, 60th IFA Annual Conference, Seoul, Korea, May 1992.
16. EPA Publication 20z-1011, U.S. Environmental Protection
Agency, Washington, D.C., September 1990.
17. EPA Informational Briefing on RCRA Reauthorization, May 1991,
U.S. Environmental Protection Agency, Washington D.C., 1991.
18. EPA Publication 9200.5-008H, U.S. Environmental Protection
Agency, Washington, D.C., November 1990.
19. Record year for environmental fines, Chemical & Engineering
News, Vol. 7, No. 46, 1992.
 Page 35

4
Ammonia Production
Mircea Turtureanu
Department of Chemical and Petrochemical Industry
Bucharest, Romania
I
Introduction
Ammonia is the principal component in the production of nitrogen
fertilizers. The world ammonia capacity for 19951996 is estimated at
123,640,000 tons of nitrogen per year [17].
Ammonia synthesis from elements was done for the first time on an
industrial scale in Germany in 1913, based on the process developed
by Fritz Haber and Carl Bosch. The nitrogen required for ammonia
synthesis is taken from the air, whereas the sources of hydrogen vary,
and over time this has led to the development of several processes for
syngas production.
Production of ammonia by the steam-reforming process is relatively
clean and presents no special problems from an environmental point
of view. However, the impact on the environment should be analyzed
not only during normal operation of the plant, but also taking into
account prevention of accidents that may be caused by factors that are
sometimes subjective. This concept is valid for any type of plant, no
matter how small the risk to operating personnel and the environment.
In the case of ammonia plants we are concerned with the handling of
hazardous substances (toxic, flammable, explosive) within a wide
range of temperatures and at relatively high pressures. For this reason
the concept of environmental protection should consider the steps to
be taken starting with plant design and continuing through
construction and operation.
In the case of old plants still in operation, the steps for minimizing
environmental pollution should be accomplished gradually, depending
on local regulations and taking into account the investment costs
needed to put these steps into practice. It is probable that in the case of
old, small-capacity ammonia plants, the additional expenses for
process
 Page 36
modernization and environmental protection could not be justified on
the basis of economics thus leading to plant shutdown.
II
Factors Related to Ammonia Production
Ammonia plants are high energy consumers, and selection of the
feedstock is the most important factor in determining the capital
investment and production costs. The availability and cost of raw
materials are factors to be taken into account when deciding on the
construction of a new ammonia plant. The primary feedstocks for
ammonia production are discussed briefly below.
1. Natural gas. Natural gas is the raw material that ensures minimum
investment and production costs, a plant that is easy to operate, and
minimal environment problems. Steam reforming of natural gas has
become the most widespread process for ammonia production.
2. Naphtha. A naphtha-based ammonia plant uses a process similar to
that starting from natural gas. The high cost of naphtha as well as its
utilization in refineries has limited its use as a feedstock for ammonia
production. Furthermore, naphtha utilization requires feed storage and
treatment; hence the plant cost is higher than that for a natural
gasbased ammonia plant.
3. Heavy residual oil. Heavy residual oil as a feedstock in an ammonia
plant requires the partial oxidation process for syngas production. The
energy consumption of a heavy residual oilbased ammonia plant is
higher than that for steam reforming of natural gas due to the
necessary air separation plant and supplementary utility consumption
for the gas purification.
4. Coke gas. Coke gas contains important quantities of H2, CH4, and
CO. Ammonia production based on coke gas is not as widespread, due
to the high costs of gas purification and compression.
5. Coal. Coal is used as a raw material for ammonia production if
there is no other source of raw materials. Higher investment and
energy requirements, environmental pollution problems, and cheap
natural gas disposal has virtually halted the construction of coal-based
ammonia plants.
There are also other sources of hydrogen that may be used as the raw
material for ammonia production, but due to their limited availability
and high cost, they cannot be taken into consideration for industrial-
scale production. Among these is electrolytic hydrogen, an attractive
source only under conditions of very low energy cost. Refinery gas
containing light hydrocarbons can be used for ammonia production
only if it is in excess of the needs of the refinery.
Figures 1 to 3 are flow diagrams for ammonia production as a
function of the feedstock used. For this purpose, three raw materials
have been selected as representative for the ammonia industry: natural
gas, heavy residual oil, and coal [14].
Figure 1 is a flow diagram of ammonia production starting from
natural gas or naphtha. The main effluents produced at each process
stage have been indicated. In this case the ammonia production
process includes the following main stages: natural gas desulfurization
or naphtha evaporation and desulfurization, primary reforming,
secondary reforming, HT-CO shift conversion, LT-CO shift
conversion, CO2 removal, methanation, compression, and NH3
synthesis. The main effluents resulting from the process are the
 Page 37

Figure 1
Steam reforming of natural gas and naphtha.
flue gas from the primary reformer, the process condensate, the
carbon dioxide, and the purge gas from the ammonia synthesis loop.
Two alternative processes are available for ammonia synthesis gas
production using heavy residual oil as the raw material. These
processes have been developed independently by Texaco
Development Corporation and the Shell Group and are presented in
Fig. 2. Utilization of heavy oil is based on a noncatalytic partial
oxidation process, but the following can be used as feedstock: natural
gas, naphtha, fuel oil, vacuum residues, and asphalt. As shown in Fig.
2, the process is more complex than that of natural gas steam
reforming, and in the alternative A it includes oil gasification, quench
cooling and soot removal, carbon extraction, CO conversion, H2S and
CO2 removal, N2 wash, syngas compression, and NH3 synthesis. In
this case CO conversion takes place immediately after gasification
and a sulfur-resistant CO shift catalyst must be employed. Alternative
B includes oil gasification,
 Page 38

Figure 2
Heavy oil gasification.
gas cooling by steam production, H2S removal, CO conversion, CO2
removal, N2 wash, syngas compression, and NH3 synthesis. The
principal effluents resulting from the process are wastewater from
carbon extraction, H2S and CO2 from the syngas purification sections,
a mixture of CO and N2 from the N2 wash, and small quantities of tail
gas from NH3 synthesis.
Coal-based ammonia production predominated until World War II.
Later, the
 Page 39

Figure 3
Pulverized coal gasification process.
availability of cheap natural gas and development of the steam-
reforming process halted construction of coal-based plants. Coal
gasification processes for ammonia production can be classified
according to the method of gasification:
1. Fixed-bed gasification (Lurgi process)
2. Fluidized-bed gasification (Winkler process)
3. Pulverized coal gasification (Koppers Totzek process, Texaco
process)
 Page 40
Figure 3 presents a flow diagram for coal-based ammonia production
using Texaco's partial oxidation of coal, including as main stages coal
grinding, preparation of slurry for pulverization, gasification, gas
cooling and scrubbing, CO conversion, H2S and CO2 removal, N2
wash, compression, and NH3 synthesis. This process is very similar to
partial oxidation of the heavy residual oil gasification process. The
principal effluents resulting from the process are coal slags,
wastewater from the char/water handling section, H2S and CO2 from
syngas purification sections, a mixture of CO and N2 from the N2
wash, and small quantities of tail gas from ammonia synthesis.
It is clear that the use of natural gas as feedstock leads to a simplified
flowsheet for ammonia production compared to that using heavy
residual oil and coal. Under these circumstances, the effluents
resulting from an ammonia plant based on natural gas are reduced and
do not create difficult treatment problems to minimize their impact on
the environment [12].
III
Pollution Control in the Ammonia Industry
In the preceding section we reviewed the ammonia production
processes on the basis of the feedstock used, pointing out the principal
effluents resulting from these processes. From the point of view of
environmental protection it is essential to have a continuous program
for waste minimization, although this may involve relatively high
costs [1]. Because plants using natural gas as feedstock predominate
in worldwide ammonia production, various aspects regarding
pollution and environmental protection for these types of plants are
presented below in greater detail.
Treatments to decrease effluents and reduce the impact on the
environment will be analyzed, as well as the importance of proper
plant design, considering all ecological aspects from the beginning. In
accordance with the flow diagram presented in Fig. 1, ammonia
production by steam reforming is analyzed stage by stage together
with the usual steps to be taken for pollution control.
A
Design for Safety in an Ammonia Plant
Plant reliability and environmental protection are two of the most
important factors in ammonia plant construction. Operating personnel
and people living in the vicinity of the plant must be protected against
hazards. For this purpose, from the beginning, the following must be
analyzed: the hazardous materials being handled in the process, local
and state regulations, applicable codes and standards (industry codes,
standards guidelines, and practices), the system design and control
philosophy, and the fire protection and detection system [11]. It is also
necessary to have a hazards analysis and risk assessment to determine
the possibility of occurrence of various accidents.
Plant design activity should consider the fact that the integrity of the
mechanical system represents the first protective step against
accidental leakage of fluids handled in the process.
The quality assurance program implemented for equipment
manufacture should be extended during erection of the plant to avoid
any deviation from design standards and codes. The plant operation
manual should indicate normal operating procedures as well as all
steps to be taken in case of an emergency during plant operation.
Proper selection and training of operating personel are additional
factors to be considered.
For ammonia plants, one should also take into account the fact that the
process implies
 Page 41
handling of gaseous mixtures containing methane, hydrogen, carbon
monoxide, carbon dioxide, nitrogen, and ammonia, which are
dangerous to human beings and may yield flammable or explosive
mixtures.
B
Natural Gas Desulfurization
The natural gas used in steam reforming process contains sulfur
compounds which are poisons for nickel-and copper-based, reforming
and low-temperature shift catalysts. For this reason, the sulfur content
in the natural gas feedstock should be below 0.5 ppm. In addition,
sulfur compounds in the natural gas used as fuel release, by
combustion, SO2, which is discharged to the atmosphere together with
the flue gases.
The main sulfur compounds in the natural gas are hydrogen sulfide
and mercaptans. In case the natural gas contains appreciable quantities
of associated gas, other sulfur compounds, such as disulfides,
tiophenes, or carbonyl sulfide, may be present.
Mercaptans and disulfides break down easily at high temperature into
hydrogen sulfide, which is retained on zinc oxide. Other compounds,
such as thiophenes or carbonyl sulfide, are converted into hydrogen
sulfide in the presence of hydrogen, using a CoMo catalyst. Removal
of hydrogen sulfide from the natural gas feedstock on zinc oxide at a
temperature of 350 to 400°C is the current practice in most ammonia
plants.
There are, of course, other processes for desulfurization, but they
depend on the content and nature of sulfur compounds in the natural
gas. Mention should be made of sulfur compound removal by passing
the natural gas over metal-impregnated activated carbon. Active
carbon is efficient for the retention of hydrogen sulfide and light
mercaptans if the natural gas content in heavy hydrocarbons is low.
Carbonyl sulfide is not retained on active carbon and in such a case,
small amounts of ammonia are injected in the natural gas. The
reaction is as follows:

The optimum reaction temperature ranges between 25 and 35°C, and

the water requirement is ensured by natural gas saturation up to close
to the dew point.
Usually, the natural gas contains small quantities of sulfur
compounds, so that in ammonia plants desulfurization of natural gas
feedstock takes place but not that of fuel gas. If the concentration of
sulfur compounds is high, the entire quantity of natural gas (feedstock
and fuel) must be desulfurized, in which case wet processes are used.
Desulfurization of fuel gas must be considered in connection with the
SO2 content of flue gases and with state and local emission standards.
C
Steam Reforming
Steam reforming of natural gas takes place in two steps:
1. Primary reforming, where the natural gas feed is converted to H2
and CO by reaction with steam
2. Secondary reforming, where the reaction continues and air is
introduced to assure the required amount of nitrogen for ammonia
synthesis
To ensure the reaction temperature (800 to 900°C), a quantity of
natural gas is burned outside the reformer tubes. The flue gases
contain primarily carbon dioxide, water vapors, nitrogen, oxygen and
small quantities of NOx and SO2. The temperature of the flue gases
should be lowered to save fuel gas, but it has to be high enough to
prevent water vapor
 Page 42
condensation. This is done at low temperature levels by preheating the
combustion air and boiler feed water, by producing low-pressure
steam, or by saturation of feed gas with process condensate. From an
environmental point of view, CO2, NOx, and SO2 should be taken into
account [8].
Although there are no restrictions on CO2 emissions, it should not be
forgotten that the increased CO2 concentration in the atmosphere has
contributed most to the greenhouse effect and to acid rain. Some
countries have already started to take steps to tax CO2 emissions, as
they are a threat to ecology.
NOx emissions to the atmosphere began to be an important problem in
recent years. About 90% of the nitrogen oxides released to the
atmosphere result from the decomposition of organic matter; only
10% comes from industrial sources. The reduction of NOx emissions
from industrial sources has become compulsory because
concentrations are greatest in crowded areas and jeopardize human
health. In many countries, including the United States, Japan,
Germany, Sweden, and the Netherlands, there are regulations
imposing a limit on NOx emissions to the atmosphere. Generally, the
accepted value of NOx emissions released to the atmosphere from
industrial furnaces is 50 ppmv.
Nitrogen oxides are formed in combustion processes by thermal
fixation of atmospheric nitrogen and oxygen from combustion air,
resulting in thermal NOx, or by conversion of chemically bound
nitrogen from the fuel, in fuel NOx.
In reforming furnaces in the ammonia industry that use natural gas as
fuel, almost the entire amount of NOx is thermal NOx. The formation
rate of thermal NOx depends on the combustion temperature, the local
concentration of components, and the reactant's residence time at high
temperature. The nitrogen oxides content of the flue gases may be
decreased by lowering the combustion temperature, by a correct
stoichiometry as well as by reducing residence time at high
temperature. The test run for gaseous fuel combustion shows that NOx
content increases upon increase in the oxygen content of the flue
gases and in the combustion air preheating temperature. In case the
combustion air is not preheated (i.e., at ambient temperature), the
nitrogen oxides content in the flue gases is below 60 ppmv.
The main methods to reduce NOx concentration in the flue gases can
be divided into two large groups [18]:
1. Reduction of NOx concentration by modifying the parameters
having an influence on combustion, including:
a. Staged combustion, which implies a fuel-rich zone near the
burner by introducing only a part of the combustion air requirement
and by feeding the remaining air in another section of the furnace
(second stage). Another way to achieve staged combustion is by the
use of overfire air operation (OFA), which means to feed a part of
the combustion air above the burner location elevation.
b. Water injection, which has a good efficiency in case of thermal
NOx. The method consists of combustion air humidification or
direct water injection to the burners.
c. Flue gas recirculation (FGR), which has as an effect the
reduction in combustion temperature. Usually, a part of the flue gas
is mixed with combustion air prior to its introduction to the burners.
d. Utilization of low-NOx burners (LNB). There are a great number
of types of LNBs, which are based on a reduction of oxygen
content in the primary combustion zone, a reduction of flame
temperature, and a reduction of resi-
 Page 43
dence time at peak temperatures. The low-NOx burners can be
classified as follows: standard staged fuel burners, staged fuel
burners with internal FRG, low excess air (LEA) burners, and
burners generating localized off-stoichiometric conditions, both
fuel-rich and fuel-lean.
e. Reburning technology, which necessitates several combustion
zones. About 70 to 85% of the total amount of fuel is burned in the
main combustion zone. The remaining fuel is burned in the second
combustion zone, while the air excess is introduced as an overfire
air.
2. Reduction of NOx concentration by postcombustion treatment
methods based on the use of some substances that react with the
nitrogen oxides and reduce them to N2 and H2O. There are three main
processes for postcombustion NOx reduction.
a. Ammonia injection in the flue gases based on the reactions

The NOx content is lowered by 70 to 75% using ammonia injection.

b. Urea injection, which is similar to ammonia injection. Urea is fed
as an aqueous solution at the outlet of the furnace radiant zone and
breaks down to form NH2 and CO. NH2 reacts with NOx to form
N2 and H2O.
c. Selective catalytic reduction (SCR), based on ammonia injection
in the flue gas and catalytic reaction of NOx at temperatures
ranging between 300 and 400°C. The main advantage of this
process is the high-efficiency reduction level of NOx exceeding
90%. The disadvantage is the relatively high reaction temperature.
This shortcoming is overcome by the new selective catalytic
 process operating at low temperatures, ranging between 150 and
300°C, using a Ti-V oxide on a silica carrier catalyst.
The problem of nitrogen oxide reduction in ammonia plants should be
treated in relation to the legislation in force at the local or state level.
When natural gas is used as fuel, NOx reduction could be
accomplished by introducing ammonia into the reformer effluent and
by using low-NOx burners, if necessary. The SO2 content of the flue
gas depends on the sulfur content of the natural gas used as fuel.
There are many processes for removing SO2 from flue gas which use
sorbent materials based on Ca (usually limestone) or on sodium
(sodium bicarbonate). Desulfurization plants are generally of large
volume and have high operating costs. For this reason the most
economical method is to desulfurize natural gas before its use as a
feedstock or fuel.
D
CO Shift Conversion and CO2 Removal:
Process Condensate Treatment
Carbon monoxide conversion is carried out in two steps: (1) high-
temperature shift conversion, and (2) low-temperature shift
conversion. The gas from the secondary reformer is cooled down by
steam generation to about 375°C, the usual temperature for HT shift
conversion. Further on, the gas is cooled down to about 200 to 260°C
before passing to an LT shift conversion sector. There are no effluents
from CO shift conversion. After leaving the CO shift conversion unit,
the gas is cooled down again and CO2 is removed. Potassium
carbonate solution with various additives to promote absorption are
used for CO2 removal in most ammonia plants.
 Page 44
The process condensate obtained as a result of gas cooling and excess
water vapor condensation contains as main impurities ammonia,
methanol, carbon dioxide, and metals from catalysts. In older plants
the process condensate is collected and sent to a stripping column,
where the low-pressure steam is fed at the bottom (Fig. 4). The
stripped condensate is cooled and discharged as effluent outside the
ammonia plant battery limit. A mixture of steam and gases desorbed
from the process condensate is vented to the atmosphere from the top
of the stripper.
As an example, a condensate stripper for about 47 tons/h with a
temperature of 79°C is 1600 mm in diameter and has two 5400-mm-
high packing layers consisting of 2-in. metal Pall rings. The quantity
of steam used for stripping is 9072 kg/h at a pressure of 3 to 5 kg/cm2.
A total of 7382 kg/h steam and gases is vented to the atmosphere from
the top of the stripper. The condensate from the stripper bottom, at a
temperature of 116°C and a pressure of 0.7 kg/cm2, is cooled down to
38°C using recycled cooling water. The continuous boiler blowdown
is also fed to the stripper bottom to maintain heat balance equilibrium.
Steam stripping can reduce the ammonia (N) content to a nominal
level of 20 to 25 g/m3, but a nitrogen level of 40 g/m3 is more
common. This solution does not solve the problem of environmental
protection, taking into consideration the fact that liquid and gaseous
effluents are coming out of the stripper. The high-pressure stripping
systems represents an important improvement in process condensate
treatment.
Two main solutions can be adopted to solve the problem of process
condensate:
1. Treatment using ion-exchange resins, which ensures a high-quality
product so as to enable its reuse as makeup water for the cooling
towers, as process water, or as boiler feed water. Water usually flows
through a strongly acid cation exchanger and then through a weakly
basic anion-exchange filter. Finally, the water may be passed through
a mixed bed that ensures the characteristics required for its reuse as
boiler feed water [5].
2. Utilization of process condensate for natural gas saturation. The
advantages of this solution consist of the fact that the process steam
consumption is reduced and the condensate does not need to be
treated before its use. In this case, a series of minor modifications will
have to be made in the flowsheet of the syngas preparation unit. A
typical flow diagram is shown in Fig. 5. This flow diagram includes a
gasgas heat exchanger at the outlet from the desulfurizer, gas
saturation with process condensate, saturated gas preheating at a low
temperature level, usually in the convection section of

Figure 4
Process condensate stripping.
 Page 45

Figure 5
Feed gas saturation.
the primary reformer, thus making possible a decrease in flue gas
temperature prior to the discharge of flue gas to the atmosphere.
Saturation can be accomplished by process condensate injection,
directly into the gas pipe or by using a packed column as saturator. In
this last case, a recycling circuit is used, consisting of pumps, a
condensate heater situated in the convection section, and a saturation
column. As only a part of the process condensate can be used for
natural gas saturation, treatment with ion-exchange resins might be
used for the remainder.
Usually, the process condensate from the ammonia plant is mixed
with wastewater from other fertilizer plants (e.g., ammonium nitrate,
urea) and is treated for recovery of
 Page 46
pollutants as useful substances. The use of nitric acid as a regenerant
for ion-exchange resin makes it possible to obtain 20 to 28%
ammonium nitrate solutions [6].
Biological treatment can also be used in such a case. Ammonia is first
removed by a biological process by oxidizing it to nitrate
(nitrification) and then by denitrification to nitrogen gas [4].
The CO2 content in the syngas after CO-LT shift conversion is about
18 mol %, depending on the natural gas composition. Trademarks for
the main CO2 removal processes include Activated MDEA, Benfield,
Catacarb, Carsol, Giammarco-Vetrocoke, Rectisol, Selexol, etc. Most
of those processes are based on carbon dioxide absorption under
pressure in various absorbants, followed by desorption through a
pressure decrease and temperature increase. After desorbtion, the
carbon dioxide is cooled and used for the manufacture of urea or dry
ice, or for other applications. The quantity of CO2 obtained for each
ton of ammonia is 1.3 to 1.4 tons, as a function of natural gas
composition.
The pollution problems raised in carbon dioxideremoving plants are
connected to the nature of the absorbants and to the emission of
relatively large amounts of CO2 to the atmosphere. The liquid losses
from a CO2 removal plant can be minimized by the solutions
indicated in the project (drainage system, storage and reconcentration
of waste solutions), as well as by proper and careful operation of the
plant. Special care should be taken in the Giammarco-Vetrocoke
process because it uses as absorbant a K2CO3 solution activated with
As2O3. Careless operation of such plants may lead to soil-or deep-
water contamination or may cause severe intoxication of operating
personnel.
In most cases, carbon dioxide is only partially utilized in urea or dry
ice fabrication; a relatively large amount is thus discharged
continuously to the atmosphere, the results being the greenhouse
effect and acid rain.
E
Methanation, Compression, and NH3 Synthesis:
Purge Gas Treatment
The gas leaving the CO2 absorption unit goes on to the methanation
section, where it is compressed and fed to the ammonia synthesis
loop, which includes the refrigeration system. Syngas compression is
generally carried out using a centrifugal compressor driven by a steam
turbine. Pressure oil is used for shaft sealing. At high inner pressure
the syngas dissolves in oil and is released when the oil pressure
decreases. The gases released, which contain primarily H2 and N2, are
usually vented but could also be used as fuel for the primary reformer.
The main effluent resulting from the synthesis loop is the purge gas.
To maintain a constant level of inert gas concentration in the synthesis
loop, a quantity of recycled gas is discharged from the system
continuously. The purge gas flow rate and composition vary,
depending on the operating characteristics of the ammonia synthesis
loop and on the way it is operated. The composition of the purge gas
ranges as follows: H2, 55 to 65%; N2, 18 to 23%; Ar, 3 to 5%; CH4, 8
to 18%; NH3, 1 to 4%. The purge gas also contains rare gases, such as
krypton, xenon, neon, and helium.
In older ammonia plants, the purge gas is used as supplemental fuel
for the primary reformer, although is a solution of low efficiency, as
large amounts of hydrogen and rare gases are thus lost. Newer
ammonia plants retain the purge gas components, resulting in
improved natural gas feedstock utilization. The separation processes
for purge gas are also varied and include primarily partial
condensation at low temperature, cryogenic distillation, adsorption,
and permeation [9].
 Page 47
1
Partial Condensation at Low Temperature and Cryogenic Distillation
Before being fed to the separation plant, the ammonia is removed
from the purge gas and the gas is dried on molecular sieves. Water
absorption is used primarily for ammonia separation. The aqua
ammonia obtained, which has a nitrogen content of about 20%, can be
used as fertilizer. Another alternative is pressure stripping of ammonia
from the aqua ammonia, its cooling, then liquefaction, and delivery as
an end product. Separation of ammonia from purge gas can also be
achieved using a small freon refrigeration unit.
The easiest way to recover hydrogen from purge gas is by partial
condensation, where the refrigeration is produced by the
JouleThomson effect, by recovery of the heat of vaporization of
components less volatile than hydrogen, and an ammonia refrigeration
loop. Thus a stream containing about 90% hydrogen is obtained, the
rest being CH4 and Ar, as well as a stream with a low hydrogen
content but rich in CH4 that can be used as fuel (Fig. 6).
An improved system with more efficient thermodynamical behavior is
one using an auxiliary nitrogen refrigeration loop, including a
turboexpander for the adiabatic expansion of nitrogen. The more
sophisticated plants use a turboexpander, the coolant being the purge
gas itself.
A more advanced treatment of the purge gas is that having argon
recovery as its goal. In this case the plant must be equipped with
distillation columns that make it possible to obtain a separate stream
containing H2, N2, CH4, and high-purity Ar (99.999%). In such a
plant, recovery of rare gases can be provided, including helium
recovery from a hydrogen-containing stream [13]. The characteristics
of a plant recovering cryogenic H2 and Ar from purge gas are
presented as an example in Table 1.
 Figure 6.
Flowsheet for hydrogen recovery using partial condensation
at low temperature. 1, NH3 washing column; 2, dryer; 3, ammonia
cooler; 4, process heat exchanger; 5, separator; 6, cold box.
 Page 48
Table 1 Cryogenic H2/Ar Recoverya
Purge Hydrogen Argon Tail
gas fraction fraction gas
Flow rate (N · 4000 2450 180 1300
cm/h)
Pressure (bar) 35 30 1.8 7
Temperature (°C) 2530 15 182 32
Composition
(mol %)
H2 5760 min. 92 11.5
N2 1822 7 5
A2 46.2 max. 1 99,999 1.5
CH4 912 (Ar + CH4) 82
NH3 23
aAqueous ammonia, 2.2 m3/h; electric power
consumption, 900 kWh.

2
Pressure Swing Adsorption Process
The pressure swing adsorption (PSA) process is based on the fact that
the higher the pressure, the more impurities the adsorbant can retain.
When the pressure is decreased from a high value to a low value, the
adsorbed impurities are desorbed and rejected in the tail-gas stream.
The PSA process can be applied for recovery of hydrogen from purge
gas and has the advantage that no effluents are released to the
atmosphere. The purge gas can pass through the PSA unit before or
after ammonia separation. Another advantage of the PSA process is
the fact that the product can be supplied at about the same level as the
inlet pressure, while the impurities are desorbed at a lower pressure.
A typical PSA flowsheet consists of four adsorbers. The hydrogen
recovery ratio ranges between 60 and 98%, as a function of gas
composition, operating pressure, and product requirements. The final
purity of the hydrogen recovered may exceed 99%, so that it can be
fed back to the syngas compressor suction unit.
3
Permeation
The permeation rate of gas mixture across a membrane depends on the
composition of the gas, the membrane material, and the operating
conditions. Generally, the polar gases (CO2 or H2S) permeate more
quickly across the membrane than do the nonpolar ones. Gases of
small molecular size (H2, He) permeate more rapidly than do gases of
large molecular size (CH4, NH3). A process of hydrogen separation
from the purge gas is based on the difference between the permeation
rates of hydrogen and impurities across a gas-permeable polymer
membrane.
Permeation implies two consequent mechanisms: the gaseous
component first dissolves on the membrane, then diffuses through it.
Solubility depends on the chemical composition of the membrane, and
diffusion depends on its structure.
Membrane separators can be classified in two large groups: spiral
wrap type and hollow fiber type. For industrial applications multiple
stages of membrane gas separation are used [2]. Hydrogen recovery
ranges from 85 to 95%, depending on the operating conditions, and
hydrogen purity may exceed 99% [3].
The three processes described above can also be combined [7,15,16].
For example, purge gas passage across a separation membrane yields
a high-purity hydrogen flow. The membrane nonpermeated gas is rich
in N2, CH4, and Ar and can be separated further by
 Page 49
cryogenic distillation, which leads to the reduction of energy
consumption and plant volume as a result of processed gas quantity
decrease.
Recovery of purge gas components using the processes described
above demonstrate important economic advantages for ammonia
industry, among which are [10]:
1. Hydrogen is recovered, allowing an increase in ammonia
production of about 3.2% for a determined natural gas consumption.
2. Argon and other rare gases with multiple industrial applications can
be separated.
3. A methane-rich stream is recovered and can be used as fuel.
IV
Final Considerations
1. The choice of the process for ammonia production and the choice
of feedstock have important consequences from the viewpoint of
environmental impact.
2. Although ammonia plants based on steam reforming are relatively
clean, pollution control problems should not be neglected and must be
dealt with according to company strategy and national and local
regulations.
3. Liquid and gaseous effluent emissions should be controlled and
sampled within plant boundaries in the same way that the process
streams are controlled.
4. For gaseous pollutant emissions (flue gases, CO2), the vent stack
should be high enough to ensure pollutant dispersion to safeguard the
immediate vicinity of the plant.
5. Liquid pollutant emissions (process condensate) can be treated as
such, or in a mixture with emissions from other fertilizer plants
(ammonium nitrate and urea). The most common solution in such
cases is the use of ion-exchange resins. By regeneration with nitric
acid, diluted ammonium nitrate solutions can be obtained and
concentrated further in ammonium nitrate plants.
References
1. R. L. Berglund and G. E. Snyder, Minimize waste during design,
Hydrocarbon Processing, April 1990, p. 39.
2. Chen-Hwa-Chiu, Advances in gas separation, Hydrocarbon
Processing, January 1990, p. 69.
3. Commission Economique pour l'Europe, La technologie des
membranes dans l'industrie chimique, ECE/CHEM '72, New York,
1990, pp. 6264.
4. J. M. Dave, Environmental pollution from fertilizer industry,
Proceedings No. R&D (Fertilizer Association of India), 3 (FAI Group
Discussion, Ammoniacal Nitrogen Effluent in the Fertilizer Industry),
1982, pp. 107111.
5. Economic Commission for Europe, Rational use of water and its
treatment in the chemical industry, ECE/CHEM '78, New York, 1991,
pp. 116117.
6. J. Jaqust, Z. Dzendo, and N. Arion, Fertilizer Plant Waste Water
Treatment by Ion Exchange, British Sulphur Corporation Ltd.,
London, 1986.
7. Verfahren zur Gewinnung eines heliumreichen heliumwasserstoff
Gemisches, Offenlegungsschrift DE 38 13976A1, 1988.
8. V. Pachayappan, Indian fertilizer industry: environmental
considerations, Environmental Symposium, 1980, pp. 143160.
9. M. Peculea, Valorificarea purjelor gazoase de la sinteza
amoniacului în vederea recuperarii

hidrogenului si separarii* argonului si heliului, Revista de Chimie,

Vol. 8, 1989.

10. G. Pohle and K. Schübel, Abproduktarme Technologie in

hochintegrierten Einstranganlagen der chemischen Industrie,
dargestellt am Beispiel der Ammoniak-produktion, Chemie Technik,
35 Jg., Heft 1, 1983.
 Page 50
11. R. H. Roberts, Design Plants for Safety, Hydrocarbon Processing,
Sept. 1989, p. 92.
12. G. S. Roy and R. R. Poricha, Ammonia nitrogen in fertilizer
industry effluents: present status, Proceedings No. R&D (Fertilizer
Association of India), 3 (FAI Group Discussion, Ammoniacal Nitrogen
Effluent in the Fertilizer Industry), 1982, pp. 97102.
13. G. Stenzel and K. Schmidt, Valorificarea economica

* a gazelor reziduale provenite din sinteza amoniacului, Revista de

Chimie, Vol 3, 1985.

14. Fertilizer Manual, United Nations Industrial Development

Organization, New York, 1985.
15. Permeation processes for separating ammonia from a gas mixture,
U.S. Patent 4,793,829, 1988.
16. Argon recovery from hydrogen depleted ammonia plant purge gas
utilizing a combination of cryogenic and noncryogenic separating
means, U.S. Patent 4,750,925, 1988.
17. World Bank Technical Paper 174, World Nitrogen Survey,
Washington, D.C., 1992, p. 37.
18. D. Breihofer, A. Mieleny, and O. Renoty, Emission Control of So2,
NO x and VOC at Stationary Sources in the Federal Republic of
Germany, Institute of Industrial Production (HP), University of
Karlsruhe (TH), Karlsruhe, Germany, 1991.
 Page 51

5
Ammonia Pollution Control Measures
Thomas A. Czuppon, Hassan M. Gomaa, and Stan A. Knez
The M.W. Kellogg Company
Houston, Texas
I
Introduction
The production of synthetic ammonia by catalytic steam reforming of
natural gas is a relatively clean process and presents no unique
environmental problems. The environmental impact of a modern
ammonia plant on air, water, and soil is assessed in this chapter by
examining each of the ammonia process steps: desulfurization,
reforming, shift conversion, carbon dioxide removal, final
purification, ammonia synthesis, and refrigeration. The sources of
pollutants are identified and matched with cost-effective solutions for
minimization/elimination by using the best available pollution control
measure. Process design considerations are highlighted along with
future developments.
The future profitability of the nitrogen fertilizer industry is highly
dependent on energy cost and capital investment. Energy and capital
cost penalties associated with pollution control systems must therefore
be minimized as far as practicable. The first step is to reduce the
sulfur concentration in the ammonia plant feedstock to less than 0.1
ppmv to prevent poisoning the reformer catalyst. Once desulfurized,
the feedstock is partially reacted with steam in a primary reformer to
produce hydrogen (H2) and carbon monoxide (CO). To provide a
source of nitrogen, air is injected in the secondary reforming stage.
The CO in the reformed gas is reacted with steam in a two-stage shift
converter to produce carbon dioxide (CO2) and hydrogen. The shift
converter effluent is cooled and most of the unreacted steam is
condensed and separated as process condensate.
The raw synthesis gas is then scrubbed to remove CO2. The residual
CO and CO2 are catalytically reacted with H2 to form methane and
water vapor. The remaining trace amounts of CO2 and water of
saturation are finally removed by molecular sieves. The purified
synthesis gas, containing hydrogen, nitrogen, and small amounts of
methane, argon and helium, is compressed and enters the synthesis
loop. The synthesis of ammonia occurs over an iron catalyst. The
ammonia converter effluent is chilled indirectly with
 Page 52
refrigerant (ammonia) to condense out the liquid ammonia product.
Only a fraction of the synthesis gas feed is converted into ammonia;
therefore, a large amount of unreacted gas is recycled. Inerts that tend
to accumulate in the synthesis loop are purged. Hydrogen contained in
the purge gas is recovered and the balance of the purge stream is sent
to the fuel system. Figure 1 is a typical ammonia plant flow diagram,
showing the pollutants associated with each process step [1]. Table 1
lists emission factors for a typical ammonia plant.
II
Emission Sources and Control
A
Feedstock Desulfurization
Feedstocks to an ammonia plant based on steam reforming can range
from natural gas to straight-run naphthas. Being the most economical
feedstock, natural gas is the most widely used, accounting for more
than 80% of the world ammonia production. Contaminants normally
encountered in these feedstocks are sulfur, chloride, and
organometallic compounds, which need to be removed to less than 0.1
ppmv to avoid poisoning the various catalysts in the ammonia plant.
Sulfur compounds are mostly hydrogen sulfide (H2S), mercaptans (R
¢SH), thiophenes (R¢S), and carbonyl sulfide (COS).
1
Activated Carbon
Sulfur removal from natural gas by adsorption at ambient temperature
on carbon, activated with cuprous oxide, has been widely used.
Carbon physically adsorbs sulfur compounds to its surface and the
cuprous oxide reacts with hydrogen sulfide according to the equation
The activated carbon is typically regenerated every 30 days by passing
steam through the bed at a temperature of 230°C (450°F) for 8 to 10 h
while air is injected. Oxygen in the air reacts with the metal sulfide to
form the metal oxide and sulfur dioxide:

During regeneration of an activated carbon bed, SO2, CO,

hydrocarbons, and steam are vented to the atmosphere. The carbon
also adsorbs heavier hydrocarbons, which are removed and vented to
the atmosphere along with the sulfur compounds in the regeneration
gas. Hydrocarbon emissions may be high during regeneration,
reaching 3.6 g/kg of ammonia product.
Activated carbon has limited application because not all sulfur
compounds can be removed and the ill-smelling effluent can cause
discomfort to people nearby. Air pollution control agencies are now
paying more attention to emissions from such operations. Organic
sulfur compounds, hydrocarbons, and CO contained in regeneration
gases can be controlled by incineration. Alternatively, the emissions
from activated carbon regeneration can be eliminated more efficiently
by substituting ZnO for the activated-carbon desulfurization beds. The
ZnO not only eliminates emissions, but saves energy and is a more
effective desulfurization technique.
2
Iron Oxide
In this process, H2S present in the feed gas is reacted with iron oxide
to form iron sulfide. The iron sulfide can be regenerated either in situ
or outside in air. After oxidation of the
 Figure 1
Process flow diagram for a typical ammonia plant with pollutants associated with each proces
 Page 54
Table 1 Emission Factors for Typical 1000-MTPD Ammonia
Plant
Emission point Emission kg/metric lb/tona
tona
Desulfurization unit Total sulfurb,c 0.0096 0.019
regenerationd CO 6.9 13.8
Nonmethane 3.6 7.2
VOCe
Primary reformer flue NOx 2.7 5.4
gas
Natural gas fuel SOx 0.0024 0.0048
CO 0.068 0.136
Particulates 0.072 0.144
Methane 0.0063 0.0126
Nonmethane 0.0061 0.0122
VOC
Distillate fuel oil NOx 2.7 5.4
SOx 1.3 2.6
CO 0.12 0.24
Particulates 0.45 0.90
Methane 0.03 0.06
Nonmethane 0.19 0.38
VOC
CO2 solvent Ammonia 1.0 2.0
regeneration CO 1.0 2.0
CO2 1220 2440
Nonmethane 0.52 1.04
VOC
Condensate steam Ammonia 1.1 2.2
stripperf CO2 3.4 6.8
Nonmethane 0.6 1.2
VOC
Source: Refs. 2 and 3.
aEmission factors are expressed in weight of emissions per
unit weight of ammonia product.

bWorst-case assumptionthat all adsorbed sulfur is emitted

during regeneration.
cNormalized to a 24-h emission factor.
dIntermittent source from carbon beds average 10 h once
every 30 days.
eVolatile organic compounds (VOCs).
fIn new plants, high-pressure condensate stripper recycles
condensate pollutants.

sulfur in situ, the SO2 is released to the atmosphere with some diluent
gas, such as the primary reformer flue gas. Regeneration in air
releases the SO2 slowly into the atmosphere. Iron oxide has been used
only in special circumstance and is not recommended today, due to the
fact that sulfur levels acceptable for the ammonia plant are usually not
reached and a second-stage desulfurizer must be provided. Disposing
of the spent iron oxide can also pose a problem because of the large
volumes involved.
3
Zinc Oxide
New ammonia plants use zinc oxide (ZnO) beds to desulfurize
entirely or provide the final step for the feedstock and operate under
the following reaction:

The ZnO catalyst can absorb up to 20 wt % sulfur and is replaced

rather than regenerated. Elminating the regeneration step lowers the
energy consumption and minimizes the environmental impact.
 Page 55
4
Hydrotreating Organic Sulfur Compounds
Many organic sulfur compounds and carbonyl sulfide (COS) are not
removed by zinc oxide alone and must be hydrotreated to H2S.
Hydrotreating involves passing preheated feed gas over
cobaltmolybdenum (CoMo) or nickelmolybdenum (NiMo) catalysts in
the presence of hydrogen to convert the sulfur to H2S, typically by the
following reactions:

If large amounts of COS are present in the feed gas, the COS is
hydrolyzed in a separate vessel according to the reaction

The H2S generated by hydrotreating or hydrolysis is subsequently

removed by zinc oxide. In the case of naphtha feedstock having
appreciable amounts of sulfur, the H2S formed through hydrotreating
is stripped from the liquid hydrocarbon feed and the resultant stream
of H2S and hydrogen, with some hydrocarbons, is incinerated in the
reformer furnace.
5
Sulfur Removal by Organic or Inorganic Solvents
In case the natural gas feedstock contains a high level of sulfur
(typically greater than 150 ppmv), it is usually more advantageous to
remove the sulfur using a solvent followed by a final cleanup with
ZnO. Solvents such as monoethanolamine (MEA), diethanolamine
(DEA), and hot potassium carbonate solutions have been widely used
in the industry. The solution is regenerated by steam, and the sulfur
compounds released, mostly H2S, are sent to the primary reformer for
incineration. A very high sulfur level in the feed gas would require
elemental sulfur recovery by a Claus-type process.
6
Chlorides
Chlorine and other halogens are also poisons to catalysts and must be
removed to a level similar to that of sulfur. Chlorides may enter the
pipelines when the pipeline is exposed to the atmosphere, or they may
be left behind after hydrostatic testing. Chlorides are removed by
catalysts and solvents used for sulfur removal.
7
Metals
Metals or organometallic compounds such as arsenic, copper, lead,
and vanadium are also poisons to the primary reforming catalysts.
Metals are present in trace quantities and are usually removed during
normal desulfurization processes. If present, mercury is removed by
adsorption on activated carbon at ambient temperature.
B
Steam Reforming
Steam reforming is a catalytic process for the conversion of light
hydrocarbons and steam into hydrogen and carbon oxides. Most of the
side reactions are retarded by the use of excess steam. First, the
hydrocarbon feed is mixed with steam and passed over catalyst at a
high temperature. The limiting reaction can be represented by the
following equations for methane and the CO shift reaction:
 Page 56
The net reaction is endothermic and is driven toward equilibrium in
catalyst-packed tubes at high temperatures.
Next, the partially reformed gas is passed to the secondary reformer,
where the final conversion of methane takes place. Nitrogen for
ammonia synthesis is provided by the injection of air into the
secondary reformer. The oxygen in the air provides the necessary heat
through combustion to convert the partially reformed gas. The general
arrangement of The M. W. Kellogg (Kellogg) reforming section is
illustrated in Fig. 2. The reforming reactions take place in the single-
cell top-fired radiant box, which contains multiple rows of vertically
suspended catalyst-filled tubes. The hot flue gas from the radiant
section is collected in flue gas tunnels located on the furnace floor.
Here tunnel burners are used to maintain the flue gas temperature
entering the convection section. In the convection section,
hydrocarbon feed and process air are preheated and the balance of the
heat is recovered for steam generation and sometimes for combustion
air preheat. A fired auxiliary boiler is offset from the convection
section with flue gas from the boiler entering the transition duct.
Superheat burners are located at the top of the hot leg to provide
additional heat and control of the superheated steam temperature.
1
Emissions from Primary Reformer
The primary reformer is a conventional combustion emission source
for oxides of nitrogen (NOx), carbon monoxide (CO), and particulate
matter (PM).
2
NOx:
Source, Generation, Causes, and Contributing Factors
NOx is produced in all combustion processes using air. NOx is
produced primarily as nitric oxide (NO) in the hottest regions of the
combustion zone. Some nitrogen dioxide (NO2) is also formed, but its
concentration amounts to 5 wt % or less of the total NOx [4]. NOx in
combustion flue gas is derived by two mechanisms: thermal NOx and
fuel NOx. Thermal NOx is formed when the temperature of the
combustion zone is high enough to cause the nitrogen in the air to
react with oxygen to form NOx. Fuel NOx is derived from the
oxidation of free or bound nitrogen in the fuel.
As shown in Table 2, the fuel gas used in ammonia plants is normally
a mixture of pipeline natural gas and purge gas from the synthesis
loop. Note that both free and bound nitrogen are present (as N2 and
NH3), and therefore it is expected that fuel NOx will be formed. The
hydrogen content is also significant because hydrogen raises the flame
temperature and thereby increases the thermal NOx generation rate.
The degree of combustion air preheat has an effect on NOx emissions.
NOx emissions from a furnace with a combustion air preheat
temperature of 240°C (400°F) were about 1.4 times higher than they
were from the same heater with ambient combustion air [5].
There is an inverse relationship between NOx and CO formation,
which plays an important role in emission control. In complete
combustion, the carbon atoms in a hydrocarbon fuel are oxidized to
carbon dioxide as in the simple equation

However, real-world combustion processes do not bring the three T's

(time, temperature, and turbulence) together perfectly to achieve
complete combustion, and some amount of CO generation is
inevitable. Generally speaking, the higher the peak combustion
temperature, the lower the CO generation. Unfortunately, the trend is
just the reverse for NOx generation; the higher the combustion
temperature, the greater the NOx generation
 Page 57

Figure 2
Typical reformer section: ammonia production.
 Page 58
Table 2 Typical Composition of Fuel Gas
Mixed fuel gas Mixed fuel gas
before water after water
Component scrubbing scrubbing
CH4 66.11 66.24
C2H6 1.58 1.58
C3H8 0.33 0.33
N2 14.68 14.71
CO2 0.12 0.12
H2 14.76 14.79
NH3 0.20 <0.01
He 0.60 0.60
Ar 1.62 1.62
100.00 100.00

[6]. Therefore, emission control for industrial combustion sources

must be a compromise between NOx and CO control.
3
Candidate NOx Control Technologies
Several potential control technologies are available for reducing the
NOx emissions from the primary reformer. The candidate technologies
are:
1. Purge gas (fuel) treatment
2. Combustion equipment design
a. Steam injection
b. Staged-air low-NOx burners
c. Staged-fuel low-NOx burners
3. Flue gas recirculation (FGR)
4. Selective catalytic reduction (SCR)
5. Selective noncatalytic reduction (SNCR)
a. Ammonia injection (Thermal DeNOx)
6. Urea injection
4
Purge Gas Treatment
The fuel gas composition shown in Table 2 is relatively clean in terms
of its NOx-generating potential, except for the nitrogen, hydrogen, and
ammonia. Of these three NOx precursors, ammonia is the most serious
problem. Burner vendors report that the 0.20 vol % of ammonia in the
fuel gas will contribute over 80 ppmv to the NOx concentration in the
flue gas.
The sources of the ammonia in the fuel gas are the purge streams from
the ammonia synthesis loop. A water scrubber system is usually
provided in new plants for removing ammonia from these purge
streams. A scrubber system consists of two packed scrubber columns,
a steam stripper, and a water circulating loop, as shown in Fig. 3.
The high-pressure purge gas from the synthesis loop is sent to the
high-pressure (HP) purge gas scrubber. Low-pressure flashed gases
from the ammonia letdown drum and refrigerant receiver are sent to
the low-pressure (LP) purge gas scrubber. High-pressure scrubber
overhead gases are directed to the hydrogen recovery unit (HRU).
Selection of a hydrogen recovery unit (cryogenic, membrane-based, or
pressure-swing adsorption) is
 Page 59

Figure 3
Water scrubber system.
plant specific. LP scrubber overhead gases, as well as residual
offgases, are recycled as fuel to the primary reformer.
The aqueous solution leaving the scrubbers has an ammonia
concentration of about 14 wt %. Liquid ammonia reflux to the top of
the stripper is provided from the refrigeration system. The ammonia
vapor from the top of the column is condensed in the refrigerant
condenser. Distillation heat is provided by condensing medium-
pressure steam.
Since the solubility of ammonia in water is very high, the water
scrubber will remove virtually all of the ammonia in the purge gases.
However, some ammonia slip is inevitable. The scrubber system is
typically designed for 200 ppmv ammonia slip in the scrubber
overhead, which translates to 60 ppmv of residual ammonia in the
mixed fuel gas.
5
Burners
The primary reformer operates at relatively low excess combustion air
levels (10% design, 5 to 20% in practice), and it is not practical to
reduce it significantly lower. However, combustion control by burner
design is a NOx control technique that has been demonstrated
successfully on utility boilers and other stationary combustion
sources.
So-called ''low-NOx" burners have been demonstrated successfully
and have earned industry acceptance. Favored low-NOx techniques
include staging the combustion air and staging the fuel. Staged fuel is
preferred over staged air as the low NOx technique for gas fuel
burners most widely adopted. Staged air is generally the accepted
approach for liquid fuel burners. By minimizing the peak flame
temperature and controlling the nitrogenoxygen contact in the hottest
zones, NOx formation can be reduced significantly.
Early low-NOx burners were designed with staged combustion air to
produce a primary combustion zone deficient in oxygen. This creates
a primary combustion zone environment that is rich in reducing
agents. The reduced partial pressure of oxygen tends to limit NOx
formation. A major disadvantage of the staged-air burner is a loose
undefined flame pattern, which is generally unacceptable for a high-
heat-flux process heater. Staged-air
 Page 60
burners require considerably more space for mounting. Other options
to produce staged-air combustion (overfire air, burners out of service,
biased firing) have been tried in utility boilers [7] but are generally
unsuited to the practical operation of the primary reformer.
The most common low-NOx burner on the market today is the staged-
fuel burner shown in Fig. 4 [8]. The technique of staged-fuel
combustion begins with combustion of a portion of the fuel with high
excess air. The remainder of the fuel is injected in the second stage of
the reaction and combustion is completed at a level of low excess air.
Burners designed for staged-fuel combustion typically reduce NOx
emissions by about 25 to 40% compared to conventional burners, and
even greater reductions are promised by the newer second-generation
low-NOx burners.
During the first stage, high excess air allows combustion to occur at a
low flame temperature that inhibits thermal NOx formation. As the
combustion reaction goes to completion in the first zone, additional
fuel is injected. The second combustion stage begins with a reduced
partial pressure of oxygen, which tends to limit the formation of NOx.
Staged-fuel burners can provide a flame pattern and shape that is
compatible with
 Figure 4
Staged-fuel burner.
 Page 61
the primary furnace requirements. Staged-fuel burners have a history
of successful installations in process heaters. Therefore, this
technology has a proven track record and can be considered
practicable.
The inherent drawback of any technique used to control NOx
emissions by lowering flame temperature is that CO emissions will
increase. Using staged-fuel burners, the CO concentration in the flue
gas is expected to be in the range 50 to 100 ppmv, values considered
acceptable with combustion equipment designed for NOx reduction.
6
Flue Gas Recirculation
Flue gas recirculation (FGR) involves a recirculated flue gas stream
(normally about 10% of the total flow) for injection into the
combustion zone, or combustion air to reduce the flame temperature
and oxygen concentration. FGR decreases thermal efficiency, but
recent research shows that the NOx reduction potential can be
impressive. As with other NOx reduction technologies, FGR lowers
the flame temperature and lengthens the combustion zone, thereby
tending to increase the amount of CO produced. FGR has not been
fitted to a top-fired primary reformer and insufficient data exist to
determine its effectiveness. However, FGR technology appears to be
the next step to lower NOx emissions incrementally, having already
found positive application in ethylene plants.
7
Selective Catalytic Reduction
Selective catalytic reduction (SCR) processes are based on the
reaction of NOx with ammonia in the presence of a catalyst, forming
nitrogen and water vapor:
The SCR process requires a reactor, a catalyst, and an ammonia
storage and injection system. For SCR to be effective, the flue gas
temperatures must be maintained within a narrow window, which
depends upon the catalyst [typically, 315 to 400°C (600 to 750°F)].
SCR catalysts are commercially available in two types: base metal
(typically, vanadium oxide or titanium oxide) and precious metal
(typically, platinumrhodium). The base metal catalysts are bonded to a
ceramic substrate. The precious metal catalysts are bonded to a
stainless steel substrate. Base metal catalyst SCR systems are now
operating in Japan, Europe, and the United States.
Most of the demonstration work in SCR NOx control can be found in
the utility industry, and most of that activity is in Japan and Europe.
The utility industry has cited the following concerns regarding the
application of SCR to boilers: difficulty in matching the NH3 injection
pattern to the nonuniform flow of NOx in the ductwork, inability to
optimize NH3 injection rate by feedback control, unavailability of a
reliable NH3 monitor, catalyst deactivation by heavy metal poisoning,
process control under load swings, and spent catalyst disposal [9].
In addition to the concerns voiced by the utility industry, the obvious
difficulties associated with fitting a SCR unit to the reformer furnace
include complete redesign and reconstruction of convection section,
environmental impact of ammonia leakage, large capital investment
requirement, increased operating costs, and operational risks. An
example of a potential operating problem is proper control of the flue
gas temperature entering the SCR unit. For effective NOx reduction to
occur, the SCR must operate within a narrow temperature window.
Temperature excursions outside the design range can result
 Page 62
in ammonia slip to the environment, as well as thermal shock and
poisoning of the catalyst, which may require replacing the catalyst.
Maintaining operation of the furnace within the SCR's temperature
window will be difficult outside the normal design operation. During
startup or turndown operation, the firing rates change, resulting in a
different flue gas temperature profile through the entire furnace.
Ammonia breakthrough is an undesirable side effect of SCR. In
practice, some amount of the injected NH3 will always escape
reaction with the NOx. A small amount of ammonia slip (about 1 to 15
ppmv, depending on the effectiveness of the SCR unit) will be emitted
to the atmosphere.
8
Selective Noncatalytic Reduction
There are at least two competing technologies for reducing NOx
emissions by postcombustion (selective) noncatalytic reduction
(SNCR) methods: ammonia injection and urea injection. Both
technologies use a reducing agent (ammonia or urea) to react with
NOx to form nitrogen and water. There is a wide disparity in the
literature on the NOx reduction achievable with these systems. The
principal areas of concern are ammonia breakthrough, side reactions,
safety, reliability, performance during load changes, and corrosion and
fouling of heat transfer surfaces.
The ammonia injection SNCR method was developed by Exxon and
carries the trade name Thermal DeNOx. It uses ammonia to reduce
NOx emissions based on the same principles and chemical reactions as
those related to the SCR process, except that no catalyst is used. The
absence of catalyst necessitates temperatures of 930 to 980°C (1700 to
1800°F) for the SNCR reaction to take place. The reaction
temperature can be lowered to 700°C (1300°F) by the injection of
hydrogen along with the ammonia [10].
In addition to the desired NOx reduction reactions, side reactions take
place at temperatures greater than the 980°C (1800°F) temperature at
which NO is produced. Above 1090°C (2000°F), NOx concentrations
in the flue gas actually increase by the reaction

For these reasons, the injection of the ammonia into the furnace
precisely at the point of desired temperature is critical. This goal is
difficult to achieve when the furnace load changes and the firing rates
of the burners vary to optimize steam production. A properly designed
ammonia injection system is essential to achieving the desired results.
The ammonia injection SNCR process is further constrained by the
fact that the absence of a catalyst slows the NOx/NH3 reaction. This
extends the required retention time and narrows the temperature
window for the reaction. These constraints require special design
considerations of the convection section to accommodate an area
where the ammonia can be injected at the proper flue gas temperature
and with the required residence time. Ammonia breakthrough is
another undesirable side effect of the SNCR process. In practice, some
amount of the injected NH3 will always escape reaction with the NOx.
A serious concern with ammonia injection is the long-term effect of
NH3 on stainless steel at the high temperatures required for NOx
reduction to occur. The injection of NH3 into the flue gas stream can
leave a nitriding film on the convection tubes. At metal temperatures
above 1000°F, atomic nitrogen bonds with the chromium in the
stainless steel to form chromium nitrides on the metal surface, thereby
contributing to metal loss.
 Page 63
The potential fouling of the heat transfer surfaces will significantly
alter the flue gas temperature profile, which will degrade the overall
performance of the reformer.
A competing SNCR method uses urea as the reducing agent.
NOxOUT is the trade name for the postcombustion urea injection
technology offered by Nalco Fuel Tech. In the urea injection SNCR
process, coated urea is injected into the combustion gas path. In the
ensuing reaction, molecules of NO are converted to diatomic nitrogen
and carbon dioxide as follows:

The chemical reaction above indicates that 1 mol of urea reacts with 2
mol of NO. However, greater than stoichiometric quantities of urea
can be injected to improve NOx reduction and to speed the reaction
kinetics. This can result in some ammonia slippage and a slight
increase in CO2; both are generated as by-products from incomplete
thermal decomposition of the excess urea.
The NOxOUT process works best at temperatures above 800°C
(1472°F) and below 1150°C (2100°F). Unfortunately, this temperature
range occurs near the bottom of the convection section of typical
reformer furnaces and in the middle of one of the coils. While the
convection coils could be redesigned to accommodate a system of
injectors at this point, the flue gas temperature drops very fast at this
point in the convection sectionpossibly too fast for the reaction to take
place. As with ammonia injection, this SNCR process is a NOx
reduction technology that looks very promising. However, there are
technical problems that must be solved before it can be applied to
reformer furnaces.
9
Candidate CO Control Technologies:
Catalytic Oxidation
The CO abatement unit uses a catalyst with a platinum group metals
base to react the CO at high temperature with the oxygen present in
the flue gas to produce carbon dioxide. The flue gas temperature range
required for satisfactory operation is not as critical as with the SCR
process. Temperatures in the range 260 to 650°C (500 to 1200°F) are
acceptable depending on the catalyst substrate. Catalyst life is
expected to range from 3 to 5 years. Additional draft will be required
to drive the flue gas through the CO oxidation unit. The increase in
flue gas pressure drop will require a larger ID fan and driver.
10
Combustion Control
As explained previously, combustion equipment designed to minimize
the formation of NOx will produce increased levels of CO. The
opposite is also true: burners designed to produce essentially no CO
will produce the highest levels of NOx. Therefore, only limited
improvement is available using this strategy. CO emissions can be
minimized through the use of excess air trim of the combustion
control system. The maximum thermal efficiency for the reforming
furnace is achieved at about 2% excess O2 in the flue gas. Continuous
O2 and CO measurements can be used to operate the furnace at peak
efficiency and low CO emission rates [11].
C
Shift Conversion
In the shift conversion step, carbon monoxide reacts with steam to
form equivalent amounts of hydrogen and carbon dioxide. Upon
cooling of the effluent gas, most of the unreacted steam is condensed
and separated as process condensate. Modern ammonia plants utilize a
two-step, in-series shifting, carried out at high and then low
temperatures to increase conversion efficiency. Use of the dual-shift
conversion system lowers overall
 Page 64
plant steam requirements, and the lower CO leakage results in
reduction in plant feed requirements due to more complete conversion
of CO to hydrogen. Under normal operating conditions there are no
emissions from the shift converters.
1
Process Condensate Stripping
Process condensate resulting from unreacted process steam is
recovered and reused in the ammonia plant, following removal of
dissolved gases and shift reaction by-products. The condensate is
taken before the CO2 removal step and contains about 1000 ppmw
ammonia, 3000 ppmw carbon dioxide, and 1000 ppmw methanol and
higher alcohols. Before reuse, the condensate is steam stripped to less
than 25 ppmw NH3, almost nil CO2, and to a combined methanol and
higher alcohol content to less than 50 ppmw. The stripped condensate
is then sent to a mixed-bed polisher prior to reuse as boiler feed water
(BFW) or it is recycled to the cation unit in the demineralizer plant.
Traditionally, low-pressure steam was used to strip ammonia and
methanol from the condensate. The stripper overheads were vented to
the atmosphere directly or diluted with the primary refomer flue gas
[12]. Modern ammonia plant design uses a high-pressure steam
stripping column (see Fig. 5) similar in design to the low-pressure
stripper system except that it operates at reforming steam pressure
level. A portion of the high-pressure process steam is used as the
stripping medium in a packed column.
Essentially all the impurities are stripped out and the overhead vapor
is recombined with the balance of the process steam and sent to the
primary reformer. The contaminants removed from the process
condensate are recycled, thereby not only solving an emission
problem but reducing feedstock usage. The high-pressure condensate
stripper has the following advantages:
1. Better stripped condensate quality
2. Elimination of ammonia/methanol vent emission

Figure 5
Simplified flow diagram of high-pressure condensate stripper.
 Page 65
3. More Efficient (Energy) System
4. Retrofit possible
A recent innovation, the Kellogg Integrated Condensate System
(KICS), integrates the treatment and recycle of the contaminants in
condensate streams from the ammonia and urea plants, resulting in a
very cost-effective condensate stripping system for existing and new
plants. A variation of high-pressure stripping is the use of a feed gas
saturator. The saturator coil is typically located in the flue gas duct of
the primary reformer, where 25 to 100% of the plant condensate is
vaporized into the reformer feed gas.
D
Carbon Dioxide Removal
Carbon dioxide removal in ammonia plants is usually accomplished
by organic or inorganic solvents with suitable activators and corrosion
inhibitors. In a few circumstances, CO2 is removed by pressure swing
adsorption (PSA). Typical solvents used in CO2 removal systems are
given in Ref. 13.
CO2 removed is sometimes vented to the atmosphere, but in many
instances it is recovered for the production of urea and dry ice. Urea is
the primary use of carbon dioxide, and in the case of a natural gas
feed, all of the CO2 is consumed by the urea plant. Typically, 1.3 tons
of CO2/ton of NH3 is produced in a natural gasbased ammonia plant.
CO2 vented to the atmosphere usually contains water vapor, dissolved
gases from the absorber (H2, N2, CH4, CO, Ar), traces of
hydrocarbons, and traces of solvent. Water wash trays in the top of the
stripper and double condensation of the overhead help to minimize the
amount of entrained solvent. Good control systems to avoid upsets
also go a long way in avoiding solvent carryover. In some cases a
slipstream of the solvent is thermally treated to decompose
degradation products and to recover solvent. For conventional MEA
systems, the solvent reclaimer contents is neutralized with caustic
before disposal. Waste may be burned in an approved incinerator
equipped with an afterburner and a scrubber (to control NOx
emissions). Most of the old arsenic-promoted hot carbonate systems
have been converted to Benfield, Catacarb, or similar carbonate-based
systems.
The solvents themselves can be a nuisance and pose environmental
problems, if released, contaminating ground water or nearby streams
and rivers, adversely affecting vegetation growth and harming fish
life. Therefore, the CO2 removal section of the ammonia plant is well
protected from accidental spills and leaks. In case they do occur, the
paved and curbed ground surface allows collection of the solvent into
a sump located underground. From there the released solvent is
recovered after filtration. The sump also serves to make up the solvent
from fresh chemicals. CO2 removal systems can be made
environmentally acceptable by properly specifying the
instrumentation, water wash of the overhead vapors, efficient use of
demisters, routine control of the solvent physical properties and
chemical composition, and care in startup and emergency shutdown
situations.
Industry has not yet been required to install emission control devices
on carbon dioxide vent streams. There is a growing awareness of
carbon dioxide generation and emission to the atmosphere as a
contributing factor to potential global warming. Any attempt to reduce
significantly the CO2 vented to the atmosphere would require a major
shift in the nitrogen fixation process. The most effective way to
reduce CO2 production is by using natural gas raw material instead of
heavy hydrocarbons and coal. Converting the ammonia to urea
essentially eliminates CO2 emissions at the plant site.
 Page 66
E
Final Purification
Prior to introduction to the ammonia synthesis loop, oxygen-
containing compounds (CO, CO2, H2O) that poison the ammonia
synthesis catalyst must be effectively removed or converted to inert
species. In most modern ammonia plants this is accomplished
efficiently in a two-step purification process: methanation followed by
molecular sieve drying. During normal operation, no emissions are
associated with the final purification step.
F
Ammonia Synthesis and Refrigeration
Ammonia is made from synthesis gas, containing hydrogen and
nitrogen at approximately a 3:1 ratio, with some inerts (Ar, CH4, He).
The reaction is carried out over a multipromoted iron catalyst.
Ammonia synthesis occurs in the vapor state at elevated pressures. To
facilitate the storage and transportation of ammonia, product is
recovered as a liquid by condensation. Indirect chilling with
refrigeration (ammonia) is used to condense out the liquid ammonia
product. Ammonia vapors are generated in the refrigeration cycle
during flashing and letdown operations. These are collected in a low-
pressure header and sent to a low-pressure ammonia scrubbing
column for cleanup and reclamation as product.
The main potential emission source from the synthesis loop is the loop
purge. Typically, 10% of the recycle flow, containing hydrogen,
nitrogen, ammonia, and inerts, is purged from the synthesis loop, to
prevent the buildup of inerts. Historically, the loop purge was sent
directly to the plant fuel system, which contributed to plant NOx
emissions. With recent energy efficiency demands, the purge flow is
subjected to a hydrogen recovery step. A high-pressure water
scrubbing system is used to remove essentially all of the ammonia in
the purge prior to its introduction to the hydrogen recovery unit. Purge
gas treatment is described in detail in Sec. II.B.
G
Fugitive Emission Control
Fugitive emissions result from incomplete sealing of equipment at the
point of interface of process fluid with the environment. Control
techniques for equipment leaks include leak detection and repair
programs, and equipment installation or configuration.
1
Pumps
Special types of equipment or equipment configurations can be used
to eliminate or capture fugitive emissions from pump seals. There are
leakless pumps: pumps designed with no interface between the
process fluid and the environment, such as diaphragm pumps and
canned motor pumps. These pumps effectively eliminate fugitive
emissions. However, they are limited in application.
Sophisticated pump seals can also be used to capture or eliminate
fugitive emissions. Dual-seal systems, with barrier fluids at pressures
greater than process pressures or low-pressure systems vented to
control devices, may be used in some applications. Another approach
involves venting the entire seal area to a control device. Capture
efficiencies should be virtually 100% in both systems vented to
control devices. Then the overall control efficiency would be limited
by the efficiency of the control device. Another approach to reducing
(but not eliminating) fugitive emissions from pumps is leak detection
and repair programs. The efficiency of leak detection and repair
programs is dependent
 Page 67
on several factors, such as frequency of monitoring, effectiveness of
maintenance, action level, and underlying tendency to leak.
2
Valves
Fugitive emissions from valves may be controlled by special features
designed to isolate the process fluid from the environment. So-called
leakless valves, such as diaphragm valves and bellows seals, are
limited in their application. Leak detection and repair programs are
typically used to reduce fugitive emissions from valves. As indicated
above for pumps, the control efficiency of a leak detection and repair
program depends on the frequency of monitoring, the effectiveness of
maintenance, the action level, and the underlying tendency to leak.
3
Pressure Relief Valves
Fugitive emissions from pressure relief valves may be virtually
eliminated through the use of rupture disks to prevent leakage through
the seal. Fugitive emissions may also be added to gases collected in a
flare system by piping the relief valve to a flare header.
4
Open-Ended Valves
Leakage of vapors through block valves at the open ends of pipes can
be eliminated by the installation of caps, plugs, or blind flanges. The
control efficiency should be 100% as long as the plugs and caps
remain in place.
5
Compressors
Compressors are a source for emission of the compressed gas through
seals and of the lubricating and seal oils. After separation from the
seal oil, the seal gases can be vented, recycled to fuel, recovered back
into the process, or sent to a flare. Usually, a mechanical separator and
a carbon filter are provided when the seal gases are to be recovered.
Recently, dry seals have been introduced commercially for centrifugal
compressors, which greatly reduce gas emissions and eliminate the
potential for oil emissions. Lubricated reciprocating compressors are
equipped with oil traps and carbon filters in the discharge lines to
capture oil carryover. The oil is usually mixed with water and in some
cases contains toxic process gases. Therefore, they must be properly
degassed and decanted. The oil can then be put into containers and
taken to disposal sites either for incineration or for land fill. There is
increased emphasis on recycle, and there are a number of companies
that can reprocess the waste oils.
6
Sampling Connections
Fugitive emissions from sampling connections can be controlled by
returning the purged material to the process, or by sending it to a
control device. The practice of returning purged material to the
process in a closed system should achieve almost 100% control
efficiency. The control efficiency achieved by diverting the collected
purge material to a control device depends on the efficiency of the
device.
7
Flanges
Leak detection and repair programs have been used to reduce fugitive
emissions from flanges. The effectiveness of this program depends on
the frequency of monitoring and maintenance.
 Page 68
H
Catalyst Disposal
The manufacture of ammonia ultimately relies on a series of
catalytically enhanced chemical reactions. The typical 1000-MTPD
ammonia has at least eight unit operations that make use of fixed-bed
catlysts, with an overall catalyst volume of approximately 310 m3.
The catalyst operations vary in useful economic life from 2 to over 10
years, depending on service. Historically, all of the catalysts were
disposed of in sanitary landfills, since they are basically inert and pose
no environmental or health problems. Today, with stricter regulation
of the nation's landfills, under the Resource Conservation and
Recovery Act (RCRA), greater attention is being given to recycling of
catalyst for recovery of the main metal components. Today, numerous,
cost-effective processes exist to reclaim valuable metal components
from spent catalyst [14]. Complete separation of the spent catalyst
into its component parts, for reuse in the industry, leaves no
environmental liability. Table 3 outlines the currently practical
catalysts disposal techniques of a typical ammonia plant.
I
Startup and Shutdown Emissions
Startups and shutdowns contribute the majority of the process
emissions from an ammonia plant. The initial startup is usually the
most severe because of its duration. The plant is started up or shut
down in a controlled sequence, which results in process gas emissions
through various vents. The normal vent points are the desulfurizer's
outlet, high-temperature shift converter outlet, CO2 absorber inlet,
methanator inlet and outlet, the ammonia converter outlet, and the
purge from the syn loop and refrigeration system.
Minimization or elimination of these vents is important not only
because of their environmental impact, but also because of their cost
impact. They can add 5 to 10% to the normal production cost of
ammonia. The minimization or elimination of emissions due to
startups and shutdowns can be summarized as follows:
1. Minimize the time required by using interlocks and a logical startup
and shutdown sequence.
2. Use recycled inert gases for preheating.
3. Use maximum allowed preheat rates for equipment and catalysts.
4. Reduce the low-temperature shift catalyst with inert gas carrier.
5. Use prereduced ammonia synthesis catalyst to reduce activation
time.
6. Go on recycle in the syn loop as soon as possible.
7. Recover vent gases to fuel.
8. Send vent gases to a flare.
Table 3 Ammonia Plant Catalyst Disposal/Recovery Methods
Unit operation Catalyst Disposal/recovery
Desulfurization Zinc oxide, spent catalyst Landfill
Cobalt oxide/molybdenum
oxide/alumina
NG reforming Supported nickel oxide Metal recovery
High- Iron oxide/chromium Metal recovery
temperature oxide
shift
Low- Copper oxide/zinc Metal recovery
temperature oxide/alumina
shift
Methanation Supported nickel oxide Metal recovery
Ammonia Multipromoted magnetite Landfill
synthesis
 Page 69
The potential for substantial emissions can result from emergency
shutdowns due to equipment failure, upsets in process conditions,
human error, and force májeure. Therefore, it is very critical to
perform a proper risk analysis on the plant and to provide, in the
design phase, the necessary preventive or control sequence in case of
an emergency. Instrument interlocks, emergency backup equipment
such as batteries, sparing of equipment, instrument voting systems,
adequate system inventory, computer control, scrubbers, and so on,
are all used to avoid or minimize emissions during an emergency.
Proper shutdown procedures must be followed to ensure that gases
containing carbon monoxide are excluded from the methanator at any
temperature below 150°C. Below this temperature, nickel carbonyl is
formed, which is an extremely toxic gas. This procedure must be
followed for any metallic nickel-containing catalyst [15].
J
Current Plant Designs Bring Positive Emission Reductions
The 1970s and 1980s saw a dramatic upsurge in the number of
ammonia plants undergoing energy-reduction retrofits due to rising
energy costs. A vast array of energy-savings features became
commonplace in ammonia plant design. The following are the most
popular energy-saving features [16]:
1. Combustion air preheat
2. Revised reforming conditions and heat recovery
3. Reduced energy consumption in CO2 removal system
4. Makeup gas drying
5. Low-pressure-drop converter design
6. More efficient refrigeration system
7. Purge gas recovery
8. More efficient steam system
Today's state-of-the-art ammonia plant incorporates a wide range of
energy-saving features that have reduced the energy required to
around 26 MMBtu/ST of ammonia from 34 MMBtu/ST just a decade
ago. Efficient feed and fuel utilization and better process heat
integration/recovery as well as implementation of other energy-
savings features have reduced the environmental impact of ammonia
plant operation. However, the principal environmental benefits
derived indirectly as a result of energy retrofits have mostly been
attained. The challenge for the process plant designer/plant operator in
the next decade is to continue to make gains in ammonia plant
efficiency and to continue to provide design features that will meet
environmental regulations at a positive cost-to-benefit ratio.
Recent process innovations from the The M.W. Kellogg Company,
such as the Kellogg Advanced Ammonia Process (KAAP), the
Kellogg Reforming Exchanger System (KRES), and the Kellogg
Dynamic Ammonia Control (KDAC), meet this formidable challenge
[17]. Implementation of KAAP technology, which is based on the use
of a high-activity ammonia synthesis catalyst, results in significant
energy savings over current conventional plant designs. Use of a
reforming exchanger, such as KRES, is currently the most effective
way of eliminating primary reformer stack emissions. The heat
required to facilitate the reforming reaction is obtained by heat
exchange with the secondary reformer effluent instead of by burning
''fuel" in the classical sense. Although a small heater is needed for air
preheating, overall plant combustion-source emisions are lowered
significantly.
Improved plant control through the use of advanced computer
management systems
 Page 70
such as KDAC can result in improved plant reliability and
profitability. The use of an advanced control system increases on-
stream factors and gives the operator a tool to respond better to plant
upset conditions under which pollutant releases occur.
III
Summary
While many of the candidate technologies discussed in this chapter
have already been used in other hydrocarbon (HC) processing plants,
implementation into ammonia plants has been slower mainly because
the manufacturing process is relatively clean compared to other HC
manufacturing plants. Ammonia operators have relied mainly on
energy reduction features which reduced pollutant emissions
synonymously. The future will invariably put greater pressure on
ammonia plant operators to incorporate these technologies as stricter
environmental regulations take force. The challenge to process
designers is to find cost-effective ways to incorporate pollution
control technologies, both as retrofits and at the grass-roots level. The
long-term challenge is to find a more direct and less energy intensive
process for nitrogen fixation.
References
1. G. D. Rawlings, and R. B. Reznik, Source Assessment: Synthetic
Ammonia Production, EPA-600/2-77107, NTIS PB276718/AS, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.,
1977.
2. Compilation of Air Pollutant Emission Factors, AP-42, 4th ed., Vol.
1, U.S. Environmental Protection Agency, Research Triangle Park,
N.C., September 1985, Table 5.2-1.
3. Source Category Survey: Ammonia Manufacturing Industry, EPA-
450/3-80-014, U.S. Environmental Protection Agency, Research
Triangle Park, N.C., August 1980.
4. Compilation of Air Pollutant Emission Factors, AP-42, 4th ed., Vol.
1, U.S. Environmental Protection Agency, Research Triangle Park,
N.C., 1985, p. 1.4-3.
5. Nitrogen Oxides Emissions and Controls, EPA/600/7-88/015, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.,
1988.
6 A. C. Stern, R. W. Boubel, D. B. Turner, and D. L. Fox,
Fundamentals of Air Pollution, 2nd ed., Academic Press, New York,
1984, pp. 8886.
7. K. M. Bentley and S. F. Jelinek, NOx control technology for boilers
fired with natural gas or oil, Tappi Journal, Vol. 72, 1989, p. 4.
8. R. T. Waibel, D. N. Price, P. S. Tish, and M. L. Halprin, Advanced
burner technology for stringent NOx regulations, API Midyear
Refining Meeting, Orlando, Fla., May 1990.
9. G. R. Offen et al., Stationary combustion NOx control, JAPCA, Vol.
37, 1987, pp. 8648761.
10. W. D. Fellows, Application of the thermal deNOx process to glass
melting furnaces, 1989 Fall International Symposium, American
Flame Research Committee, International Flame Research
Foundation, 1989.
11. E. P. Smith, Considerations in determining O2 and CO control for
combustion efficiency and quality, Industrial Heating, Vol. 54, 1987.
12. C. J. Romero et al., Treatment of Ammonia Plant Process
Condensate, EPA-600/2-77-200, NTIS PB273069/AS, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.,
1977.
13. T. A. Czuppon, S. A. Knez, and J. M. Rovner, Ammonia, in R. E.
Kirk and D. F. Othmer, eds., Encyclopedia of Chemical Technology,
4th ed., Wiley, New York, 1992, Vol. 2, p. 665.
14. Recovery of spent catalysts, American Chemical Society, Division
of Petroleum Chemistry Symposia, Kansas City, Mo., Vol. 27, No. 3,
September 1217, 1982.
 Page 71
15. M. V. Twigg, Catalyst Handbook, 2nd ed., Wolfe Publishing,
Prescott, Ariz., 1989, pp. 182190.
16. J. R. LeBlanc, Technical aspects of reducing energy consumption
in new and existing ammonia plants, Chemical Economy and
Engineering Review, Vol. 18, No. 5, May 1986, pp. 2226.
17. T. A. Czuppon and S. A. Knez, Emerging technologies for the
ammonia industry, Fertilizer Focus, August 1991, pp. 1115.
 Page 73

6
Ammonia Storage, Loading, and Transport
Robert H. Roberts and John G. Baguley
The M.W. Kellogg Company
Houston, Texas
I
Ammonia Storage
Anhydrous ammonia is stored either in refrigerated atmospheric
pressure storage tanks or in pressurized spheres and vessels
commonly called "bullets." Large-capacity storage systems are
normally refrigerated atmospheric pressure storage tanks. These are
usually located at production plant sites and in the North American
hemisphere at central storage terminals along several pipeline
distribution systems. Pressurized or atmospheric temperature storage
(bullets and spheres) are usually located at retail outlets and are much
smaller in size. Due to the seasonal nature of the ammonia fertilizer
market, ammonia is usually stored as a liquid at near-atmospheric
pressure in large refrigerated storage tanks.
A
Atmospheric Pressure Refrigerated Storage Tanks
Storing large quantities of ammonia in atmospheric pressure
refrigerated tanks is most advantageous economically compared to
pressurized storage, due to a significant reduction in fixed capital
costs of vessels versus the operating costs, and is safer due to the low-
pressure containment relative to volume stored. The ammonia is
stored at 33.3°C (28°F) and approximately 3.5 kPa gauge (0.5 psig)
(see Fig. 1). The storage tanks are flat-bottomed, of either single-or
double-walled construction, and insulated to reduce atmospheric
heating of the stored liquid. These tanks range in size from 10,000 to
50,000 tons and usually have a dedicated vapor recovery system. The
vapor recovery system may be integrated with the production plant
refrigeration process unit. The storage facility may or may not have
secondary containment (dikes). Table 1 summarizes the type of tank
construction and containment combinations.
Health and safety precautions must be taken when storing ammonia.
In most cases there are national or local laws covering ammonia
storage and transport. If ammonia is spilled or leaks, it presents a
hazard because it is both toxic and in some cases combustible.
 Page 74

Figure 1
General arrangement of double-walled storage tank.
 Page 75
Table 1 Summary of Refrigerated
Storage Tank Construction and
Containment
Approximate
Tank type Containment % of total
Single wall None 15
Earth dike 50
Concrete 8
dike
Double wall None 15
Earth dike 9
Concrete 3
outer wall

Therefore, it is necessary to have both the physical means to contain

or control a spill and also, appropriate plans to implement in reacting
to an ammonia spill. Most tanks are provided with an earth or
concrete wall to contain the entire contents of the tank if total tank
failure occurs. Concrete walls are typically used in Europe to surround
large ammonia storage tanks. The most effective design includes high
walls and a minimum surface area exposed to the atmosphere.
Single-walled tanks are flat-bottomed cylindrical shell domed roof
tanks constructed of low-temperature carbon steel materials. The
bottom foundation is usually compacted soil/sand with a ring wall and
heat elements to prevent subsoil freezing. Some tanks are supported
on an elevated pile cap foundation. The tanks are insulated with
sufficient insulation to reduce heat leakage and prevent moisture
buildup on the insulation surface. The insulation is sealed to prevent
frost buildup under the insulation. Frost buildup would cause tank
shell corrosion, insulation deterioration, and increased heat leak (boil-
off of contents). Insulation thickness for these tanks is typically 100
mm (4 in.) of cellular glass or foamed-in-place polyurethane. The
insulation must be properly anchored to the tank wall and contain
expansion joints for the difference in material expansion coefficients
of the tank wall and insulation. The insulation load should be
considered in the tank shell and foundation design.
Double-walled tanks are designed as flat-bottomed cylindrical shell
domed roof tanks comprising an outer tank bottom, an outer shell, and
a roof to enclose the insulating space around the inner tank. The inner
tank is designed to contain the stored liquid in a refrigerated state. The
annular space between the outer and inner tank shells is usually filled
with loose perlite (heat-expanded volcanic rock) as the insulation and
an inert purge gas (nitrogen) to maintain a dry gas atmosphere. The
tank bottom foundation is made up of a reinforced concrete ring wall
with compacted sand and heating elements. The compacted sand is to
provide a stable base for the inner tank bottom. The heating elements
are to prevent frost buildup and subsequent frost heaving of the
foundation. The ring wall provides support for the inner and outer
tank walls, stability for thermal movements, and weight to resist uplift
loads due to internal pressure and wind loads. The inner tank metal
bottom is usually supported by a load-bearing insulation system, of
expanded polystyrene concrete or foamglass that is sealed by a
bitumen system and covered by a layer of dry sand. The outer tank
bottom is covered with a layer of leveling material, usually lean
concrete, and acts as a vapor barrier.
The contents of both single-and double-walled tanks are maintained in
a refrigerated state, 33°C (28°F), by retaining near-atmospheric
pressure, approximately 3.5 kPa
 Page 76
gauge (0.5 psig) (see Fig. 2). Pressure is maintained by withdrawing
any boil-off vapors and processing these vapors through a
refrigeration system. Storage systems located at production plants
route the vapors primarily to the ammonia unit refrigeration system
and secondarily, to a dedicated packaged refrigeration system. An
atmospheric vent and/or flare is used if the two systems described
above fail to operate. Storage systems located remote from production
plants are equipped with a packaged refrigeration system and an
ammonia flare. Some storage tank facilities do not have a flare
system; instead, the vent gases are routed to a water scrubbing unit
that removes the ammonia gas by absorption in the water. Emergency
overpressure protection is provided by pressure relief valves, usually
pilot-operated valves, set to open at the tank design pressure and sized
to prevent the tank pressure from rising 10% over design pressure.
Vacuum relief valves are also used to prevent collapse of the tanks
during maximum withdrawal of the liquid contents.
B
Pressurized Storage Vessels
This method of storing ammonia is usually used for small remote
retail locations. These storage vessels are commonly called "bullets."
Each vessel volume is typically between 68 and 115 m3 (18,000 to
30,000 U.S. gallons). The minimum design working pressure should
be 1720 kPa (250 psig). Ammonia stored in these vessels should be
limited to 0.5 ppm oxygen and maximum 2000 ppm water to reduce
the possibility of stress corrosion cracking. These vessels should be
fully stressed relieved. They should be fabricated, designed, and
inspected per ASME Section VIII, Division 1. Since these vessels are
located at remote sites, a regular inspection program should be
developed to check all welds, especially head-to-shell welds, at least
once every 5 years.
 Figure 2
Should vapor flow exceed the refrigeration system's
capacity, the excess goes to flare via storage tank pressure
control valves or refrigeration system relief valves.
 Page 77
Large quantities of pressurized, semirefrigerated ammonia may be
stored in spheres at 350 kPa (60 psig) and 1°C (34°F). These spheres
are typically 3000-ton capacity and provide storage for transfer to
truck or railcar tanks. A refrigeration system must be provided to
recondense boil-off vapors and to maintain the ammonia contents at
the desired pressure and temperature. These vessels must be insulated
and are usually designed to ASME Section VIII, Division 1
requirements. A pressure control system and overpressure protection
are required. A remotely operated emergency isolation valve is
typically located on the liquid outlet nozzle. The materials of
construction are carbon steel meeting the ASTM requirements
specified in the ASME Code. These vessels may be located at the
production plant or at remote locations.
II
Ammonia Loading Facilities
Anhydrous ammonia is shipped by pipelines, tank trucks, tank cars,
barges, and ships. Loading equipment for these transport carriers can
be classified by facilities to load land transport and those for water
transport. Facilities to load land transport are those that supply
distribution pipelines and those that supply mobile transports (truck or
rail). Facilities to load water transport are similar but larger than land-
load transport systems.
A
Fixed Land Transport Facilities
Fixed land transport facilities contain a vaporization or heating system
to bring ammonia from the atmospheric pressure storage temperature
to the required pipeline temperature. The pumps are usually capable
of approximately 6200 kPa (900 psig) discharge pressure. The
pipeline maximum pressure is nominally 9600 to 10,500 kPa (1400 to
1500 psig). In the United States there are several pipelines serving
many midwestern agricultural state customers. These pipelines
connect the production facilities located near the natural gas supplies
with users throughout the midwest.
B
Mobile Transport Loading Facilities
Loading facilities for mobile transport can be separated into those
handling refrigerated atmospheric pressure liquid at approximately
33°C (28°F) (for loading ships and barges) and those handling
ammonia liquid/vapor at atmospheric temperature 1°C (34°F) to 40°C
(100°F) (for loading transport trucks and railcars). The transfer of
refrigerated ammonia is through large high-volume insulated piping
systems at low pressure. Atmospheric-temperature ammonia is heated
and pumped through smaller, lower-volume piping systems.
1
Refrigerated Ammonia Loading Facilities
These facilities transfer refrigerated ammonia from refrigerated
storage tanks to barges and ships (see Fig. 3). The principal equipment
is transfer pumps, insulated transfer line, flexible loading arms, and a
displacement return system to return vapor from the barge or ship.
These systems all require cooldown prior to liquid transfer and
inerting after liquid transfer. The cooldown is done with the liquid at a
controlled rate and the flashed gas is returned to the storage tank and
refrigeration system. The transfer pumps are usually vertical canned
pumps to accommodate the suction liquid at its saturation pressure
and temperature. The transfer lines are similar to those in an ammonia
plant refrigeration system but are larger due to the high pumping rates.
Emergency shutoff valves are provided
 Page 78
to shut off flow during an emergency condition. The loading arms
contain swivel joints that allow the arm to be extended and retracted
for mating up with the barge or ship connections and also provide
flexibility for movement of the barge or ship. A piping system is also
provided to transfer the barge or ship tank-displaced vapors to the
storage tank and/or shore-side refrigeration system. Both the liquid
and gas lines should have emergency disconnect coupling at the
shore-to-vessel connection. These disconnects are utilized if the vessel
movements are greater than the loading arm's allowable movement or
if other emergencies arise that may be disastrous to the barge or ship.
A nitrogen header should be provided at the loading station. A
nitrogen line should be provided to the apex of each loading arm. This
line should contain a block valve at grade and a check valve minimum
distance from the loading arm connection. Other nitrogen purge
connections should be provided to inert or blanket the loading arms
and connections. Provisions should be made to displace the ammonia
liquid in the loading arms subsequent to loading completion.
The materials of construction for the liquid ammonia system should
be impact-tested carbon steel rated for 300 ANSI class. Valving and
valve trim should be suitable for refrigerated ammonia service. The
ammonia vapor return line can be carbon steel rated for 150 ANSI
class and be suitable for warm ammonia service.
2
Atmospheric Temperature Mobile Transport Loading Facilities
These facilities transfer ammonia at medium pressures and
atmospheric temperature (see Fig. 4). The ammonia is maintained in a
liquid state by the pressure in the containing system. The loading
facility will be comprised of a refrigerated tank or tanks, refrigerated
liquid pumps, a heater (heat exchanger), pressurized spheres or
bullets, warm transfer pumps, liquid loading arms for connecting to
railcars or trucks, and a vapor return line to the pressurized storage.
The refrigerated ammonia is pumped by high-pressure pumps,
approximately 1725 kPa (250 psig), through the heater and then to
storage spheres or bullets. The spheres and/or bullets provide surge
capacity in loading facility operation.
When railcars or trucks are to be loaded, the connections are made to
the vehicle tanks, transfer pumps started, liquid transferred to the
vehicle tank, and vapor returned to the storage spheres or bullets. The
loading system should have a strainer, flow measurement instrument,
flow totalizer, and a flow control valve in the liquid line. The vapor
return line should have a back-pressure control valve to maintain the
pressure in the vehicle tanks. The loading connections, liquid and
vapor, to the vehicle tanks may be flexible hoses or swivel-jointed
arm connections. The final connecting couplings should have block
valves and nitrogen purge connections.
A nitrogen purge header should be provided at each railcar and truck
loading location to provide nitrogen to purge loading connections
prior to and after loading and to purge the vehicle tanks if required.
All lines and connections that could contain air or be susceptible to air
ingress should be purged and/or blanketed with nitrogen. An
emergency eyewash and shower should be provided at each vehicle
loading station for personnel should personnel contact spills or large
leaks. Pressure relief valves should be placed on systems that may be
blocked in during preparation for loading operations to accommodate
thermal expansion of the liquid. The materials of construction can be
carbon steel rated for 150 ANSI class. Valve trim and accessories
should be for standard warm ammonia service.
 Page 79

Figure 3
Barge and ship loading of refrigerated ammonia. ESD, emergency shutdown.
 Page 80

Figure 4
Railcar and truck loading of warm ammonia.
 Page 81

7
Nitric Acid Production
Eugen-Petre Cristescu
Eurochim S.R.L.
Bucharest, Romania
I
Introduction
Nitric acid is manufactured in large amounts because it is an important
intermediate product for the production of chemical fertilizers. It is
manufactured at concentrations ranging from 55 to 65%, known as
''diluted acid." The tail gases resulting from dilute nitric acid
manufacture contain nitrogen oxides (NOx) and oxygen (2 to 4 vol
%), the remainder being nitrogen and water vapors. In addition to
their toxicity for plants and animals, nitrogen oxides are corrosive as a
result of their conversion into acids when in contact with
precipitation.
Recent worldwide pollution control regulations set the limitations for
nitrogen oxides released at the stack; thus for 1 ton of manufactured
acid, the amount of nitrogen oxides released should not exceed 1.9 kg
(expressed as HNO3). This is 0.03 kmol of NOx, leading to a
concentration in the tail gases of about 200 ppm. NOx is a mixture of
nitrogen monoxide (NO) and nitrogen dioxide (NO2) without a
defined NO/NO2 ratio. It is to be noted that the monoxide is colorless
and the dioxide is brownish red. This is important, as the regulations
mandate a colorless tail gas. To cope with this requirement, NO2 in
the tail gas should be below 150 ppmv.
II
Chemistry of Nitric Acid Production
The raw materials used for nitric acid production are ammonia,
oxygen from the air, and water. The chemical process develops in two
stages:
Stage 1: ammonia oxidation in the presence of PtRh catalyst

Simultaneously, there are undesired secondary reactions of ammonia

oxidation, with release of N2 and N2O as well as decomposition as N2
and O2 of a small part of the oxide
 Page 82

Figure 1
NO oxidation in a gaseous medium.
formed. Practically, the NO/NH3 conversion efficiency is between 94
and 98%, depending on the reaction conditions.
Stage 2: oxidation of nitrogen monoxide to nitrogen dioxide, followed
by water absorption:

Nitrogen monoxide formed according to reactions (4) and (5) is

reoxidized to dioxide according to reaction (2). Reaction (2) develops
relatively slowly, the reaction speed increasing while the temperature
decreases. It is obvious that pressure favors equilibrium displacement
to the right, with nitrogen dioxide release. Dimerization (3) is a fast
reaction; the equilibrium (i.e., displacement toward formation of
N2O4) is influenced by a pressure increase and a temperature
decrease. Absorption of nitrogen oxides (4) and (5) is highly favored
by increasing pressure in direct proportion to the partial pressure of
NO2 at the third power. The reaction mechanism is not known
completely. Nitrous acid is also produced as an intermediate.
In addition to oxidation in a gaseous medium as in reaction (2),
oxidation is also possible in a liquid medium, with the help of
dissolved oxygen:

For high concentrations of nitrogen oxides, NO oxidation in a gaseous

medium is predominant, the process taking place as shown in Fig. 1.
For very small concentrations of nitrogen oxides, when the oxygen
dissolved in water is quantitatively of the same order of value for both
dissolved oxides and nitrous acid, oxidation in a liquid medium
becomes significant (Fig. 2).
In a liquid medium, oxidation is very important for pollution control.
The most

Figure 2
NO oxidation in a liquid medium.
 Page 83
advanced conversion into nitric acid is in relation to both the partial
pressure of oxygen and nitrogen oxides and the NO/NO2 ratio.
III
Manufacture of Dilute Nitric Acid
Differences among industrial processes are based on the various
pressures at which the two stages described above take place. Two
basic groups of processes exist (Fig. 3):
1. Monopressure processes, in which both ammonia oxidation and
nitrogen oxide absorption take place at the same pressure: 10 to 12
atm
2. Dual-pressure processes, with ammonia oxidation from 3 to 5 atm
and nitrogen oxide absorption at 10 to 12 atm
Detailed economic calculations do not show a clear advantage for
either of these groups. One process may be more economical than the
other only when judged from the point of view of the specific
conditions under which it operates.
Nitric acid factories established years ago when there were no
regulations for NOx emission control were built to operate at lower
pressures. Thus there are both mono-and dual-pressure plants
operating at pressures below 9 atm. These plants should be equipped
with special purification units because otherwise they cannot comply
with most current pollution control regulations (Fig. 4).
IV
Methods for Pollution Control in Nitric Acid Plants
Generally, five areas were studied with regard to reducing nitrogen
oxide emissions in nitric acid manufacture: (1) absorption, (2)
absorption accompanied by chemical reactions, (3) adsorption, (4)
catalytic reduction, and (5) selective catalytic reduction. On an
industrial scale, catalytic reduction procedures were utilized,
especially for ammonia selective catalytic reduction, which was
common over a wide range of applications.
It should be noted that chemical adsorption pollution control was
achieved by direct steps taken in the nitric acid manufacturing
process. Thus, under certain conditions, by

Figure 3
Mono- (a) and dual-pressure (b) processes.
 Page 84

Figure 4
Mono- (a) and dual-
pressure (b) processes completed with special purification units.
further chemical absorption in water, very low NOx concentrations in
residual gas are reached. Similarly, methods were developed to
diminish pollution during unsteady operation at startup and shutdown.
A
Absorption
One purification procedure [14] suggests NOx absorption in cold
dilute 25 to 30% nitric acid, followed by stripping desorption and
recycling of the gas recovered in the process. NOx concentration at the
residual gas outlet is below 100 ppmv. The same principle applies
when concentrated nitric acid is used for absorption [5]. Being made
of stainless steel, the units have a relatively high volume and level of
energy consumption.
B
Absorption Accompanied by Chemical Reactions
Both acid and alkaline chemical absorptions were considered.
1. As outlined in Refs. 610, oxidant chemical acid absorption has as
its goal NO-to-NO2 oxidation (e.g., with oxygen, ozone, or hydrogen
peroxide), with further water absorption to form nitric acid.
2. Reducing acid chemical absorption [11] removes NOx from
residual gas by nitric acid (20%) solution and ammonia (pH < 2)
treatment:

This process occurs at 50°C and may consist of one stage only.
3. In oxidant chemical alkaline absorption [12,13] NOx is removed
from the residual gas using 0.5% NaOCl and 10% NaOH solution
spraying. A high degree of purification is achieved, NOx content in
the residual gas being 100 ppm.
4. When reducing chemical alkaline absorption, gas is treated with an
acid solution containing 50 to 500 g/l urea at pH 0.5 to 1, with further
alkaline purification to pH 11 to 14 [14].
 Page 85
Due to high raw material consumption and especially to the
impossibility of using products obtained by the reactions, application
of the chemical absorption method on an industrial scale was limited.
C
Adsorption
The method suggests adsorption on molecular sieves and desorption
by heating, the oxide-containing gas being recycled. As shown in
Refs. 1518 molecular sieves, clinoptylollyte, mordenite, and generally,
zeolites with a high Si/Al ratio can be used, which, by hydrochloric
acid treatment, increase their absorption capacity. NOx concentration
in residual gas is reduced to <200 ppmv.
Activated carbon impregnated with sulfur (0.25%) and/or metals
reduces NOx content by pressure adsorption. Regeneration occurs by
heating [19]. Dilute HNO3 activated carbon also adsorbs NOx. After
oxidation, NO2 is water-absorbed and converted to nitric acid, which
is recirculated [20].
D
Catalytic Reduction
Catalytic reduction [2124] can take place with natural gas, other
hydrocarbons, or hydrogen, according to the reactions

Since residual gas contains oxygen, the latter reacts with methane:

Industrially, platinum, palladium, chrome salt, and oxide catalysts

were used, with two-stage combustion.
 Figure 5
Pollution control by catalytic procedure.
 Page 86
As shown in Fig. 5, residual gas is heated to reaction temperature (300
to 450°C) and then, together with methane gas, it is fed to the first
stage of catalysis, where the temperature rises to 550 to 700°C.
Further on, gas is transferred to a heat exchanger, where it is cooled
down, mixed with methane, and fed to the second stage of catalysis.
The system allows for NOx reduction to 15 to 20 ppmv. The
procedure has several disadvantages: high investment costs, major
alterations of the process, and high fuel consumption.
E
Selective Catalytic Reduction of Nitrogen Oxides
The process develops according to the following reactions:

The catalyst consists of vanadium pentoxide (V2O5) on an alumina

carrier. In addition to the vanadium oxide catalyst, the following were
considered:
1. Mixed MnCu, MnCr, or MnCo metallic oxides with 0.5% graphite
addition [25,26]
2. Al2O3 carrier compounds with MeFe(CN)2NO and Me2Fe(CN)6,
where the metal (Me) can be Mn, Fe, Ni, or Cu [27]
3. TiO2 mordenite mixed with such metals as Mo, V, W, Cr, or Te with
Al2O3 binder [28]
4. Highly stable zeolites with rare earth metals and/or a V2O5
combination [29]
In Ref. 30, three catalysts are suggested: (1) noble metal or Al2O3
carrier at 215 to 250°C temperature, (2) noble metal + metal oxides on
Al2O3 carrier at 310 to 370°C, and (3) noble metal with metal oxides
on TiO2 carrier at 410 to 480°C. Amorphous V2O5 on Al2O3 carrier is
commonly used on the industrial scale.
As shown in Fig. 6, the residual gas leaving the absorption plant with
500 to 3000 ppmv NOx is preheated, mixed with ammonia gas, and
fed to the catalytic reactor. The NOx content in the residual gas
depends on the type of absorption plant used, but it can be stated that
this content is in reverse proportion to pressure. Values between 500
and 3000 ppmv are generally met in industry for pressures between 2
and 8 atm.
A good conversion efficiency is obtained at an equimolecular
NH3/NOx ratio. This means that preferably, the process develops
according to reaction (14), which also presupposes an equimolecular
NO/NO2 ratio. The most favorable conditions are obtained when the
degree of oxidation of NO2/NO + NO2 is between 0.3 and 0.5.
The optimum temperature at which the process develops has been
established to be in the range of 200 to 360°C; between these limits
the conversion efficiency exceeds 95%. Below 200°C the activity of
the catalyst decreases and ammonium salts (nitrite and nitrate) might
occur. Above 360°C the selectivity decreases and ammonia
consumption increases [see reaction (15)]. The purified gas contains
50 to 200 ppmv NOx, a maximum of 50 ppmv NO2, a maximum of 20
ppmv NH3, and is colorless.
The process is relatively economical, requiring only a modest
investment that does not involve important modifications of the
existing plant, the catalyst is longer lived (>5 years), and the ammonia
consumption does not exceed 3 kg/ton HNO3 for each 1000
 Page 87

Figure 6
Flowsheet of pollution control plant
by ammonia catalytic reduction.
ppmv NOx in the residual gas. In fact, many older plants for nitric
acid manufacture are equipped with such catalytic reactors.
Practically, this process is the most appropriate on an industrial scale
and is the only one used on a large scale [3134].
V
Pollution Control in Modern Nitric Acid Plants
A
Pollution Control by Absorption Extension
Absorption columns used in nitric acid plants are equipped with sieve
trays and cooling coils submerged in a foaming layer. Nitrous gases,
nitric acid condensate, and demineralized water are fed to the
equipment. Recycled water is used for cooling. As shown in Fig. 7,
three zones are noted:
Zone I, the lower part of the column, is characterized by a high flow
rate of absorbing liquid and a high concentration of nitrous gases. In
this zone a massive absorption of nitrous gases takes place, the largest
amount of nitric acid being obtained here. From the construction point
of view this zone is characterized by a large heat transfer surface and a
high foaming layer where the cooling coils are submerged.
Zone II is a zone of low nitrogen oxide concentration. It is
characterized by a larger free volume between the trays required by
reaction (2) for NO2 regeneration. The heat carrier used here is
somewhat colder and in most cases is recycled water that first passed
through the ammonia evaporator.
Zone III, the upper zone of the column, is the adiabatic zone, where
the concentration of nitrogen oxide is reduced (£6000 ppmv NOx) and
where absorption efficiency is diminished considerably. This zone,
covering about two-thirds of the column height, is meant for pollution
control (i.e., a decrease in NOx concentration in tail gases below 200
ppmv).
 Page 88

Figure 7
Absorption column.
A study of chemical reactions has revealed equilibrium and kinetic
conditions for absorption under favorable conditions. Thus the
reaction volume required for obtaining a certain amount of nitric acid
is determined through use of a computer-based program. By means of
such programs, absorber operation can also be observed. In pollution
reduction, oxidation of N3+ to N5+ takes place primarily in a liquid
medium, as shown in Figs. 2 and 8.
In Fig. 9 the curves are plotted for NOx
concentration variation in the tail gas as a function of pressure and
oxygen content as calculated and checked by analysts at the nitric acid
plant Nitramonia SA in Fagaras

* Romania. In attempting to obtain a colorless natural gas, pressure is

very important; for this purpose it is necessary that the total system
pressure exceed 10 atm. This condition is not unique for Nitramonia
SA. The problem is very well known and generally accepted by
specialists. All dilute nitric acid plants that use absorption expansion
for pollution control are built and operated with a minimum pressure
of 10 atm for this stage. Consequently, all plants that do not operate at
a minimum

Figure 8
Nitrogen oxides absorption in pollution
control zone of the absorption column.
 Page 89

Figure 9
Tail-gas concentration at the nitric acid
plant Nitramonia SA in Romania.
of 10 atm cannot use only absorption expansion as pollution control
but must use other methods as well.
B
Preventing Pollution during Plant Startup
A nitric acid plant is commissioned by startup of the turbocompressor
train, which at first operates with air. The thermal and hydrodynamical
parameters of the airstream in the plant are different from those in
normal operation with nitrous gases. For this reason, an air flow rate
initially different from the rated flow is fed. Following that, after
ammonia oxidation reaction priming, the air flow rate is increased
incrementally with the turbocompressor train speed to reach the rated
parameters. Under these transient conditions, due to low pressure at
the absorption stage, the nitrous gases are discharged to the stack and
pollute the atmosphere.
This situation can be prevented by equipping the axial compression
machine with adjustable-pitch blades. These adjustable blades allow
for operation at various flow rates at the same pressure and under
conditions of constant efficiency within a wide control range. An
example is a plant operated at two different pressures where the air
compressor and expansion turbine have adjustable-pitch blades (see
Fig. 10). The turbocompressor train is started with the adjustable
blades in a nearly closed position, at a minimum flow rate, but at
values close to the desired pressures (i.e., about 4 atm for the air
compressor and 11 atm for the gas compressor). After ammonia
oxidation reaction priming, by alternate control of the adjustable blade
position of the air compressor and the expansion turbine, the desired
flow rate is reached, keeping the pressure constant. Keeping the
pressure at the desired value of about 11 atm, pollution is prevented
during the startup period.
 Page 90

Figure 10
Exemplification curves of turbocompressor train operation
with blades adjustable for minimum, normal, and maximum
blade opening from the first stage of the expansion turbine.
C
Preventing Pollution during Accidental Shutdowns
During accidental shutdown, the nitrous gases in the plant are
suddenly discharged to the atmosphere by automatic opening of the
antisurge valves. In such conditions pollution is prevented by gas
recirculation through the compressor after release of shutdown
impulse control.
To understand the method, an example is presented for an absorption
plant whose
 Page 91

Figure 11
Flow diagram of absorption plant for nitrous oxides.
flow diagram is shown in Fig. 11. This plant is usually shut down by
the sudden opening of valves 1 and 2 and closing of valve 3, which
causes a red cloud around the gas discharge stack. The phenomenon is
prevented by mounting an additional valve 4, which at shutdown
protects the compressor from the surge phenomenon but retains the
pressure in the system as discharge valves 1 and 2 remain closed. The
gas in the plant goes to the atmosphere and passes through the
absorption column, where it is largely absorbed, thus reducing
pollution.
References
1. G. Carto, Nitrogen oxides scrubbing with nitric acid solutions,
Chemical Engineering Communications, Vol. 42, No. 13, 1986, pp.
157170.
2. D. W. Bolme, Humphreys and Glasgow/Bolme process for nitric
acid manufacture, Ammonia Plant Safety, Vol. 21, 1979, pp. 184188.
3. A. Horton, A new non-polluting process for nitric acid
manufacture, Fertilizer Acids: Proceedings of the British Sulphur
Corporation 3rd International Conference on Fertilizers, 1979, Vol.
1, Paper 5.
4. D. W. Bolme, Humpheys and Glasgow Bolme process for nitric
acid manufacture, Chemical Engineering Progress, Vol. 75, No. 3,
1979, pp. 9598.
5. G. D. Grab et al., Process for the removal of nitrogen oxides from
nitric acid plant tail gas, U.S. Patent 4,562,052, December 31, 1985.
6. P. J. Schmid, Process for removal and utilization of nitrogen oxides
from tail gases, Eur. Patent Appl. EP 213,117 (Cl.B01D53/34), March
1987.
7. G. D. Grab, Removal of nitrogen oxides from the nitric acid plants
tail gases, U.S. Patent 4,562,052 (Cl.423235; C01B21/00), December
1985.
8. A. F. Sudak et al., Removal of nitrogen oxides from tail gases in the
presence of ozone in excess, Promyshlennoe Sanit. Ochistka Gozov.,
1982.
9. C. L. Winter, Removal of nitrogen oxides from industrial off gases,
Brit. Patent Appl. 2,056,420 (ClB01D53/34), March 1981.
 Page 92
10. J. Hofmann, Equipment and method for nitrous gases removal,
Ger. Offen. 2,818,473 (Cl.B01D35/34), November 1976.
11. R. Miller, Conversion of nitrogen oxides to nitrogen for pollution
abatement, U.S. Patent 4,663,135.
12. M. Glomba et al., Removal of nitrogen oxides from tail gases,
Pol.Pl.123,035 (Cl.301D53/14), September 1982.
13. M. A. Gostmezyk, (1981), Process study for purification of gases
from nitric acid manufacture, Ochrona Powietrza, Vol. 15, No. 3,
1981, p. 63.7.
14. Y. P. Dyens, Process and equipment for wet purification of
nitrogen oxides from the tail gases, Eur. Patent Appl. EP 217,437
(Cl.B01D53/34), April 1987.
15. J. R. Alvarez Gonzales et al. (1989), Adsorption of nitrogen
oxides on natural zeolites with high silica oxides content, Afinidad,
Vol. 46, No. 419, 1989, pp. 4248.
16. J. R. Alvarez Gonzales, (1986), Removal and recovery of nitrogen
oxides from tail gases from nitric acid plants by adsorption on natural
and acid forms of erionite and carbazite, Anales de-la Quimica Series
A., Vol. 82, No. 3, Suppl. 1, 1986, pp. 523533.
17. H. Ringel, Separation and recycling of gaseous nitrogen oxides by
adsorption and desorption on molecular sieves, Fr. Demande FR
2,530,156 (ClB01D53/04), January 1984.
18. G. Klopp, Recycling of nitrogen oxides, Hung. Teljes 19,743
(Cl.C01B21/20), April 1981.
19. L. Marzo et al., Catalytic reduction of nitrogen oxides for the
nitric acid plants, Actas 9th Simp. Iberoam. Catal., 1984, Vol. 2, pp.
826835.
20. L. Roudier et al., Recovery of nitric acid from nitrous tail gases by
Cofaz process, Fertilizer Nitrogen: Proceedings of the British Sulphur
Corporation 4th International Conference on Fertilizer Technology,
1982, Vol. 2, pp. 561569.
21. A. D. Tereschenko, Study of crust type catalyst used to remove
nitrogen oxides from the tail gases, Khim. Tekhnol. (Kiev), Vol. 3,
1985, pp. 1315.
22. S. I. Bryazgina, Testing of APK-1,4 catalyst at industrial scale,
Khim. Prom. St. (Moscow), Vol. 8, 1981, pp. 504505.
23. Yu. V. Knyazev, Purification of nitrogen oxides containing gases
for AK-72 plants, Vestn. Kiev Politekh. Inst. Khim Mashinosti.
Technol., Vol. 24, 1987, pp. 3536.
24. H. Semel et al., Decrease of nitrogen dioxide content in the gases
mixture, Ger. Offen. DE 3,326,639 (Cl.B01D53/36), 1985.
25. J. Straszko, Catalytic reduction of nitrogen oxides in the tail gases
from nitric acid plants, Pol. Pl. 145,139 (Cl.C01B21/02), August 31,
1989.
26. C. Hamon, Catalytic composition for the selective reduction of
nitrogen oxides in oxygenated tail gases and tail gases purification
process, Eur. Patent Appl. EP 286,507 (Cl.B01J29/18), October 12,
1988.
27. V. Svasenko, (1988), Heterogeneous catalysts from metallic
complexes for treatment of tail gases from nitric acid manufacture,
Zurnal Prikladnoi Khimii (Leningrad), Vol. 61, No. 3, 1988, pp.
475479.
28. G. R. Lester, Catalyst for pollution control, U.S. Patent 4,663,300
(Cl.502-66; B01J29/24), May 1987.
29. S. G. Yaeras, Catalyst and method for nitrogen oxides removal
from tail gases, Eur. Patent Appl. EP 214,942 (Cl.B01 J29/08), March
1987.
30. L. Marzo et al., Catalytic reduction of nitrogen oxides for nitric
acid plants, Actas 9th Simp. Iberoam. Catal., 1984, Vol. 2, pp.
826835.
31. H. Dittmar, H. Catalytic reduction of NOx in nitric acid plant tail
gas: the BASF process, Nitrogen, British Sulphur Corp. Ltd., London,
NovemberDecember 1984.
32. J. Roiron, Resolution des problèmes de pollution NOx dans les
ateliers de fabrication d'acid nitrique, JFA Port-El-Kantaoui, Tunisia,
October 1986.
33. F. Luck and J. Roiron, (1989) Selective catalytic reduction of NOx
emitted by nitric acid plants, Catalysis Today, Vol. 4, 1989, pp.
205218.
34. H. Dittmar, (1985), Catalytic reduction of nitrous gases at nitric
acid manufacture, VDJ-Berichte, No. 525, pp. 147161.
 Page 93

8
Impact of NOx Abatement on the Fertilizer Industry
Jumpei Ando
Chuo University
Tokyo, Japan
I
Introduction
NOx emissions from artificial sources have increased to 100 million
tons yearly. Emissions are increasing primarily in developing
countries. In several countries, including the United States, Japan, and
Germany, emissions have started to decrease (Fig. 1) but are still at
high levels resulting in acid rain and photochemical smog. Since
nearly all the NOx is derived from fossil-fuel combustion, combustion
modification to reduce NOx by 20 to 50% has become popular in
many countries. For further NOx abatement, ammonia has been used
to convert NOx to N2 by the following reactions:

In 1991, more than 500,000 tons of ammonia was consumed for NOx
reduction. Much larger amounts of ammonia may be consumed in
future as the NOx abatement will become more and more important in
many countries. To remove 10% of the total NOx by selective
catalytic reduction (SCR), 7 million tons of ammonia will be
consumed annually. Urea has also been used to reduce NOx. Processes
to convert NOx in flue gases to nitrogen fertilizers have also been
tested in several countries. If these become commercially feasible,
large amounts of nitrogen fertilizer may be produced from NOx. In the
present chapter we outline NOx removal processes that consume
ammonia and also those that yield nitrogen fertilizers as a by-product.
II
NOx Removal and Ammonia Consumption
Among various NOx removal processes, a selective catalytic reduction
(SCR) process that uses a catalyst and ammonia at 300 to 400°C is the
most advanced. SCR is used mainly for flue gases containing 150 to
400 ppm NOx at electric power plants burning oil and
 Page 94

Figure 1
Annual NOx emissions. (From Ref. 1.)
coal. Nearly all the NOx in the combustion gases is in the form of NO.
Usually, a ratio of 0.7 to 0.9 mol of NH3 to 1 mol of NOx is injected
into the flue gas leaving the boiler at 350 to 400°C to remove 70 to
85% of the NOx, yielding less than 5 ppm unreacted NH3 (leakage
NH3, Fig. 2). SCR is also used for tail gases in nitric acid plants and
for waste gases in nitrophosphate plants containing 600 to 3000 ppm
NOx. The approximate capacities of SCR plants in eight countries are
shown in Fig. 3.
At 900 to 1000°C, ammonia reacts readily with NOx without a
catalyst. This reaction occurs to some extent at a nitric acid plant in
oxidizing ammonia to produce NO at about 1000°C, causing ammonia
loss. The reaction can be utilized to remove NOx in gases at 900 to
1000°C. The removal process is called selective noncatalytic
reduction (SNCR), or Thermal DeNOx (trade name of Exxon Corp.).
For SNCR, ammonia is injected into the flue gases at around 900 to
1000°C. Usually, a ratio of 1 to 2 mol of NH 3 to 1 mol of NOx is used
to remove 30 to 60% of NOx, giving
 Figure 2
Efficiency of SCR and SNCR. (From Ref. 1.)
 Page 95

Figure 3
Capacities of SCR plants (logarithmic scale). (From Ref. 2.)
10 to 20 ppm leakage NH3 (Fig. 2). SNCR is utilized primarily in the
United States and Japan at relatively low levels of gas only, because
although it is simple, it is less efficient and consumes a large amount
of ammonia.
An aqueous solution of urea is also used in SNCR to replace
ammonia. Injection of 1 mol of urea to 1 mol of NOx usually removes
50 to 60% of NOx. Urea readily decomposes at high temperatures and
may give 2 mol of ammonia from 1 mol of urea, which reacts with
NOx. The use of urea in SNCR has been increasing because urea is
easily handled and stored.
III
NOx Removal at Fertilizer Plants
NOx concentrations in tail gases from nitric acid plants usually ranges
from 500 to 1000 ppm NOx in the pressurized absorption process and
from 1000 to 3000 ppm in the atmospheric absorption process. Both
NO and NO2 are contained in the gases. At some nitric acid plants, the
tail gas is subjected to scrubbing with a sodium hydroxide solution to
recover sodium nitrite, but the demand for nitrite is limited.
SCR of the NOx in fertilizer plant waste gases has become popular in
many countries. Most plants use 0.8 to 1 mol of NH3 to 1 mol of NOx
to remove 75 to 90% of NOx at 250 to 300°C. In China, SCR units
were constructed in a nitric acid plant, using Chinese technology to
treat a 122,000-N·m3/h tail gas containing 3000 to 3500 ppm NOx. As
shown in Fig. 4, the tail gas at 50°C is heated to 250 to 300°C by
natural gas combustion, injected with ammonia, and introduced into
two reactors in parallel. Each reactor is 4.2 m in diameter and 5 m
high and contains 7 m3 of an alumina-based Cu2Cr2O5 pellet catalyst.
By using 1.1 to 1.4 mol of NH3 to 1 mol of NOx, 70 to 95% of the
NOx can be removed. The gas leaving the reactor is sent to a boiler for
heat recovery and is emitted at 180°C. The units have operated well,
usually removing about 90% of the NOx when using about 1.3 mol of
NH3 to 1 mol of NOx. However, the plant consumes a large amount of
ammonia. It may be more economical in the long run to lower the
NOx concentration of the tail gas by improving the absorption system
of the nitric acid plant. Ammonia consumption may be reduced
further by using more efficient catalyst.
Five nitrophosphate plants in the Commonwealth of Independent
States (CIS) are equipped with SCR units manufactured by Mitsui
Toatsu Chemical, Japan, to remove
 Figure 4
Flowsheet of SCR plant for nitric acid plant tail gas (manufactured in China). (From Ref. 2.)
NOx in the waste gases leaving the digester of phosphate rock with nitric acid.
The catalyst in these units is resistant to fluoride in the waste gas. In some nitric
acid plants, molecular sieve, a synthetic zeolite, is used to adsorb NOx in the tail
gas. By heating the zeolite loaded with NOx, NOx in high concentration is
desorbed and returned to the nitric acid plant. Molecular sieve is fairly costly
but has a long life and the process may be economical in the long run. Zeolite
may not be useful for SOx-rich gas because it may be damaged.
IV
Fertilizer Production from SOx and NOx in Combustion Gases
Various simultaneous SOx and NOx removal processes that yield fertilizers as a
by-product have been tested in several countries, including Japan, the United
States, and Germany. Although the processes have not been used yet
commercially, further development is desired because both SOx and NO
fertilizer sources, while SCR and SNCR convert both NOx and ammonia to N
causing the loss of energy and resources.
A
Electron Beam Process [2]
A waste gas at 70 to 120°C containing SOx and NOx is added with ammonia
and subjected to electron beam radiation. Fine particles of ammonium sulfate
and nitrate, which are formed rapidly, are caught by a bag filter or an
electrostatic precipitator. The product consists of double salts, 3NH4NO
(NH4)2SO4 and 2NH4NO3 · (NH4)2SO4, and is useful for fertilizer.
The process was originated in Japan and has been developed further in several
countries, including the United States, Germany, Poland, and China. The
process may be the only process so far developed to utilize SOx, NOx, and
ammonia to produce nitrogen fertilizer by a simple procedure. About 90% of
SOx and 80% of NOx can be removed. A major reason for the delay in
commercial application is that the demand for the by-product ammonium nitrate
sulfate is limited in industrialized countries that require SO2 and NOx removal,
whereas agricultural countries that require ammonium nitrate sulfate do not yet
require extensive NOx reduction.
 Page 97
B
Wet Process [3]
Two wet processes have been developed in Japan to remove SOx and
NOx converting them into useful by-products. In those processes, NOx
oxidizes SO2 into a sulfate, while NOx is reduced to NH3. One of the
processes uses an aqueous ammonia solution containing ferrous and
chelate compounds as the absorbent. Although various reactions take
place on SOx and NOx absorption, the overall reaction is

There is another wet process in which NO is first oxidized to NO2 by

adding ozone to the gas. The gas is then treated with a limestone
slurry to produce gypsum and ammonia by the following overall
reaction:

Those wet processes have been tested at pilot plants. The major
drawback of the processes are the complexity and difficulty in the
treatment of wastewater containing various compounds.
References
1. J. Ando, SO2 and NOx abatement: The impacts on acid rain and
greenhouse gases, Chemrawn VII World Conference on the Chemistry
of the Atmosphere, IUPAC and ACS, Baltimore, Md., December
1991.
2. J. Ando. Flue Gas Cleaning Technology of the World, Coal Mining
Research Center, Japan, August 1990 (in Japanese).
3. J. Ando, NOx Abatement for Stationary Sources in Japan, EPA-
600-7/79-205, August 1979, and EPA-600-7-83-027, May 1983, U.S.
Environmental Protection Agency, Washington, D.C.
 Page 99

9
Ammonium Nitrate and Nitrochalk
Robert E. Nitzschmann and Johannes G. A. Reuvers
BASF Aktiengesellschaft
Ludwigshafen, Germany
I
Introduction
Calcium ammonium nitrate (CAN, nitrochalk) is a standard solid
fertilizer, used widely as a versatile source of nitrogen [1]. Granulated
ammonium nitrate, usually containing 33.5 wt % N, is also often
utilized. The principal starting materials for the production of CAN
are ammonium nitrate (AN) and calcium carbonate (CaCO3, lime).
CAN and AN both provide nitrogen to the soil in the form of nitrate
and ammonium ions and, as such, show both fast-and slow-acting
properties [1]. AN can be synthesized directly from nitric acid and
ammonia [2] but can also be produced in the course of the
nitrophosphate process [3]. Lime is also used as a solid fertilizer. It
serves to reduce soil acidity and supplies calcium as a plant nutrient.
Limestone and dolomite are most often used as naturally occurring
sources of lime.
Calcium ammonium nitrate and ammonium nitrate are usually
supplied as granular fertilizers. The advantages of granulates are well
known: free-flowing materials allow mechanized handling and
distribution, higher bulk density, reduced dust formation, uniform
application, and diminished leaching losses [1]. Granulation can be
achieved using various types of equipment: prilling towers, pug mills,
and drum granulators. Also, pan granulation processes have been
developed and operated commercially [2]. Prilling methods were used
widely up to about 1975. Since then prilling has become less favored,
due in part to stricter standards regarding safety, environmental
controls, and product quality. The high melt temperatures as well as
the huge quantities of air required in a prilling tower gave rise to
problems involving materials of construction, pollution control, and
the like.
BASF now uses primarily drum granulation processes for AN and
CAN. We have chosen our CAN process using ammonium nitrate wet
CaCO3 to show the various measures taken by us to achieve high
environmental standards, to minimize air pollution, and to eliminate
process wastewater. Before describing these aspects of the BASF
CAN process we set forth the environmental state of the art in the
production of ammonium nitrate.
 Page 100

II
Production of Ammonium Nitrate
A
Neutralization and Evaporation
The neutralization of ammonia gas with nitric acid is normally
performed in a single-or two-stage neutralizer at ambient or elevated
pressure. A single-stage neutralizer may involve, for example, a
forced-circulation system or a pipe reactor and is inherently simpler in
design and lower in investment cost. Using a two-stage neutralizer the
total NH3 emission can be reduced. Many different types of
neutralizers are in use [2].
Temperature, pressure, and final concentration of the ammonium
nitrate solution are interdependent, with only two of the three
parameters free determinable. Operating at elevated pressures allows
the production of steam at higher temperatures. Temperature and pH
as well as impurities (Cl, Cu, organic material) must be strictly
controlled by reliable equipment/measuring devices. Deviations of
these parameters beyond the normal operating ranges should be
indicated by audiovisual means.
Reaction products of the neutralization of NH3 with HNO3 are an
ammonium nitrate solution and steam/process vapors. Depending on
the type and design of the neutralizer, the concentration of the nitric
acid, and the actual operating conditions, the heat of neutralization
may be used for (1) the concentration of the ammonium nitrate
solution, (2) the production of export steam, (3) the evaporation of
liquid ammonia, and (4) the preheating of ammonia and/or nitric acid.
Process vapors and/or steam can be used for evaporation of the
ammonium nitrate solution to the desired final concentration (e.g., 93
to 94 wt % and/or 97 to 99.5 wt %). Favorably falling film
evaporators are employed to concentrate the ammonium nitrate
solution under vacuum. The process vapors from neutralization and
evaporation will contain small amounts of ammonium nitrate and
ammonia or nitric acid. By proper design of the neutralization and
evaporation section, the amount of these contaminants can be reduced.
The process vapors may be freed from these impurities by means of
droplet separators (demister pads, wave plate separators) and/or
scrubbing systems (packed columns, venturi scrubbers). Emissions of
ammonium nitrate mists, which may also occur from this section, are,
however, difficult to remove due to their submicron size (Sec. II.B).
The process vapors are normally condensed, thereby exploiting the
heat of condensation for the concentration of the ammonium nitrate
solution. The resulting process condensate may be purified using
stripping techniques (air/steam), ion exchange, or reverse osmosis or
may be used as makeup water [5].
The treated waste air emitted from the neutralization and evaporation
section will normally contain less than 30 mg/N·m3 of ammonia and
less than 30 mg/N·m3 of dust, respectively. In case the process vapors
are condensed, the remaining waste air will merely contain small
quantities of inerts, such as methane, hydrogen, and nitrogen. These
inerts originate from the ammonia gas and contain only minor
amounts of NH3. Practically, they may, for example, be vented to a
combustion chamber for the preheating of air used in the drying of
granulated fertilizers.
B
Granulation and Prilling
In prilling, a granular product (prills) forms by the solidification of
droplets that fall down a tall tower. The droplets are usually formed
using a rotating perforated bucket and are cooled by a countercurrent
stream of ambient air. The amount of air required for the solidification
of droplets is far larger in prilling than by using rotating drums or
fluidized
 Page 101
beds. In particular, in hot and humid climates the amount of air
required in prilling and the prill tower height become almost
prohibitive.
The particle size of prills (usually <2.5 mm) is much smaller than the
size of granules (2 to 5 mm), the crushing strength of prills is lower,
their porosity is higher, and their bulk density is lower. These physical
properties imply unfavorable agronomic effects regarding fertilizer
field distribution and application techniques. Prills with a particle size
in the range 2.5 to 3.0 mm can be realized only with increased prill
tower height (i.e., at higher investment costs).
The ammonium nitrate melt fed to the prill tower must be essentially
anhydrous (i.e., the water content of the melt or slurry is less than
about 1.0%). This is also true for the prilling of CAN or NP(K)
fertilizers. Using rotating drums, fluidized beds, pug mills, or
combinations of these, a final water content of about 3 to 8% is
maintained in the slurry. This simplifies slurry handling and allows
utilization of the heat of solidification for the further evaporation of
water, which is not possible in prilling.
Rotating drum granulation means lower slurry temperatures with
reduced safety risks, reduced requirements for the materials of
construction, and the absence of the necessity to preheat additives
(e.g., calcium carbonate for CAN or potassium salts for NPK). It is
because of these technological, economical, and environmental
advantages combined with the superior product quality of granules
that rotating drum or fluidized-bed granulation has taken the place of
prilling as the preferred solidification technique.
The waste air from prilling or granulation contains ammonia and
ammonium nitrate. Normally, a wet scrubbing system is employed to
remove ammonia (Sec. IV.C), although such equipment becomes
rather largeand thus expensivein prilling. The waste air leaving the
scrubbing tower contains less than about 15 mg/N·m3 of ammonia
under normal operating conditions. Dust (ammonium nitrate particles)
is first separated in cyclones and is further removed by the acidic
scrubbing solution (Sec. IV.C).
Of particular importance to the prilling technique is the removal of
submicron ammonium nitrate particles from the waste air. These
particleslost from the surface of the prills or formed in the reaction of
gaseous ammonia with gaseous nitric acidare not absorbed by
applying the usual scrubbing methods and are quite visible as a
persistent blue haze. The environmental perception of such a haze by
the general public can have tremendous consequences.
To avoid this haze and to reduce the emission of submicron
ammonium nitrate aerosols, fiber mist eliminators can be installed.
Fiber mist eliminators such as Brink filters or irrigated candle filters
show high collection efficiencies even for particles below 0.5 mm.
Brownian diffusion is the predominant mechanism for the collection
of submicron particles at low gas velocities: a high pressure drop is
not required, which saves power costs. At higher waste gas velocities
the mechanism for fiber mist collection shifts from Brownian
diffusion to impaction. As the collection efficiency at impaction is
enhanced by higher gas velocities, the size of the mist eliminatorsand
therewith investment costscan be reduced by proper design of the
abatement system.
Other systems for the control of emissions from prill towers have also
been developed. For example, the Beco emission control system uses
a new ammonia-based fume suppression technique in combination
with wet scrubbing. This system can control stack opacity with
recovery of ammonia and appears particularly well suited for
ammonium nitrate prill towers. After granulation or prilling the
product needs further cooling. This is normally realized in a rotating
drum or fluidized-bed cooler. Dust in the waste air from cooling may
be removed using cyclones or bag filters (Sec. IV.D).
 Page 102

III
Production of Calcium Ammonium Nitrate (CAN):
Process Description
Before considering the environmental aspects and related technical
features of the BASF CAN process, a brief overview will be useful
(Fig. 1). Ammonium nitrate (94 wt %) is produced from NH3 and
HNO3 in an AN synthesis section or obtained by evaporation. In the
latter case, dilute AN (60 wt %) results from the CN conversion
reaction, which is part of the Odda process for manufacturing
nitrophosphates [3]. AN-94 wt % is concentrated up to about 97.0 to
98.5 wt % by means of a falling film evaporator. The concentrated AN
is not stored but is used immediately for AN/lime mixing.
Ammonium nitrate wet lime produced in the Odda process is neither
washed with
 Figure 1
Production of calcium ammonium nitrate using ammonium
nitrate wet lime originating from the Odda process.
 Page 103
water nor dried. The calcium carbonate produced in CN conversion is
merely filtered on a belt filter and treated with neutral ammonium
nitrate. The resultant ammonium nitrate wet CaCO3 contains up to 25
wt % of AN solution [4]. Without further treatment this wet lime is
used in the AN/lime mixing section. Inorganic additives can be added
to enhance product quality (Sec. IV.C).
Proper design of the mixing section allowed us to produce a very
homogeneous CAN slurry, at the same time minimizing the residence
time of the AN/lime mixture. In this manner the extent of the reverse
reaction between AN and lime leading to Ca(NO3)2 is limited and
therewith the hygroscopicity of the final product reduced. The
conditioned melt, containing additives and having the set N content
(e.g., 27 wt %) is sprayed directly into a granulation drum. After this
granule-forming step, the granulation loop includes several processing
operations: drying, screening, crushing of oversized material,
recycling of oversized and undersized material, cooling, and finally,
conditioning of the withdrawn on-size CAN product. The recycle ratio
can vary but normally is about 2:1. Drying and conditioning are
usually realized in rotating drums, for cooling, a rotating drum or
fluidized-bed cooler can be employed.
IV
Environmental Aspects of the BASF CAN Process:
Pollution Control in Waste Air and Wastewater
A
Lime Handling
Utilizing ammonium nitrate wet lime has some significant
environmental advantages. In the first place, there is no need to wash
with water the lime resulting from the CN conversion. As a result, less
water is introduced into the Odda process and a smaller quantity of
contaminated condensates results when the AN is evaporated.
Moreover, the energy input is reduced. Most important, however, is
the fact that lime drying has become obsolete. This eliminates the
need for natural gas energy for lime drying and reduces investment
costs. It also eliminates waste air and wastewater from a lime drying
unit. The BASF granulation process is autothermic [4,5]; thus no
natural gas is required in the CAN process, not even during startup.
B
AN Concentration and AN/Lime Mixing
The process vapors produced in the concentration of AN from 94 wt
% to about 98.5 wt % are normally condensed indirectly. The
resulting vapor condensates contain up to 2 g/L of nitrogen (as
ammonium nitrate). This condensate is used first to clean equipment
and floors and subsequently, as makeup water for the scrubbing tower
(Sec. IV.C). As a result, no (or very little) process water is produced.
The condensate may also be combined with condensates from the AN
evaporation (in the Odda process). If required, the combined
condensates can be freed from N-NH4 and/or N-NO3 by ion exchange
or reverse osmosis processes [5]. The waste air passing through the
condenser contains mainly inert gases. The waste air from the
AN/lime mixing section is treated in the scrubbing tower provided for
the waste air from granulation and drying (Sec. IV.C).
C
Treatment of Waste Air from Granulation and Drying
During granulation the hot CAN melt is sprayed onto recycled
granules. To remove this sensible heat as well as the heat of
crystallization released during the forming process,
 Page 104
large quantities of air are drawn through the granulating and drying
drums (Fig. 2). The granulator drum can be operated in a
countercurrent or cocurrent fashion. In the drying drum CAN material
and ambient air move countercurrently. The quantity of air drawn
through each drum can be adjusted individually by means of valves.
In addition to water vapor, the air leaving each drum contains CAN
dust and ammonia. Before the dust-laden air is treated in a battery of
cyclones, the two airstreams from granulation and drying are
combined. Among other advantages, this reduces the formation of
encrustations. The dust separated in the cyclones falls onto screw
conveyors. The connection between each cyclone and screw conveyor
can include a short piece of flexible transparent material to allow for
ready control of the proper functioning of the cyclone. Moreover, such
a gadget reduces the mechanical treatment often inflicted on these
connecting spools.
 Figure 2
CAN granulation.
 Page 105
The screw conveyors collect and transport the CAN dust to double
pendulum flaps. The design and construction of these flaps have been
modified to improve the transport of dust and to avoid the
introduction of false air due to incomplete closure of the flaps. False
air is to be avoided throughout the granulation loop: not only will the
sheer amount of additional air carry more dust and ammonia, but in
particular, the efficiency of the cyclones will be affected.
The air leaving the cyclones still contains about 35 to 75 mg/N·m3
(STP) of dust as well as ammonia and is subsequently scrubbed in a
packed column. The scrubbing liquid is circulated in a closed circuit,
while the pH value is adjusted through the addition of HNO3. The
resulting ammonium nitrate solution is continuously returned to the
process. The basic principles for scrubbing NH3-laden waste air are
well known: (1) minimum pH value at the top of the scrubbing tower
to avoid the formation of aerosols, and (2) a sufficiently high
circulation rate to avoid the breakthrough of ammonia. Given the
quantity of waste air to be treated, the operating temperature, and the
NH3 (peak) concentrations, the dimensions of the scrubbing tower can
be set [57]. Waste air and circulating dilute ammonium nitrate solution
flow in a countercurrent or cocurrent fashion.
The air leaving the scrubbing tower contains less than 10 mg/N·m3 of
ammonia under normal operating conditions and in a corresponding
design even lower values are obtainable. Dust is well within the limits
set forth by the German regulations (£50 mg/N·m3 as specified by the
TA Luft). To avoid droplets being carried away and emitted from the
scrubbing tower with the clean air, a droplet separator is usually
installed. During operation these separators become encrusted, which
reduces their efficiency, and finally they must be cleaned, often
mechanically. The design of such separators was extended by BASF
to allow for fast in-line cleaning, therewith raising the overall
efficiency of scrubbing column and droplet separator [8].
In the scrubbing tower the calcium carbonate contained in the CAN
dust is converted to Ca(NO3)2. The scrubbing solution is usually
returned to the AN concentration or to the AN/lime mixing section.
Owing to its hygroscopic properties, Ca(NO3)2 has an adverse effect
on the quality of the final CAN product. Often, the maximum
tolerable Ca(NO3)2 content in CAN is stipulated to be less than, for
example, 0.5 wt %, Less dust arising from the handling of ammonium
nitrate wet lime, from AN/lime premixing, and from CAN granulation
and drying thus not only prevents emissions but also improves CAN
quality.
The special internal treatment developed by BASF, involving the
addition of certain additives, including inorganic sulfates, increases
CAN quality: crushing strength, abrasion stability, and resistance to
breakdown during thermal cycling are improved. As such, these
additives reduce dust formation, not only during granulation and
drying but also during handling of the final product. In particular, the
adverse effects of returning the scrubbing solution containing
Ca(NO3)2 to the AN concentration or AN/lime mixing section are
offset. It is also beneficial to return the scrubbing solution to the CN
conversion section [5].
Product safety is also enhanced. To avoid the formation of a solid
fraction excessively rich in ammonium nitrate, we take due care to
realize a homogeneous mixing of ammonium nitrate, lime, and
additives [9]. The inorganic additives offset the hygroscopicity of the
Ca(NO3)2: caking, disintegration of granules, and dusting of the final
product are minimized. The resistance to breakdown during thermal
cycling avoids an increase in porosity. Reduced fines and reduced
porosity result in a more than sufficient level of detonation resistance
of the CAN product [9]. Oil retention tests also show favorable
properties.
 Page 106
D
Treatment of Waste Air from Cooling and Conditioning
Cooling of the granulated and dried but still hot on-size CAN product
can be achieved by using rotary drums and/or fluidized-bed coolers.
Advantageously, a rotary drum is used to bring about slow initial
cooling to about 70 to 80°C. In a second step a fluidized bed is
operated to obtain the final product temperature, about 33 to 36°C. A
fluidized-bed cooler (FBC) may comprise two or three cooling stages
and usually can be switched between summer and winter operating
conditions. Ambient air is drawn in; depending on the climatic
conditions, this air may be preconditioned to reduce its relative
humidity and temperature. In some cases the waste air from the final
stage of such a FBC can be reused for cooling in the rotary drum. On
the other hand, this air could also be reused in the first stage of a
three-stage FBC. The dust in the waste air leaving the first stages of
the FBC as well as in the waste air from the rotary drum is removed in
sufficient quantities using cyclones as described above, and emitted.
This air need not be scrubbed.
The final CAN granules are conditioned externally to reduce the
caking tendency. External conditioning is carried out on the cooled
product after police screening to avoid returning organic material to
the granulation section. It is preferentially realized in a rotary drum to
reduce the contamination of waste air. Coating can be done with fine
inorganic powders such as talcum or kaolin, with nonionic organic
sealants such as polyethylene waxes, or with hydrophobic surfactants.
We prefer to use a rotary drum to apply surfactants such as fatty
amines, which yield excellent anticaking properties. The limited
amount of waste air from a rotary drum can then be emitted without
further treatment.
IV
Summary
Ammonium nitrate and calcium ammonium nitrate are standard
granular fertilizers. Ammonium nitrate is produced by the
neutralization of ammonia gas with nitric acid. The heat of
neutralization is used favorably to concentrate the ammonium nitrate
solution. Process vapors and/or process condensates can be purified
by known techniques.
Granulation rather than prilling has become the favored solidification
process. This is due to technological, economic, and environmental
advantages of rotating drum granulation combined with the superior
quality of granules over prills. Waste air from the granulation or
prilling of ammonium nitrate may be treated by means of cyclones,
scrubbing systems, and fiber mist eliminators.
We have focused here on the environmental aspects of the BASF
process for producing granular calcium ammonium nitrate (CAN). In
this process ammonium nitrate wet calcium carbonate is utilized. The
absence of a lime drying unit drastically reduces emissions into air
and water during lime handling. The use of additives further reduces
the formation of dust at the onset. The quality of process off-gas is
enhanced and encrustations avoided by proper design of ducting and
equipment and by the treatment of waste air in cyclones and scrubbing
towers. With a CAN plant designed according to these guidelines,
undisturbed operation at capacity load can be achieved over a period
of more than 10 days.
References
1. Ullmann's Encyclopedia of Industrial Chemistry, 5th ed., Vol. A10,
VCH Verlagsgesellschaft, Weinheim, Germany, 1987, pp. 323431.
 Page 107
2. C. Keleti, ed. Nitric Acid and Fertilizer Nitrates, Fertilizer Science
and Technology Series, Vol. 4, Marcel Dekker, New York, 1986.
3. L. Diehl, K. F. Kummer, and H. Oertel, Nitrophosphates with
variable water solubility, paper read before the Meeting of the
Fertiliser Society of London, London, April 17, 1986.
4. R. E. Nitzschmann, K. H. Zapp, and H. Pottgieber, Verfahren zur
Herstellung von Ammoniumnitrat und Calciumcarbonat enthaltenden
Düngemitteln, Ger. Offen. DE 3,607,994.
5. J. G. A. Reuvers, L. Diehl, and R. E. Nitzschmann, The BASF
nitrophosphate concept: experience and environmental aspects related
to three recently commissioned NP(K) plants, Proceedings of an
International Workshop on Phosphate Fertilizers and the
Environment, Tampa, Fla., March 26, 1992, pp. 241250.
6. R. Wunder, R. E. Nitzschmann, and B. Meyer, Verfahren zur
Entfernung von Ammoniak aus Abgasen, Ger. Offen. DE 4,000,540,
Eur. Patent Appl. EP 440,932.
7. L. Diehl and R. Wunder, Prozebintegrierte Abgasreinigung bei der
Duengemittelherstellung nach dem Nitrophosphat-Prozeb, VDI
Berichte, No. 730, 1989, pp. 395415.
8. O. Meib und E. Hajok, Verfahren zur Abloesung von
Verkrustungen in Tropfenabscheidern, Ger. Offen. DE 3,738,737.
9. B. Purucker, Safety requirements in fertiliser plants. IFA Technical
Conference, Edmonton, Canada, September 1216, 1988,
TA/88/15/117.
 Page 109

10
Urea Production
Hidetoshi Uchino
Toyo Engineering Corporation
Chiba, Japan
I
Urea Process
Urea processes are classified as noted below, based on how the
unreacted materials are separated from urea synthesis solution and
recovered. A typical block flow diagram of the urea process is shown
in Fig. 1.
1. Once-through process. Unconverted raw materials, CO2 and NH3,
are discharged to other plants, such as an ammonium sulfate plant or
ammonium nitrate plant, to recover the NH3.
2. Partial recycle process. Unconverted raw materials, CO2 and NH3,
are partially separated in the first-stage decomposer and recovered in
the first-stage absorber and the remainder is delivered to other plants,
such as an ammonium sulfate plant or ammonium nitrate plant, to
recover the NH3.
3. Total recycle process. Unconverted raw materials, CO2 and NH3,
are totally separated through multistage decomposers, recovered in the
corresponding multistage absorbers, and recycled back to the reactor
for full recovery of the NH3 and CO2.
II
Description of Urea Process
The process flow of the total recycle stripping process is introduced
below based on the ACES process as a typical example. ACES
process flow is shown in Fig. 2.
A
Synthesis Section
Urea is synthesized from NH3 and CO2 in a reactor operated at 175
kg/cm2G, 190°C, and a NH3/CO2 molar ratio of 4.0 to perform 68%
of a one-pass CO2 conversion rate. Liquid NH3 is fed directly to the
reactor by an NH3 feed pump. Gaseous CO2 compressed in a CO2
compressor is sent to the bottom of the stripper as a stripping medium.
Air is fed to
 Page 110

Figure 1
Urea process.
the stripper together with CO2 for passivation of the construction
materials of the equipment. The urea synthesis solution, a mixture of
urea, excess NH3, ammonium carbamate, and water, is led to the top
of the stripper. The stripper separates excess NH3 in the feed urea
synthesis solution through the trays and then falls down to the falling
film heater, where unconverted ammonium carbamate and excess
NH3 in the solution are decomposed and separated by CO2 stripping
and external steam heating.
Overhead gas from the top of the stripper is introduced to the
carbamate condenser,
 Page 111

Figure 2
Process flow sheet.
 Page 112
two units operated in parallel, where the gaseous mixture is condensed
and absorbed into the carbamate solution fed by a carbamate pump
from the high-pressure absorber in the recovery section. The heat
formed in each carbamate condenser tube side by carbamate
formation and NH3 condensation is utilized to generate low-pressure
steam at 5 kg/cm2G in the No. 1 carbamate condenser and to heat urea
solution from the stripper in the No. 2 carbamate condenser by direct
heat exchange between the carbamate solution and the urea solution
coming out of the stripper after pressure reduction to 18 kg/cm2G
from 175 kg/cm2G. The noncondensed gaseous mixture and
carbamate solution from the carbamate condensers are recycled back
to the reactor by gravity.
The inerts in the synthesis sectionN2, H2, CH4, O2which are
contaminated by NH3 and CO2 and cause a pollution problem are
purged to the scrubber from the top of the reactor for recovery of NH3
and CO2 accompanied with the inerts. The urea solution leaving from
the stripper with a NH3 content of 12 wt % is further purified in the
subsequent decomposition section.
B
Decomposition Section
NH3, CO2, and water in the urea synthesis solution are further
decomposed and separated through the stepwise decomposers and
absorbed and condensed in the respective absorbers operated at 18 and
2 kg/cm2G, respectively. The urea solution with 80 wt % of urea
concentration is concentrated up to 99.7 wt % for urea prill production
or 98.5 wt % for urea granule production in the concentration section.
C
Concentration Section
The aqueous urea solution from the decomposition section is first fed
to the vacuum concentrator for preliminary concentration of the urea
solution by utilizing heat formation in the high-pressure absorber for
evaporation of water. The urea solution is then concentrated to the
required level, depending on the type of the product, through the
vacuum evaporator.
D
Finishing Section
The urea melt formed in the concentration section is solidified for the
production of either urea prill or urea granule by mean of air cooling
through the prilling tower or the spouted bed urea granulator. The urea
dusts in the effluent air from the prilling tower and granulator are
scrubbed by sprayed aqueous urea solution through the packed-bed
dust scrubber to reduce the urea dust emission below 30 mg/N·m3 air.
E
Recovery Section
The overhead gaseous mixtures from the decomposers in the
decomposition section are absorbed and recovered in the respective
absorbers using the process condensate as an absorbent and are finally
recycled back to the synthesis section.
F
Process Condensate Treatment Section
The water vapor evaporated while processing aqueous urea solution to
the final concentration in the concentration section is condensed in the
surface condensers under vacuum along with accompanying urea
mists, gaseous NH3, and CO2 to form the process
 Page 113
condensate. The process condensate is then sent to the process
condensate stripper and the urea hydrolyzer for treatment. The
condensate after treatment (clean condensate) is discharged from the
battery limits of the urea plant for use of either makeup water to the
cooling water or the boiler feed water, depending on the demand.
III
Sources of Pollution
In the urea plant, the following major sources of pollution always
exist during urea plant operation.
1. Evaporating water formed in urea synthesis as a by-product
2. Cooling anhydrous urea melt in forming solid urea product
3. Purging inerts accompanied by the raw materials of NH3 and CO2
Countermeasures for avoiding gaseous and liquid pollution are
introduced below.
A
Liquid Pollution
Urea is synthesized in the reactor in accordance with the equation

NH3 and CO2 are supplied to the urea plant from the ammonia plant
in the liquid phase and the gaseous phase, respectively. Both
feedstocks unavoidably contain inert gases such as H2, N2, and CH4.
As shown in the equation, 1 mol of H2O is formed as a by-product of
1 mol of urea. Therefore, water formed as a by-product and water fed
to the recovery section as an absorbent for recovery of unreacted NH3
and CO2 are evaporated in the concentration section to produce
unhydrous solid urea as the final product. In the course of
concentrating urea solution, water is evaporated under vacuum and
forms a process condensate by cooling the evaporated water vapor,
which is contaminated with NH3 and CO2 and entrains urea mists.
When daily maintenance of reciprocating and centrifugal pumps is not
done properly, NH3, urea, and ammonium carbamate solution leak
through the seals and gland packings into the cooling water.
B
Gaseous Pollution
Unhydrous urea melt is cooled by direct contact with a massive
amount of air to produce solid urea product in the form of either prills
or granules. The air coming out of the prilling tower or granulator is
the major concern regarding gaseous pollution. Without treatment, air
from the prilling tower or granulator contains about 100 to 150 mg
urea dust/N·m3 air. The internationally accepted emission levels is 20
to 50 mg urea/N·m3 air. Inerts fed to the urea plant together with NH3
and CO2 and air for passivation are purged from the urea plant at
medium pressure to atmospheric pressure. NH3 in the inerts is another
source of gaseous pollution.
Maintenance work in a urea plant can also be a source of pollution
problems. When the urea solution in high-pressure synthesis
equipment such as the reactor, stripper, and carbamate condenser is to
be transferred to the atmospheric (or low-pressure) storage tank while
the urea plant is shut down for maintenance work in any part of the
plant, NH3 is released from the solution due to the reduction in
pressure.
 Page 114

IV
Pollution Control
For pollutant abatement in a urea plant, the pollution control system
described below can be adopted, in addition to various technical
innovations designed to eliminate sources of pollution.
A
Liquid Effluents
The largest source of liquid pollution in a urea plant is process
condensate formed in the concentration section. An overall pollution
control system used in a typical urea plant is shown in Fig. 3. Water
formed by urea synthesis and fed to the recovery section as an
absorbent is evaporated under vacuum in the concentrator/evaporator.
Evaporated water unavoidably entrains the mists of urea. The
evaporated water vapor is first condensed in the multistage surface
condenser to form a process condensate. Uncondensed gas from the
surface condenser is vented to the atmosphere through the offgas
scrubber, where NH3 and CO2 accompanied by inerts are finally
absorbed into the process condensate under atmospheric pressure.
Inerts purged at the medium-pressure level and gas released from the
storage tank are also led to the final absorber to remove
accompanying NH3. Thus the emission of NH3 is reduced to less than
0.15 kg NH3/ton of urea product.
Most of the process condensate formed in the first-stage surface
condenser is used for absorbent in the recovery section because its
concentration of urea is higher than that in the other part of the
process condensate. The other part of the condensate is first treated in
the process condensate stripper, operated at 3 kg/cm2G, for separation
of NH3 and CO2 by steam stripping. Then the process condensate is
extracted from the middle stage of the process condensate stripper, a
part of which is delivered to the dust scrubber of the prilling tower (or
granulator), and the rest is fed to the urea hydrolyzer, operated at 16 to
20 kg/cm2G. Figure 4 shows relation between residual urea
concentration and retention time of the process condensate in the urea
hydrolyzer. The overhead gaseous mixture from the process
condensate stripper is sent to the low-pressure decomposer in the
decomposition section for recovery of heat, NH3, and CO2.
Another source of liquid pollution is leakage of process fluid into
gland cooling water through the stuffing boxes of the reciprocating
NH3 feed pump and carbamate pump. The circulating cooling water is
contaminated with urea, NH3, and lubrication oil. This can be
eliminated, however, when centrifugal pumps are used for high-
pressure NH3 and carbamate services instead of reciprocating pumps,
because cooling water in circulation is not used for these pumps. The
leakage to gland cooling water through the seals of the centrifugal
pumps in the concentration and finishing sections causes a high
concentration of urea and that in the recovery section causes a high
concentration of NH3 in gland cooling water, depending on the type of
seal and the degree of maintenance.
To abate liquid effluent pollution problems caused by contamination
with urea and NH3, a closed circulating cooling water system can be
adopted for gland cooling. The lubrication oil in the gland cooling
water is removed through an oil separator installed on the cooling
water circulation line and circulated back to the pumps after cooling.
Blow-down of this cooling water with concentrated NH3 and urea is
discharged to the process condensate treatment section for purification
of the cooling water after further removal of the lubrication oil
through another on-line oil separation unit.
B
Gaseous Effluents
A major source of gaseous effluent is urea dust in the effluent air from
the prilling tower or granulator, which, without treatment, contains
urea dust in the range 100 to 150 mg
 Page 115

Figure 3
Pollution control system.
 Page 116

Figure 4
Urea hydrolysis.
 Page 117
urea/N·m3 air. This means that the amount of urea emitted to the
atmosphere from a 1700-metric ton/day urea plant reaches about 1.7
to 2.5 metric tons/day. Recent trends in strict environmental regulation
make a dust recovery system a standard installation.
Technical and economical factors involved in the development and
use of the dust removal systems are (1) the fact that the amount of air
to be handled is massive; (2) that dusts contain extremely fine
particles, under 1 µm; and (3) that a high-pressure drop in the dust
removal system increases the cost of operation. Therefore, maximum
dust collection efficiency with minimum pressure drop through the
system and low installation cost are essential.
A packed bed type of dust recovery system that is compact and
lightweight and can be installed on the top of a prilling tower was
developed and has been adapted for use with a prilling tower and
granulator. Such a dust recovery system is illustrated in Fig. 5. A
packed bed installed in a scrubbing zone increases the effective
interfacial area where the urea dusts in the effluent air coming from a
prilling tower or granulator are scrubbed and caught by direct contact
with a sprayed aqueous urea solution of approximately 20 wt %. The
mists of sprayed solution are caught by a demister located over the
spray nozzles. Urea dusts recovered as urea solution are returned to
the urea plant.
C
Integrated Pollution Control System
A combination of the pollution abatement systems described above
has resulted in a practically nonpolluting urea plant with the following
reasonable emission of pollutants:
 Figure 5
Urea dust scrubber.
 Page 118
1. Air pollution
Urea dust: 3050 mg urea/N·m3 air
2. Water pollution
Urea content: 15 g urea/m3H2O
NH3 content: 15 g NH3/m3H2O
The utility requirements for the foregoing system are estimated as
follows:
1. Steam: less than 20 kg/ton of product
2. Electricity: less than 6 kWh/ton of product
 Page 119

11
Control of Water Pollution and Water Treatment in the
Nitrogen Fertilizer Industry
Lucian Negulescu
Polytechnical University
Bucharest, Romania
I
Effects on Natural Waters of Pollutants from the Nitrogen Fertilizer
Industry
Depending on the operation and maintenance of nitrogen fertilizer
plants, nitrogen loss is estimated at 0.5 to 1.0% in well-operated
modern plants and at 3 to 4% in older plants [1]. The waste effluents
from the industry can be classified into four groups [2]:
1. General manufacturing effluents resulting from the contact of water
with gas, liquids, or solids. Generally, these effluents are continuous
and are known as pollutant contributors.
2. Particular effluents likely to be separated either for a specific
treatment, eventually accompanied by recovery of useful substances,
or to be stocked and reinjected to the treatment circuit at a controlled
flow rate.
3. Effluents from general services such as steam blow-down,
pretreatment sludge, and washings.
4. Occasional effluents, fugitive leakages of products while being
handled or stocked, platform washings, and polluted precipitations.
Wastewater from the nitrogen fertilizer industry is an active hazard to
public health, and especially to aquatic life. The ways in which
fertilizer industry pollutants get into natural waters are diverse.
Pollutants get into rivers or stagnant water through uncontrolled
disposal of wastewater and by precipitation washing on in-plant
platforms. Underground water gets contaminated by polluted seepage
coming from precipitations or washings from industrial platforms and
buildings, pump leakages, pipe leaks, sewage, and outseepage from
treatment ponds or storage areas.
The nature and composition of wastewater varies depending on the
manufacturing technologies used. The composition may include
suspended solids (SS), ion dissolved material (acids, bases, salts),
mineral oils, fats, and a wide range of toxic and poisonous substances
[14,15].
 Page 120
Once in water, the nitrogen fertilizer pollutants NH+ and NO3 affect
the quality of natural water, as shown first by the increase in
biological oxygen demand (BOD) and by unpleasant organoleptic
effects such as odor and taste [3].
A special phenomenon is the eutrophication of nitrogen, especially of
phosphorus receiving water. Eutrophication depends on the buildup of
nutrient substances in waters, rivers, and lakes. It is especially
outstanding in lakes and in sluggish rivers, in off-coast canals and
lagoons, in bays, and generally in all waters of slow change subject to
ample disposal of nutrients (N and P) [4].
II
Sources of Wastewater from the Nitrogen Fertilizer Industry
An ammonia plant is the core of a nitrogen fertilizer complex. Plants
producing HNO3, NH4NO3, (NH2)CO, and other complex fertilizers
work in conjunction with the ammonia plant. The pollutants present in
wastewater from the manufacture of the three products noted above
are listed in Table 1. To this must be added pump and piping leakage,
the source of additional pollutants in precipitations, washings, and
platforms.
A
Ammonia Plant
As shown in Fig. 1 of Chapter 5, wastewater originates from the
stripping of ammonia from the process condensate following low-
temperature conversion. For 1000 tons of NH3 per day, the following
effluent loss occurs at this stage: ammonia, 0.57 kg/h; methanol, 0.28
kg/h; and metals, traces of Na, Fe, Cu, Zn, Ca, and Al [2]. This does
not include random loss. Process condensate forms during synthetic
gas cooling.
To bring waste effluents to standard levels, several manufacturers
have incorporated condensate stripping to reduce the ammonia and
methanol concentrations to levels acceptable for disposal in the
receiving stream [2]. Thus a condensate with an average content of
0.5 g/kg ammonia amounts to 4.75 g/kg in the overhead and 10 ppm
in the effluent. Table 2 shows the average concentration of metals
discarded into the stripping effluent and their ratios to the product unit
[7].
B
Ammonium Nitrate Plant
As shown on the flowsheet for ammonium nitrate manufacture (Fig.
1), wastewater comes from (1) the neutralization process, (2) the
evaporation unit, and (3) air cooling equipment.
Table 1 Process Water in the Nitrogen Fertilizer Industry
Section Wastewater Pollutants
Ammonia 1 m3 condensate/ton 14 g/L NH4HCO3
product 0.21 g/L CH3OH
Ammonium 0.5 m3 basic 23 g/L NH4NO3
nitrate condensate/ton product
Washings from 0.20.8 g/L NH4OH, NH4NO3,
platforms, workshops suspended solids
Urea 0.6 m3 condensate/ton 0.10.5 g/L NH3
product 0.52 g/L (NH2)2CO
Source: Ref. 2.
 Page 121

Figure 1
Ammonium nitrate process scheme.
In addition, leakage from pumps, tanks, piping, and valves, and
leakage of both ammonia and prilling powder carried into the sewage
system by precipitation, constitute other potential sources.
About half of all U.S. ammonium nitrate plants experience condensate
process leakages; the other half either recycle the condensate to the
process or combine it with fertilizer liquors. Consequently, only 23%
of the plants subject the process condensate to treatment (Fig. 1).
C
Urea Plant
In a urea plant the primary wastewater source is the condensate from
the vacuum condenser (Fig. 2), and where there are no air strippers,
from the prilling powder.
Table 2 Metal Traces in Condensate
Stripping Effluent
Specific
Average disposal
Metal concentration (g/kg
(ppm) product)
Chrome 0.2 4 × 105
Copper 0.02 4 × 106
Iron 0.1 2 × 105
Zinc 0.02 4 × 106
Nickel 0.2 4 × 105
 Page 122

Figure 2
Urea flowsheet.
D
Admissible Limits for Disposal of Nitrogen Fertilizer Industry
Pollutants
As discussed above, most of the wastewater from nitrogen fertilizer
manufacture consists of clear solutions, condensates, which in
moderate concentrations contain products and intermediates from the
process, but a second source, which although much smaller cannot be
ignored, is that of leaks from gaskets, piping, and valves. Platform
washings and accidental precipitations constitute a category worth
considering separately. In addition to specific contaminants and
suspended solids, the latter may contain oils, vegetative remains, and
possibly, washing agents. Table 3 shows wastewater from a variety of
plants.
Table 4 shows the U.S. Environmental Protection Agency standards in
force as of July 1, 1984 for existing and new ammonia, ammonium
nitrate, and urea plants.
III
Physicochemical Treatment of Wastewater from the Nitrogen
Fertilizer Industry
A
Pretreatment
When platform washings, precipitation from various industrial wash
processes, and other leakage are included in the total amount of
wastewater (process condensate), the waters are subjected to settling
and filtration (by sand filters and active carbon filters) when it is
desirable to demineralize them by ion exchange. As these operations
are conventional and their characteristics are well known, they are not
discussed further here.
B
Treatment of Process Condensate from Ammonia Manufacture
Over the past few years special interest has been directed toward
condensate treatment with regard to the prevention of air and water
pollution. Condensate treatment and
 Page 123
Table 3 Types of Wastewaters
Process Specific wastewater
Ammonia Condensates from steam stripping and
plant scrubber blow-down that may contain
ammonia, ammonium bicarbonate,
methanol, and other contaminants (metal
traces)
Ammonium Condensates from neutralization,
nitrate and evaporation, blow-down from air scrubbing;
calcium and may contain HNO3 and NH4NO3
ammonium
nitrate plant
Urea plant Discharge from crystallization and
evaporator condensate that may contain
ammonia, ammonium carbonate and
carbamate, and urea
Nitric acid Pump, valve, and piping emission and
plant leakage; normally, there is no wastewater
Cooling Blow-down from cooling towers may
system contain all the contaminants from the entire
fertilizer complex, including cooling-water
conditioners (corrosion inhibitors, scale
removers, biocides)
Leakage Gasket, piping, pump, and tank leaks
(uncontrolled
release)
Repair and Prior to major repair, all piping and tanks
maintenance are drained and the platform areas are
affected
Leakage, Indoor and outdoor platform surfaces
release contain nitrogen compounds that may be
washed off by precipitations

reutilization, especially as makeup water for steam boilers, has

considerably reduced the amount of demineralized water being used.
As mentioned earlier, there are two major sources of process
condensate: (1) the effluent circuit, which contains dissolved gas,
large quantities of ammonia, CO2, methanol, and light organic
compounds, and (2) effluents from the methanator, which contain
large quantities of dissolved gas and little ammonia.
Table 5 shows the average composition of process condensate.
Condensate also contains traces of metals, which for the most part
come from catalytic entrainments. The major contaminantsNH3, CH3
OH, and CO2should be removed before the condensate is considered
fit for feeding steam boilers.
Figure 3 illustrates an older condensate treatment method. This
method is based on a surface evaporator cooler. In this system the
condensate is entirely or partially evaporated in the heat exchanger,
the contaminants being released to the atmosphere. CO2 from the
monoethanolamine (MEA) system serves as a heat carrier. The
condensate is sprayed over the exchanger tubes, and air blown in at
the same rate of flow acts as a carrier of steam and contaminants.
Because of the corrosive nature of the condensate, the heat-exchanger
tubes and tube
 Page 124
Table 4 Daily and 30-Day Effluent Valuesa
Effluent limitsb (kg/1000
kg)
Effluent Daily 30-day
Plant variables average average
Ammonia Ammonia N 0.050.185 0.0250.0625
pH 6.09.0 6.09.0
Urea liquorc Ammonia N 0.530.95 0.270.48
Organic N 0.450.61 0.240.33
pH 6.09.0 6.09.0
Solid urea Ammonia N 0.531.18 0.270.59
Organic N 0.861.48 0.370.80
pH 6.09.0 6.09.0
Ammonium Ammonia N 0.060.73 0.040.39
nitrated Nitric N 0.120.67 0.070.37
pH 6.09.0 6.09.0
aThe two values, expressed as a range, represent the best
conventional control technology and the best available
control technology.
bBased on 100% product.
cDispatch loss, cooling-tower blow-down, washing, and
precipitation are not included.
dValid for plants that produce liquid and solid fertilizers,
for precipitation, for cooling-tower blow-down, and for
plants with integral use of overhead condensation from
neutralization.

sheets are made of stainless steel. The freshwater makeup depends on

the amount of condensate to be treated. As the water contains
dissolved salts, which by concentration may cause deposits, blow-
down is necessary. Owing to the relatively high volatility of the
contaminants in the process condensate, they will be present in the
blow-down stream within allowable limits. Figure 4 shows another
alternative of an older scheme. This process uses a reboiler
condensate stripper coupled with a CO2 removal system using a hot
K2CO3 solution. In this case the stripper overhead vapor, containing
considerable water and most of the condensate contaminants, is used
as part of the heat carrier for the CO2 regenerator. The contaminants
are finally eliminated along with the CO2 stripped from the recycled
solution. The amount of condensate processed by this system depends
on the requirement for the reboiler selected for the condensate stripper
and on the heat balance of the CO2 regenerator. In this system CO2 is
partly released to the atmosphere, along with the stripping steam and
condensate contaminants.
Table 5 Average Composition of
Process Condensate in Ammonia
Plants
Compound Concentration
(ppm)
Ammonia 8001000
Methanol 10002000
CO2 15002000
Organic 3550
matters
 Page 125

Figure 3
Condensate treatment system with wet cooler.

Figure 4
Alternative condensate treatment.
 Page 126

Figure 5
Use of condensate for medium-pressure steam generators.
Another scheme uses the condensate as makeup water for the
medium-pressure steam boilers, as illustrated in Fig. 5. In this process
the dissolved gas in the condensate is removed through a deaeration
stage. The steam developed in this system, the steam containing the
condensate contaminants, is used for the reforming process.
Another scheme, used in Kellogg plants (Fig. 6), includes a
condensate stripper fitted with Pall ring packings [8]. The
contaminants are stripped out of the condensate using
 Figure 6
Single stripping system for process condensate.
 Page 127
low-pressure steam. The overhead product, containing ammonia,
carbon dioxide, methanol, and stripping steam, is released to the
atmosphere, thus transferring pollution from water to air.
In the stripped condensate the ammonia content is reduced to about 40
to 50 ppm, and that of methanol to 20 to 100 ppm. The CO2 residual
gas content depends on the ammonia concentration. The condensate
may be used as makeup water to the cooling towers and makeup water
to feed the steam boilers, after some finishing to remove traces of
metals.
While plain stripping of the condensate proceeds satisfactorily, the
process has quite a number of disadvantages, the worse being that it
does not permit full use of the condensate and requires vast amounts
of steam, with no provision for the recovery of waste heat. Based as it
is on contaminant release to the atmosphere, under certain conditions
this process becomes a source of pollution. A substantial reduction in
utility costs is possible if the stripper overhead heat is employed as a
carrier for a CO2 regenerator reboiler (Fig. 7). It should be noted that
after heat recovery in the CO2 removal system, the stripper overhead
condensate is diverted to a seawater cooling system for plants located
on the coast. This process is designed to be utilized in coastal areas to
save on high energy costs but it can also be useful in water-short
zones.
Two processes have been developed for conditions that require
removal of pollutants from water and from air, as well as reutilization
of the condensate. Either process uses the condensate refluxed by a
stripper, running with high heat recovery. The first alternative (Fig. 8)
provides refluxing of concentrated contaminants in the condensate as
an overhead vapor product, which is then incinerated or processed in a
separate ammonia recovery unit. The stripping effluent may be
recycled to feed the steam boiler after flowing through a metal-
removing system, or may serve as makeup to the cooling towers [16].
Residual methanol in the stripped condensate is not removed, but the
condensate, along with fresh water, is allowed to enter the boiler feed
system. Past experience has revealed no operating difficulties caused
by the very low methane concentrations in feedwater.

Figure 7
Modified stripping system for process condensate.
 Page 128

Figure 8
Process condensate stripping system with overhead reflux.
The efficiency of contaminant removal (NH3, CO2, CH3OH) by the
process shown in Fig. 9, as well as by the processes shown in Figs. 7
and 8, is 95%. Better efficiencies are achievable by increasing the
amount of stripping. The amount of steam used for such operations is
96 to 240 kg steam per ton of condensate treated. The quantity of
steam can be reduced by preheating the influent condensate with the
hot stripper effluent.
The second alternative (Fig. 9) improves the performance of the
preceding alternative. This process enables 100% recycling of the
process condensate. It also uses a better system to recover the product
and the top reflux. The condensates originating from various sources
are directed, after being preheated, to the condensate stripper, which is
equipped with a reboiler to minimize the water inventory. The stripper
is fitted with an overhead refluxing system coupled with the CO2
regenerator reboiler.
The stripper pressure ratio is determined in this scheme so as to attain
the condensation temperature at the top corresponding to the
requirements of the regenerator reboiler, which normally runs at 115
to 120°C. These temperatures comply with MEA or K2CO3 processes
used to remove CO2. To accomplish this process feature, higher-than-
normal pressures, on the order of 1.4 to 2.1 bar or equivalent to the
dew point (127 to 135°C), have to be used to achieve a thermal power
of 11 to 14°C for the regenerator boiler. The small quantity of liquid
withdrawn from the condensate stripper, containing water, ammonia,
CO2, methanol, and traces of organics, serves as a recycling flux. As
all contaminants,
 Page 129

Figure 9
High-efficiency process condensate recycling system.
including CO2, NH3, and CH3OH, are reformable, the catalyst
encounters no difficulty. The condensate effluent from the stripper
may be employed for several purposes: as makeup water for cooling
towers, as feed water for the water jacket of the secondary reformer (if
any), or, after pretreatment, as makeup water fed to the steam boilers.
Briefly, operation is based on the use of condensate in the saturation
column for the raw natural gas. Using this technique, the entire
process condensate, with contaminants included, is vaporized and thus
serves as a reboiler feed source. The use of a saturator for recycling
the process condensate offers several advantages:
1. No reforming steam is necessary other than that normally used in a
reforming operation, due to use of the sensible condensate heat.
2. Makeup water to feed the steam boilers associated with reformer
operation is reduced by more than 50% for an ammonia plant.
3. Problems related to condensate storage and treatment are eliminated
completely.
The schemes shown in Figs. 9 and 10 are worth considering for
locations governed by severe environmental protection laws and/or
where there is a water shortage.
To conclude, we note that many of the schemes shown are now being
used in large ammonia plants. Some have been used to eliminate
water and/or air pollution without condensate recycling. Today the
situation has changed somewhat and economics is playing a major
role. Recycling the condensate reduces the number of water
demineralization stations. The contaminants contained in the
condensate may easily be removed by steam stripping and the system
can be adapted to heat recovery [7].
 Page 130

IV
Biological Treatment of Wastewater from the Nitrogen Fertilizer
Industry
A
Nitrification
As mentioned earlier, nitrogen is the principal element found as
ammonia in wastewater. Ammonia nitrogen is oxidized successively
(always biologically) according to the following simplified reaction:

and then, also biologically, nitrites are oxidized to nitrates:

Cumulatively, the process may be expressed by a simplified reaction:

It is obvious that the mechanism making modifications in the nitrogen

compounds in wastewater is the cause of consistent fractioning of
BOD, which is added to the carbonic fraction considered for
biologically activated sludge treatment stations. Ammonia and its
organic compounds are converted to nitrates by the conjugated action
of nitrifying bacteria, autotrophic bacteria, and strictly aerobic
bacteria, acting sequentially.
Nitrosomonas bacteria, which oxidate ammonia to nitrites following
reaction (1), and Nitrobacter bacteria, which oxidate nitrites to nitrates
according to (2), participate in this process. The nitrates resulting from
oxidation appear as solubles in the effluent or in the sludge.
Nitrosomonas bacteria reproduce far more slowly than do Nitrobacter
bacteria; therefore, the growth rate of Nitrosomonas is the limiting
factor in the process of nitrification.
Experiments conducted by Downing and his collaborators in England
have revealed that the growth of Nitrosomonas bacteria is
characterized by a well-determined daily rate, which in its turn is
lower than that of the bacteria that cause oxidation of the carbonic
fraction of the organic matter contained in wastewater. When the
sludge growth rate gets higher than that of Nitrosomonas bacteria (in
the case of too young a sludge), the bacteria are fully eluted into the
effluent together with the sludge excess (which is removed faster than
their growth rate); consequently, there can be no nitrification.
Downing's experiment has shown that the daily growth rate of
Nitrosomonas is governed by pH, dissolved O2 concentration, and
temperature [4]. Slightly alkaline pH favors growth. By nitrification,
pH diminishes; thus suitable alkaline buffering becomes necessary.
The effect of dissolved O2 concentration is illustrated convincingly in
Fig. 10, where it can be seen that maximum nitrification is obtained at
3 ppm O2 and that below 0.5 ppm O2 it is practically null.
To understand the temperature effects one should examine Fig. 11,
which is based on Downing's experimental data. The graph shows the
daily growth rate of Nitrosomonas and the minimum age of the sludge
versus temperature. The minimum sludge age versus temperature for
nitrification purposes can be expressed by the formula

where A is the sludge age in days and T is the temperature in °C.

Bearing in mind the fact that the nitrification entity depends on the
sludge age and
 Page 131

Figure 10
Evolution of nitrification rates at
varying dissolved oxygen concentrations.
that the latter is related to the organic load factor Fc, it is clear that the
nitrification entity is strictly dependent on this factor. Assuming
sufficiently low factor values Fc, highly efficient nitrifications can be
obtained, provided that the temperature is fairly high and the
dissolved oxygen is of suitable concentrations.
Unoxidized nitrogen compounds as ammonia exert many undesirable
influences once they have entered the receiving stream: (1) they
consume dissolved oxygen, (2) they are toxic to aquatic life, and (3) if
the stream of receiving water downstream from the discharge point is
made potable, the nitrogen compounds cause a high chlorine
consumption (1 kg of ammonia requires 10 kg of Cl2) for sterilization.
When dimensioning aerators, a conservative estimate is that ammonia
loads in peak hours coincide at least temporarily with peak biological
oxygen demand (BOD) hours of the carbonic fraction. The peak
coefficients of the ammonia load may become as high as 1.7 to 1.8 of
the average values (Table 6). This criterion was adopted by Vasloo
[8], who to evaluate the nitrogen oxygen demand (NOD) peak for
nitrogen compounds used in the complex oxygen demand calculation
used the same incremental coefficient as that of the daily average
coefficients listed in Table 6.

Figure 11
Nitrosomonas bacteria growth curve and minimum sludge
age to avoid washing off at varying temperatures.
 Page 132
Table 6 Maximum and Minimum
Correction Coefficients for Fc
Organic load factor Fc Correction
(kg BOD/kg SSAMa coefficient
× g) MaximumMinimum
0.10 or lower 1.50 0.50
0.15 1.60 0.50
0.20 1.70 0.50
0.25 1.80 0.50
0.30 1.90 0.50
0.40 2.00 0.50
0.60 2.20 0.50
0.80 2.40 0.50
1.00 or higher 2.50 0.50
aSSAM, suspended solids in air mixture.

Taking into account the BOD concentrated peaks for the carbonic and
nitric fractions, one can thus dimension aerators to meet the peak
oxygen demand, which may be three times the average figure and up
to one-third of the hourly average. It is this requirement that should be
considered when adopting aerators flexible enough to withstand easily
ample load variations. Figure 12 shows one type of activated sludge
unit, an aeroaccelerator, with sludge recycling [4].
In conclusion, one must say that nitrification is applicable where
wastewater from in-plant units mixes with organic mattercontaining
water (domestic water and the like) and when nitrification is followed
by nitrification.
B
Denitrification
As discussed in Sec. IV.A, nitrates formed by the oxidation of
ammonia compounds make up the oxygen stock not only in receiving
water but even within the same treatment station.

Figure 12
Activated sludge unit with a single pond, with central
aeration sections, peripheral settling section, aeration
and air inblowing by immersed turbine, natural recirculation.
 Page 133
In cases when the sludge remains without aeration for a longer period,
to satisfy their oxygen demand bacteria may attack the oxygen in the
nitrate molecules, thus causing the nitrates to act as a secondary stock
of the oxygen required for biological reactions, in the absence of
sufficient oxygen supplied by the aerating devices in exchange with
the atmosphere.
The absorption of oxygen from nitrates is similar to that in a
nitrification process in the sense that the reduction reaction caused by
bacteria leads to the formation of molecular nitrogen, released from
the liquid according to following simplified reaction:

This denitrification reaction is effected by numerous heterotrophic

bacteria (Pseudomonas, Micrococcus, Achromobacter, Bacillus
spirillum). These are optional: in the presence of dissolved oxygen,
they use oxygen as such, while in the absence of dissolved oxygen in
anaerobic conditions, they use oxygen from nitrates, thus releasing
nitrogen according to reaction (5) and ultimately affecting
nitrification.
In addition to oxygen, to grow, these bacteria need food, which
comprises a sublayer made up of organics such as methyl alcohol,
saccharose, and glucose that are to be added thereto. As shown in Fig.
13, the growth rate of bacteria differs according to the availability of
food, as does the denitrification rate. On the other hand, denitrification
entails an increase in pH with the alkalinity produced, about 3.5 parts
of CaCO3 to 1 reduced part of NO3N.
Figure 14 shows a scheme for controlled denitrification in a slurry
culture reactor. The denitrification reactor is followed by an aeration
channel that serves to strip off nitrogen bubbles stuck to sludge flakes;
in this manner an aerobic condition is reestablished in the liquidsludge
mixture, and it is possible to avoid nitrification in the final settling
stage, where the bottom-settled sludge may float up, which can cause
problems. As shown in the figure, upstream from the denitrification
pond there is a methanol supply system feeding outside carbon as a
nutrient to bacteria. The settling stage should be dimensioned

Figure 13
Evolution of denitrification rate using various
organic sublayers for denitrifying bacteria.
 Page 134

Figure 14
Controlled denitrification process in anaerobic reactor with suspended growth.
for an upward speed of 1.5 m/h maximum, similar to that of activated
sludge ponds; the recycling flow rate is higher, attaining 50 to 200%
of the influent flow rate (recycling ratio r = 0.5 to 2).
Next we discuss an original process that has been applied to an
industrial site for the treatment of wastewater containing ammonium
and nitric ions. The technology used started from economical
considerations: lower material consumption (organic carbon,
neutralizers) and lower power consumption and investment costs. The
originality of the method lies in the treatment of NO3 ion directly
from NH4NO3 by activated sludge biological denitrification in a
homogeneous system or immersed biofilter followed by the stripping
off of NH4 +.
In the activated sludge alternative in a homogeneous system work was
carried out under the following conditions:
1. The activated sludge concentration was kept at 2 to 3 g/L (dry
substance sludge).
2. To influent water was added a biogenic solution in an N-NO3/P
ratio of 200:300, where N-NO3 is nitric acid and P is inorganic
phosphorus.
3. The organic carbon sourcemethanolwas varied in proportions.
4. The temperature was maintained at about 20°C. In cold zones, a
source of heat was necessary.
5. The pH in the denitrifier was maintained at about 7 by CO2
bubbling.
6. The dissolved oxygen was below 1 ppm.
7. The retention time was 20 hs.
8. The treatment efficiency for NO3 ion was 99.0%.
After denitrification the effluent subjected to stripping contains on the
average: NH4, 780 ppm; NO2 , below 0.1 ppm metal hydroxides,
about 0.5 ppm; COD, about 50 ppm; HCO3 , 1000 ppm; and CO3 ,
400 ppm. Denitrification was conducted in a self-mixing tank similar
to that used in the biomethane technique.
VI
Ion-Exchange Treatment
Ion-exchange treatment of polluting effluents from the nitrogen
fertilizer industry is successful due to lower treatment costs, which
implies recycling of pollutants and water (demineralized water) for
technological purposes or to feed steam boilers [12].
Under certain conditions ion-exchange treatment may be the most
economical method available because it permits:
 Page 135
1. Full recovery of the nitrate products contained in wastewater.
2. Regeneration of ion resins with products available in the plant
(NH3, HNO3).
3. Utilization of ammonium nitrate liquor from the regeneration of ion
resins by recycling to the manufacturing circuit.
4. Production of demineralized water, good for immediate use.
These features make ion-exchange treatment perfect for use in a
nitrogen fertilizer complex. For technoeconomical reasons the ion-
exchange procedure should meet the following demands:
1. The waste effluents should be treatable at varying low rates and
concentrations (ion charge from 2 to 15 eq/h, concentrations from 12
to 100 eq/m3).
2. The eluates produced should have the highest possible regeneration
rates, to minimize evaporation concentration costs.
3. Investment and operating costs should be as low as possible.
A
Theoretical Considerations
Ion-exchange treatment involves two major operations: (1) retention
on the resin of desirable ions from the water, and (2) elution of
retained ions by regeneration and restoration of the resin to its initial
ion form. During the ion-retention operation, waste effluent
containing ammonia and ammonium nitrate flows through two contact
columns, one equipped with strongly acid cation resin, the other with
weakly basic anion resin. In the strongly acid cation resin column the
following reaction occurs:
The effluent of the former column, which contains nitric acid, flows
into the weakly basic anion column, where the following reaction
takes place:

The quality of the effluent treated following the two columns depends
on ion leakage, which to be minimized. It is important that the first
resin bed the liquid flows through should as well regenerated as
possible.
In the ion-exchange process equilibrium is created between H+ ions of
the active resin group and NH4+ ions in the liquid. The equilibrium
constant is expressed as

Where the overbars indicate concentration in resins and the unbarred

symbols indicate concentration in liquids. In terms of the leaking ion
(respectively, ammonium) for the cation resin, the following equation
is derived:

For minimum leakage the relevant factor is certainly the resin

concentration ratio, For it to be as low as possible necessitates
that the resin be very well regenerated so that [H+] can be as high as
possible.
The exchange is different in the weakly basic anion resin since the
active groups therein (NH2) get neutralized and nitric acid forms
through cation exchange, as shown in
 Page 136
equation (7). Given the low dissociation of the active groups of
weakly basic anion resins, ion exchange depends to a great extent on a
pH value above 12 (respectively, on the order of magnitude of NH4 +
ion leakage) into the cation column effluent. This leakage is
influenced by [H+] concentrations in the wastewater flowing into the
cation column, which may cause partial self-regeneration. It may
sometimes be necessary to dilute such water. For the reasons noted
above, countercurrent regeneration is mandatory in ion-exchange
treatment to ensure a perfectly regenerated bed (Fig. 15). If in the
exhaustion stage, the treatment of waste effluents with ammonium and
nitric ions is identical to water demineralization in conventional
plants, regeneration is altogether different from that for exhausted
resins in an effluent treatment station in the nitrogen fertilizer industry
and requires special techniques and equipment.
The motivation supporting process efficiency is to produce as
concentrate regeneration products as practicable. During regeneration
the eluate concentrations vary according to the curve shown in Fig.
16. Curve 1 represents the concentration variations of the eluate from
the regeneration process, where P is the maximum concentration and
the S1 surface is the total regeneration eluate. Curve 2 represents the
concentration variations of the excess regenerating agent (above the
stoichiometric requirement) and S2 the excess reagent.
While S1 depends on the useful ion-exchange volume capacity of the
resin, the eluate volume V depends on several factors: resin type,
regenerating agent concentration, and the operational and constructive
features of the column, which include resin bed depth, regeneration
direction, distribution system, and recovery of regeneration eluates.
The regeneration eluate is split into three portions (Fig. 16):
 Figure 15
Countercurrent ion-exchange system with upward fixing and
downward regeneration. 1, exhausted resin zone; 2, ion-
exchanging upward zone; 3, regenerated resin zone;
4, polishing zone; 3¢, regenerated resin zone; 2¢, ion-
exchanging descending zone; a, analyzer; b, breakpoint.
 Page 137

Figure 16
Elution curves for ion-exchange resin. (From Ref. 11.)
1. F1: cut for volume V1 which has a low concentration of product (3
to 5%). This cut is directed to the waste effluent storage tank.
2. F2: cut for volume V2, which contains most of the eluate (i.e., 85 to
95% and about 50% of the excess regenerating agent). This
regeneration cut is neutralized and concentrated and then returned to
the fertilizer manufacture.
3. F3: cut for volume V3, which contains the excess regenerating agent
and a small amount of product and can be used for the preparation of a
regenerating solution or be sent to the waste effluent storage tank.
B
Parameters Governing Regeneration Eluate Concentrations
One of the most significant features of ion-exchange resins is their
exchange capacity. In conventional water treatment stations,
relinquishing some exchange capacity in favor of enhanced
regeneration efficiencies leads to larger resin volumes and bulky
equipment. However, lower regenerating agent quantities often
compensate for the supplementary equipment.
Treatment of waste effluents from nitrogen fertilizer plants requires
large amounts of regenerating agents to enhance the exchange
capacity while obtaining the highest concentration of regeneration
eluates together with minimized ion leakage. Experiments have shown
that quadrupling the amount of regenerating agent (i.e., from 10 to
45%) may cause concentrations to rise only twofold. On the basis of a
theoretical calculation of the material balance, Harries [11] has
concluded that a regenerating agent concentration of 5 to 6 eq/L is
best to attain maximum eluate concentrations; concentrations above 6
eq/L have no influence on regeneration eluate concentrations.
To use strong nitric acid (47 to 60%) as a regenerating acid, Rohm &
Haas (formerly Diaprosim) has developed a special resin grade known
as Duolite C265, which is completely resistant to such concentrations.
Figure 17 shows the behavior of this resin to nitric acid of various
strengths. During regeneration the resin should briefly be put into
 Page 138

Figure 17
Influence of temperatures and concentrations
of nitric acid on Duolite C265 resin.
contact with the regenerating agent at low temperature. This is a
requirement for system safety and the long service life of the resin.
Since it has been determined that in regeneration, the best velocities
are small (0.9 to 3 m/h) and to avoid the disadvantages of such
velocities (i.e., less uniform distribution of the liquid and appearance
of preferential channels), the regeneration geometry has been
optimized to obtain very deep beds. In practice, velocities of 10 to 20
m/h and beds 3 to 9 m deep have been achieved. Countercurrent
regeneration in special equipment is preferable, to keep the resin bed
compact and to avoid fluidization.
External regeneration can also be used where both exhausting and
regeneration are effected by downward current. Modifications have
been made in the distribution and discharge systems of liquid leaving
the column, in terms of uniform distribution across the filtering
section and avoidance of regenerating agent dilution.
C
Industrial Applications
Of the best known industrial systems, two are considered to have the
widest range of applications: the short-cycle and continuous systems.
The continuous system [12] shown in Fig. 18 uses a contactor, from
which resin is extracted in small portions. These portions are
regenerated with 20 to 22% strong HNO3 for the cation resin and with
7% strong NH3 for the anion resin. Regeneration occurs continuously
in the contactor, which contains four or five portions of resin in
different regeneration stages. Washing is also effected nonstop in an
outside area and covers three to four portions. The regeneration
eluates obtained by this process contain 15 to 18% strong NH4NO3.
D
Short-Cycle Systems
Short-cycle systems have been conceived especially for the treatment
of waste effluents from nitrogen fertilizer plants. In one Arion system
the exhaust phase is effected by an upward flow over short cycles
lasting 20 to 80 min. Exhausted resin is regenerated
 Page 139

Figure 18
Continuous ion-exchange system.
externally in a downward flow by 47 to 60% strong HNO3 for
cationite and by 15 to 25% strong NH3 for anionite. Regeneration
eluates contain over 25% strong NH4NO3 [12].
Another Arion system (Fig. 19) has been implemented by Bran-Lubbe
in Germany. This system has been used since 1982 in Inakutina's
plants in Yugoslavia at a flow rate of 190 to 285 m3/h, a cation charge
of 8000 eq/h, an anion charge of 5200 cal/h, and an influent
concentration of 18 to 37 eq/m3. The resins used are strong Duolite
C265 cationite, weak Duolite A374 anionite (Rohm & Haas, France),
and the regenerating agents are 60% HNO3 and 16% NH3 [14]. Arion
short-cycle systems are highly efficient, as they produce regeneration
eluates of concentrations in the range 25% (for cationite) to 30% (for
anionite).
Figure 20 shows a complex treatment station for waste effluents,
based on a UFD conception put into practice by the Degremont
Company in France [2]. The same company has developed the
continuous-process FCI shown in Fig. 21. A particular case is the
treatment of condensate from the manufacture of ammonium sulfate,
illustrated in Fig. 22. By regeneration of the cationite with sulfuric
acid, the regeneration effluent may reach a concentration of about
27% (NH4)2SO4; this is recycled to the fertilizer manufacturing
process.
VI
Other Treatment Methods
In a complex that produces urea and ammonium nitrate
simultaneously, if ion exchange is the primary method for recovering
NH3 and NO3 , stripping of free NH3 can be
 Page 140

Figure 19
Short-cycle ion-exchange system.
considered useful provided that its concentration is fairly high and that
steam recovery is possible. Urea effluent hydolysis should also be
used.
Ultimately, biological oxidation of ammonium and potential removal
of nitrates (a contribution of organic nutrient requirement) should be
ensured. The scheme in Fig. 23 shows a possible arrangement of
treatments and pretreatments [2].
A recently patented German method consists of ion exchange to
remove NO3 from water subjected to potabilization. The Carix
(carbon dioxide regeneration of ion exchangers) process of Rohm &
Haas in France uses carbon dioxide as the regenerating agent [15].
Process originality resides in the selection of resins that should be
regenerable with a weak acid such as CO2. The resins used have been
developed by Rohm & Haas: IMAC HP334, a weakly acid carboxylic
resin, selective for
 Figure 20
NH4NO3 recovery.
 Page 141

Figure 21
NH4NO3 recovery, ECI Degremont process.

and IMAC HP494, a strongly basic acrylic resin with selectivity for

Their performance in terms of nitrates removal is illustrated in Fig.

24.
In conclusion, it is clear that considerable effort has been made to
develop processes for depollution of nitrogen fertilizer effluents. Most
of these processes aim at recovery as well, which renders them quite
advantageous.
 Figure 22
Treatment of effluents from a urea/nitrate complex.
 Page 142

Figure 23
Treatment diagram for effluents from urea/nitrate complex.

Figure 24
Efficiency of NO3 ion fixing on IMAC resin, Carix process.
1, Ion-exchange filter; 2, dissolved CO2 tank; 3, degasifier; 4,
CaCO3 dosing; 5, treatment water; 6, purified water; 7, waste;
8, CO2 storage; 9, CO2 recovery; 10, degasifier.
 Page 143

References
1. R. D. Roland, The recovery of ammonium nitrate from fertilizer
factory wastes, in Physicochemical Methods for Water and Waste
Water Treatment, Pergamon Press, Elmsford, N.Y., June 1979.
2. Degrémont, Memento technique de l'eau, 1989.
3. A. K. Iyer, J. N. Sayal, J. N. Trehan, and R. S. Kachawa, Industrial
effluents at Nangal, Fertilizer News, February 1974, pp. 1531.
4. L. Masotti, Depurazione delle Acque, Caldarini, Rome, 1991.
5. M. Hassan and J. R. LeBlanc, Environmental considerations for the
nitrogen fertilizer industry, Fertilizer News, June 1985, pp. 2338.
6. R. Banks, Hydrogen recovery unit ups NH3-plant efficiency,
Chemical Engineering, Vol. 84, (No. 21) 1977, pp. 9092.
7. O. J. Quartulli, Revue of methods for handling ammonia plant
process condensate, Environmental Symposium, New Orleans, 1976.
8. P. B. B. Vosloo, Some factors relating to the design of activated
sludge plants, Water Pollution Control, 1970, pp. 486495.
9. Continuous process and treatment plant for waste water with
NH4NO3 content and/or ammonia, Rom. Patent 87,294, 1985.
10. N. Arion and H. Uebersax, Récupération des engrais azotés et de
l'eau dans les effluents des usines d'engrais, Informations Chimie, No.
274, March 1984, pp. 149157.
11. R. R. Harries, The role of pH in ion exchange kinetics, Ion
Exchange for Industry, SCI, 1988, pp. 314325.
12. J. Jaqust, Z. Dzendo, and N. Arion, Fertilizer Plant Waste Water
Treatment by Ion Exchange, British Sulphur Corporation Ltd. London,
1986.
13. Rohm & Haas France, CARIX, Special Pollutec-Procédés,
Produits et Marques, 1990, pp. 4849.
14. C. Keleti, Nitric Acid and Fertilizer Nitrates, Marcel Dekker, New
York, 1985.
15. I. Moldovan, N. N. Popovici, and G. Chivu, The Technology of
Mineral Fertilizers, British Sulphur Corporation Ltd., London, 1961.
16. L. Negulescu, Considerations on the chemical regime
maintenance in a water cooling system, Revue Roumaine de Chimie,
No. 4, Academia Romana, Bucharest, 1989.
17. Water Pollution Control Federation, Operation of Wastewater
Treatment Plants: A Manual of Practice, Lancaster Press, 1976.
18. A. Bolto and L. Pawlowsky, Reclamation of wastewater
constituents by ion exchange, Effluent and Water Treatment Journal,
No. 5, 1983, pp. 208213.
 Page 145

12
Introduction to the Phosphate Fertilizer Industry
Manny Sierra*
Consultant
Clearwater, Florida
I
Introduction
Phosphorus is an essential factor in the existence of all living
organisms. The phosphate industry supplies phosphate fertilizers to
farmland areas throughout the world. It is estimated that the supply of
phosphate rock is more than adequate to meet future identified
demand through the balance of the twentieth century, after which U.S.
supplies could be inadequate to meet domestic demand unless new
sources and new mining processes are developed [1].
Future phosphate industry capacity is limited not only by capital cost
increases caused by inflation, but also by local, state, and federal
regulations. Environmental problems such as calcium sulfate disposal;
fluorine, cadmium, uranium content; and radon emissions are of
primary importance in industrialized countries, often outweighing the
decision to implement a new project. The major issue can be the time
required to obtain exploration and installation permits. On private land
and for new manufacturing facilities, overcoming obstacles to the
granting of government permits frequently takes 5 or more years and
many millions of dollars before capital can be committed for future
facilities. Whatever the growth rate, strong competition between
mine-integrated government-backed industries in developing
countries and environmentally burdened producers in developed
countries has initiated a nonreversible trend to move the phosphate
industry toward rock sources located in areas with fewer economical
and environmental constraints.
Phosphate fertilizers consume at least 90% of the world's phosphate
industry. Statistics published in the May 1982 issue of Chemical
Industry indicate the following use of the world phosphate supply:

Fertilizer 90.2%
Detergent 4.5%
Animal feed 3.3%

*Retired.
 Page 146

Food and beverage 0.7%

Surface agents 0.6%
Water treatment 0.25%
Dentistry 0.22%
Fire extinguishers 0.11%
Others 0.12%

II
Phosphate Rock:
Raw Material for Fertilizers
Phosphate rock is an essential raw material in the manufacture of
various chemical fertilizers:
1. Normal superphosphate. Finely ground phosphate rock treated with
sulfuric acid to convert insoluble tricalcium phosphate to
monocalcium phosphate. No longer in high demand, superphosphate
normally contains 18 to 21% P2O5.
2. Wet-process phosphoric acid. Comprises the acidulation of
phosphate rock with sulfuric acid to transform all the calcium
contained in the rock from calcium phosphate to calcium sulfate.
3. Triple superphosphate (TSP). Phosphate rock acidulated with
phosphoric acid to produce a granular fertilizer with 44 to 47%
available P2O5.
4. Ammoniated superphosphate. Products resulting from
neutralization of the free acid in superphosphates (both single and
triple) with aqueous or anhydrous ammonia. Ammoniacal solutions of
nitrates and urea are also used. Ammoniated superphosphates, used
extensively as the ''base" for NPK fertilizer, have been displaced by
more highly concentrated phosphates.
5. Ammonium phosphates. The product of neutralization of
orthophosphoric acid with anhydrous ammonia. Impurities in the
phosphoric acid limit the amount of plant food to the following, the
most popular grades of ammonium phosphates:
a. Monoammonium phosphate (MAP): 10% N, 50% P2O5; 11% N,
48% P2O5.
b. Diammonium phosphate (DAP): 18% N, 46% P2O5.
In all of the phosphate products listed above, acidic fluoride
compounds evolve from the reaction vessels. Emissions have to be
scrubbed, recovered, and combined to produce salable or at least
harmless compounds. The most abundant and controversial by-
product from phosphate fertilizer manufacture is gypsum, calcium
sulfate separated by filtration during production of wet-process
phosphoric acid. Gypsum is impounded immediately after filtration,
but it is regarded as a strong pollutant not only because of its acidic
characteristics, but particularly because it is the carrier of a large
percentage of the radon contained in the phosphate rock [2].
Environmental problems related to the manufacture and impounding
of phosphogypsum are reviewed in full detail in Chapter 18.
III
Phosphate Rock:
Raw Material for Industrial Products
Industrial phosphates are produced predominantly in developed
countries (Table 1). Total consumption is restricted primarily by
environmental problems; detergents, which use the largest amounts of
industrial phosphates, are frequently responsible for river and lake
eutrophication. The most common industrial phosphates in the United
States include the following [3]:
 Page 147
Table 1 Industrial Phosphates Consumption, 1987
(Thousands of Tons of P2O5)
Water
Countries DetergentsFood treatment OtherTotal
Western 437 36 22 60 555
Europe
Eastern Europe 233 13 13 18 277
North America 342 72 64 85 563
Latin America 207 13 10 10 240
Asia and 275 14 11 13 313
Oceania
Africa 76 5 2 4 87
Total 1570 153 122 190 2035

1. Defluorinated phosphate; animal feed. Supplementary, nontoxic

animal feed made available from heated phosphate rock. High
temperatures make phosphorus and calcium available and eliminate
fluoride toxicity.
2. Calcium phosphate. Produced by interaction of fluorine-free
phosphoric acid with milk of lime to yield feed grades of 18.5 or 21%
P2O5.
3. Sodium tripolyphosphate (STPP). Controlled calcination of a
mixture of sodium carbonate and phosphoric acid. Used for
detergents, water softening, and food additives.
4. Phosphoric acid; furnace grade. Produced by heating phosphate
rock, coke, and silica in an electric furnace. Gaseous elemental
phosphorus is then oxidized and hydrated. It is used for the
manufacture of soaps and detergents, pickling and rust-proofing
metals, pharmaceuticals, sugar refining, water treatment, animal feed,
and food-grade phosphates.
IV
Phosphate Rock:
Environmental and Health Impacts
Like many other industrial activities, production and use of phosphate
fertilizers can impinge on the local environment. Chemical
composition and process requirements frequently result in different
environmental concerns. Handling, storage, and application of the raw
material or the final product affect environmental conditions in one or
more of the following ways:
1. Directly, through emissions such as dust, fluoride, sulfur and
nitrous oxides, acid mist, radon, or other radioactive emissions,
industrial noise, and so on.
2. Solid waste and contaminated fluids, such as gypsum, acidic
waters, and phosphate slimes.
3. Food; some phosphate rock contaminants, such as cadmium, are not
only concentrated during the process of producing fertilizers but are
absorbed from the soil by plants and incorporated into manufactured
food [4].
Despite the general belief and some particular claims, the main
environmental concerns associated with fertilizers are not at the
production site but on the farm. The soil provides the principal natural
source of nutrients; fertilizers provide the additional nutrients required
to increase crop yields. If hunger is to be avoided, the total tonnage of
crops must continue to grow in the future as the world population
increases. Emission of
 Page 148
dust and undesirable gases from the handling and processing of
phosphate rock is always more visible and less desirable than that of
other environmental offenses and constitutes a matter of immediate
concern for the industry. However, regulations, technology, and
process adjustments have created a complex but successful system of
control, and a reasonably safe environment prevails at most plant sites
[5].
Dust is controlled by reducing drying and grinding to avoid significant
amounts of fine particles. Storage, transportation, grinding, and
processing of "wet rock" (about 12% moisture) have been promoted
intensely in the last decade to save energy and to improve
environmental control. Some process specifications still demand dry
rock; the necessity to save on freight and volume in large transoceanic
shipments still requires dry rock. Conveying and loading/unloading
systems handling dry rock (<3% moisture) are normally equipped
with enclosures to operate under negative pressure to facilitate
removal and collection of suspended fine particles from the air.
Separation equipment (cyclones) reduces phosphate losses and
performs well to control and maintain particle emissions within
environmental regulations.
If process specifications demand finely ground dry rock, the
efficiency of the air-handling system must be improved to operate
under conditions where the air is finally filtered before being released
to the atmosphere. Fume emissions are equally successfully abated in
manufacturing facilities using state-of-the-art scrubbing systems.
Modern technology is used to design the geometry carefully and to
select materials of scrubber construction to control air velocity,
improve gas abatement and particle separation, and perform as
required to satisfy clean air emission regulations. Solid waste,
impounding, and handling of contaminated fluids resulting from
processing phosphate rock are also regulated and controlled to
minimize their impact on the environment.
The biggest environmental concern and health menace is related to
specific elements not necessary for plant, animal, or human nutrition,
but present in the composition of phosphate rock. These elements,
normally referred to as heavy metals, are found in small amounts in
phosphate fertilizers (Table 2). It is estimated that, on average, 75 to
80% of
Table 2 Heavy Metals Content of
Phosphate Rock Compared to Typical
Content in Soil
Average content (mg/kg)
Element Rocka Soilb
Arsenic 7 6
Cadmium 25 0.35
Chromium 150 70
Cobalt 2 8
Copper 30 30
Lead 6 35
Manganese 30 1000
Mercury 0.0 0.06
Molybdenum 5 1.2
Nickel 35 50
Zinc 100 90
Source: Ref. 6.
aThe content in phosphate rock varies
widely with origin and type of rock.
bMedian values.
 Page 149
Table 3 Heavy Metals Content of Phosphate
Rock
Country of Type of Cadmium content
origin rock (ppm)
Former USSR Dola- 1
volcanic
Palfos, South Volcanic 1.5
Africa
Florida, U.S. Sedimentary 615
Morocco Sedimentary 875
Israel Sedimentary 1530
Tunisia Sedimentary 2535
Togo Sedimentary 4282
Senegal Sedimentary 6690
Idaho, U.S. Sedimentary 40340

the trace elements present in the rock become part of the phosphate
fertilizer. With the exception of cadmium, the total input of heavy
metals through fertilization is small compared with that naturally
present in the average soil.
The main external sources of soil cadmium are from metallurgical
smelting dust and inorganic phosphate fertilizers. Along with various
other heavy metals, phosphate rock contains varying concentrations of
cadmium, depending on its origin, as indicated in Table 3. Cadmium
is highly toxic to human beings but less toxic to plants. After being
taken up by plants and passing into food, average daily consumption
of cadmium in food is currently said to be less than that absorbed
from tobacco smoke by an average smoker.
During the last few years, some attention has been paid to the
concentration of cadmium and the significance of this source of
environmental damage. Recent reviews in the European Community
(EC) have focused on estimating the flow of cadmium associated with
the manufacture and application of phosphate fertilizers on
agricultural land. Worldwide, there is concern that the use of
phosphate fertilizers is slowly increasing the cadmium content of
arable land to levels that will necessitate its removal. There would be
a financial incentive to recover the cadmium if its current high market
price remains in force.
Cadmium is thus definitely an undesirable element in fertilizers,
although current fertilizer use presents no immediate danger. The
development of a process for cadmium removal should be given high
priority.
References
1. Fantel, Peterson, and Stowasser, The worldwide availability of
phosphate rock, 1985 Natural Resources Forum, United Nations, New
York, 1985.
2. A. Fouratry and Gy. Faludi, Changes in Radioactivity of Phosphate
Rock During the Process of Production, Department of
Radiochemistry and Physics, University of Chemical Engineering,
Hungary, January 1988.
3. G. G. Hawley, The Condensed Chemical Dictionary, 8th ed., Van
Nostrand Reinhold, New York, 1971.
4. M. Hutton, The environmental implications of cadmium in
phosphate fertilizers, Phosphorus and Potassium, February 1983.
5. C. Sine, Phosphate: report from Bone Valley, Farm Chemicals,
February 1980.
6. O. Bockman, Agriculture and Fertilizers, Norsk Hydro, Oslo,
Norway, 1990.
7. Cadmium in phosphates: one part of a wider environmental
problem, Phosphorus and Potassium, JulyAugust 1989.
 Page 151

13
Phosphate Rock Storage, Transportation, and Processing
Manny Sierra*
Consultant
Clearwater, Florida
I
Introduction
Phosphate rock, also called apatite and phosphorite, belongs to the
apatite class of minerals. In trade markets, apatite refers to
macrocrystalline fluorapatite of igneous origin, and phosphate rock
refers to microcrystalline calcium fluorophosphate of sedimentary
origin [1]. The phosphate industry operates under a plethora of
regulations concerning surface mining and industrial plant waste.
Specific rules and regulations are enforced to minimize the effect of
mining and processing of rock on the environment. Some regulations
have been developed without taking full account of the cost-benefit
ratio or the availability of technology to implement the standards. This
causes problems in achieving a balance between phosphate rock
production and environmental goals. Outside the scope of this book
are environmental concerns related to the mining and separation of
phosphate concentrate, such as factors related to vegetation and
wildlife, drainage, water supply, wetland preservation, floodplains,
solid waste disposal, and historical and archaeological sites.
Common phosphate rock contaminants qualified as potential risks to
the environment include fluorine, uranium oxide, radon, cadmium,
and thorium. However, phosphate rock is so insoluble that it can be
classified as "neutral" and not "water degradable" while it is being
transferred, stored, and processed (physical changes) [2]. The mineral
needs to be treated with mineral acids (sulfuric, nitric, phosphoric) to
make it suitable for the production of chemical fertilizers. As the
phosphate rock is decomposed with an excess of acid, the
contaminants listed above can be released or separated, resulting in
elements of concern for environmental damage. This subject is
covered elsewhere in the book.
A
Definition
Following industry practice, the term phosphate rock is defined as the
beneficiated product of phosphate ore. Phosphate rock consists of
calcium phosphate mineral apatite, with
*Retired.
 Page 152
quartz, fluoride, calcite, dolomite, clay, and ironaluminum oxide as
gang constituents. Various chemical formulas are commonly called
phosphate rock: Ca5(PO4)3F, fluorapatite; Ca5(PO4)3Cl, chlorapatite;
and Ca5(PO4)3(OH), hydroxyapatite.
Most of the phosphate rock used for chemical fertilizer comes from
sedimentary deposits. Kola apatite, mined in northern Europe, is a
significant exception since it is of igneous type. The principal
producers, the United States, the former USSR, Morocco, and China,
account for over 80% of the worldwide production of phosphate rock
(Table 1). Between 85 and 90% of the phosphate rock consumed is
used to produce phosphate fertilizer [3].
B
Chemical Evaluation
The phosphate content or grade of phosphate rock is normally
expressed as a percentage of P2O5; it may also be expressed as a
percentage of bone phosphate lime (BPL)reminiscent of the times
when bones composed the principal source of phosphate in fertilizer
manufacture:

The trade value of phosphate rock is normally assessed by the BPL

(P2O5) content. After beneficiation, phosphate rock ranges from 26 to
34% phosphorus pentoxide (P2O5). The following criteria relate to the
most common impurities. Final appreciation of rock quality depends
on the final product to be manufactured. To produce chemical
fertilizer, the main criteria for phosphate rock evaluation include:
Table 1 World Phosphate Rock
Installed Production Capacity,
December 1989a
Country Percentage
North America
U.S. 30.0
Mexico 0.5
South America 5.0
Europe
Former USSR 18.0
Other 0.5
Africa
Morocco 16.0
Tunisia 5.0
Other 6.5
Asia
China 9.5
Jordan 4.0
Israel 2.0
Other 1.5
Oceania 1.5
Source: Ref. 4
aWorld total estimated capacity:
200.5 million metric tons/year.
Listed capacity of operating plants
and plants on standby basis.
 Page 153
1. P2O5 or BPL grade: current grade: 26 to 33% P2O5 (57 to 72
BPL); noncommercial grade: below 26% P2O5 (56 BPL)
2. CaO: affects sulfuric acid consumption; impurity has an economic
impact
3. Fluorine: approximately 10% of P2O5 content; a corrosive
component of environmental concern if not separated or neutralized
properly
4. SiO2: reactive silica, needed to combine with fluorides and prevent
corrosion; nonreactive silica: dead-weight impurity
5. Al2O3 and Fe2O3: impurities that will result in sludge formation
and increase of insoluble components in the production of phosphate
fertilizer
6. Sulfates: reduce the corresponding amount of sulfuric acid in the
acidulation process
7. Chlorides: an undesirable impurity; above 0.1% becomes very
dangerous, as it mixes with HF and/or H2SO4
8. MgO: soluble in sulfuric, stays in liquid phase in phosphoric acid;
higher than 0.4% causes manufacturing difficulties
9. Na2O and K2O: create scaling on pipes and equipment as Na2SiF6
is precipitated
10. CO2 and organics: responsible for foam formation during
phosphate rock reaction
11. Radioactive elements and heavy metals: radon and cadmium
especially are highly toxic and devalue rock quality
II
Storage
Phosphate rock is not strategically stockpiled. Raw material demand
for phosphate rock is closely associated with production levels in the
world's fertilizer industry. Market trends and industry operating plans
dictate the cost and profitability of phosphate rock stock. Storage of
rock at the mine site will seek the best utilization of expensive mining
equipment. "On-hand" supply of rock at users' facilities is always
related to the method of transportation between the mine and the
processing plant.
A
Wet Rock Storage at the Mine Site
Phosphate rock is stored over reclaiming conveyor belts for proper
dewatering; it is accumulated in outdoor piles, classified by BPL
grade and/or by impurities level (Fig. 1). Phosphate rock is discharged
from dewatering bins at the mine to the storage area with
approximately 20% moisture. By shipping or reclaiming time, the
moisture content of the rock will be reduced to 10 to 12% free water.
Proper mixing of rock from different reclaiming ports in the storage
area will provide the operator with the ability to meet user
specifications.
B
Wet Rock Storage at User's Site
Rock to be transported short distances to the user or for shipping is
normally transported and stored wet (10 to 12% moisture) (Fig. 2). A
moisture level in the rock is desirable to avoid environmental
pollution caused by dust and particles released when handling dry
rock. Storage of wet rock at the user or shipping location is very
similar to mine storage, but significantly smaller, with tonnage limited
to cargo sizes or only a few days of consumer demand.
 Page 154

Figure 1
Stacker and conveyers used for storage of wet phosphate rock at mine site.
(Courtesy of Cargill Fertilizer, Fort Meade, Florida.)

Figure 2
Wet phosphate rock pile and conveying system at chemical plant site.
(Courtesy of Cargill Fertilizer, East Tampa Chemical Plant, Riverview, Florida.)
 Page 155
C
Dry Rock Storage
If economics or specifications make it necessary to dry the phosphate
rock to less than 3% moisture, the rock is normally stored in bottom-
discharge silos similar to those used for grain storage. The silos are
maintained with a closed air circulation system at slight negative
pressure, returning dust and particles via cyclones, that will clear any
air emission. Bulk dry rock can also be stored in buildings or hoppers
properly designed for a 30° angle of repose.
D
Dry Ground Rock Storage
If the dried rock to be stored is also ground, the pollution control
system will include air bags or other air filtration methods to avoid
dust being discharged to the atmosphere.
III
Transportation
A
Transportation of Wet Rock
Whenever possible and economical, phosphate rock is transported
with 10 to 12% moisture to eliminate damage to the environment
resulting from dust or dry particles normally released from dry rock.
Depending on the distance to be transported and the use of available
equipment, transportation of wet rock can be by one of two methods:
1. Moving conveyor belts. This system allows large volumes to move
rapidly improving the removal of excess water. Conveyor belts are
normally designed for hundreds of tons of rock per hour, resulting in
very efficient transport. For long distances (over 100 km) belt
transportation systems are used efficiently in the Spanish Sahara.
2. Railroad or trucks used to transport wet rock from the mine to
users' or shipping facilities. Dump trucks, bottom-discharge gondolas,
and/or self-unloading barges are commonly used.
B
Phosphate Rock Slurry:
Pipelines
A more recent trend in the industry is the use of slurry pipelines,
which require lower installation and operating costs than that those of
a transport belt, truck, or railroad system once distances exceed more
than a few miles. They are environmentally superior to dry ore
handling systems since they practically eliminate dust and particle
emissions and reduce discharge of NOx, F, and CO2. Pipeline is
generally made of carbon steel (API 5L × 52 or 5L × 56). It is usually
located underground and has both protective coating and electronic
protection. Pipeline placement is such that the grade does not exceed
15% and curves are not less than 9 in radius. Pumps are variable-
speed positive-displacement type and can be piston and/or
pistondiaphragm equipped. Operating pressure is up to 220 kg/cm2.
The ore grind is very important; the average screen size should be 150
mesh, with very low +65 mesh and less than 25% passing 325 mesh.
Data concerning phosphate slurry pipelines throughout the world are
provided in Table 2.
C
Transportation of Dry Rock
Environmental control requirements limit the use of dry rock to
processes that cannot handle the extra moisture. Transportation of dry
rock is restricted even more by the fact that drying is preferably done
at the user's site. Transportation is confined to bottom-discharging
trucks and/or totally enclosed bottom-discharge hopper railroad cars.
For
 Page 156
Table 2 Worldwide Phosphate Slurry Pipelines
Solids Flow
Pipe size (% velocity Length Operating
Owner (cm ID) conc.) (m/s) (km) date
Fosfertil, Brazil 23 61 1.66 120 1979
Goiasfertil, 15 65 2.1 14.5 1982
Brazil
Simplot, Smokey 20 60 1.83 43 1982
Canyon
Majon, 6 60 1.55 10.5 1983
Hindustan, India
Chevron, Vernal, 26 5560 1.7 155 1986
Utah
Simplot, Canada 20 60 1.83 93 1991
Wengfu, China 23 68 1.8 47 1993
Source: Geho Pumps of Holland.

transoceanic transportation in large bulk carriers, loading and

unloading facilities are equipped to prevent particulates from being
discharged to the atmosphere (Fig. 3). Dust is abated by the addition
of a small amount of oil to the surface of the rock after it is dried
properly.
D
Transportation of Dry Ground Rock
Higher risk of particulate environmental pollution occurs in
transporting dry ground rock. Transport is normally done using a
pneumatic system, with air as the moving agent and with the solids
fully contained in pipes. To improve air and solids separation,
continuous fluidization and pneumatic transport are generally used.
When the fluid (air) velocity through the bed of solids (rock) becomes
large enough, all the particles are entrained in
 Figure 3
Port terminal at Jacksonville, Florida. Storage silos and
transportation of dry phosphate rock and dry fertilizer.
(Courtesy of Occidental Chemical Company.)
 Page 157
the fluid and are carried along with it to give continuous fluidization.
Its principal application is in transferring solids from point to point in
a factory or plant [5].
IV
Processing:
Physical Changes
A very small amount of phosphate rock is used for direct application
in agricultural or industrial products. Reportedly, permanent crops
(such as rubber trees cultivated in acidic soils in Australia and
Indonesia) have responded favorable through the years to the
application of finely ground phosphate rock. However, many, if not all
of the chemical processes where phosphate rock is a basic raw
material have their own set of specifications that make it necessary to
submit the rock to one or more of the physical changes described
below.
A
Sizing:
Wet or Dry Screening
Before chemical attack, phosphate rock is frequently classified and
valued by particle size. Wet or dry screens are used as a preparation
step to enhance chemical reaction and to improve P2O5 recovered
value in the final product. For most cases the ideal phosphate rock is
sized to be smaller than 35 mesh and larger than 325 mesh. If the
percentage of fine particles (100 mesh) is high, it can produce possible
particle and dust problems when loading, unloading, and handling the
material. If the percentage of coarse particles (+65 mesh) is high, the
rock reactivity index can be low, resulting in a lower yield of P2O5.
B
Grinding
Rock delivered from mines seldom has the desired particle-size
distribution. Size reduction is necessary for most processes. The
primary purpose of rock grinding is to create and expose sufficient
surface area to enhance reaction and dissolution of the rock in the
shortest possible time. Rock reactivity and conversion are correlated
with the external surface area available.
Variables affecting the degree to which phosphate rock needs to be
subdivided include:
1. The rock origin; feed size
2. Recovery required
3. How the rock will be processed
4. Retention time and designed circulation
5. Level of agitation in reactors
The design of a grinding system must be integrated with the overall
raw material/process requirements. Energy utilization by all system
components is a prime consideration. In a typical phosphoric acid
plant, approximately 40% of the electrical energy is used for grinding.
Present industry practices give preference to the use of wet grinding
over dry grinding. Both energy saving and easier pollution control are
major advantages of wet grinding circuits. Process technologies are
directed toward the use of coarser particles; wet grinding is favored
since it normally provides larger capacity. But if size distribution
control demands a large percentage of fine particles, dry grinding
could be necessary. Grinding circuits can also be designed with an
open cycle if size distribution is not critical, or with a closed loop
when high grinding efficiency is required.
The heart of the grinding circuit is the mill. Among those used by the
phosphate industry are rod mills, ball mills (Figs. 4 and 5), and spring-
loaded roller mills. Selection
 Page 158
of the type of mill to be used depends on the configuration of the
grinding circuit, the nature and size of the feedstock, and the control
required to satisfy particle sizing.
C
Drying
Phosphate rock dryers are normally designed to reduce the moisture
content of wet stored rock (up to 15% moisture) to users or according
to shipping specifications, usually to less than 3% free water. Rotary
dryers are used if a high tonnage rate is desired. Fluid-bed reactors are
preferable when higher energy efficiency is desired. Bunker C fuel oil
or natural gas provides the heat required to dry rock. A drying plant
includes auxiliaries for fuel storage and handling; it must also be
equipped with an efficient dust collecting system.
D
Calcination
When phosphate rock is calcined, it is heated to a temperature below
its melting point to bring about a state of thermal decomposition and
for destructive distillation of organic compounds. Calcination is also
used to improve beneficiation of the ore. Calcination systems are
similar to those used for drying, but protected to resist temperatures
between 820 and 950°C.
 Figure 4
Wet grinding ball mills of capacity 200 tons/h (dual system).
(Courtesy of Cargill Fertilizer, Riverview, Florida.)
 Page 159

Figure 5
Phosphate rock ball mill converted for wet rock grinding.
(Courtesy of Farmland Industries, Bartow, Florida.)
E
Defluorination
Defluorination is used to increase the available P2O5 in phosphate
rock by fusion at elevated temperatures. Removal of fluorine occurs
through volatilization [6]. Certain reagents may be added, depending
on the rock origin, to facilitate removal of flourine without fusion or
by actually melting the phosphate rock. In most cases, for effective
defluorination, water vapor must be present; fluorine is volatilized as
HF. The process requires a temperature of at least 1320°C.
To obtain high solubility and to prevent the P2O5 from reverting to an
insoluble form, quick cooling is essential. This is usually achieved by
quenching the hot product with air or water. The final product is
marketed on the basis of total P2O5 present and is substantially free of
fluorine. Fused tricalcium phosphate is a source of phosphorus and
calcium and is used as a supplement for livestock and poultry feed.
Because of the concentration of gaseous HF, effluent recovery of that
product can be commercially attractive. Dust must also be removed
from the stack; gases and fluorine can then be absorbed at
temperatures below 480°C.
References
1. W. H. Waggaman, Phosphoric Acid, Phosphates and Phosphatic
Fertilizers, Reinhold, New York, 1952.
2. A. V. Slack, Phosphoric Acid, Vol. 1, Marcel Dekker, New York,
1968.
3. P. Becker, Phosphates and Phosphoric Acid, 2nd ed., Marcel
Dekker, New York, 1989.
 Page 160
4. Minerals Yearbook: Phosphate Rock, Bureau of Mines, U.S.
Department of the Interior, Washington, D.C., October 1990, 1983,
1974.
5. W. L. McCabe and J. C. Smith, Unit Operations of Chemical
Engineering, McGraw-Hill, New York, 1967.
6. G. G. Hawley, The Condensed Chemical Dictionary, 8th ed., Van
Nostrand Reinhold, New York, 1971.
 Page 161

14
Sulfuric Acid Production
Teodor Ciobanu*
IpranIprochim
Bucharest, Romania
I
Introduction
Acid rainfall, which is more and more frequent and very harmful to
the environment, is due primarily to atmospheric pollution by
significant quantities of SOx and NOx, released by various point
sources. The most important SOx source is considered to be the sulfur-
containing solid and liquid fossil fuels used in large quantities in
thermopower stations. Second in degree is the chemical industry,
particularly the sulfuric acid industry, which creates adverse effects
primarily in atmospheric pollution, but rivers and soils can be
contaminated as well.
The annual world production of sulfuric acid in 1990 was almost 160
million tons [1]. Considering an average SO2-to-SO3 conversion yield
of about 98% and an absorption efficiency of 99.8%, large quantities
of SO2 and H2SO4 are still released in the atmosphere: 2.13 million
tons/yr of SO2 and about 320,000 tons/yr of H2SO4, arising from
unabsorbed SO3 when expressed as H2SO4 totaling about 3.6 million
tons/yr.
In recent years developments in the sulfuric acid industry show the
strong influence of environmental pollution phenomena, such as:
1. Very high-capacity single-stream units now implement up to 2400
tons/day using sulfur as the raw material and up to 1500 tons/day
using pyrites.
2. Plants are sited based on the economic advantages of having many
units in the same industrial area, leading to a production capacity of
1000 to 10,000 tons/day in a single factory.
3. The process and construction aspects of the plants differ somewhat
in accordance with the raw materials used, elemental sulfur being the
leading material, followed by pyrites and gases from nonferrous
metallurgy. Plant operation and pollution control are affected as well
[2].
4. To reduce the adverse effects of noxious agents, efficient methods
have been developed as part of the process itself, as well as in the
equipment characteristics.
*Retired.
Factors involved in eliminating or diminishing pollution generated in the sulfuric
this chapter.
II
Manufacturing Processes Using Raw Materials
The most important raw materials used in the sulfuric acid industry worldwide are
containing gases from nonferrous metallurgy. Consequently, we limit the focus of
plants using these raw materials. The principles involved in the materials flow and
corresponding to the three raw materials are presented in Fig. 1.
III
Pollution Sources in Various Process Steps
The most adverse pollution effects in the sulfuric acid industry are generated when
materials, the next most noxious effects when using gases from nonferrous metall
when the process is sulfur-based.
In Table 1 we list the polluting agents that may appear in various stages of the thre
processing, corresponding to the respective raw materials containing sulfur that ar
numbered/lettered paragraph designations below correspond to the respective sect
IV
Technical Antipollution Measures in the Sulfuric Acid Industry
1.A. Solid sulfur (dust particles, granulated, or pellets) is delivered to storage area
sulfur is stored in either open or closed areas or buildings. The advantages of the g
over dust are:
 Figure 1
Manufacture of sulfuric acid, with main stages for processing gas from sulfur burning, pyrite r
 Page 163
1. Greatly improved properties of flow
2. Capability to be stocked in large metallic or concrete silos
3. Very low dust release
To diminish the release of dust to the atmosphere when handling
sulfur in particle form, open storage areas are bordered by high panels
made of lightweight construction materials. In dry areas where strong
winds carry sand dust, sulfur should be stored in a closed building,
because sand covering the sulfur surface may lead to extensive
erosion of equipment in the sulfur-melting area (mixers, pumps,
valves, filters, etc.). In the storage area, sulfur is handled by cranes,
belt conveyors, or payloader.
Liquid sulfur is usually stored in metallic tanks heated by steam coils.
Liquid sulfur can be shipped over long distances in specially designed
ships or barges, or by rail or auto tanker, handling being by pump.
Using liquid rather than solid sulfur as raw material, some significant
advantages can be noted, such as:
1. The sulfur melting phase is eliminated, resulting in important
savings in capital and operational costs.
2. Sulfur storage and handling are much easier, less expensive, and
less polluting.
1.B. Flotation or granulated (ground) pyrite is transported to the
factory by rail, truck or barge and is stored in closed buildings. Both
flotation pyrite having less than 8% humidity and natural pyrite can be
stored and handled easily in a sulfuric acid plant. In continental and
cold-climate zones, particularly during the cold season, flotation
pyrites with greater than 3.5% humidity may become frozen, making
handling and feeding rather difficult [4]. Usually, sulfuric acid plants
are supplied with pyrite in the hot season so that they deplete their
stock during the winter. Inside the storage facility the pyrite is handled
by crane, belt conveyor, or payloader.
2.A. Solid sulfur (dust or granulated) is fed to the melting station by
belt conveyors. During the melting process, both water and sulfur
vapors are released. This mixture is collected in a stack and released
to the atmosphere. Bitumen is often present as an impurity in sulfur
(so-called ''dark-colored" sulfur) and when burned releases water
vapor and slag. The slag clogs the sulfur sprayers in the furnace when
generating sulfuric acid by making a mist in the absorption towers. To
avoid this harmful effect, liquid sulfur is filtered in steam-jacketed
leaf filters that have been precoated with material to trap bitumen and
solid impurities (ash).
3.A. Sulfur is burned using air previously dried by spraying with
98.5% sulfuric acid in a drying tower. Combustion takes place in a
furnace previously heated by fuel up to 550 to 600°C, where liquid
sulfur is injected through sprayers activated by submerged pumps.
During this operation two important processes occur: (1) liquid sulfur
evaporates, and (2) gaseous sulfur oxidizes to SO2. The burning
temperature in the furnace must be over 1000°C for complete
decomposition of COS and HCN into less harmful components. The
temperature of the exit gases out of the combustion furnace is 1000 to
1050°C and the SO2 volume content is 10.5 to 11 %. The SO2
optimum concentration for the oxidizing stage is controlled by
injecting secondary dried air.
3.B. The pyrite roasting process, first realized in mechanical
multistaged furnaces in 1867, was improved step by step in the 1950s
when a real technical revolution took place through the recuperation
of a significant amount of heat from sulfur gas in the form of steam
[3]. In the 1930s, suspension roasting furnaces were introduced,
providing additional improvements in temperature increase and heat
recovery. Fluidized-bed
Table 1 Possible Polluting Agents in Sulfuric Acid Processing
Raw material

Technological stage A. Elemental sulfur B. Pyrite

1. Raw material storage Sulfur dust (powder) Pyrite dust
SO2-containing gases originating
from self-ignition
2. Raw material preparation Sulfur aerosol with water vapor
sulfur melting from sulfur melting
Gases with SO2 originating from
sulfur self-ignition
3. SO2-carrying gases from SO2-containing gases Pyrite dust from the feeding po
burning of raw materials SO2-containing gases and cind
dust
especially at startup of burning
4. Waste heat recovery from SO2-containing gases SO2-containing gases
SO2 gases Pyrite cinder dust
5. Dry purification of SO2- Cinder dust during mechanical
containing gases handling
6. Collection, cooling, Pyrite cinder dust
wetting, and handling of SO2-containing gases from the
pyrite cinder during cinder
transport to stockpile cooling drum
7. Wet purification of SO2- Acid sludge with 535% H
containing gases SO2-saturated, containing As,
Te, Fe, and other compounds
8. Drying of SO2-containing Possible leakage of concentrated Possible leakage of concentrat
gases; air drying in case A sulfuric acid because of reduction or technical defects; H
in system tightness or technical 175250 mg/cm saturation of th
defects with SO2

(table continued on next page)

(table continued from previous page)
Raw material
Technological A. Elemental C. Gases from
stage sulfur B. Pyrite nonferrous metallurgy
9. Conversion SC/SA (single conversion/single absorption) process
of SO2 to
SO3
SO2-to-SO3 conversion: 9898.2%
Untransformed SO2: about 13.3 kg SO2/ton H2SO4
When vanadium catalyst with cesium promotor is used, the transforming degre
SO3 could reach
99%; when not transformed, SO2 decreases to 6.6 kg SO2/ton H2SO
DC/DA (double conversion/double absorption) process
SO2-to-SO3 conversion: 99.599.7%
Untransformed SO2: about 2.6 kg SO2/ton H2SO4
10. SO3 SC/SA process
absorption
Final absorption recovery: 99.9%
Possible forming of mist in the tower due to the absorption process disturbing a
mg/N · m3 gas
Nonabsorbed gas to be released in the atmosphere, equivalent to 2 kg H
DC/DA process
Final absorption recovery: 99.9%
Forming of mist in the intermediary tower due to the absorption process disturb
5001750 mg/N · m3
gas
Nonabsorbed SO3 to be sent to the atmosphere, equivalent to 2 kg H
11. Sulfuric Leaking of concentrated H2SO4 during loading and unloading, or due to mecha
acid storage the equipment
Forming H2 accumulated under the tank cover, due to the corrosion process
Release of SO2 dissolved in the drying acid in cases B and C
12. Reduction Pollution To be considered for plants operating using the SC/SA process, w
of SO2 + SO3 reduction for gases could be important and where plant changeover to the DC/D
content the DC/DA would be difficult, especially in cases B an C
of final gases process 2 + 2
with H2O2 or
H2SO5
By applying one of the known methods to catch SO2
gas, one can expect to reduce the pollutants in the atmosphere by
9097%
13. Dispersion Final gases with SO2 and SO3 traces have to be dispersed in the atmosphere at
of final gases ground that will ensure lower concentrations than those of federal local standar
containing
SO2 and SO3
Determination of the dispersing stack: height is a specific calculation based on
of SO2 and SO3 released, possibly added to similar emanations in the area, clim
conditions, etc.
 Page 166
furnaces appeared at the beginning of the 1960s, developed by BASF
in Germany and designed and implemented in practice by the Lurgi
company (Frankfurt-on-Main).
The principal advantages of fluidized-bed pyrite roasting furnaces are
both technical and economical, with corresponding important
improvements in environmental conditions through decreased release
of SO2 and cinder dust from pyrites. At present, fluidized-bed
furnaces are generally used in modern plants for pyrite roasting. In a
few countries, such as Canada and Italy, older suspension roasting
kilns are still in operation, particularly when roasting blends and in the
cellulose industry.
When starting up the furnace, even during fluidized-bed heating,
important quantities of dust are released; after reaching temperatures
of 500 to 550°C, pyrite is fed in, thus also releasing large quantities of
SO2. As power cannot be supplied to the dry electrofilter until the
temperature of the first electric fields has reached 270 to 300°C, the
dust and SO2-containing gases should be discharged in the
atmosphere.
To avoid release of dust and SO2 to the atmosphere during startup
periods, preheating of the dry electrofilter by means of fuel
combustiongenerated hot gases or electrically heated air has been
advocated. Therefore, before fluidized-bed heating is begun, the heat
recovery boiler is monitored to operate at 25 bar, the required
quantities of steam being supplied from outside battery limit sources.
The boiler water feeding and recirculating pumps are started up at this
time, as is the pyrite cinder collecting and conveying system [4].
Some engineering companies recommend water or lime milch
washing in a closed loop of the sulfur-containing gases exiting the
roaster. This is carried out in a venturi scrubber that has a settler in its
lower part. Under normal operation of a fluidized-bed roasting
furnace, gases that contain 12 to 14 vol % SO2 at 900 to 950°C result.
In the United States, the Dorr-Oliver process is frequently used. This
process does not require a heat recovery boiler or dry electrofilter.
Water is added to the flotation pyrite so that a 65% solids slurry is
obtained, which is fed directly into the fluidized bed. The sulfur-
containing gases coming out of the furnace at a temperature of about
900°C and 12 to 13 vol % SO2 are introduced into a cyclone battery
for dust trapping and are subsequently fed to the wet purification
section. The pyrite cinder formed in the furnace and the cyclones are
water cooled, producing a 10% solids slurry that is fed to a clarifier.
The clarified material is returned to the cinder cooling system, and the
settled slurry is pumped to a well-insulated pond to protect the
groundwater and soil from contamination.
The Dorr-Oliver process has a few important disadvantages, and
includes the fact that great quantities of water must be pumped; the
suspension and sludge pumps suffer very severe errosion effects;
cinders with very high water content (18 to 20%) are delivered; the
height of the settling ponds is rather limited; more land is required for
cinder disposal; and there is an additional cost to drive the pumps.
Recently, heat recovery boilers have been introduced in the United
States, but dry electrofilters are still not used.
4.A. Sulfur-containing gases produced in sulfur-burning furnaces at a
temperature of over 1000°C and 10.5 to 11 vol % SO2 should be
cooled rapidly to 700 to 400°C, to avoid as much SO3 formation as
possible. This reaction is favored by the presence of Fe2O3 on the heat
recovery boiler tubes. Such operational conditions make it possible for
small quantities of SO3 to react with equally small quantities of
volatile, unburned sulfur, giving SO2 according to the chemical
reaction [3]

Usually, heat recovery from gases resulting from sulfur burning is

realized in natural-circulation boilers when 40 to 60 bar pressure
steam is generated. Steam
 Page 167
superheating up to 450°C is done in a superheater located between the
first and second converter passes. Some companies use forced-
recirculation boilers, but the disadvantage of this system is the
additional electric power consumption necessary to drive the
recirculating pumps. In the heat recovery boiler the sulfur-containing
gases are cooled to 440°C before being circulated to the converter's
first catalyst layer.
4.B. The sulfur-containing gases resulting from the roasting furnace at
900 to 950°C and great quantities of pyrite cinder dust are introduced
into the waste heat recovery boiler. To reduce SO3 formation made
possible by the catalytic properties of pyrite cinder (when the
optimum temperature is 650°C), as well as to avoid pyrite cinder
sulfurization by the SO3 thus formed (when the optimum temperature
is 500°C), a sudden cooling of the gases is mandatory (0.3 to 0.4 s)
starting at 900 to 950°C, going down to 450 to 400°C.
For waste heat recovery, forced-recirculation boilers are used, divided
transversally into several compartments. Superheated steam at 400 to
450°C and 40 to 60 bar is produced in the boiler.
5.A. Sulfur-burning gases exiting the heat recovery boiler are
introduced downward to a gas filter packed with a layer of 6 to 10-
mm refractory material granules meant to trap mechanical impurities
from the furnace and recuperating boiler.
5.B. For dust removal the gases from pyrite roasting exiting the heat
recovery boiler are sent to cyclones and to a dry electrofilter operated
in a slight negative pressure regime.
6.B. The pyrite cinder from the roasting process is collected from the
roaster, the heat recovery boiler, the cyclones, and the dry electrofilter
by tight, water-cooled, metallic chain conveyors. To avoid dust release
in the atmosphere while cinder is being conveyed to the pile, the
cinder is wetted in a rotating drum provided with water spraying
nozzles. To allow steam discharge to the atmosphere, a stack with
inside water spraying nozzles is provided. When leaving the rotating
drum the cinder has 8 to 10% humidity and is further conveyed to the
cinder pile by covered belt conveyors.
In addition to Fe2O3, the pyrite cinder contains various impurities,
such as iron, copper, zinc, and cadmium sulfates, iron arsenates, SeO2,
and so on, all of them more or less water soluble. To avoid soil and
groundwater contamination, the pile area has to be very tight, with
trenches provided on its perimeter to collect rainwater. Rainwater that
collects on the pile surface should be neutralized with chalk milk and
drained to the main sewage system. If arsenic and nonferrous metals
are not present in the cinder, they may be used in iron production;
cinder containing small quantities of arsenic and nonferrous metals
can be supplied to cement factories. During storage these are sprayed
with water to maintain the surface layer at 8 to 10% humidity. Due to
their high content of arsenic and nonferrous metals, commercial
cinders are stored in piles which for environmental protection reasons
are covered with soil and vegetation.
In Germany, Japan, and Romania, plants have been built for the
recovery of noble metals (gold and silver), copper, zinc, lead, As2O3,
and so on, from pyrite cinders. After elimination of these components,
the cinders can be used in iron production.
7.B. The gases resulting from pyrite roasting have 0.1 to 0.15 g/N·m3
dust containing mainly Fe2O3 and various quantities of metallic
oxides; low boiling points (e.g., As2O3, b.p. 193°C; SeO2, b.p. 337°C;
and TeO2, b.p. 450°C; and 0.1 to 0.2 vol % SO3. It should be noted
that during pyrite roasting in fluidized-bed furnaces, important
quantities of the arsenic contained in the pyrite, after being
transformed into As2O3, are reacted with the main component of the
pyrite cinder (Fe2O3). Iron arsenate is obtained according to the
reaction.
 Page 168

The iron arsenate is present in the pyrite cinders whereas the other
oxides (SeO2, TeO2, and the balance of As2O3) are carried in the
gases, their concentration being related primarily to the temperature of
the gas when it is discharged from the dry electrofilter. The higher the
temperature, the higher the oxide content of the gases. (To illustrate
this statement, the SeO2 case is discussed later [4].) When the gas
discharge temperature is 300 to 320°C in the purification section,
sludge is obtained that contains only 34% of the selenium in the raw
material. In cases where the temperature is as high as 380 to 400°C,
sludge containing up to 60% of the selenium is obtained.
7.C. The sulfur-containing gases produced in nonferrous metallurgy
have a dust content as high as 0.2 g/N·m3, and their composition is
related to the quality of the raw material. CuO, ZnO, PbO, CdO, and
other compounds may be formed in various quantities. In addition,
various quantities of As2O3, SeO2, TeO2, HF, and SiF4 may be found
(particularly when ZnS concentrate is roasted), as well as PbS (when
the PbS concentrate is roasted or smelted). Up to 0.03 to 0.05% vol %
SO3 is also present.
To process the sulfur-containing gases from pyrite roasting or
nonferrous metallurgy manufacturing into sulfuric acid by catalytic
conversion, dust removal is mandatory as well as the removal of
volatile substances such as SO3, As2O3, SeO2, TeO2, HF, and SiF4.
The presence above certain limits of such components in the gas flow
makes their processing difficult, and dust can clog the installation.
As2O3 is a poison for the vanadium catalyst because it blocks the
catalyst active area and diminishes the V2O5 content primarily in the
first layer, due to the formation of vanadyl arsenate (VOAsO4), a
volatile substance, according to the reaction
The reduction in the active component of the catalyst leads to a
decrease in SO2-to-SO3 conversion efficiency, resulting in increased
environmental pollution. Chlorine in the gas flow leads to the
formation of volatile vanadyl chloride, which also lowers the
catalyst's active component. When HF and SiF4 are present, the SiO2
(catalyst support) is affected. As a direct consequence, the main
stability factor of the active component (alkaline sulfovanadates) is
destroyed which leads to decreased catalyst activity.
Reaction of the SO3 in the sulfur-containing gas with water vapors
leads to sulfuric acid mist generation. This mist has a corrosive effect
on heat exchanger tubes, and the ferrous sulfate solid layer hinders
heat exchange. Some of the ferrous sulfate is carried by the gas flow
into the converter's first catalyst layer. Where clogging of catalyst
granules occurs, there is an increased pressure drop.
The harmful components mentioned above should be removed from
the gas flow operation in the wet purification phase. Up-to-date wet
purification sections utilize three main purification steps:
1. Adiabatic cooling of the gas flow to as low as 70°C in a washing
tower or venturi tube
2. Further gas cooling down to 30 to 40°C in a washing tower
3. Trapping mist formed during previous cooling steps in wet
precipitators that have two-step retention fields
In recent years, tube-shaped two-stage-design electroprecipitators of
great efficiency, manufactured by Beltran Associates Inc., Brooklyn,
New York [8], have been introduced. In previous cooling steps, the
mist may be generated in two ways [6]:
 Page 169
1. By sudden gas cooling when the partial pressure of volatiles in the
gas flow becomes higher than the vapor pressure of the pure
component (as for As2O3, SeO2, and TeO2),
2. By a chemical reaction in the gaseous phase when at the reaction
temperature, the product so formed has a low vapor pressure (as for
SO3)
In practice, particles carried by a gaseous stream having less than 10
µm are considered mist.
During the washingcooling steps, the following take place:
1. Trapping of coarse dust particles
2. Partial As2O3 and SeO2 dissolution in the washing acids; As2O3
and SeO2 dissolution in the sulfuric acid mist formed earlier, and
trapping of the airborne dust in the mist
3. HF trapping by absorption in the washing acid (particularly for
gases from nonferrous metallurgy, and primarily when ZnS is roasted)
When a higher fluorine content is noted, Na2CO3 or NaOH is added
to the second washing step so that insoluble Na2SiF6 will be formed.
Where wet purification of gases is complete, the impurities content in
the gas flow should not exceed: As2O3, 0.005 mg/N·m3; sulfuric acid
mist, expressed as 100% H2SO4, 0.005 g/N·m3; chlorides, expressed
as chlorine, 1.2 mg/N·m3; fluorides, expressed as fluorine, 0.25
mg/N·m3; and dust, 1.2 mg/N·m3. In Fig. 2, a typical process
flowsheet is presented for the gas wet purification phase.
When operating wet purification plants, the sludges of rather varied
composition that result are collected in a weak acid clarifier. Variable
quantities of SO2 related to acid concentration and temperature, are
dissolved in the spraying acids. (Fig. 3). Due to the SO3 in the gases,
some quantities of weak sulfuric acid are produced as well. The
sulfuric acid effluent should be eliminated from the plant. To avoid
dispersing into the environment

Figure 2
Flowsheet for SO2 gas wet cleaning process. 1, Scrubbing tower; 2,
cooling tower; 3, mist precipitators; 4, buffer tanks; 5, weak acid
stripper; 6, weak acid clarifier; 7, filter; 8, tank; 9, cooler. (From Ref. 7.)
 Page 170

Figure 3
Solubility of SO2 in sulfuric
acid versus temperature.
(From Ref. 9.)
the SO2 dissolved in the weak sulfuric acid discharged from the plant
(from the spraying loop of the first washing tower), the acid flow is air
stripped in a weak acid stripper and is then sent to a clarifier.
When using gases from pyrite roasting, the sludges resulting from the
gas purification step are beneficiated mainly for nonferrous
metallurgy, where components such as arsenium and selenium are
recovered. When a client for this sludge is not immediately available,
the settled sludge is filtered in a pressure leaf filter. The cakes are
stored in a covered and tightly sealed area. When such a client does
not exist, the settled sludge is pumped into a multicompartment, well-
lined pond, secured by a fence. After natural drying, it is delivered to
metallurgical plants of various types. If the settled sludge resulted
from nonferrous metallurgy gas purification, it is pumped directly
back to the metallurgical plant.
The weak, clarified sulfuric acid overflowing in the clarifier and
exiting the filter is collected in a tank. The clear 30 to 35% sulfuric
acid from pyrite-roasting purified gases can be used either to dilute
the acid in the absorption tower circuit or to manufacture copper,
aluminum, and other sulfates. The 5 to 10% sulfuric acid resulting
from the nonferrous metallurgy gas purification step is heavily
contaminated with various impurities (HF included), so it must be
treated by one of the following:
1. Chalk milk, neutralization, or solids settling, the settled sludge
being pumped back to the metallurgy section
2. Pumping together with the principal raw material, in the fluidized-
bed furnace
In the Dorr-Oliver roasting process, the gas dust content when
entering the specific purification step is about 100-fold more
concentrated (10 to 15 g/N·m3) than that of processes provided with
dry electrofilters. The purified sulfur-containing gases, saturated with
water vapors and at a temperature of 30 to 40°C, are directed to the
drying phase.
8,10.A,B,C. The air and SO2-containing gases that result from pyrite
roasting or from nonferrous metallurgy are examined later when we
discuss the SO3 absorption phase. These process steps are strongly
interconnected, due primarily to the necessary acid transfer related to
the quantity of water introduced into the system when gases enter the
drying tower.
The water vapors present in the air used for sulfur combustion, as well as in the
gases exiting a pyrite roasting or nonferrous metallurgy process, particularly
when discharged from a conversion phase, generate sulfuric acid mist. We should
also mention that once formed, the mist cannot be trapped by absorption. For
these reasons, the humidity present in air or in sulfur-containing gases should be
trapped as completely as possible.
In sulfuric acid plants where very high SO2 concentration gases are used, very
simple DC/DA flowsheets can be implemented. A flowsheet such as that shown
in Fig. 4 is applicable with gases from pyrite roasting as well as for gases in
nonferrous metallurgy when their SO2 content is as high as 8 vol %. When drying
air for sulfur combustion, an acid stripping tower is not necessary. As noted in the
Fig. 4 flowsheet, both the drying and absorption towers are sprayed with 98.5%
sulfuric acid, which, as shown in Fig. 5 (see the 98.7% acid concentration curve),
has the lowest vapor pressure and therefore generates the least mist. In these
plants, 98.5 to 99% sulfuric acid can be produced. For this reason, automation is
applied in order to keep optimum and constant the concentration of acid both for
drying and absorption.
Specific problems occur with gases generated in nonferrous metallurgy, such as
gases from molybdenum sulfide partial roasting/agglomeration, which are low in
SO2. The very sulfuric acid plants that deal themselves with gases low in SO
environment depollutants.
Acceptable economic parameters are also necessary. Still, there are practical
limitations in using standard conversion processes directly when sulfur-
containing gases of less than 4 vol % SO2 are to be used. The lower limits are
dictated to a large degree by the water balance in the acid system as well as by the
autothermal conditions in the conversion phase [10]. Usually, the water vapor
content in the entering gases must be kept at the maximum admissible value
related to the SO2 content in gases, starting with the wet purification phase, when
they are cooled in the dew-point field.
In Fig. 6 are shown the conditions necessary to get sulfuric acid of a certain
 Figure 4
Flowsheet of dryingabsorption stages in DC/DA process for gas from pyrite
and nonferrous metallurgy. 1, Drying tower; 2, air filters; 3, acid stripping
tower; 4, buffer tanks; 5, interabsorber; 6, final absorber and coolers; 8, blower. (From Ref. 7.)
 Page 172

Figure 5
Equilibrium vapor pressure over sulfuric
acid at 80°C and 760 mm Hg. (From Ref. 11.)
concentration as against the SO2 content in gases (vol %), and the
temperature of gases entering the drying tower. The graph shows that
when using low-SO2-containing gases, in the gas wet purification
phase low temperatures can be reached only by water cooling if
commercial-grade sulfuric acid is to be produced. To solve the water
balance problem even when low-SO2-containing gases are to be used
and 98.5% sulfuric acid is to be produced, the Lurgi Company
developed a special system combining a predrying step with acid
preconcentration in the predrying tower.
In Fig. 7 is shown a flowsheet for a sulfuric acid plant able to use low-
SO2-containing gases (but over 4%) by the DC/DA process to obtain
98.5% H2SO4 sulfuric acid. Before entering the drying tower of the
sulfuric acid plant, the water-vapor-saturated gases exiting the
purification phase are sent into a predrying system provided with a
venturi scrubber and an empty tower, where the excess water vapor is
retained by medium-concentration (30 to 60%) sulfuric acid spraying.
The weak acid so obtained is sent in a closed loop to

Figure 6
Maximum temperature of SO2 containing
gas at drying tower inlet (at 500 kgf/m2)
versus SO2 content and concentration
of acid product. (From Ref. 10.)
 Page 173

Figure 7
Flowsheet of dryingabsorption stages in DC/DA process with predrying
preconcentrating systems for processing of low-content-SO2 gas. 1, Predrying
system; 2, drying tower; 3, interabsorber system; 4, acid heat exchanger; 5,
final absorber; 6, preconcentrating system; 7, coolers; 8, blower. (From Ref. 10.)
a preconcentration system (similar to a predrying system), where the
gases in the final absorption tower (a dry tower) remove the excess
water from the weak acid by vaporization. The water vapor remaining,
as well as the tail gases, are discharged after being demisted to the
atmosphere by a dispersion stack. But as the caloric capacity of the
final gases is usually not high enough to evaporate great water
quantities, the recirculated acid flow is heated additionally by an
indirect system. The heat used is that eliminated from the intermediate
absorption unit. For this purpose a heat exchanger built mainly of
PTFE tubes is provided. In the DC/DA system, the intermediate
absorption loop is similar to the predryingconcentration loop. Using a
cocurrent system for gasacid circulation instead of a countercurrent
system, the off-gases have a 20 to 30°C higher, temperature, so that an
autothermic regime in the second conversion step is also possible.
Fig. 8 is a flowsheet for a sulfuric acid plant using gases of content 1.5
to 4% SO2 by a SC/SA system, based on the removal of excess
humidity in predryingpreconcentration units. In this process, both
venturi scrubbers and tower absorbers are sprayed with sulfuric acid.
For drying tower spraying, 98.5% sulfuric acid is recommended, but if
the plant produces oleum as well, 96% sulfuric acid may be used. In
any case, where made possible by the water balance, the drying tower
is sprayed with 98.5% sulfuric acid.
But even if 98.5% sulfuric acid is used for both drying and in the
absorption phase, because of its low vapor pressure (see Fig. 5), when
suddenly cooled in the acid coolers, sulfuric acid mist may be formed.
Where water vapor is present in great quantities in drying towers,
reaction with SO3 destroys the vapor equilibrium, making possible the
formation of sulfuric acid, and by condensation in the gaseous phase
generates mist as well. In absorption towers where SO3 is present,
mist can also be formed, by its reaction with water vapors [11]. In the
drying tower the mist content reaches 175 to 350 mg/N·m3 with a
particle dimension of 0.2 to 10 µm [12]. When the tower is sprayed
with 92 to
 Figure 8
Flowsheet of drying absorption stages in SC/SA process with predryingpreconcentrating
systems for processing of low-content-SO2 gas. 1, Predrying system; 2, drying tower; 3,
absorption tower; 4, acid heat exchangers; 5, preconcentrating systems; 6, coolers; 7, blower.
(From Ref. 10.)
93% acid, the mist content may increase to 3500 mg/Nm3. In the intermediate
absorption tower (the final absorber when low-SO2-containing gases are used
and when oleum is not also produced in the plant), the mist content is between
500 and 1750 mg/N·m3, with particle dimensions of 0.25 to 10 µm. When
oleum is produced in the plant, the particle dimensions will be 0.15 to 10 µm. In
the final absorption tower, the mist content is 175 to 350 mg/N·m3, with particle
dimensions of 0.35 to 10 µm.
In the most up-to-date large-capacity plants, where the gas velocity is as high as
1.2 m/s, the trapping of acid mist and droplets is extremely important, especially
where there are very stringent environmental pollution regulations. Brink-type
mist and droplet eliminators developed by Monsanto Enviro-Chem and mist
eliminators manufactured by Koch Engineering (Fleximex and Flexifiber) may
also be used.
The construction principle of a Brink-type element is represented in Fig. 9. The
eliminator's elements (separators) are cylindrical in shape and are made of two
concentric screens, the space between being filled by special glass, PTFE, or
metallic fiber pads. Each tower is equipped with a certain number of such
standardized elements, as required by the flow of gas. These elements are
located either in a vertical position at the upper part of the tower, or sustained by
a plate in the upper or lower part of the respective tower.
Mist elimination takes place as follows [12]:
1. For particles over 3 µm, by internal impact with the fibers
2. For particles between 1 and 3 µm, by direct trapping on H2SO4 prewetted
fibers
3. For particles under 1 µm, by Brownian diffusion followed by coalescence
Depending on the size and distribution of the mist particles to be trapped, the
mist eliminator elements may be of several types: CS (Cost Saver)-I,
recommended for drying towers; CS-II, fitted for the final absorption towers in
the DC/DA process; and ES (Energy Saver), for interabsorption towers in the
DC/DA process as well as for final absorption towers in the SC/SA process. In
Fig. 10 the retaining efficiency of mist eliminators for sulfuric acid mist
particles having different dimensions is shown. Table 2 presents
 Page 175

Figure 9
Brink mist eliminator element. (From Ref. 12.)
standardized mist eliminator element dimensions as well as the
corresponding pressure drop used in the sulfuric acid industry.
For sulfur-containing gases resulting from pyrite roasting or
nonferrous metallurgy, and frequent heating of solids, which can lead
to fouling of Brink-type eliminators, mesh or screen pad mist
eliminators are recommended for drying towers. These are
constructed of porous horizontal pads, metallic or PTFE in texture,
100 to 150 mm thick with a void content
 Figure 10
Collection efficiency versus particle size. (From Ref. 12.)
 Page 176
Table 2 Dimensions of Mist Eliminator Elements
Pressure drop
Mist eliminator (mm water Diameter Length
type column) (mm) (mm)
CS-I 75100 760 1650
CS-II 125200 660 1000
ES 100500 610 18006100
Source: Ref. 12.

of 94 to 95%. Usually, in the upper, small-diameter region of the

drying tower, two such pads are located some distance apart. Their
retaining efficiency is shown in Fig. 10.
The corrosion resistance related to temperature in 98% and 93%
sulfuric acid of various grades of steel is shown in Figs. 11 and 12.
The entire metallic portion of the mist eliminators is made of Sx steel,
due to its high corrosion resistance in both 98% and 93% sulfuric
acid.
For acid cooling in the dryingabsorption section, either coil or shell-
and-tube coolers made of austenitic, highly resistant steel such as Cx
steel (silicon austenitic stainless steel) have been used so far. In more
recent developments, kemetics-type shell-and-tube coolers with
anodic protection are highly feasible. The basic construction
principles of such a cooler are shown in Fig. 13. A 316L steel is used
for construction and the cooler can be erected either horizontally or
vertically. When positioned vertically, less space is required, resulting
in an economy of piping.
For acid-handling vertical submerged pumps, 304L or Cx steel is
used. For strong acids the pipelines are made of 0.4 to 0.7% Cr cast
iron. The drying and absorption towers are made of acid-proof brick-
lined steel. The tower spraying device is made of Cx steel.
As packing in the towers, ceramic Rashig or Pall fixtures, ceramic
saddles, or more recently, cascaded ceramic minirings (CMRs) are
used. If the Rashig ring efficiency is taken as 1, the saddles have an
efficiency of 1.3 and the value is 1.6 for the same gas velocity [14].
It is known that the SO3 absorption process is highly exothermic, so
that a lot of heat

Figure 11
Sx, Hastelloy C, Alloy 20, 316 stainless steel
corrosion rates in 98% H2SO4. (From Ref. 12.)
 Page 177

Figure 12
Sx, Hastelloy C, Alloy 20, 316 stainless steel
corrosion rates in 93% H2SO4. (From Ref. 12.)

Figure 13
Anodic passivation measurement and control circuit for shell-and-
tube coolers. (From Ref. 7.)
is lost in the cooling water (about 500,000 kcal/ton). Lately, a great effort has
been made to recover as much of the heat as possible. The most important heat
sources are considered to be the interabsorber in the DC/DA process and the
final absorber in the SC/SA process.
Usually, SO3 absorption takes place at an inletoutlet acid temperature of
60/90°C. To make possible at least a partial heat recovery, hot absorption has
been introduced, at an acid temperature of 110/140°C. In this temperature
range, the heat may be recovered as energy by using either a Rankine Freon
cycle for the recovery of 82°C hot water, or low-pressure saturated steam
(developed by Lurgi Company).
Very high absorption heat recovery (about 95%) is performed by the Monsanto
process [15] when the acid temperature reaches 165/200°C, and the heat is
recovered as saturated steam at 3 to 10 bar pressure. For equipment
construction, Monsanto recommends austenitic-type steel whose corrosion
index (CI) follows the relation

9.A,B,C. The most important process phase is conversion (oxidation of SO

SO3). The gas purification steps described above are of sufficient scope to
prepare gas so that oxidation can be performed under optimum conditions. The
entire eeconomic and environmental protection structure of sulfuric acid plants
is highly dependent on conversion operating conditions.
Wherever it comes from, gas should have a temperature of 440°C at the
converter inlet. The converter usually has four vanadium catalyst layers. In
addition to vanadium (V2O5), the catalyst has a promoter K2O (K2SO4
silica (SiO2)-based support.
In traditional processes the gases are treated by single contact/single absorption
(SC/SA), and an SO2-to-SO3 conversion grade of 98% is obtained, after which
longer operation can decrease the grade to 95 to 96% [16]. A flowsheet of the
SC/SA oxidation phase using pyrite roasting or nonferrous metallurgy gases of
maximum SO2 concentration 7.5% is presented in Fig. 14.

Figure 14
Flowsheet of SC/SA contact process for processing SO2 gas from pyrite roasting
and nonferrous metallurgy. 1, Blower; 2, gas heat exchangers; 3, converter. (From Ref. 17.)
 Page 179

Figure 15
SO2-to-SO3 conversion versus
O2/SO2 ratio in DC/DA process.
(From Ref. 24.)
In DC/DA processes, much more SO2 concentrated gases are used.
For gases arising from pyrite roasting or nonferrous metallurgy, the
SO2 content may reach 8.5 to 9.0%, but for sulfur-burning gas the
SO2 content may be as high as 10 to 11% SO2; the degree of
conversion can reach 99.5 to 99.7%. The DC/DA processes can be
grouped into several categories based on the catalysis layers between
the first and second passes: 3 + 2; 3 + 1; 2 + 2. Each of these has its
own advantages and drawbacks, related to conversion yields, capital
cost, or operation cost, as shown in Figs. 15 to 17 for sulfur-burning
gases. Used most frequently is the 3 + 1 alternative [19].
The oxidation flowsheet for DC/DA (3 + 1) when pyrite roasting or
nonferrous metallurgy gases are used is shown in Fig. 18, and that
when sulfur combustion gases are used is shown in Fig. 19. For this
case, in the oxidation phase important heat recovery at 450°C/40 to 60
bar steam is possible as well.
A different situation takes place when sulfur-containing gas involved
in copper metallurgy is used. When the suspension melting phase of
calcopyrite (CuFeS2) is carried out in a separate plant that releases
gases continuously, this is associated with the mats conversion. When
the mats conversion is fulfilled in a cyclic-operated Pierce Smith
converter, the gas is released for only 66% of the operating time and
contains 4 to 6% SO2. Obviously, this gas can be processed only by
SC/SA.

Figure 16
Relative capital cost versus SO2
gas content in DC/DA process.
(From Ref. 24.)
 Page 180

Figure 17
Specific operating cost versus SO2
gas content in DC/DA process.
(From Ref. 24.)

Figure 18
Flowsheet of DC/DA contact process for processing SO2
gas from pyrite roasting and nonferrous metallurgy.
1, Blower; 2, gas heat exchangers; 3, converter. (From Ref. 17.)
 Figure 19
Flowsheet of DC/DA contact process for SO2 gas from
sulfur burning. 1, Converter; 2, steam superheater; 3,
gas heat exchangers; 4, economizer. (From Ref. 18.)
 Page 181
When the suspension smelting plant is associated with a continuous-
gas-releasing Hoboken converter, the SO2 content is as high as 8 to
9% and may be DC/DA processed. When both phases (smelting and
conversion) take place on the same piece of equipment, such as the
Mitsubishi or Noranda-Canada type of reactor, which continuously
releases gases having 8 to 9% SO2, the DC/DA process may be used.
When necessary, the SO2 content may be lowered as convenient, by
the air supply in the drying tower. As for the gas from sulfur burning,
the air/SO2 ratio is adjusted in the sulfur-burning furnace.
A very specific situation is that when sulfur-bearing gases of very low
SO2 content have to be processed. The water balance problem in this
situation was mentioned above.
In the conversion stage, the primary goal is to make possible an
autothermic process [17]. This condition can be met only by lowering
the temperature of gases leaving the conversion stage at the absorber
inlet.
Heat exchange is realized in the gas/gas heat exchangers of the
conversion stage. Therefore, the lower the SO2 content, the larger the
gas/gas heat exchange area should be. In Fig. 20, the higher
percentage of heat exchange specific area to SO2 content in the gas is
shown and compared with that of a conventional plant (when the SO2
content is within normal limits). Due to the excessive enlargement of
the heat exchange area, the actual possibilities for processing low-SO2
gases are limited to 1.5 to 4.0% SO2 for the SC/SA method and
possibly over 4% SO2 for DC/DA. In such situations the SC/SA
converter has only three catalyst layers. When the last layer is
provided with vanadiumcesium-promoted catalyst, the conversion
yield may be as high as 99% [20].
When an inlet gas temperature of at least 400°C cannot be reached in
the first layer (due to a low SO2 content), the first catalyst layer
should also be cesium promoted. That type of catalyst is delivered by
Monsanto and Haldor-Topsoe [2022] and is much more expensive
than the K2O-only-promoted catalysts.
To avoid heat exchanger corrosion of the intertube area (when cold
gas is introduced, possibly carrying small mist quantities), the entire
area is aluminized using a method developed by Alon Processing, Inc.
[23]. All pipelines between the converter and the heat exchangers are
aluminized as well.
11.A;B;C. The product acid is usually removed as it is leaving the
drying tower, where its contained SO2 is air stripped. In some plants
the process is different; the SO2 is

Figure 20
Specific heat exchange area for gas versus SO2
content and versus flowsheet of contact unit.
(From Ref. 10.)
 Page 182
released in the sulfuric acid storage tanks. To trap it, the aeration duct
is connected to a vessel packed with granulated sodium carbonate
hydroxide + calcium hydroxide. When an oleum storage tank is
present, the aeration duct is provided with an ES type of mist
eliminator. Pumps and valves have to be maintained in good
condition, to avoid acid leakage. For avoiding difficulties related to
the acid freezing with corresponding pollution, the acid is stored at
98.5% (freezing point + 1.8°C) in warm climates and diluted before
storing to 96% (f.p. 12.6°C) in continental climates or even 94% (f.p.
31.9°C) in colder climates.
12.A,B,C. To reduce the SO2 content of the tail gas developed in the
DC/DA process from 300 ppm SO2 to 50 ppm SO2, Lurgi
recommends spraying with hydrogen peroxide (H2O2) or
supersulfuric acid (H2SO5) [24]. Such a plant's flowsheet is shown in
Fig. 21. SO2 oxidation by H2O2 or H2SO5 takes place according to
the following chemical reactions:

The H2O2 consumption for 1 kg of SO2 is 0.6 kg.

The residual SO2-containing gases released in various processes and
having an SO2 content as low as 0.8 to 1.5 vol %, which cannot be
processed by conversion, are treated with nitrogen oxides according to
the method developed by Ciba-Geigy, which utilizes equipment for a
final acid reconcentration. The flowsheet for the method is presented
in Fig. 22. With this method, 70 to 75% H2SO4 is produced and the
SO2 tail-gas content is under 100 ppm. The NOx losses in the tail gas
do not exceed 1 g/N·m3 and should be compensated by adding HNO3
or nitrogen oxides.
To reduce the polluting effects of residual gases of 0.2 to 0.8 vol %
SO2 content, absorption processes are used, with the absorbing agent
a solution of Na2CO3, NaOH, NH3, or Ca(OH)2. These kinds of gases
may arise from an SC/SA process or other source.
Generally, the residual gases are absorbed in one or two series-
arranged towers when sodium or ammonium sulfitebisulfite solutions
result, as well as insoluble calcium sulfite (that by air flow is
transformed into CaSO4). The solutions so obtained are boiled
(WellmanLord process) when SO2 is released in gaseous phase. Later,
the SO2 is used in cooling and sulfuric acid drying and a final product,
100% SO2, is obtained. By compressing and cooling it is transformed
into commercial-grade liquid SO2. The SO2 gas may be returned to
the sulfuric acid plant for conversion to SO3. The calcium sulfate

Figure 21
Lurgi Peracidox tail-gas scrubbing system with H2O2
or H2SO5. 1, Final absorber; 2, Peracidox tail-gas
scrubber; 3, buffer tank; 4, cooler; 5, feed tank. (From Ref. 24.)
 Page 183

Figure 22
Ciba-Geigy processing of low-content-SO2 gas with nitrogen oxides. 1, Venturi
scrubber; 2, gas cooler; 3, gas drying; 4, acid denitrifier; 5, blower; 6, acid
denitrifier; 7, acid heat exchanger; 8, SO2 reactors; 9, N2O3 absorber;
10, acid reconcentrating tower; 11, coolers; 12, product tank. (From Ref. 10.)
suspension is clarified in a settler and the settled sludge is pumped to
a fence-secured pond and may be used, if required, in the construction
materials industry.
Some processes used to treat sulfitebisulfite solutions involve sulfuric
acid in the absorption step, when the entire quantity of absorbed SO2
is released. The solution that results is air stripped, releasing a gas
flow of 18 to 20% SO2. The concentrated Na2SO4 or (NH4)2SO4
solutions formed at the same time are concentrated to get the
respective commercial products. The wet SO2 gas is directed to the
sulfuric acid drying tower.
In some processes, ammonium sulfitebisulfite solutions are
decomposed using phosphoric acid, resulting in MAP + DAP
solutions as well as SO2 gas [25]. The respective phosphoric acid
should be fluorine-free and its concentration controlled so that low-
viscosity easy-handling solutions will result. Mono-diammonium
phosphate solutions may be pumped to a phosphorus fertilizer plant
[26].
Generally speaking, the pollution-abatement processes are not cost-
efficient; the operating costs exceed the value of resulting commercial
products. These costs should be included in the general costs of plants
generating the gas.
At least two conditions have to be considered when choosing a
nonpolluting process:
1. The least expensive process (accounting for the specific local
circumstances) should be adopted.
2. The product resulting from the process adopted, must have the best
opportunity to be processed further in the factory, and the resulting
products should have good sales potential.
When considering the foregoing aspects, it must be noted that even if
having minimum costs, the final product of the Ca(OH)2 absorption
process is itself a nuisance. Whether sodium sulfate or sulfite, the
market conditions predominate. As such products are obtained in
other industrial processes, it is sometimes difficult to market them. In
a fertilizer complex, it may be better to produce ammonium sulfate. In
this process, SO2 absorption takes place in an ammonium
sulfitebisulfite solution, with ammonia solutions
 Page 184
added continuously. In Fig. 23 a typical Ammsox (Monsanto) process
flowsheet is shown [7]. The SO2 absorption process is carried out at
temperatures of 35 to 40°C. The absorption efficiency depends strongly
on the NH3/SO2 molar ratio in spraying solutions. The higher this ratio,
the higher the SO2 absorption efficiency. Nevertheless, when the ratio is
over 1.3, the ammonia losses in the tail gas rise as well, so that a 1.25 to
1.3 ratio has to be observed (when the spraying solution pH ranges
between 5.5 and 5.65). When sulfuric acid mist is present, in the inlet
gas, important equipment corrosion occurs.
In the Ammsox process, SO2 absorption takes place in a ceramic packed
tower with a mist eliminator in the upper part. The 0.3 to 10-µm mist
particles formed are trapped by an ES type of Brink eliminator. The tail
gas never contains more than 300 ppm SO2. The excess solution quantity
desolved in the absorption step is contacted by sulfuric acid, at which
time SO2 is released and (NH4)2 SO4 is formed. When the solution
enters an SO2 stripping ceramic packed tower, air flowed, 18 to 20% SO
gas as well as 40% ammonium sulfate result. The SO2-containing gas is
sent to the inlet of the drying tower of a sulfuric acid plant, and the
ammonium sulfate solution is sent to a fertilizer plant, evaporated, and
processed to a commericial-grade product.
13.A;B;C. The tail gas is dispersed into the atmosphere by acid-proof
stacks (whose protection is dictated by their humidity content). The
dispersion stack height is calculated in relation to the climatic conditions.
Dispersion of stack tail gas into the atmosphere when the gas contains
SO2 as well as sulfuric acid mist should be realized in such a way as not
to exceed unity, according to the relation

where C1 and C2 represent component concentration in the atmosphere,

and M1 and M2 are the concentration levels permitted by local
regulations.

Figure 23
Ammosox process (Monsanto Enviro-Chem) for tail gas scrubbing. 1,
Scrubbing tower; 2, acidifying chamber; 3, stripping tower; 4, air fan. (From Ref. 7.)
 Page 185

V
Fundamental Considerations in Design, Construction, and Operation
of Sulfuric Acid Plants
In sulfuric acid plants, under the high temperatures (900 to 1050°C) at
which SO2 gas is produced, sulfuric acid of varying strengths is
handled, which is especially corrosive at the temperatures generated in
several of the process phases. Because of this, special precautions
should be observed during plant design, construction, and operation,
to minimize environmental pollution.
1. Construction materials and equipment should be carefully selected
during the design period to ensure the highest degree of integrity for
the entire installation.
2. The plant should be of absolutely tight construction, to completely
avoid SO2 or sulfuric acid leakage.
3. In the wet purification of gas, the dryingabsorption process and the
acid storage areas must be located in an acid proof area provided with
sloped sides and basins where acid rainwater and any sulfuric acid
spills are collected, to be returned to the process or pumped to a
neutralizing station.
4. The primary cooling agent is recirculated cooling water whose pH
is controlled continuously. Whenever sulfuric acid is present in the
recirculated water, an alarm signal operates, the polluted water is
discarded, and the problem is remedied.
5. The entire plant is equipped with the instrumentation necessary to
completely monitor the process. The buffer tank has level transmitters
with an alarm attached, to prevent acid overflow.
6. During plant operation, continuous maintenance activity is very
important, to reduce downtime and repair costs.
7. Whenever a plant undergoes repair or upgrading, great care must be
exercised to ensure material selection at the same high level as that
provided initially.
8. First-class operating conditions are essential; personnel should be
carefully instructed as to correct process parameters and taught how
best to maintain them.
References
1. British Sulphur Corporation, World sulphuric acid production
19881990, British Sulphur, No. 216, SeptemberOctober 1991.
2. B. T. Vasiliev, Present state of sulphuric acid industry in the USSR
and abroad: technical progress in sulphuric acid industry, lecture for
the In-Plant Group Training Program Organized by UNIDO, 1985.
3. B. Waeser, Schwefelsaure Fabrication, Frieder Vowet,
Braunschweig, Germany, 1961.
4. K. M. Malin, Spravocinik Sernokislotcika, Izdatelistro Himia,
Moscow, 1971, pp. 295296, 387388, 500.
5. G. K. Boreskov, Kataliz v proisvodstve sernoe Kisloti, Goshim,
1954, pp. 167174.
6. A. G. Amelin, Proizvodstvo sernoi Kisloti, Gosudarstvennoe
naucino-techniceskoe izdarelistro himiceskoi literaturi Moscva, 1956
pp. 111112.
7. Sulphuric Acid Plants Prospectus, Simon Chem. Ltd., Stockport,
England. 1985.
8. H. S. Surati, Two stage precipitator for hydrocarbon emission
control, IEEE-IAS Annual Meeting, Chicago, 1976, pp. 1971. 340345.
9. D. A. Kuznetov, J. Him. Prom., No. 22, Jurnal himicescoi
promislenosti, 1941.
10. U. Sander and V. Fattinger, Verfahrenstechnische und
wirtschaftliche Aspecte der verarbeitung verdunnter SO2: Gaze zu
Schwefelsaure, Chemie Ingenieur Technik, Vol. 55, 1983.
 Page 186
11. D. R. Duros and E. D. Kennedy, Mist used to be eliminated
because it damaged the environment, but now there is another reason,
to protect downstream equipment from corrosion, Chemical
Engineering Progress, September 1978.
12. Monsanto Enviro-Chem., Mist eliminators guardians of the
sulphuric acid plants, fabric, performance and economics, British
Sulphur, JanuaryFebruary 1990.
13. British Sulphur Corporation, Kemetics coolers, British Sulphur,
SeptemberOctober 1991.
14. British Sulphur Corporation, Packing for sulphuric acid plants,
drying and absorption towers, British Sulphur, MayJune 1977.
15. Monsanto Enviro-Chem., Recent advances in H2SO4 production
processes, British Sulphur, SeptemberOctober 1990.
16. T. Ciobanu, Teza de doctorat, Catalizatori de vanadiu PAV,
Institutul Politehnic Bucuresti, 1965.
17. T. Ciobanu, (1976). Relatiile de autotermicitate ale procesului de
oxidare catalitica a gazelor sulfuroase, Revista de Chimie, No. 10,
Romania.
18. Romanian Patent 68, 130, 1975.
19. J. R. Donovan, I. S. Palermo, and R. M. Smith, Sulphuric acid
converter optimization, Chemical Engineering Progress, September
1978.
20. Monsanto Europe SA, Cesium-promoted sulphuric acid catalyst,
prospectus, 1990.
21. Monsanto Enviro-Chem., Monsanto Catalyst Bucharest 0.3.04,
1991.
22. Haldor Topsoe Company, VK 38 catalyst, British Sulphur,
MayJune 1991.
23. W. A. McGill and M. J. Weinbaum, The use of the aloziming
process in sulphuric acid plant construction, UNIDO Expert Group
Meeting on Minimizing Pollution from Fertilizer Plants, Helsinki,
1974.
24. A. Reinhard, Possibilities of improving the conversion efficiency
of double catalysis sulphuric acid plants based on sulphur combustion,
Proceedings of the Regional Seminar on Pollution Problems in
Fertilizer Plants in Arab States, Bahrain, 1982.
25. N. Popovici, Fertilizer industry: environment pollution source,
UNIDO Expert Group Meeting on Minimizing Pollution from
Fertilizer Plants, Helsinki, 1974.
26. U.S. Patent 3,421,848, 1967.
 Page 187

15
Sulfuric Acid Storage and Transport
Robin W. Strickland
Crescent Technology Inc.
New Orleans, Louisiana
Commercial grades of sulfuric acid are stored and transported in steel
vessels. For concentrated sulfuric acid at ambient temperatures, a
protective layer of iron sulfate forms on the vessel walls which retards
further corrosion. This protective layer can be removed by the high
liquid velocities experienced in pipes, pumps, and valves. Steel pipes
are satisfactory to transport sulfuric acid if the pipe diameter is large
enough to yield a liquid linear velocity under 2.5 ft/s. Cast iron, alloy,
and plastic-lined pipe are widely used to transport sulfuric acid.
Pumps and valves are generally manufactured from a variety of acid-
resistant alloys [1].
Leakage of liquid sulfuric acid provides the only significant threat to
the environment. Sulfuric acid is classified by U.S. federal regulations
as an Extremely Hazardous Substance and sulfuric acid spills of 1000
lb or more are Reportable Quantities to state and federal agencies.
Soils contaminated with sulfuric acid must be treated to neutralize the
acid. Storage tanks are vented to the atmosphere, but the vapor
pressure of sulfuric acid at ambient temperature is too low to cause
significant sulfur emissions from venting. The very fine acid mist that
can form in improperly controlled sulfuric acid plant absorbers cannot
be produced in storage tanks or from leaksnot even from spraying
leaks. Small amounts of hydrogen gas are vented from sulfuric acid
storage tanks vents.
Sulfuric acid storage tanks are built of flange-quality steel with all-
welded construction. Horizontal tanks with dished heads are often
used for quantities of acid up to 30,000 gallons. For larger quantities
of acid storage, vertical tanks with flat bottoms and conical or
spherical roofs are used. For both types of tanks, their installation
should provide for ease of inspection of the entire tank, including the
bottom. Horizontal tanks are installed on saddle supports. Vertical
tanks are placed on steel grillage supported by steel beams or concrete
piers. Placement of sulfuric acid storage tanks directly on sand or
gravel should be avoided, as this makes it difficult to locate and repair
leaks. Storage tanks should never be buried.
 Page 188
Dikes and/or ditches around sulfuric acid storage tanks are commonly
used to contain spills of sulfuric acid. All sulfuric acid storage tanks
should be equipped with automatic shutoff valves to prevent
overfilling. The overflow line from the tank should be routed to a
secondary containment vessel. The ground within the diked area is
often covered by a concrete or asphalt layer or other impermeable
membrane barrier to avoid contamination of the underlying earth by
sulfuric acid. Lined ditches convey the spilled acid to a secondary
containment vessel for neutralization and disposal. In fertilizer plants,
it is often possible to use any spilled acid directly in certain process
units and thus avoid neutralization and disposal concerns.
Sulfuric acid is transported in 50-and 75-ton rail tank cars with acid
tanks of steel construction. The acid is usually unloaded by
pressurizing the tank car with air at 30 psig or by pumping from pipe
connections at the top of the car. The unloading station at the process
facility often has pans or a paved curbed section under the rail track to
catch any acid spills from the railcar. Spills are then conveyed to a
secondary containment vessel for disposal. If a railcar is found to be
leaking while being transported, the railroad company will isolate the
car and contact the owner of the acid. Information on the contents of
the railcar is displayed on the car and on the railroad's bill of lading.
Many sulfuric acid processors retain chemical spill specialist
companies that will respond to leaking railcars by containing the spill,
removing the leaking acid, and performing remedial site cleanup.
Transport of sulfuric acid by truck is common for relatively short
hauls. The trucks have steel tanks with piping connections in the top
similar to those of railcars. The acid is usually unloaded using
compressed air from an air compressor on the truck. Acid spills are
handled similar to those for railcars.
Ships and barges in sulfuric acid transport service employ multiple
tanks or compartments for storing the acid. The acid is usually
pumped onto and from the barge or ship. Not infrequently, the acid
pipeline for loading/unloading the barge or ship will cross a public
road. When this occurs, double-walled pipe is used for the section
over the roadway. Leaks from the inner pipe are contained within the
outer pipe and conveyed to a secondary containment vessel away from
public access.
References
1. W. W. Duecker and J. R. West, The Manufacture of Sulfuric Acid,
Van Nostrand Reinhold, New York, 1959.
 Page 189

16
Impact of SO2 Abatement on the Fertilizer Industry
Jumpei Ando
Chuo University
Tokyo, Japan
I
Introduction
Flue gas desulfurization (FGD) is now popular in the United States,
Japan, Germany, Austria, the Netherlands, Sweden, Taiwan, and
elsewhere, and has begun to be used in many other countries. Still,
annual world SO2 emissions have increased to 140 million tons due to
the increase in most developing countries (Fig. 1). To alleviate acid
rain problems, a further SO2 reduction of about 40 million tons/year
may be desired. SO2 may be reduced considerably by the switch from
high-sulfur to low-sulfur fuels, but the need for FGD may become
greater as coal consumption increases. In the past, nearly 30 million
tons/year of SO2 has been emitted in the United States (Fig. 1). Under
the Clean Air Act, SO2 from coal-fired boiler flue gas is to be reduced
by 10 million tons/year.
Gypsum, the major FGD by-product in Japan and Europe, has had an
appreciable impact on the market for phosphogypsum and natural
gypsum. Sulfuric acid, elemental sulfur, and ammonium sulfate by-
products may also influence the fertilizer industry. In the present
section we describe FGD processes that may have an impact on the
fertilizer industry.
II
Status of FGD and FGD By-Products
The total world FGD capacity reached 200 GW equivalent in 1990,
involving treatment of more than 600 million N·m3/h (cubic meters at
0°C, 1 atm pressure per hour) of flue gas to remove about 10 million
tons of SO2 annually. Nearly 90% of the FGD plants are in three
countries: the United States, Japan, and Germany (Fig. 2). FGD
processes in those countries are shown in Table 1.
The wet throwaway sludge process has been used predominantly in
the United States.
 Page 190

Figure 1
Yearly SO2 emissions.
It has seldom been adopted in other countries because of the large
amount of land required for sludge disposal. The primary process in
Japan, Germany, Austria, and the Netherlands has been wet scrubbing,
producing as a by-product gypsum useful for cement, wallboard, and
plaster, replacing natural gypsum and phosphogypsum.
The capacities and selling prices of FGD by-products in Japan are
shown in Figs. 3 and 4. The production of gypsum by-product
increased rapidly in 1973 and 1974, to exceed 2 million tons/year,
causing an oversupply and a reduction of the selling price to zero.
During that period, the total gypsum demand in Japan was about 3
million tons/year and the wet process phosphoric acid industry had
difficulty in selling the phosphogypsum. The selling price has
recovered as the demand for gypsum has increased to 6 million
tons/year.
 Figure 2
Increase in FGD capacity.
 Page 191
Table 1 Approximate FGD Capacity (1000 MWe, 1989)
U.S.a Japanb Germanya
Wet lime/limestone process
Throwaway sludge 54 0 0
Gypsum by-production 6 34 36
Other wet process 3 19c 2
Dry and semidry lime/limestone 7 6 5
Total 70 60 43
Source: Ref. 1.
aFor utility boilers.
bIncluding other uses.
cMainly sodium scrubbing by-producing waste liquor.

III
FGD Processes with Fertilizer Materials as By-Products
A
Ammonium Sulfate By-Production Process
A flowsheet of the ammonia scrubbing process is shown in Fig. 5. The
flue gas is washed with ammonium sulfite solution at a pH of about 6.
Ammonium sulfite reacts with SO2 to form the bisulfite.

The solution is sent to a neutralizer and added with ammonia to form

the sulfite. Most of the sulfite solution is returned to the absorber, but
a portion is oxidized by air bubbles to ammonium sulfate. The sulfate
solution is concentrated to produce ammonium sulfate for fertilizer.
The process is relatively inexpensive but has not yet become popular,
for the following two reasons: (1) a considerable volume of fumes due
to fine particles of NH3SOxH2O compounds is generated in the
scrubber and released from the stack; and (2) ammonium sulfate is in
oversupply in most industrialized countries that use FGD. Many
developing countries have a demand for the sulfate but little need for
FGD. The fume problem can

Figure 3
FGD by-product capacity in Japan.
 Page 192

Figure 4
Change in selling price of FGD by-products in Japan.
be solved by improved technology. Ammonia scrubbing may become
available in developing countries. Three FGD plants were recently
constructed in China to produce the sulfate by-product.
B
Sulfuric Acid and Elemental Sulfur By-Products
Several processes have been used commercially to produce sulfuric
acid or elemental sulfur from SO2 in flue gases. SO2 is absorbed by
wet or dry absorbent. In most cases, the absorbent loaded with SO2 is
heated to generate concentrated SO2 gas, which is used to produce
sulfuric acid or elemental sulfur. Sulfuric acid is produced from the
SO2 by conventional processes. This process is conveniently used at
chemical factories with existing sulfuric acid plants.
Elemental sulfur is conveniently produced at oil refineries by
introducing the recovered SO2 into existing Claus furnaces to react
with H2S generated from oil refining.

For electric power stations and similar facilities that have no H2S
source, SO2 is reduced in a reactor by coal, coke, or natural gas and
the product gas containing S, H2S, COS, and so on, is introduced into
a Claus furnace. Such processes have been used commercially but are
fairly costly and energy consuming.

Figure 5
Flowsheet of ammonia scrubbing for sulfate production.
 Page 193
1
Sodium and Ammonium Scrubbing
SO2 is absorbed in a sodium sulfite solution to give a bisulfite
solution.

By heating the bisulfite solution, the reverse reaction takes place,

giving a sulfite solution and concentrated SO2 gas that can be used to
produce sulfuric acid or sulfur. The process has been used in the
United States, Japan, and Germany (Table 2) but has the disadvantage
of high energy consumption for heating the bisulfite liquor. Another
drawback is the partial oxidation of the sulfite to sodium sulfate,
which cannot absorb SO2.
At a fertilizer factory in Japan that has a sulfuric acid plant, SO2 in the
tail gas of the acid plant is subjected to ammonia scrubbing.

The resulting bisulfite solution is heated to release concentrated SO2

gas, which is returned to the acid plant, while ammonium sulfate
formed by the partial oxidation of the ammonium sulfite is used to
produce NPK fertilizer.
2
Magnesium Scrubbing
SO2 is absorbed in a magnesium hydroxide slurry to form magnesium
sulfite and sulfate, which are dried, added with carbon, and calcined
around 900°C to give MgO and SO2.

The calciner gas, containing about 7% SO2, is used to produce

sulfuric acid or sulfur. MgO is hydrated to form the hydroxide, which
is returned to the scrubber. This process has been used commercially
but has not become popular (Table 2) because of the high investment
cost and energy required for drying and calcination.
3
Activated Carbon Adsorption (Water Wash)
Below about 200°C, SO2 is readily adsorbed by activated carbon,
activated char, or coke together with oxygen and moisture to form
sulfuric acid. By washing the carbon with water, sulfuric acid at about
10% concentration can be obtained. This process causes almost no
carbon loss, but the product acid is too weak for commercial use. In a
Japanese chemical factory, the dilute acid is concentrated to 60% by
spraying into a boiler flue gas at around 250°C. The product acid
containing a small amount of impurity has been used at the factory.
Flue gas cooled below 200°C is subjected to FGD by the carbon
adsorption. Sulfuric acid of 60% concentration and a small amount of
impurities from the flue gas may be useful at fertilizer plants.
Table 2 Numbers and Approximate Capacities
of FGD Plants with Sulfuric Acid and Elemental
Sulfur By-Products (1989)a
U.S. Japan Germany
Sodium scrubbing 7 (9600)19 (8400) 2 (2100)
Magnesia scrubbing 3 (2300) 1 (460) 0
Carbon adsorption 0 4 (1430) 3 (1440)
aCapacities are shown in parentheses in 1000 N
· m3/h of treated gas.
 Page 194
4
Activated Carbon Adsorption (Heating Process)
To obtain concentrated sulfuric acid of high purity, carbon loaded with
sulfuric acid is heated at 350 to 400°C to release concentrated SO2 gas
useful to produce sulfuric acid or sulfur [2]. However, this process
causes chemical loss of the carbon.

The production of low-cost activated coke from coal started recently,

making the process commercially useful (Table 2). By adding
ammonia to the flue gas, up to about 80% of NOx with more than 90%
SO2 can be removed while passing through the reactor. The process is
not suitable for SO2-rich gas because of the high coke consumption.
5
SNOX Process
In the SNOX process, ammonia is added to a flue gas containing SO2
and NOx. After dust removal, the gas is introduced at 300 to 400°C in
an SCR reactor containing an active catalyst, in which most of NOx is
reduced to N2 while a considerable portion of SO2 is oxidized to SO3.
The gas is then introduced into a converter with an oxidation catalyst
for further SO2 oxidation. The gas, containing SO3 and moisture, is
cooled to condense sulfuric acid. More than 90% of both SO2 and
NOx can be removed to give concentrated sulfuric acid. The process
has been developed in Denmark, used recently at several plants in
Europe, and is being tested in the United States by the Department of
Energy as one of the Clean Coal Technology Projects.
IV
Evaluation of FGD Processes in Relation to the Fertilizer Industry
The ammonium sulfate by-production process described in Sec. III.A
may be useful to treat a relatively small amount of flue gas at a
fertilizer factory. To treat a tail gas from sulfuric acid plants in a
fertilizer complex, ammonia scrubbing (Sec. III.B.1) may be adequate
because by-product ammonium sulfate may be used at the complex.
To treat a large amount of gas, processes other than ammonia
scrubbing may be preferred because of the limited demand for
ammonium sulfate.
A typical coal-fired power station with three 500-MW boilers using
3% sulfur coal gives about 5 million·m3/h of flue gas at full capacity.
The annual SO2 emission may be about 220,000 tons at a moderate
operating load. If 90% of the SO2 is recovered, about 300,000 tons of
concentrated sulfuric acid or 400,000 tons of ammonium sulfate may
be produced yearly. Most of the sodium and magnesium scrubbing
plants listed in Table 2 were constructed before 1980. Not many have
been constructed since because of the high cost and large energy
consumption. The processes currently may require $300 to 400 per
kilowatt for capital charge and 20 to 25 mills per kW·h or 6 to 8 mills
per N·m3 for desulfurization, including depreciation in 10 to 15 years.
Wet limestone scrubbing may currently cost about half that of the
figures noted above. Before 1980, the limestone scrubbing process
suffered severe operating problems, but nearly all of the problems
have now been solved, resulting in widespread use of the process. On
the other hand, the activated coke adsorption plants listed in Table 2
have been constructed since 1985 as the coke price has been reduced
and NOx removal has become more important. The cost of the process
may be close to the total cost of wet limestone scrubbing and NOx
removal provided that the sulfur content of the fuel is about
 Page 195
1%. The process may cost considerably more with high-sulfur fuels
because of the increased coke consumption.
The activated coke process with a water wash (Sec. III.B.3) may be
less costly and useful for high-sulfur fuels provided that the product
acid can be utilized at fertilizer factories. The SNOX process produces
sulfuric acid containing a small amount of impurities and may be used
more widely provided that the product acid is consumed at fertilizer
factories.
References
1. J. Ando, Abatement of ozone and acid rain precursors,
International Symposium on the Environmentally Sound Technology
and Their Transfer to Developing Countries and European Economics
in Transition, UN and ENEA, Milan, October 1991.
2. J. Ando, Coal Utilization and Environmental Protection, Recent
Developments in Japan, Center for Coal Utilization, Japan, November
1991.
 Page 197

17
Phosphoric Acid by Wet Process:
Recovery of Fluorine-Containing Gases
W. R. Parish
Westinghouse Savannah River Company
Aiken, South Carolina
I
Introduction
The regulations of the Clean Air Act of 1991 in the United States are
recorded in the congressional Federal Register [40 CFR Ch. 1 (7-1-91
Edition)], Subpart T, Standards of Performance for the Phosphate
Fertilizer Industry: Wet-Process Phosphoric Acid Plants. Some of the
requirements are:
1. The regulations apply to any wet-process phosphoric acid plant
built or modified after October 22, 1974 with a capacity of more than
15 tons of equivalent feed of P2O5 per day.
2. No more than 10 g of fluorine per metric ton (0.020 lb/short ton) of
equivalent P2O5 feed to the plant may be discharged into the
atmosphere.
3. The phosphorus feed to the plant must be monitored with
equipment accurate to ±5% over the operating range.
4. The scrubber must have a continuous monitor that records the
pressure drop across the scrubber (to ±5% accuracy).
5. The testing procedure is specified in Appendix A of Section 60.204,
''Test methods and procedures," of the Federal Register (7-1-91
Edition).
6. The procedure developed by the Association of Official Analytical
Chemists (AOAC), Method 9, is required by the Environmental
Protection Agency (EPA) for determining the P2O5 content of the feed
to a wet-process phosphoric acid plant. The method is printed in
Section 60.17 of the Federal Register (7-1-91 Edition).
7. A performance test of the pollution control equipment must be
made within 60 days of achieving the maximum production rate at
which the plant will be operated, but not later than 180 days after
startup. The EPA administrator must be furnished with a written report
on the test of the pollution control equipment. The administrator must
 Page 198
be given 30 days notice before the test is performed so that the tests
may be witnessed, if desired.
8. The federal regulation allows states to make their own stipulations
as long as they are not less stringent than the federal regulations.
a. Some Typical Regulations Added by the States
The concern over fluorides released into the air by phosphoric acid
plants started in the late 1950s or early 1960s when flowers grown
commercially were reported damaged by fluorine and cattle developed
fluorosis of the teeth. The air pollution regulations apparently have
been successful in reducing fluoride pollution in the central Florida
area. No case of fluorosis of the teeth in cattle has been reported since
the middle 1970s (W. C. Thomas, Florida Department of
Environmental Regulations, personal communications).
Some stipulations added to the federal regulations by the state of
Florida relating to the design of fluorine scrubbers are:
1. The total operating time of the plant is limited (i.e., the plant must
not operate more than a specified number of hours per year).
2. The equivalent P2O5 feed to the plant must not exceed a specified
number of tons per day.
3. Performance testing of the fluorine scrubber must be done only
when the plant is operating within 90% of permitted operating
capacity.
In effect, the state of Florida has added total mass of fluoride
emissions in pounds per day to the rate requirement by the EPA. This
is a severe restriction since it limits the production capacity of the
plant and in effect limits the profit potential of the plant if an
opportunity arises for increased product sales. If the operator wants to
increase plant production capacity, a new permit has to be obtained.
We have indicated only a few of the typical regulations that have to be
considered when designing fluorine scrubbers for wet-process
phosphoric acid plants. It would be impossible to cover all the rules
and regulations of every country and district in this book. The
following discussion covers design criteria and tips concerning U.S.
and state of Florida regulations. Hopefully, the text will be helpful in
the design of fluorine scrubbers that meet requirements in other states
or districts.
II
Areas of the Plant That Have to Be Vented to Fluorine Scrubbers
The EPA states that "the affected facility includes any combination of:
reactors, filters, evaporators, and hot wells" [Section 60.200 of the
Federal Register (7-1-1991 Edition)]. The EPA does not monitor all
wet-process phosphoric acid plants. It subcontracts to individual states
certain enforcement requirements, such as plant inspections.
The EPA requirements have been defined to include venting to a
fluorine scrubber every vessel that contains less than 54% P2O5. The
54% acid storage tanks are not required to be scrubbed, unless they
are kept hot, because of the low vapor pressure of fluorine (W. C.
Thomas, Florida Department of Environmental Regulations, personal
communications).
III
Design of Fluorine Scrubbers
Reaction (1) to (4) are basic to the discussions that follow:
 Page 199

The requirement that every vessel which contains wet-process

phosphoric acid be vented to a fluorine scrubber usually demands the
use of more than one scrubber within the plant area because of the
length of ductwork from vessels into the scrubber. Most plants have at
least two fluorine scrubbers, generally one scrubber for the reactor
area and the other scrubber for the remaining vessels.
A
Ductwork Design
The design of vapor ductwork is based on experience rather than on
calculating the size of the duct. SiF4 gas is liberated from the
phosphoric acid and reacts with water vapor that has condensed,
forming hydrated SiO2 which builds up inside the ductwork and
eventually causes blockage of the duct.

One engineering company makes all its ducts out of 15-cm fiberglass-
reinforced plastic (FRP) and installs spray nozzles along the ducts
leading into the fluorine scrubber. The ducts are designed with access
ports that allow 100% of the duct to be cleaned. Typically, the ducts
are inspected every 2 weeks and cleaned during downtime, if
necessary. Each duct has a damper that can be adjusted to keep the
flow of air into the scrubber to a minimum.
Not all of the ducts have sprays. Only the vessels that contain "hot"
acid require spray nozzles in the ductwork to reduce the rate of silica
buildup. The reactor and the filter feed tank contain the hottest acid.
Vapor ducts from these vessels have water sprays inside the ductwork.
The ducts from the filtrate receivers under the vacuum filter do not
need spray nozzles. Cleanout ports are required on all ducts (J. Wing,
HiTech Solutions, Inc., personal communication).
There are no federal or state regulations that specify how much
venting from a vessel is required. They leave the design up to the
plant owner. When the plant has been constructed and is operating, the
plant is inspected. If fluorine fumes are escaping from any vessel, the
problem must be corrected.
B
Comments on Scrubber Sizing
From a review of the scrubbers presently installed in wet-process
phosphoric acid plants, the size of the plant has no relation to the gas-
handling capacity of fluorine scrubber(s). The gas-handling capacity
of these scrubber(s) depends on the type of wet-process plant and the
number of vessels in the plant that contain phosphoric acid.
Engineering contractors have found that a certain-size scrubber meets
federal and state requirements and they install a scrubber size (gas-
handling capacity) based on experience (J. Wing, HiTech Solutions,
Inc., personal communications). Some notes about scrubber design
may be helpful.
Plants using reactors that operate under vacuum (such as the Badger
isothermal reactor) send all the vapors from the reactor to a
barometric condenser and thus eliminate
 Page 200
the need to vent from the reactor to a fluorine scrubber. These plants
should require smaller fluorine scrubbers than dihydrate plants, which
operate the reactors at atmospheric pressure and are required to vent
the reactors to a fluorine scrubber.
Hemihydrate plants operate at higher temperatures and liberate more
fluorine from the reactor than do dihydrate plants. Seemingly,
hemihydrate plants would require fluorine scrubbers with a larger gas-
handling capacity. This is not necessarily true, as we will see later.
At no time does the acid actually boil in the vessels exposed to the
atmosphere in a wet-process acid plant, so the vapor pressure of the
acid is always less than 1 atm. Therefore, in vessels that contain acid,
except those in which CO2 is liberated, vapors are not "pushed" out
the top of the vessels. Boiling does take place in the flash cooler,
which is under vacuum, but the vapors from the flash cooler are
condensed in a barometric condenser into the pond water system. CO2
in phosphate rock is liberated in the reactor and will carry fluorine
fumes out of the vessel. A vent system is needed to draw the CO2
from the reactor to the fluorine scrubber to prevent "puffing" around
the vessel. For a plant with a capacity of 1000 tons/day of P2O5, the
volume of the CO2 released would be about 1.2 m3/s from phosphate
rock containing 5% CO2. The size of the fluorine scrubber attributed
to the reactor should include a substantial safety factor over the m3/s
estimate.
Another factor that may displace fluorine fumes from a vessel is the
rise of liquid level inside the vessel, which is controlled by the
maximum rate of liquid that is fed into the vessel. For 63 L/s this
amounts to only a nominal 0.064 m3/s that would have to be vented
from the vessel. A design using 0.14 to 0.24 m3/s from each vessel
should provide ample venting. The filter area requires a hood and
venting from the point where the acid slurry is fed to the filter to the
point where the first dry cake appears. The filtrate receivers have to be
vented to the scrubber.
C
Design Philosophy
1
Think Small
Government regulations require removal of fluorine to a level of 10 g
of fluorine per equivalent ton of P2O5 fed to the plant. The amount of
fluorine leaving the stack is calculated by multiplying the
concentration of the gas sample from the scrubber stack by the
volumetric flow of dry gas in the stack: in other words, milligrams of
fluorine per standard dry cubic meter of gas multiplied by the standard
cubic meter of gas (dry basis) per hour. The EPA defines this standard
at 293 K (20°C) and 101 kPa. The amount of fluorine per hour
calculated leaving the stack gas is divided by the equivalent feed rate
of P2O5 into the plant to determine the amount of fluorine evolved per
ton of P2O5 fed to the plant.

The regulations allow

Most wet-process phosphoric acid plants use recirculated pond water

for scrubbing liquid. The pond water contains fluorides and therefore
has a vapor pressure of fluorine.
 Page 201
The fluorine scrubber can actually strip fluorine from the pond water
into the air. If the scrubber is designed with a large gas-handling
capacity, it may be impossible to meet the EPA requirements even if
the inlet gas contains no fluorine.
Figure 1 shows the maximum allowable SCMS (standard cubic meters
per second) when 1000 tons/day of P2O5 is fed to a wet-process acid
plant and pond water is used for scrubbing. Figure 1 was developed
by calculating the amount of fluorine in equilibrium with air over a
solution of dilute fluorosilicic acid. The total 416 g F/h allowed for
1000 tons/day P2O5 fed to the wet-process acid plant was divided by
g F/SCMS (dry basis) at the temperature specified. Note the small
effect of concentration and the large effect of temperature on the
SCMS allowed. For a 1% H2SiF6 solution, a 10-SCMS scrubber
operating at 65°C would strip the 416 g of fluorine allowed from the
pond water even if no fluorine were in the inlet gas. At 60°C, the
scrubber could operate at 15 SCMS before stripping the 416 g/h of
fluorine allowed for a 1000-ton/day P2O5 plant. (Note: The vapor
pressures and calculations shown above are based on fluorosilicic acid
solution, not pond water. The actual vapor pressure of fluorine over
pond water varies and should be measured for the particular pond
water that the fluorine scrubber will use.)
From Fig. 1 we can see that two important criteria should be
considered in the design of a fluorine scrubber: (1) the scrubber
should be sized with as low a gas rate as necessary to allow complete
venting of fumes from the vessels that contain phosphoric acid to the
fluorine scrubber; and (2) the temperature of the scrubbing solution
should be kept as low as possible to reduce the vapor pressure of
fluorine from the scrubbing solution.
2
Cooling the Scrubbing Liquor
Figure 1 shows that the scrubbing liquor, even with as little as 0.2 wt
% H2SiF6, should be no hotter than about 55°C to use a scrubber that
handles 10 to 15 SCMS (dry basis).

Figure 1
SCMS (dry basis) that strips 416 g F/h from a wet-process
acid plant scrubber using fluorosilicic acid solutions. See the
text for specific gas and liquid conditions. , 0.2% H2SiF6; D,
1.0% H2SiF6; , 2.0% H2SiF6.
 Page 202
Figure 2 was developed to show that a large quantity of liquid is
required to prevent the scrubbing solution from reaching the
saturation temperature of the gas. Figure 2 should not be used for
design purposes because of its limited scope. It is based only on air
entering the scrubber at 70°C and 80% saturation (0.2 kg water/kg air)
and pond water entering the scrubber at 30°C. It also assumes that the
exit gas temperature is 3°C higher than the leaving liquid temperature.
Figure 2 shows that 3.6 L/s per 1.0 SCMS is about the minimum
makeup to the scrubber (at 32°C) to keep the scrubbing liquor at
55°C. Public records at the Florida Department of Environmental
Regulations show that the liquid volume to fluorine scrubbers in
Florida ranges between 3.4 and 11.2 L/s per 1.0 SCMS.
3
Staged Scrubbers
In meeting governmental requirements for fluorine removal, using
more than one scrubbing stage provides very little help. In other
words, if a low input of liquid with a high circulating rate through a
pump tank is used to scrub the fluorine vapors, the temperature of the
scrubbing liquor would approach the saturation temperature of the
gas, which (from the reactors) is normally greater than about 60°C.
Even if three scrubbing stages were used, the scrubber would not meet
the required degree of fluorine removal. Of course, external cooling
using heat exchangers could be used, but this would be expensive
compared to adding more pond water.
4
Types of Scrubbers Used
Almost every type of scrubber imaginable has been used to scrub the
fluorine fumes from wet-process phosphoric acid plants. They include
cross-flow scrubbers with packing, cross-flow with the packing
replaced by plastic mesh pads, venturi scrubbers, cyclonic scrubbers,
and some homemade scrubbers. Several articles [13] discuss the
design of fluorine scrubbers for phosphoric acid and granulation
plants. The common design factors in the scrubbers are high liquid-to-
gas rates and excellent demisters to remove mist that contains
fluorine.

Figure 2
Liters per second of 32°C pond water required to maintain
a pond water temperature in the scrubber. See the text
regarding use of this figure.
 Page 203

IV
Recovery of H2SiF6
Earlier we discussed the removal of fluoride vapors from vessels open
to the atmosphere in a phosphoric acid plant. The amount of fluoride
released from such vent systems is too small to be recovered
economically. Sources of vapor with fluoride concentrations large
enough to recover are:
1. Phosphoric acid recycle from the filter to the reaction section
2. Vapors released from the flash cooler in the reaction section
3. Vapors from the evaporators, including superphosphoric acid
evaporators
The following terms apply to the discussions that follow:
"fluorosilicic acid" refers to a solution of 15 to 25% H2SiF6 in water;
and "H2SiF6" refers to 100% H2SiF6 or dry basis. H2SiF6 does not
exist as a compound. In solution the compound is a mixture of HF and
SiF4. If no silica precipitate is obvious in the solution, there is an
excess of HF [4].
A
Recovery of Fluorosilicic Acid from Recycle Phosphoric Acid
USS Engineers and Consultants Inc. offers a process for producing
fluorosilicic acid [5]. This process mixes recycle phosphoric acid from
the gypsum filter with strong sulfuric acid. The heat of dilution of the
sulfuric acid, as well as supplemental steam and hydrated silica, are
used to strip fluorine from the recycle phosphoric acid. The vapors are
scrubbed in two stages using venturi scrubbers. The scrubbers are
designed to recycle fluorosilicic acid through a filter for removal of
hydrated silica. The silica does not have to be washed since it is
returned to the phosphoric acid upstream to promote more fluorine
evolution. A plant using this process was constructed in 1970.
B
Recovery from Flash Cooler Vapors
In dihydrate phosphoric acid plants, only 5 to 10% of the fluorine in
the phosphate rock is evolved from the reactor system. Most of the
fluorine is released in the flash cooler and goes to the barometric
condenser. Chevron Chemical Company in Wyoming has a fluorine
scrubber located between the isothermal reactor and the barometric
condenser that makes about 2% fluorosilicic acid. This system
precipitates silica because almost all of the fluorine liberated from the
reactor is SiF4 [reaction (1)].
P. T. Petrokimia Gresik Company in Indonesia recovers fluorosilicic
acid from the vapors in the No. 1 flash cooler in a Nissan C hemi-
dihydrate plant. The plant recovers about 5% H2SiF6 from fume
scrubbers and feeds the weak fluorosilicic acid as makeup into the
fluorine scrubber on the flash cooler. A product containing about
15.8% fluorine is produced. The product contains precipitated silica,
which is removed on a vacuum drum filter. This plant also recovers
fluorosilicic acid from the evaporators (M. Walters, BCI, personal
communication).
C
Recovery from Evaporators
The method most used for recovery of fluorosilicic acid in wet-
process phosphoric acid plants is scrubbing of the vapors evolved
during evaporation of phosphoric acid [6]. From a dihydrate type of
phosphoric acid plant, as much as 60% of the fluorine originally in the
phosphate rock is evolved from the phosphoric acid during
evaporation from 28% P2O5 to 54% P2O5 [7]. The amount of fluorine
evolved is dependent on the type of
 Page 204
phosphoric acid process and the source of the phosphate rock. It is
easy to determine how much fluorine is available from the
evaporators. Simply multiply the weight F/weight P2O5 into the
evaporators minus the weight F/weight P2O5 out of the evaporators by
the weight P2O5/hour fed to the evaporators. Depending on the
design, about 85 to 90% of this fluorine can be recovered as
fluorosilicic acid. Using a dihydrate process and Florida rock, about
50 to 65 kg of H2SiF6 is evolved per ton of P2O5 evaporated from 30
to 54% P2O5.
Most vendors who design evaporator systems for phosphoric acid
plants offer designs for recovery of fluorosilicic acid from the
evaporators. At least one system was installed to recover fluorosilicic
acid from a superphosphoric acid evaporator. The system for recovery
of fluorosilicic acid from evaporators consists of a P2O5 entrainment
separator located between the evaporator and the fluorine scrubber, a
fluorine scrubber (normally, a void spray tower), and most important,
a fluorosilicic acid demister (Fig. 3). The P2O5 entrainment separator
is normally a cyclonic type because solids deposit on the walls of the
entrainment separator and have to be removed periodically. Fisons has
patented use of a prescrubber ahead of the fluorosilicic acid scrubber,
which circulates liquid through the prescrubber to remove P2O5 [8].
The H2SiF6 builds up to such a high strength that fluorine is not
recovered in the prescrubber, but phosphoric acid entrainment is
recovered. The purge from the prescrubber is recycled to the
evaporator. This system can produce fluorosilicic acid with lower than
50 ppm of P2O5 in 25% H2SiF6.
The spray tower normally has been designed with three layers of
spray nozzles. A total of 6.6 L/s·m2 of tower cross section (2.2 L/s per
layer) is sprayed into the tower using a 35-to 70-kPa pressure drop
across the nozzles. A higher pressure drop across the spray nozzles is
not desirable because it would produce too much mist and flood the
mist eliminator. A higher pressure drop across the spray nozzles
would also increase the energy of pumping the circulating liquid
through the spray chamber.
The diameter of the spray tower typically has been the same diameter
as the evaporator vapor body, which operates with vapor velocities of
2 to 3 m/s. There is no need to use such a large-diameter spray tower.
If a mesh-pad demister is used, the design of the demister requires a
higher vapor velocity to impinge the fluorosilicic acid mist on the
mesh pad. Demister vendors have their own design equations,
depending on the density of the pad and other considerations, which
use vapor velocities through the mesh pad between 3 and 6 m/s. The
diameter of the spray tower should be about the same diameter as the
mesh pad. A spray tower diameter smaller than the evaporator
diameter would reduce the circulating load of fluorosilicic acid
through the spray tower, subsequently reducing the power
requirements. For example, a 3-m-diameter spray tower operating at a
vapor velocity of 2.4 m/s would require 7.4 m2 × 6.6 L/s·m2 = 49 L/s
circulating through the tower. Using a vapor velocity of 3.6 m/s
instead of 2.4 m/s would reduce the area of the spray tower and the
required circulating rate of fluorosilicic acid to about 33 L/s, a 33%
saving in circulating rate.
The fluorosilicic acid demister can be either a mesh pad or a
centrifugal type of demister. It is important to have a well-designed
demister since, depending on the evaporator size, 30 to 160 L/s of
liquid is passing through the spray tower and the production is only
0.6 to 2 L/s of fluorosilicic acid solution. A mist loss of 1 to 2% of the
circulation through the tower would drastically reduce the production
of fluorosilicic acid.
The controls needed are a specific gravity controller and a level
controller. Circulating tanks are used to hold the volume of
fluorosilicic acid solution held in the barometric leg
 Page 205

Figure 3
Equipment for recovery of fluorosilicic acid from evaporators in wet-
process phosphoric acid plants. LC, level control; DC, density control.
when the vacuum is applied. On startup and shutdown, the level in the
circulating tank changes rapidly to fill the volume of the circulating
pipe. To prevent short-circuiting, makeup water should be added in
the pipe just ahead of the circulating pump and not into the circulating
tank. Relatively large swings in product concentration (±2%) are
experienced if short-circuiting is allowed.
When more than one stage of evaporation is used to concentrate
phosphoric acid, it is important to understand how the fluorine is
evolved in order to optimize the design of the fluorine recovery
system. Figure 4 shows that very little fluorine is evolved until about
 Page 206

Figure 4
Approximate percentage of the total fluorine evolved during
concentration of wet-process phosphoric acid at 88°C. Vacuum
is adjusted during concentration to maintain 88°C boiling point.
33% P2O5 is obtained. A diagram of this type is simple to make for
any particular phosphoric acid. Simply set up a small evaporator in the
laboratory and sample the acid periodically during evaporation.
Adjust the vacuum to maintain a constant temperature during the run.
Since the gas leaving the spray chamber is nearly at equilibrium with
the scrubbing solution, the percent recovery of the fluorine evolved is
almost a direct function of the concentration of the inlet fluorine
vapors. Calculated recoveries are in the range 95 to 99%, but mist
losses drop the actual recovery to about 90%.
Figure 5 shows that during concentration of wet-process phosphoric
acid, SiF4 is evolved first and HF evolution begins later. If only the
first stage of a two-stage evaporator had fluorosilicic acid recovery,
the solution would contain a precipitate of hydrated silica [reaction
(1)]. Eventually, the silica would be so thick that the system would
have to be shut down and cleaned. A slurry of as little as 4% SiO2 is
almost too thick to pump. If only the second-stage scrubber were
operated, the fluorosilicic acid solution would contain about 4% HF in
25% H2SiF6. If a scrubber were installed on a superphosphoric acid
evaporator that concentrated the phosphoric acid from 54% to 70%
P2O5, the solution recovered would have a weight ratio of about 85%
HF and 15% H2SiF6. A rubber-lined scrubber with a Hastelloy inlet
pipe has been operated successfully to recover the mixture of HF and
H2SiF6 evolved from a superphosphoric acid evaporator. Although the
temperature of the superphosphoric acid was about 205°C, the
circulating fluorosilicic acid solution was only about 44°C. The
temperature of the circulating fluorosilicic acid solution corresponds
almost exactly with the temperature of water at the pressures listed in
the steam tables (e.g., 46°C at 10 kPa absolute pressure). This happens
because 25%
 Page 207

Figure 5
Typical curves showing that SiF4 is evolved before HF during
evaporation of wet-process phosphoric acid solution.
H2SiF6 has only about a 2°C boiling point rise. Storage of 20 to 25%
H2SiF6 solution is no problem in cold weather since the solution does
not freeze at 20°C (30.2% H2SiF6 freezes at minus 35°C [9]).
There are many options for the design of fluorine recovery systems on
an evaporator system, depending on the number of evaporators and
how they are arranged.
1. Three-stage evaporation system. Fluorine scrubbers are added only
to the second-and third-stage evaporators. There are several reasons
for this. First, the first-stage evaporator operates at higher pressure so
that the fluorosilicic acid scrubbing solution is hot (about 60°C).
Under these conditions, it is possible to strip fluorine from
fluorosilicic acid solution instead of recovering H2SiF6. Second, the
concentration of fluorine in the vapors from the first-stage evaporator
is so small that the efficiency of recovery is very low. Third, the
fluorine leaving the first-stage evaporator is mainly SiF4, which
results in silica precipitation during scrubbing of the vapors.
2. Two-stage evaporation. Normally, fluorine scrubbers are added to
both evaporators. Makeup water is added to the scrubber on the
second-stage evaporator, and fluorosilicic acid product is removed
from the first-stage evaporator.
3. Single-stage evaporation. Fluorine recovery is added. If more than
one evaporator is used (single stage in parallel), makeup water is
added to one fluorosilicic acid circulating tank and product is removed
from the other circulating tank. This allows one scrubber to operate
with a low-strength fluorosilicic acid scrubbing solution, increasing
the overall recovery of H2SiF6.
 Page 208

References
1. A. O. Hansen and R. J. Danos, The design and selection of
scrubbers for granulation plants, paper presented at the Central
Florida Section of the American Institute of Chemical Engineers,
Clearwater, Fla., May 24, 1980.
2. A. O. Hansen and R. J. Danos, Scrubbers for phosphoric acid
plants, Chemical Engineering Progress, March 1982, pp. 4045.
3. A. O. Hansen, Phosphate plant scrubbers, paper presented at the
Central Florida Section of the American Institute of Chemical
Engineers, Clearwater, Fla., May 24, 1991.
4. Superphosphate: Its History, Chemistry, and Manufacture, U.S.
Department of Agriculture and Tennessee Valley Authority, 1956, p.
225.
5. R. G. Smaltz et al., Recovery of fluorosilicic acid and P2O5 from
recycle streams and pond water, U.S. Patent 3,498,746, March 3,
1970.
6. W. R. Parish, U.S. Patent 3,091,513, 1963.
7. Fluorine recovery by the phosphate fertilizer industry. Part 1. Wider
interest in conversion to salable fluorides, Phosphorus and Potassium,
No. 53, MayJune 1971.
8. M. English, U.S. Patent 3,512,348, 1968.
9. I. G. Ryss, The Chemistry of Fluorine and Its Inorganic
Compounds, Clearinghouse for Federal Scientific and Technical
Information, U.S. Department of Commerce, Washington, D.C., AEC
No. 3927, Part 1, 1956, p. 372.
 Page 209

18
Phosphoric Acid by Wet Process:
Phosphogypsum, Transport, Storage, and Utilization
Jumpei Ando
Chuo University
Tokyo, Japan

John E. Cameron
IMCAgrico Company
Mulberry, Florida
I
Introduction
For each ton of P2O5 produced, approximately 2 tons of fertilizer
product is produced, principally calcium and ammonium phosphates.
In addition, approximately 5 tons of by-product phosphogypsum also
results. Simply stated, more by-product is produced than product.
Consequently, the transport, storage, and utilization of
phosphogypsum are of major importance to a phosphate fertilizer
producer.
It should be noted that the information presented here results primarily
from the author's experience with phosphogypsum hydraulically
transported and stored in gypsum stacks via the rim ditch method, as
practiced at the New Wales plant of IMCAgrico Company in
Mulberry, Florida. Other operations have different circumstances and
have developed different methods and/or variations of methods to
transport and store gypsum in a cost-effective manner.
II
Transport of Phosphogypsum
New Wales is one of the largest phosphate fertilizer plants in the
world, with production capacity in excess of 1,500,000 tons (metric)
of P2O5 annually. By-product phosphogypsum is produced in
dihydrate form. The gypsum stack is the single largest land use,
occupying an area of approximately 160 ha at its base. A new storage
area of comparable size is presently under construction, and it is
projected that yet another stack expansion will be needed within 10
years.
 Page 210
A
Transport to the Gypsum Stack
Gypsum is generally discharged from filters as a wet cake and
immediately slurried with pond water, that is, water from the process
water cooling pond. The slurry is then gravity fed via pipelines and/or
open launders toward the gypsum storage area. Most transportation
systems are designed for a slurry of approximately 20% solids. This
results in a slurry flow of about 1260 L/s at the New Wales plant,
which produces up to 22,000 tons of gypsum (dry weight) each day.
At plant startup the slurry was discharged by gravity directly to the
gypsum storage area. However, as the storage area filled and
increased in height, a surge tank and pumping system were soon
required to lift the slurry to the top of the gypsum stack. As the stack
increased further in height, the pumping system evolved from a 6 ×
105 W single-stage pump to a 15 × 105 W two-stage pumping system.
The present system consists of a fixed-speed primary pump, with a
variable-speed secondary pump controlled by the level in the surge
tank; both pumps are 0.53 × 0.51 m (21 × 20 in.) ASH GH-9-5. Many
in the industry do not use a variable-speed pump, but instead,
maintain level control in the surge tank by a controlled addition of
makeup pond water. Whichever method of flow control is used, the
slurry is then pumped to the top of the gypsum stack via a pipeline.
As with any slurry, pipeline design is of utmost importance to
minimize friction loss, yet maintain transport velocity and prevent
plugging in case of pumping system failure. This is especially true
when the storage area is often a mile or more distant from phosphoric
acid plants.
B
Distribution to the Settling Areas
Once the gypsum slurry is lifted to the top of the stack, it is generally
distributed around the stack perimeter for deposition into the settling
area(s) via pipelines and/or rim ditches. A cross section of a typical
perimeter dam and rim ditch is shown in Fig. 1. Rim ditches are
established by building a berm between the settling area and its
perimeter dam. The rim ditch method for slurry distribution has two
advantages:
1. Rim ditches assure capture of the coarsest gypsum particles and
provide gypsum above the elevation of the settling area. This results
in the coarsest and best-drained gypsum, which can be efficiently
excavated and used to construct perimeter and cross dams, as well as
the berms required to maintain the rim ditches.
2. Because rim ditches can provide a much greater quantity of gypsum
suitable for construction, larger settling areas can be maintained,
allowing increased water storage and thereby assisting with pond
water management.

Figure 1
Cross section of perimeter dam and rim ditch.
 Page 211
A series of field experiments were conducted at the New Wales
gypsum stack several years ago to better understand the hydraulics of
slurry flow in gypsum ditches and to optimize the rim ditching
methods. The following characteristics were determined [1]:
1. 0.014 to 0.018 roughness coefficient, as used in Manning's
equations for open-channel flow
2. 0.9 to 1.5 m/s transport velocity for a slurry flow of 1260 L/s at
20% solids
3. 0.003 to 0.004 downslope in the rim ditches to sustain transport
velocity
The rim ditch, which is initially excavated to a relative flat bottom,
fills with gypsum very rapidly. Sufficient downslope is quickly
developed in the ditch to attain transport velocity, and the gypsum is
conveyed to the settling area. It is important to maintain design flow
even when all phosphoric acid plants are not running. If flow rate is
reduced, hydraulic radius and velocity will also be reduced, and the
downslope will increase to compensate until transport velocity is
again achieved.
The slurry is generally directed to the furthest reach of the settling
area being used. Openings are cut successively through the berm in
about 120-m increments, until the entire settling area is beached with
gypsum along the length of the rim ditch being used, back to the point
of slurry introduction.
C
Decanting Water from the Settling Area
Some of the slurry water, perhaps even most, will percolate through
the gypsum stack to collection ditches at the base of the stack and/or
to the process water cooling pond. This is sometimes assisted by drain
pipe systems on the stack's face and/or under drain systems beneath
the stack. Many gypsum stacks have a spillway system with pipes to
decant and transport excess water from the settling areas to the
cooling pond. At New Wales excess water is decanted by excavating
through the perimeter dam and allowing the water to discharge down
the outside face of the stack directly to the cooling pond. Water has
been decanted in this fashion for over 10 years without erosion
problems on the stack's face.
III
Gypsum Stack Maintenance via the Rim Ditch Method
Unless encumbered by freezing temperatures, the perimeter of the
gypsum stack is generally raised in small increments of 0.6 to 1.2 m
as additional storage volume is needed. The gypsum itself is
excavated from the edge of the settling area to construct the perimeter
dams. This is the major expenditure in operating the gypsum stack.
Wherever rim ditches are used, the berm separating it from the settling
area will need to be raised similarly.
Many producers subcontract stack maintenance; at New Wales, the
staff performs this task. Three draglines with 24-m booms and 2.3-m3
buckets, have done the majority of the work involved in raising the
dams and berms with material excavated from the rim ditches. The
draglines work essentially around the clock except for downtime as
required for equipment maintenance. Some producers use backhoes,
and more recently, New Wales as well. Two bulldozers are also used
intermittently, to level the material cast by the draglines on the dams
and to assist with the ditch excavation. In view of the level of effort
and cost involved, a fairly extensive analysis of the rim ditch method
of stack
 Page 212
maintenance has been done. For a relatively simple activity, there are
a number of variables involved that affect the quantity of material that
must be moved.
A
Slope of the Stack's Perimeter Face
The outside slope of a gypsum stack is essentially a geotechnical
consideration, based on drainage conditions, stability of the geological
formation underlying the site, and design height. It is obviously
desirable to build the slopes of the stack face as steeply as possible to
minimize stack maintenance and to maximize storage capacity.
Geotechnical considerations at the New Wales site dictate a minimum
slope of 3H (horizontal):1V (vertical) on the perimeter face of the
existing stack, for a design height of 60+ m. Because of gypsum
consolidation as the stack increases in height and weight, the face will
settle to a flatter slope than that to which the perimeter dam is initially
constructed with each stack lift. Although many producers are able to
stack gypsum initially as steep as 1H:1V, at New Wales the gypsum is
generally stacked at about 2H:1V to avoid localized sloughing
problems, which have been experienced in the past. After stack
consolidation, this generally results in an overall slope of 4H:1V at the
existing stack.
B
Rim Ditch Downslope
The downslope in the rim ditches at New Wales generally ranges
between 0.003 and 0.004. The down slope is essentially self-
sustaining: whatever is required to maintain transport velocity for the
particular slurry conditions. A survey crew is not necessary to monitor
this slope. If the perimeter dams and ditch berms are built to a greater
slope than necessary, the rim ditch will not fill at its upper reaches,
and less construction will be necessary for the next run to the settling
area. If the slope is not sufficient, the rim ditch will fill in its upper
reaches before slurry can be directed to the entire settling area, and
more construction will be required before the next run to the settling
area.
C
Quantity of Settling Areas and Geometry
A plan view of a gypsum stack is shown in Fig. 2. It should first be
recognized that a minimum of two settling areas are recommended:
one that is in use and one that is being readied for use by raising its
perimeter dams. The stack maintenance cycle consists of four
activities related to the four rim ditches required to service two
settling areas:
Ditch 1. In use for the current run of gypsum slurry to a settling area.
Ditch 2. The ditch last used, which needs time to drain and
consolidate before it can be excavated efficiently.
Ditch 3. The ditch currently being excavated to raise its dam and berm
for the next stack life.
Ditch 4. This ditch should already be excavated and its dam and berm
raised, ready to be used for the next run.
At New Wales there have been as many as eight settling areas in the
existing stackthree areas are depicted in Fig. 3. It is planned to have
only two settling areas in a new stack just constructed. More settling
areas require more linear meters of dams and berms to maintain, but
the dams/berms need not be as high in smaller areas and will not
require as much digging per linear meter. Furthermore, the extra dams
required for multiple areas are cross dams, that is, dams that cross the
stack between settling areas.
 Page 213

Figure 2
Plan view of gypsum stack top.
These dams do not require as much digging as do perimeter dams,
since they can be lifted straight up and do not have to be stepped in as
do perimeter dams, which must follow the slope of the stack's
perimeter face. Overall, analyses indicate that digging quantity is
essentially the same regardless of the number of settling areas
maintained. The geometric shape of the settling areas is much more
important from a maintenance point of view. Long, narrow settling
areas will require much more digging for a given storage volume than
will relatively round or square ones.
D
Dam and Berm Widths, Slopes, and Freeboards
The operating philosophy with respect to the variables discussed
above is to use the minimum necessary to accomplish the task at hand,
with due consideration to the safety of equipment and personnel. It is
also necessary that the ditch width/depth provide sufficient material to
construct the dam and berm for the stack lift being implemented.
At New Wales the draglines require a minimum width of 8 m to
construct the dams on which they are supported. There has been
continuing success with berms constructed as steeply as the gypsum
will stand, generally 1H:1V, and as narrow as 0.3 m at the top. A dam
freeboard of about 0.3 m above the top of the berm is maintained to
assure that if the ditch does overflow with slurry, it will overtop the
berm and not the dam. Occasionally, the berms are overtopped before
a slurry run is complete, but another ditch is always already prepared
to address that eventuality.
Many producers construct a wider berm to support excavation
equipment, obviously requiring considerably more material, on the
order of 40% more.
E
Stack Lift
It is readily apparent that a higher stack lift will require more dam
construction per lift but will also provide more storage volume.
Construction material quantity is essentially proportional to stack lift,
but it is also proportional to dam and berm widths,
 Page 214

Figure 3
Aerial view of gypsum stack facilities, New Wales, Fl
(Courtesy IMCAgrico).
which due to sloping faces on everything being constructed, are also a
function of the lift dimension. To this extent, a portion of dam and
berm construction is proportional to the lift squared.
Based strictly on quantity, analysis indicates that about 15% more
work would be required to lift a stack of two 70-ha settling areas in
one 2-m lift rather than three 0.7-m lifts. Moreover, the ditch width
required to provide material for one 2-m lift may exceed the reach of
the draglines to construct the berm. All the construction at New Wales
is from the dams; this neccessitates giving careful attention to the
cut/fill balance to assure that stack lift and other parameters do not
require a ditch too wide for the draglines to reach the berm to be
constructed.
Nevertheless, we presently lift the existing stack, in 1.3-to 2-m lifts.
With smaller areas and reduced berm heights as the stack approaches
its design height, ditch widths do provide sufficient material within
the draglines' reach to construct higher lifts.
 Page 215
Generally, the ditches are maintained as wide as the equipment can
reach, and lift the stack as high as this material will allow.
F
Quantifying Dam and Rim Ditch Maintenance
A computer program has been developed to quantify the rim ditch
digging necessary to maintain the dams and berms of the existing
stack. In addition to the variables discussed above, the programs will
accept varying quantities of settling areas and resultant rim ditch
lengths. Such a program can be most helpful to evaluate proposed
changes to stack geometry and to forecast future dam and rim ditch
maintenance as stack geometry changes with increasing height.
IV
Sizing the Gypsum Storage Area
A gypsum stack is, of course, constructed from the bottom up. Its
design, however, should begin at the top, where the settling areas are
of minimum size. The following two approaches should be considered
when determining the minimum size of the top area.
A
Top Area Based on Minimum Settling Area
The first set of criteria should include the minimum size of an area
required to settle the gypsum [2]. A rigorous study of this has not been
made at New Wales. The smallest settling area tried was 12 ha in size,
and we did experience some gypsum carryover with the decant water.
The clarity of the decant water is of utmost importance, at least with
respect to suspended gypsum. Otherwise, the cooling pond to which it
reports will fill with gypsum. In the absence of better information, we
suggest 12 ha minimum per settling area to prevent gypsum carryover.
As already dicussed, a minimum of two settling areas are
recommended for the rim ditch method of stack maintenance. An
allowance also needs to be provided for the perimeter and cross dams,
ditches, and berms necessary to maintain the gypsum stack. At New
Wales 30 m is allowed all around for perimeter dams and ditches, and
46 m between areas for cross dams with ditches on both sides.
Overall top dimensions for two settling areas are then a function of
minimum settling area dimensions plus these allowances. Assuming
two square settling areas of 12 ha minimum (i.e., 346 m by 346 m)
plus the allowances discussed above, a minimum top area of 406 m by
797 m, or 32 ha, would be required.
B
Minimum Top Area Based on Production Rate
The second set of criteria to consider are stack lift, the density of
hydraulically deposited gypsum, and the stack maintenance cycle, to
assure that the top area will accommodate the appropriate rate of
gypsum production. As already discussed, the gypsum stack
operation/maintenance cycle consists of four activities related to the
four rim ditches required to service two settling areas. Equipment
resources aside, the controlling activity is the time it takes for the
ditch to drain after each use, before it can be excavated efficiently. At
New Wales it takes 4 to 5 days for the ditch to drain sufficiently. An
allowance of 20 days is made for the overall cycle.
For a production facility of this size, producing 1,500,000 tons of
P2O5 annually and gypsum at a rate of 5 tons per ton of P2O5, about
420,000 tons of gypsum might be
 Page 216
produced in a 20-day period. Based on an initial dry density of 720
kg/m3 for hydraulically placed gypsum, this quantity would require
approximately 583,000 m3 of storage space. Assuming a somewhat
conservative 1.3-m stack lift, a minimum top area of 45 ha would then
be desirable using this approach.
C
Sizing the Initial Stack
It is recommended that gypsum stacks and subsequent expansions be
constructed in increments to provide for a storage life of at least 10
years each. Recent permitting regulations in Florida require an
impermeable liner beneath new gypsum stacks. In view of the
tremendous expense involved to install liners, it may not be possible
for a facility to finance an initial gypsum stack or expansion that
would last much more than 10 years.
The principal elements of stack design that will dictate its size are
height, side slopes, and gypsum production rates. To maximize
storage volume, the stack should be designed with slopes as steep and
as high as possible, restricted only by pumping capacity or other
limitations that may be applicable at a particular site.
Stack height is especially important because it will also affect the
average in situ dry density of the storage pile. Although the gypsum
initially settles to a dry density of 720 kg/m3, it does consolidate and
become much denser as the stack increases in height and is surcharged
with the weight of additional gypsum. Core samples and periodic
aerial surveys of the existing gypsum stack at New Wales indicate an
average in situ dry density of approximately 1200 kg/m3 pcf when the
stack was 30 m high, and perhaps an average density of 1360 kg/m3
when the stack is 60 m high.
To demonstrate the effect of stack height and gypsum production rates
on land space requirements using the New Wales facility as an
example [3], given a phosphogypsum production rate of 7,700,000
tons per year, the storage facility would need to accommodate
77,000,000 tons of gypsum for a period of 10 years. Based on a
design height of 60 m, the proposed facility would then require a
storage volume of 5.66 × 107 m3 at a gypsum dry density of 1360
kg/m3.
The simplest way to estimate the required stack geometry is to neglect
the corners, which are not significant. The stack's volume V can then
be estimated by the following relationship:

where H represents the height, L the top length, W the top width, and
S the side slope (H:1V). Assuming that the stack's top length will be
twice its width with two settling areas of similar size, this relationship
can then be simplified as follows:

or

Using the quadratic equation to solve for the positive root of W, for H
= 60 m and S = 3, top width W = 558 m, and top length L = 2W = 1117
m. Bottom dimensions at the stack's ''footprint" would be greater by
2SH = 366 m. Finally, some allowances should be made for
equipment access and drainage ditches at the bottom of the stack, say
 Page 217
30 m all around, resulting in overall bottom dimensions of 985 m ×
1544 m. Design for a 30-m-high stack can be determined similarly,
resulting in the following comparisons:
Stack height (m) 30 60
Gypsum dry density (kg/m3) 1200 1360
Storage volume (m3) 6.42 × 107 5.66 × 107
Top area (ha) 184 62
Stack footprint (ha) 240 137
Total bottom area (ha) 260 152

A check of the top area indicates that the criteria are satisfied for a
minimum area, as discussed previously.
D
Land Planning for Gypsum Storage for the Life of the Facility
It would be shortsighted to restrict the height and/or size of a gypsum
stack without giving due consideration to the effect of such limitations
on the total land area that may be required for gypsum storage over
the total life of a production facility. Whatever size is selected for the
initial gypsum stack, it is not likely that it will suffice for the entire
life of the facility, which may exceed 50 years.
Two basic choices are available relative to stack expansions: whether
to build them contiguous to the initial stack or to construct separate
multiple stacks. If separated, total land planning for 50 years of
gypsum storage can easily be estimated by increasing the areas,
developed previously for 10 years of storage, by a factor of 5.
The overall storage volume capacity required for continguous stacks
is the same as that for separate multiple stacks. However, much less
land space will be used, because the land and air space that would be
left empty between multiple separate stacks can now be fully utilized
for gypsum storage. Land area can be estimated in the same fashion as
that for a stack with a 10-year life, increasing the required volume by
a factor of 5.
For a facility with an annual P2O5 capacity of 1,500,000 tons, the
following comparisons would result for the assumed 50-year life:
Contiguous stacks 60 m high 585 ha
30 m high 1161 ha
Five separate stacks 60 m high 760 ha
30 m high 1302 ha

A very significant premium in land space of up to 717 ha could be

paid if the most restrictive criteria considered were imposed.
Many factors should be considered when sizing gypsum storage areas,
including water balance, operating costs, geotechnical limitations,
production rates, reclamation, and other site-specific criteria that may
be applicable. From a land-planning point of view, gypsum storage
facilities should be constructed as high and as steep as practical to
minimize the land area required. Furthermore, when stack expansions
are required during the life of a production facility, they should be
placed contiguous to the existing stack(s) whenever possible.
 Page 218

V
Phosphogypsum Utilization
Phosphogypsum formed by the conventional dihydrate process consists of
fine crystals, contains considerable amounts of impurities, and has little use.
Better-quality phosphogypsum useful for wallboard and cement has been
produced by improved processes in Japan and other countries that have
limited land space for the disposal of the useless gypsum. A portion of
phosphogypsum has been used in a few countries to produce sulfuric acid
and cement by calcination with clay. In Japan the use of phosphogypsum has
been started for fertilizer or conditioner for certain types of soils. In this
section we describe the chemistry to improve the quality of phosphogypsum
and methods of utilization.
A
Impurities in Phosphogypsum and Ways of Reduction
By the conventional dihydrate process, pulverized phosphate rock is digested
at 60 to 70°C by a mixture of sulfuric and phosphoric acids to produce
phosphoric acid containing about 30% P2O5. The following reaction occurs
rapidly and results in the formation of fine needlelike gypsum crystals
smaller than about 50 µm, which contain considerable amounts of impurities
derived from the rock (Fig. 4, Table 1).

The hemihydratedihydrate process has been used widely to improve gypsum

quality. By the process, phosphated rock is first treated by the mixed acid
around 90°C to form calcium sulfate hemihydrate CaSO4 · 1/2H2O (Fig. 5
and 6). The hemihydrate slurry is then kept around 60°C to permit it to
convert slowly into the dihydrate, consisting of much larger platelike crystals
(Fig. 4). The crystals can be grown to have certain thickness so that the
product contains fewer impurities (Tables 1 and 2).
The dihydratehemihydrate process has also been used to increase gypsum
purity. By the process, gypsum slurry obtained by the conventional dihydrate
process is filtered and filter cake rich in phosphoric acid is added with the
concentrated sulfuric acid. The temperature of the mixture is raised by the
heat of dilution of the acid, resulting in the conversion of gypsum into
hemihydrate, which is filtered and washed with water. The by-product
hemihydrate is usually added with water and is granulated to produce
gypsum

Figure 4
Gypsum crystal geometry: (A) phosphogypsum (dihydrate process); (B) phosphogypsum
(hemihydratedihydrate process); (C) FGD gypsum (limestone scrubbing).
 Page 219
Table 1 Examples of Phosphogypsum Composition (%)
P2O5
Water
Process CaO SO3 Total soluble F SiO2 R2O3
Dihydrate 30.75 44.71 0.81 0.17 0.88 2.23 1.33
31.43 44.60 0.29 0.09 0.24 0.98 0.59
Hemihydrate 32.22 44.11 0.31 0.09 0.84 0.63 0.23
dihydrate
32.50 45.53 0.20 0.02 0.77 0.33 0.15
31.93 44.03 0.32 0.12 0.34 1.43 0.63
Hemihydrate 30.50 44.32 0.40 0.10 0.42 1.12 1.14

pellets for cement setting retarder. P2O5 content of the product is

usually less than 0.1% (Table 2). Recently, the hemihydrate process
became popular to by-produce concentrated phosphoric acid
containing 40 to 45% P2O5. After acid filtration, the hemihydrate is
washed and hydrated to give useful gypsum (Table 2).
Wallboard producers normally accept phosphogypsum with less than
25% moisture. The moisture content often exceeds 25% with the
dihydrate process using a continuous belt or traveling pan filter. The
moisture can be reduced to about 10% with the improved processes
and by using a centrifuge.
1
Effects of Impurities on Wallboard and Cement Setting Retarder [4]
P2O5 is undesirable for wallboard production because it delays the
setting of the hemihydrate obtained by calcining the gypsum to reduce
the strength of the board. Figure 6 compares the setting of typical
samples of hemihydrate made from natural gypsum and
phosphogypsum. The delayed setting of the latter may be due mainly
to P2O5.
P2O5 in the gypsum is present in various forms. Water-soluble P2O5 is
derived from phosphoric acid remaining on the surface of the crystals,
while water-insoluble P2O5 is

Figure 5
Stable range of dihydrate in phosphoric acid
containing sulfuric acid.
 Page 220

Figure 6
Hydration and setting of hemihydrate prepared from natural
gypsum and phosphogypsum by dihydrate process. (The
hydration ratio after 9 days was assumed to be 100%.)
mainly in the form of HPO4 2+ replacing SO4 2+ in gypsum. Small
amounts of unreacted phosphate rock and other impurities are also
present. Water-soluble P2O5 is undesirable and usually maintained
below 0.5% for wallboard production. It can be reduced by improved
water wash or lime addition. Unreacted phosphate rock may not have
an appreciable adverse effect.
On the other hand, P2O5 in the form of HPO4 2 is hardly reduced by
water wash and delays the hydration of the hemihydrate obtained by
the calcination of the gypsum. Some wallboard producers restrict the
P2O5 in the form of HPO4 2+ to below 0.3%. The HPO4 2 content can
be reduced by increasing the sulfuric acid content of the product
phosphoric acid but cannot be eliminated entirely.
It has been known that the combination of Al and F reduces
hemihydrate activity. The hydration of the hemihydrate proceeds
fairly quickly in phosphoric acid made from rock from Morocco,
which contains little Al, but slowly in acid made from rock from
Florida, which is relatively rich in both Al and F. Possibly the
combination in the form of complex ions such as AlF4 and AlF5 2
may play a role. Al and F in the product
Table 2 Impurity Content of Phosphogypsum (%) by Various
Processes
1 2 3 4 5 6 7 8 9
Process D D D HD HD DH DH H H
Rock F M K F M F K F M
Total P2O5 0.73 1.26 1.6 0.35 0.34 0.2 0.14 0.50 0.45
Water-soluble P2O5 0.42 0.28 0.20 0.05 0.05 0.03 0.18 0.10
F 0.57 0.64 1.7 0.45 0.66 0.15 0.46 0.40
Organic 0.18 0.15 0.05 0.03
D, Dihydrate; HD, hemihydratedihydrate; DH, dihydratehemihydrate;
H, hemihydrate; F, Florida; M, Morocco; K, Kola.
 Page 221
gypsum may delay the hydration of hemihydrate to some extent
during wallboard production.
The same impurities are also undesirable in gypsum for cement setting
retarder. Cement producers generally want to use gypsum containing
less than 0.05% water-soluble F. An appreciable portion of F in the
gypsum may be in water-insoluble forms such as Na2SiF6, Na3AlF6,
and unreacted phosphate rock. In cement setting retarder, water-
soluble P2O5 is usually kept below 0.1% and the moisture content
below 15 to 20%. When needed, the by-product gypsum is repulped in
water and centrifuged to reduce water-soluble P2O5 below 0.1% and
the moisture below 15%. Sometimes the gypsum is stored in an
outdoor open pile for months so that the water-soluble components
may be washed away by rain.
Organic compounds in phosphogypsum are present in various forms
and may affect the quality to some extent but have not been studied
extensively. Rare earth elements such as Sr, Ce, and La are contained
in phosphate rocks of igneous origin such as Araxa (Brazil) and Kola
(Russia) and also in gypsum produced with them. Those elements
may delay the hydration of hemihydrate.
B
Production and Use of Gypsum in Japan [5]
Figures 7 and 8 show the supply and use of gypsum in Japan. Since
Japan has no natural gypsum resources, nearly all by-product gypsum
has been utilized. Phosphogypsum has been used mainly for
wallboard and partly for cement setting retarder and plaster. Many of
the wallboard plants are located close to phosphoric acid plants. Some
of the acid plants supply gypsum to wallboard plants in the form of a
slurry.
Phosphoric acid production in Japan has decreased since 1980 with
the increase in diammonium phosphate import. Several smaller acid
plants have been shut down. Until 1985, the hemihydratedihydrate
process was the most popular. Currently, the hemihydrate process to
by-produce 40 to 45% acid is used primarily at larger plants with a
daily capacity of about 200 tons of P2O5.

Figure 7
Yearly gypsum supply in Japan.
 Page 222

Figure 8
Yearly gypsum consumption in Japan.
The selling price of the by-product gypsum was about 3000 yen/ton in
1970 and dropped to zero in 1975 because of the oversupply due to
the rapid increase in flue gas desulfurization (FGD) gypsum (see
Chapter 16, Fig. 4). Since 1978 the demand for gypsum has exceeded
the supply of by-product gypsum due to the increase in wallboard
production and stabilization in the FGD gypsum production. The
selling price was about 1000 yen/ton in 1978 and has been about 3000
yen/ton since 1982.
C
Production of Sulfuric Acid and Cement [6]
Gypsum is useful in producing sulfuric acid and cement by calcination
with clay and coal around 1450°C. Gypsum is decomposed to form
CaO and SO2. CaO reacts with clay to produce cement, while SO2 in
the gas is used to produce sulfuric acid by the conventional catalytic
process.
Natural gypsum has long been used for production. Major problems
with phosphogypsum are the high moisture content, which increases
energy consumption, and also the impurities, which have adverse
effects. A portion of F in phosphogypsum goes into the gas and may
poison the catalyst, necessitating gas cleaning. The phosphorus in
phosphogypsum goes into cement and tends to delay the setting. It is
generally required that the P2O5 in phosphogypsum be kept below
0.5%.
A commercial plant consuming 350 tons/day phosphogypsum has
been operated in Austria since 1969. A commercial plant consuming
350 tons/day phosphogypsum has been operated in South Africa since
1972. Tests have been made in Germany and the United States to
improve the thermal efficiency by using a fluidized-bed furnace to
replace the conventional rotary kiln. A large commercial plant may be
constructed in the United States using the improved furnace.
China has limited sulfur resources and has been interested in sulfuric
acid production from phosphogypsum. In 1984, Yunan Phosphate
Fertilizer Co. succeeded in producing acid with a good quality of
cement. In 1986 the company constructed a commercial plant
 Page 223
to produce 60,000 tons/year of sulfuric acid with 100,000 tons/year
cement using phosphogypsum. Additional plants may be constructed
in China.
D
Fertilizer and Soil Conditioner [7,8]
Gypsum is useful as a fertilizer for Ca-and S-deficient soils, such as
those found in Brazil and other countries. It is also effective as a soil
conditioner for acid soils rich in active Al in Al3+, which tends to
have adverse effects on plant roots. Acid soils may predominate in
nearly 30% of the total farmland in the world. Pulverized limestone
and lime are widely used to neutralize acidity. Those are usually
topdressed and neutralize the surface of soils up to 5 cm deep but not
the deeper part, because those are almost insoluble in water. On the
other hand, water-soluble calcium compounds such as calcium
chloride penetrates the soil too far.
Gypsum is slightly soluble in water and thus can penetrate 10 to 30
cm or even deeper according to the amount of gypsum and water
applied and thus can supply Ca to plants and can also reduce the
adverse effects of Al3+ by forming AlSO4+ . Phosphogypsum is
advantageous because it is powdery, requiring no grinding. In
addition, fluorine in the phosphogypsum may combine with A13+ to
weaken the adverse effect. In Japan, commercial application of
phosphogypsum to certain volcanic soils rich in active Al has begun
to show appreciable effects.
References
1. J. E. Cameron and R. T. McGillivray, Field measurements and
analyses to evaluate open channel hydraulics of gypsum slurries,
Environmental Proceedings of the 1984 Joint South Florida/Florida
Annual Meeting, American Society of Civil Engineers, Fort
Lauderdale, Fla., 1984.
2. J. E. Cameron and J. J. O'Connor, Design and operating criteria for
a gypsum stack/cooling pond from an owner/operator's point of view,
Proceedings of the 2nd International Symposium on Phosphogypsum,
Miami, Fla., 1986.
3. J. E. Cameron, Land planning for phosphogypsum stacks in central
Florida, 1988 Florida Section Annual Meeting, American Society of
Civil Engineers, Clearwater, Fla., 1988.
4. K. Murakami, M. Sekiya, J. Ando, K. Uemura, and S. Konno,
Gypsum and Its Utilization, Soft Science Co., 1976 (in Japanese).
5. Information given by the Sulfuric Acid Association, Japan, in 1992.
6. Developments in phosphogypsum utilization, Sulfuric Acid and
Industry, Vol. 45, No. 10, 1992 (in Japanese).
7. I. Shenberg, M. E. Sumner, W. P. Miller, M. P. Farina, M. A. Paven,
and M. V. Fay, Use of gypsum on soils: a review, Advances in Soil
Science, Vol. 9, 1989, pp. 111.
8. M. Saigusa, Plant growth at low-pH soils: a review, Niohn
Dojohiryo Kagaku Zasshi, Vol. 4, 1991, pp. 451459 (in Japanese).
 Page 225

19
Phosphoric Acid by Wet Process:
Pond Water Management
John E. Cameron
IMCAgrico Company
Mulberry, Florida
I
Introduction
Pond water is defined as whatever water is contaminated by the
phosphate fertilizer manufacturing process, including rainfall runoff,
generally ending up in the cooling pond or gypsum stack. With
respect to pond water balance and inventory, the active gypsum
storage facility is an inseparable extension of the cooling pond.
Successful pond water management is the ability to control pond
water inventories not only during normal weather conditions and at
normal production rates, but also to deal with major variances from
the assumed norms. It is obvious that makeup water could correct for
pond water deficits, but it may not be readily available. Water
treatment and reuse or discharge could deal with the reverse situation
of too much pond water, but this could be a very expensive solution.
In the case of a zero-discharge facility, discharge of treated pond water
is not permitted.
In this section we develop the primary inputs and outputs to the pond
water system and some methods that can be used to change these
inputs or outputs to achieve pond water balance. Some suggestions
will also be made to determine when to implement these methods.
Finally, a model will be suggested to determine the ability of these
methods to achieve long-term pond water balance.
II
Major Elements of Pond Water Balance
Aside from pond water recirculation used for cooling and other
process purposes, the principal water inputs and outputs to the pond
water system are illustrated in Fig. 1. Following is a development of
each component to determine its relative impact on pond water
balance, based on assumed parameters as indicated in Table 1.
 Page 226

Figure 1
Pond water system diagram.
A
Process Heat Evaporation
For an assumed pond water recirculation rate of 88,000 gal/min, an
average temperature drop of 19.5°F, and a heat of evaporation of 1050
Btu/lb, the process heat evaporation loss is estimated at 2696 acre-feet
of pond water per year.
B
Solar Evaporation and Evapotranspiration
Based on tests of solar pan evaporation rates of process pond water
versus fresh water, we estimate that solar evaporation of process water
ponds is 3% higher than that of lake evaporation rates, or about 50.5
in./yr. Annual evaporation on the 280-acre cooling pond would then
be 1179 acre-feet. An additional annual evaporation of 758 acre-feet
would
Table 1 Parameters Affecting Pond Water Balance:
Assumptions
P2O5 production rate 1.0 million
tons/yr
Gypsum production rate 5.0 million
tons/yr
Gypsum stack average dry density 75 pcf
Pond water recirculation rate 88,000
gal/min
Cooling pond temperature drop 19.5°F
Plant process water input 3000 acre-
feet/yr
Rainfall 54.2 in/yr
Lake evaporation 49.0 in./yr
Evapotranspiration 20.0 in./yr
Equivalent plant area reporting to the 60 acres
cooling pond
Gypsum stack/cooling pond watershed 775 acres
Cooling pond water surface 280 acres
Two gypsum stack settling areas 90/30 acres
maximum/minimum water surface each
Groundwater seepage net input 350 acre-
feet/yr
 Page 227
occur from the water surface on the gypsum settling ponds, assuming
that full water coverage of 180 acres is the normal operating
condition.
Evapotranspiration accounts for evaporation from the ground surface,
including transpiration of water through plants to the atmosphere.
Evapotranspiration rates on a gypsum stack/cooling pond system are
difficult to estimate. Some areas, such as nonponded portions of the
gypsum stack settling areas, are still generally very wet and will
experience relatively high rates; other areas, such as access roads, are
generally dry and will experience relatively low rates. A consultant
has judged an evapotranspiration rate of 20 in./yr as representative at
our facility. This rate is applied to the gypsum stack/cooling pond
watershed, less water surface areas, or 775 280 180 = 315 acres.
Annual evapotranspiration would then result in a pond water loss of
526 acre-feet.
C
Gypsum Pore Water and Crystallization Water
At a specific gravity of 2.33, the theoretical dry density of gypsum
with no pore spaces would be 2.33 × 62.4 = 145.4 pcf. Gypsum at 75
pcf occupies only 51.6% of the actual volume of a gypsum storage
facility. The remaining pore space is essentially saturated with water.
For a facility producing 5 million tons of phosphogypsum per year,
the gypsum stack would increase in volume by 3061 acre-feet. Of this
volume, 48.4%, or 1482 acre-feet, will be filled with water. It should
be noted that this is an oversimplification, in that dry density will vary
with height as the gypsum storage facility consolidates with increased
height and weight. This will result in much more pore water during
initial stages, and much less as the stack reaches design height.
The dihydrate gypsum crystal (CaSO4 × 2H2O) consists of 20.9 wt %
water. At a gypsum production rate of 5 million tons per year, water of
crystallization can be estimated at 769 acre-feet per year. Although
both pore water and crystallization water are internal to the pond
water system, they are treated as outputs with respect to pond water
stored in the cooling pond and gypsum settling ponds.
D
Rainfall
The major source of water entering a pond water system is usually
rainfall, on the gypsum stack, the cooling pond, and portions of the
plant that report to the cooling pond. The equivalent plant area is
determined by multiplying the actual plant area reporting to the
cooling pond by whatever runoff coefficient is judged appropriate. For
the stated equivalent plant area of 60 acres, 271 acre-feet of rainfall
runoff will enter the cooling pond with an assumed annual rainfall of
54.2 in. The gypsum stack/cooling pond watershed of 775 acres
would contribute an additional 3500 acre-feet per year to the pond
water system.
E
Groundwater Seepage
The majority of the cooling pond at the IMCAgrico New Wales plant
in Mulberry, Florida, is below the groundwater table. Net seepage has
been estimated at about 350 acre-feet per year into the cooling pond
system.
F
Plant Process Water Input
We have done a thorough water balance at the New Wales plant,
starting with actual well water usage, estimating cooling tower
evaporation and blow-down, vented steam and
 Page 228
other plant water vapor emissions, and even accounting for moisture
content in raw materials and products. The latest study indicated a net
process water input to the cooling pond system of about 3000 acre-
feet per year per million tons of P2O5 production rate.
Consultants checked this estimate of process water input by a
somewhat different approach, which they called a specific gypsum
disposal water balance study. This study considered water inputs and
outputs to the gypsum stack and cooling pond as described previously,
based on specific records over a 4-year period, including actual
production rates, rainfall, water surface areas, and soon. The specific
records were used to predict change in pond water inventory over a
period of four calendar years, and when checked against actual
changes, they yielded the following comparisons (in acre-feet):
1983 1984 1985 1986
Water inputs + 8196 + 6681 + 9845 + 8106
Water outputs 7716 8454 9217 8728
Predicted change + 480 1773 + 628 622
Actual change + 531 1595 + 547 358
Error 51 178 + 81 264

Given the magnitudes involved, we judged the indicated error to be

acceptable.
G
Apparent Pond Water (Im)Balance
Next we summarize all of the foregoing results for the set of
conditions prescribed. Inputs are represented as positive, outputs as
negative, the boundary being the water bodies available for
discretionary pond water storage, that is, the cooling pond itself and
the settling ponds atop the gypsum stack.
Inputs:
Rainfall + 3771
Plant process water + 3000
Groundwater seepage + 350
Subtotal (acre-feet/year) + 7121
Outputs:
Process heat evaporation 2636
Solar evaporation 1937
Evapotranspiration 526
Gypsum porewater 1482
Crystallization water 769
Subtotal (acre-feet/year) 7350

The total of inputs and outputs represents the change that is predicted
to occur in the pond water storage system. Thus a net pond change of
7121 7350 = 229 acre-feet is predicted to occur each year for the
parameters assumed in Table 1. Although this loss may be considered
significant, we prefer to operate on the negative side of pond water
balance:
1. Solar evaporation will inevitably decrease as the gypsum settling
ponds become smaller with stack height.
2. There is a ready source of makeup water from a large freshwater
pond.
 Page 229
H
Sensitivity of Pond Water Balance to Different Conditions
A few examples will demonstrate how sensitive pond water balance is
to weather and production rates, and help to place the importance of
pond water management into perspective.
Example 1: Annual rainfall of 64.2 in., exceeding the norm by 10 in.
An additional rainfall of 10 in. on the 835-acre watershed would result
in an additional water input of 696 acre-feet.
Example 2: Annual rainfall of 44.2 in. 10 in. less than the norm. This
would result in a decreased input to the cooling pond of 696 acre-feet.
Example 3: Ten percent drop in production rates. Pond water
inputs/outputs related to production rate are reiterated as follows:
Plant process water + 3000
Process heat evaporation 2636
Gypsum pore water 1482
Gypsum crystallization 769
Subtotal 1887

Assuming that production-related pond water inputs/outputs are

directly proportional to production rates, a 10% drop in rates for one
year would increase pond water inventory by 189 acre-feet.
III
Managing the Pond Water System
The inputs and outputs affecting pond water balance must be such that
balance is indeed feasible, at least during average weather conditions
and at normal plant operating rates. If not, even optimum pond water
management will only delay an inevitable imbalance that cannot be
managed successfully.
A
Plant Operation
In that excess pond water is generally more of a problem than pond
water deficits, following are a few suggestions that might be
considered to reduce water input to the cooling pond.
1. Reduction of plant watershed reporting to the cooling pond. Plant
areas not in risk of process contamination, such as parking lots, can be
directed to a separate freshwater pond.
2. Retrofit of pumps with mechanical seals, in lieu of seal water,
which invariably reports to the cooling pond.
3. Partial use of pond water in ball mills, if used in lieu of fresh water.
For every 10 acres of plant area runoff removed from the cooling
pond, water input can be reduced by 45 acre-feet during normal-
rainfall years, even more during high-rainfall years when it is most
needed. Every 10-gal/min reduction of water input equates to 16 acre-
feet per year. The cumulative effect of such changes can be of very
significant benefit to successful pond water management.
 Page 230
B
Cooling Pond Operation
As a minimum, there should be sufficient operating room between
maximum and minimum operating levels to at least accommodate
seasonal variations in rainfall. Assuming maximum/minimum
operating levels of ±2.5 ft from normal operating level, a 280-acre
pond would then have a discretionary storage volume of ±700 acre-
feet. Level control structures can be used to provide the desired range
in operating levels, or increase it if necessary.
If the cooling pond can be compartmentalized, portions of it can be
isolated and pumped out during drought periods to maintain minimum
cooling pond levels. In this way, water below minimum operating
level, normally at least 5 ft to assure adequate cooling, can be used for
makeup to the main pond. If an 80-acre portion of the pond could be
so isolated, 400 acre-feet of pond water could then be pumped to the
main pond if needed.
C
Gypsum Stack Operation
Both short term and long term, the method of operation of the gypsum
storage facility can have the primary impact on effective pond water
management. Water surface on the gypsum settling ponds can be
increased or decreased to change evaporation rates. Water depth in the
stack ponds can be manipulated to accommodate large water transfers
to and from the cooling pond. The gypsum settling ponds have the
greatest potential to accommodate large fluctuations in pond water
inventory which are beyond the capacity of the cooling pond.
Three alternative cross sections are developed in Fig. 2 for gypsum
settling ponds. Assuming that the ponds are relatively square in shape
and that a ''normal" condition of full water coverage exists as
indicated in Fig. 2B, the dimensions of a 90-acre pond can be
estimated at 1980 ft square. Assuming a "normal" underwater slope of
0.010 on the settled gypsum, depth can be estimated at 10 ft, and
storage volume at 300 acre-feet per pond.
During periods when cooling pond levels are projected to be too low,
water can be displaced with gypsum until the water surface is reduced
to the minimum required to settle the gypsum, say 30 acres, and the
underwater slope flattened to 0.005, as indicated in Fig. 2A. Storage
volume would then be reduced to 30 acre-feet, resulting in a
displacement of 270 acre-feet per pond to the cooling pond.
Furthermore, the reduced water surface of 30 acres per pond, even
allowing that evapotranspiration will still occur, will result in an
annual net reduction in evaporation of 30 acres × (50.5 20.0) in. = 76
acre-feet per pond.
Conversely, during periods when cooling pond levels are projected to
be too high, perimeter dams around the settling areas can be
constructed at an accelerated pace, more than that required to
accommodate gypsum disposal. The rim ditch method of gypsum
stacking is especially well suited to accomplish this. Rim ditches
allow short runs to the settling areas for the primary purpose of filling
the ditches with gypsum, thereby providing the material necessary to
accelerate the dam construction. It has been our experience that
underwater slopes steeper than 0.020 can be achieved in this manner.
At this slope, as illustrated in Fig. 2C, the storage capacity of each
settling pond can be increased to 600 acre-feet and thereby
accommodate a transfer of 300 acre-feet from the cooling pond.
It should be noted that the capacity of the gypsum settling ponds to
store and evaporate
 Page 231

Figure 2
Gypsum settling pond cross sections: (A) minimum water surface and depth;
(B) normal water surface and depth; (C) maximum water surface and depth.
pond water is reduced as the ponds become smaller with increases in
stack height. Although pond water balance may be achievable with
several settling areas during initial stages of gypsum stacking, it may
be desirable to reconfigure the gypsum stack to no more than two
settling areas as the stack approaches design height, to recover their
capacity to store water.
D
Effectiveness of Methods Suggested
One way to demonstrate the effectiveness of some of the methods
suggested above is to see if they could deal with the example
problems described earlier.
Example 2: Rainfall of 44.2 in., 10 in. less than normal, net pond
water deficit of
 Page 232
696 acre-feet. By displacing water with gypsum, it has been suggested
that 270 acre-feet could be recovered per gypsum settling pond. If
both ponds were managed in the same way, 540 acre-feet of water
could be displaced to the cooling pond. The reduced water surface on
the gypsum settling ponds would reduce water loss by 76 × 2 = 152
acre-feet per year. The combined effect would address 692 acre-feet of
the projected pond water deficit, resulting in virtually no change in
cooling pond level.
Example 1 and 3: combination of the circumstances described in
Examples 1 and 3, net pond water excess of 885 acre-feet. By
increasing water depth in the gypsum settling ponds as suggested, 600
acre-feet of additional water storage could be accommodated in the
two ponds. Assuming that the cooling pond was initially at midlevel,
the balance of the 285 acre-feet excess could be taken up with a 1.0-ft
rise in its operating level, still 1.5 ft below its maximum level.
IV
Determining the Status of the Pond Water System
A typical pond water balance has been developed for the prescribed
conditions based on normal production rates and rainfall.
Additionally, some examples of upset conditions have been given, and
some methods suggested to maintain pond water balance.
Changes in conditions related to plant operating rates are generally
planned, and appropriate adjustments can be similarly planned and
implemented to manage pond water inventory. Weather, however,
cannot be predicted reliably for more than a few days in advance. A
means is therefore required to determine the status of the cooling
ponds, what action should be taken and when, to change inputs and/or
outputs to the pond water system, or to move water between the
cooling pond and the gypsum settling ponds.
A
Cooling Pond Level Forecast
Before sufficient historical records were available, we developed a
computer model of pond water inputs and outputs to project daily
cooling pond level 12 months in advance, assuming normal rainfall.
This forecast was updated monthly, and if pond levels were predicted
to be approaching minimum or maximum levels, decisions were made
to implement the appropriate action.
The model assumed a uniform daily distribution of production related
inputs and outputs for the 12-month forecast. Weather-related inputs
and outputs were adjusted to reflect the normal monthly distribution
of rainfall and lake evaporation rates, based on locally available
weather records. As can be seen by the data in Table 2, rainfall and
evaporation can vary widely from month to month.
B
Historical Cooling Pond Level
At the New Wales facility, a wide variety of operating rates and
weather have been dealt with for almost 15 years without having to
treat pond water or use the emergency holding pond. On occasion
makeup water has been required. Monthly high and low cooling pond
levels for each month have been extrapolated from this track record to
determine a seasonally adjusted range of normal cooling pond levels.
Depending on whether the current cooling pond level is nearer the
historic high or low for the current month, the appropriate action can
be selected to adjust inputs, outputs, and/or storage capacity in the
gypsum settling ponds.
 Page 233
Table 2 Monthly Distribution of Rainfall and
Evaporation, IMCAgrico Company, New Wales
Plant
Pond solar
Rainfall evaporation Evapotranspiration
(in.) (in.) (in.)
January 2.4 2.5 1.0
February 3.1 2.6 1.0
March 3.4 3.9 1.5
April 2.5 5.4 2.1
May 4.1 6.3 2.5
June 7.9 6.1 2.4
July 8.4 5.7 2.3
August 8.7 5.3 2.1
September 6.8 4.5 1.8
October 3.0 3.4 1.4
November 1.7 2.5 1.0
December 2.2 2.3 0.9
54.2 50.5 20.0

C
Records
Adequate records are necessary to manage pond water effectively. In
addition to normal production records, the following minimum
records and frequency are recommended for this purpose:
1. Cooling pond level: daily.
2. Site-specific rainfall: daily.
3. Pond water inventory survey (cooling pond and gypsum settling
ponds): annual.
4. Gypsum stack water surface area: quarterly.
5. All other occurrences that affect pond water balance, such as
significant changes in watershed, plant process water input, makeup
water, and so on, should also be recorded.
D
Emergency Situations
During potential emergency conditions, such as when threatened by a
hurricane, there is normally sufficient freeboard in the gypsum settling
ponds to accommodate at least 2 more feet of water and 360 acre-feet
of emergency storage. Pond water can be moved quickly from the
cooling pond to the settling ponds simply by discontinuing decant
water and/or using the spare gypsum pumping system. With a pump
of 20,000-gal/min capacity, 176 acre-feet of pond water can be
transferred within 48 h, and storage increased in the cooling pond by
the same amount, to accommodate anticipated storm water runoff.
This precaution should be implemented before the storm; waiting until
after the storm is in progress may be too late if a power failure occurs
and the pumps are down.
Additionally, an emergency holding pond can be maintained,
connected to the cooling pond through a spillway, to receive pond
water overflow if all else fails. We maintain such a holding pond
approximately 80 acres in size, with 20 ft of empty storage depth
available below the maximum cooling pond level. This provides a
capacity of 1600
 Page 234
acre-feet of water, more than enough to accommodate a 12-in. storm
on our watershed. For this purpose, we calculate watershed as the
plant equivalent portion which drains to the cooling pond, the cooling
pond itself, and the gypsum stack, less the stack ponds. For the
conditions stated, (60 + 775) 180) × 12 in. = 655 acre-feet would be
sufficient for this purpose. The emergency spillway is designed to
gravity discharge this quantity of water from the cooling pond to the
emergency holding pond within 24 h.
V
Forecasting Long-Term Water Balance
The ultimate goal of pond water management is to achieve long-term
water balance, again with minimum make-up water or treatment and
discharge. It has already been stated that changes in pond water
balance due to production rates can generally be predicted with some
degree of accuracy. It has also been stated that weather is not
predictable, at least not rainfall. However, local records are usually
available to at least determine a long-term historical record of rainfall.
If it can be demonstrated that pond water balance could be managed
successfully over an historical record of rainfall, it is reasonable to
assume that it could be managed successfully in the future.
A
Developing a Computer Model to Forecast Long-Term Water Balance
At the New Wales facility a consultant was retained to modify inputs
to the computer model for forecasting cooling pond level as follows:
1. Reduce the surface area and storage capacity of the gypsum settling
ponds annually, as the gypsum stack increases in height.
2. Change lake evaporation and evapotranspiration outputs to reflect
the change in water surface area of the gypsum settling ponds.
3. Select an appropriate range of normal operating levels in the
cooling pond, to dictate when water should be transferred to or from
the gypsum settling ponds.
4. Select a minimum cooling pond level at which makeup water would
be made available, assuming that all available water has already been
transferred to the cooling pond from the gypsum settling ponds.
5. Select a maximum cooling pond level at which pond water would
be discharged to the emergency holding pond, assuming that the
gypsum settling ponds were already filled to their maximum storage
capacity.
6. Select an appropriate rate of water treatment for pond water stored
in the emergency holding pond, assuming that this water would be
preferentially returned without treatment directly to the cooling pond
whenever space is available.
7. Account for rainfall and evaporation at the emergency holding pond
whenever it is in use.
Actual rainfall can then be input from historical weather data for a
period matching the life of the gypsum storage facility. Evaporation
and projected production rates and related inputs and outputs will
complete the balance. For example, if 93 years of rainfall records
were available, and if the projected life of the gypsum storage facility
were 11 years, 83 different sets of historical rainfall data can be
iterated.
 Page 235
B
Computer Model Output
The model can calculate the following outputs for each of the 83 sets
of rainfall data:
1. Monthly record of pond water inventory
2. Monthly record of cooling pond levels
3. Number and quantity of makeup water occurrences
4. Number and quantity of overflow occurrences to the emergency
holding pond, and quantity of water treatment
5. Number and quantity of overflow occurrences that exceed the
capacity of the emergency holding pond
Judgments can then be made as to the acceptability of the probable
occurrences and quantities of makeup water and pond water treatment
that will be necessary to maintain pond water balance.
Overflow occurrences that exceed the capacity of the emergency
holding pond, however, are probably not acceptable. The risk of such
an occurrence can be predicted statistically; the more sets of rainfall
iterations, the greater the reliability. We suggest a 99% probability of
"zero discharge" as acceptable, that is, a 99% probability that no
occurrence would exceed the capacity of our emergency holding
pond.
C
Using the Results of the Computer Model
A computer model can be used to forecast long-term pond water
balance and a determination made as to its acceptability. If the
prediction is not acceptable, changes such as those already suggested
can be made to improve pond water balance and tried in the model
until an acceptable scenario is reached. The model can also be useful
to evaluate proposed changes that would have a significant impact on
pond water balance, such as operating rates or gypsum stack
expansions.
References
1. S. J. Babb and G. M. Hebbard, Computer modeling for cooling
pond management, paper presented to the Central Florida Section
Annual Meeting, American Society of Chemical Engineers, Lakeland,
Fla., 1985.
2. J. E. Cameron and J. J. O'Connor, Design and operating criteria for
a gypsum stack/cooling pond from an owner/operator's point of view,
Proceedings of the 2nd International Symposium on Phosphogypsum,
Miami, Fla., not published, 1986.
 Page 237

20
Phosphoric Acid by Wet Process:
Radioactive Components in Phosphoric Acid and Their
Removal
Florin T. Bunus
Institute of Atomic Physics
Bucharest, Romania
I
Radioactivity of Natural Phosphates
The occurrence of radioactivity in natural phosphates was established
by Strutt in 1906 [1] and uranium was found to be the main
component. The interest shown in this element began with the nuclear
era in the 1940s, when all new uranium sources were screened. The
uranium content of natural phosphates of sedimentary origin was
determined to be within the range 0.005 to 0.02% U and is attended
primarily by its decaying products [2].
A
Uranium in Sedimentary Phosphates
Radioactivity surveys confirm the presence of uranium in sedimentary
phosphates, where thorium is essentially absent, whereas in igneous
apatites, thorium is prevalent over uranium. Various studies have
suggested that uranium substitutes for calcium in the apatite structure
[3,4]. Autoradiographic methods with emulsion films have shown
dispersed points rather than clusters or segregations [2]. The evidence
of substitution of uranium for calcium seems to be supported by the
fact that uranium occurring naturally in the tetravalent form has an
ionic radius almost identical to that of calcium [5]. Some of the
principal natural phosphate deposits are summarized below.
The Florida Central Bone Valley Formation is characterized by 30 to
36% P2O5 and 0.014% U. A distinction is necessary between the
Bone Valley Pellets Formation, with 0.009% U, and Bone Valley
Pebbles, with 0.016% U.
The Phosforia Formation (Permian) from the Rocky Mountains has an
average content of 0.009% U and 30% P2O5, but appreciable
fluctuations are also possible and values of 0.02 to 0.021% U are
sometimes found [6].
Moroccan phosphate deposits [7], characterized by average content of
0.014% U and 34% P2O5,differ in character. Uranium concentration is
dependent on deposit age and
 Page 238
various layers: Ypresian layer of Khouribga, 0.013% U; second
Thanatian layer, 0.016% U; and Maestrichtian layers III and IV, 0.019
and 0.025% U, respectively.
Israeli phosphates are heterogeneous [7,8]. The Zefa deposit is
characterized by 0.012 to 0.018% U and 20 to 30% P2O5 content. For
the Saraf deposits, larger variations are involved, with 0.0067 to
0.0132% U and 21.2 to 29.4% P2O5. The Mishas Formation has
0.011% U and 20 to 25% P2O5 [9].
The Jordanian Ruseifa, Al Hasa, and Shadya deposits have a variable
uranium content of 0.0014 to 0.0156% U and 5 to 34% P2O5. There
are similar deposits in Egypt and Syria of the same origin. Tunisian
deposits have a lower uranium content (0.004 to 0.006% U),
comparable to that of neighboring North African deposits (R.
Dumitrescu, personal communication, 1984).
Central African phosphates are totally unlike, especially the Bakouma
deposits, which has a high uranium content, 0.166 to 0.56% U, and
corresponding P2O5 of 8.75 to 33.6%. Similarity is noted in the case
of phosphate deposits situated in lacustrine depressions in Siberia (R.
Dumitrescu, personal communication, 1984), Utah, and Wyoming
(Eocene) [10], with 0.3% U. Pacific Island deposits, with phosphate
derivatives from guano, were reported to contain low uranium
concentrations. Aluminous phosphates [9] from Florida leached zone
deposits are important due to large quantities with a uranium content
of 0.004 to 0.04%.
Russian volcanic deposits of apatites from the Kola Peninsula have a
very low uranium content, 0.001 to 0.002% [3]. South African
volcanic ores are similar, but the Araxa deposits of Brazil, of the same
origin, are completely different, with a 0.02% U content. This
enrichment is explained by the influence of external factors and
alteration processes [11].
The presence of uranium in natural phosphates of marine origin is
explained by its existence in seawaters as [UO2(CO3)3]4. Uranium
retention in apatites has occurred via organic matter, which is
abundant on the seabed. Organic matter exerts a reduction effect on
UO22+ and during the mineralization process will enter the
fluoroapatite as U4+ . Some of the UO22+ , which did not undergo the
reduction process, will be left as such and absorbed on the mineral
surface. A reductive phenomenon may be induced by light and
perhaps by nuclear radiation in the presence of organic matter [12].
Due to the presence of U4+ in minerals, by isomorphous replacement
of Ca2+, a linear relation between uranium and P2O5 is assumed. Such
relations are described in the literature, but nonobservance of this rule
is also mentioned [2]. Igneous apatites such as those from Kola do not
show such a linear relationship.
B
Radioactive Equilibrium
Uranium, thorium, and potassium are the only known radioactive
elements in nature. The half-lives involved have permitted these
elements to survive through the geologic ages. Uranium, by far the
most radioactive element, is found in natural phosphates. In the last
years of the past decade, 100 to 150 million tons/yr of natural
phosphates (80% of sedimentary origin) were processed, mostly for
chemical fertilizers. Assuming an average content of 0.01% U in
sedimentary phosphates, it is estimated that 10,000 tons/yr of uranium
is lost, with fertilizers produced and spread on agricultural lands being
a source of radioactive pollution.
With the exception of the United States, Belgium, and Taiwan,
uranium is not eliminate from phosphates (fertilizers) and is lost at the
same time that fertilizers are
 Page 239
spread on agricultural lands. By emitting nuclear radiation, uranium
decays to a chain of radioactive elements (radioisotopes).
Uranium as an element has two isotopes (ignoring 234U), both
radioactive: 238U with 99.28% abundance and 235U with 0.72%. The
half-lives involved are 4.47 × 109 and 7.04 × 108 years, respectively.
Therefore, uranium isotopes are emitting a and g rays. Spontaneous
fission is also present but at a reduced rate. 235U is actually the most
important naturally occurring isotope that undergoes fission with
thermal neutrons.
238U emitting a particles decays to a 234Th b emitter which decays to
234Pa. By b radioactive disintegration, 234Pa decays to 234U, which
by successive a emission is transformed to 226Ra [13]. 226Ra is a very
strong a emitter that decays on excited levels of 222Rn [13].
Radium (226Ra) is the main contaminant of phosphate fertilizers. One
ton of phosphate with 0.02% U will have 70 pCi/g Ra (0.07 mg Ra).
Therefore, 238U is transformed to 226Ra by successive decays, and
this process continues to the production of 206Pb, which is stable.
Similar behavior occurs in the case of 235U; thus 238U and 235U are
the parents of two radioactive series.
238U is present in natural sedimentary phosphates with its series of
decaying products. Radioactive equilibrium is often encountered
within the members of the series, where 226Ra in equilibrium with
222Rn are the main nuclides.
226Ra decays by energetic a particles on excited levels of 222Rn and
the process is attended by strong g emission. 226Ra has a long half-
life, comparable with that of 222Rn radioactive gas, which inhaled in
the lungs, decays at a high rate to a solid a and g emitter that will be
retained in the lung tissue. It is known that a particles release all the
energy in the tissue (3 to 5 MeV); therefore, 226Ra and 222Rn are the
most hazardous nuclides.
Radioactive equilibrium in natural phosphates is maintained in the
238U series but is not the general rule. Due to natural alterations,
leaching, and other processes, this equilibrium can be disturbed. When
radioactive equilibrium takes place, the following relation holds:
NRalRa = NRnlRn, where N denotes the number of nuclei and l is a
disintegration constant. Thorium is present in igneous phosphates as
232Th.

II
Phosphate Rock Solubilization by Sulfuric Attack
In the process of fertilizer manufacture, phosphorus from phosphate
ore is transformed in a soluble-assimilable form. This process is
carried out by sulfuric acid attack of the rock, resulting in either
simple superphosphate (SSP) or phosphoric acid. In the first case, the
actual tendency is to replace it with more concentrated products such
as triple-superphosphate (TSP), monoammonium phosphate (MAP),
or diammonium phosphate (DAP). In the first case, a soluble
phosphate form is obtained but gypsum is also present. In the second
alternative, gypsum is separated. Most of the phosphoric acid is
produced in the chemical reaction of rock dissolution by dihydrate
process. The acid obtained in this way, denoted WPA, is usually 26 to
28% P2O5 and must undergo a concentration step to attain 47 to 54%
P2O5.
Uranium in phosphate rock, predominantly in the tetravalent state, is
best solubilized concomitantly with an oxidation process. However, in
a phosphoric acid medium, uranium behavior depends on the redox
potential of the couple Fe(II)Fe(III). When H2PO4 is absent, Fe(III)
oxidizes U(IV) because standard redox potentials are Fe(II)Fe(III)
0.77
 Page 240
V and U(IV)U(VI) 0.33 V. In the presence of H2PO4 , oxidation by
Fe(III) is slowed down, and at concentrations higher than 2 M H3 PO4,
reverse reaction is possible [1416]. This behavior is due to strong
complexation of Fe(III) by H2PO4 .

The redox potentials of the couples involved are affected very much
by concentration of H2PO4 and Fe(II) while U(VI) reduction occurs.
In the dihydrate process dissolution of the phosphate rock takes place
in the presence of atmospheric oxygen, creating an oxidative medium
with a strong influence on Fe(II). Thus Fe(II) is oxidized to Fe(III) by
O2 [17]. The reaction rate is strongly dependent on H2PO4 ,
temperature, and O2 pressure.
In the dihydrate process, oxidation conditions are prevalent in most
plants. Uranium dissolved in WPA is found in the hexavalent state
[18]. If the Fe(II) concentration is higher than 0.4 g/L at 6 M H3PO4, a
reduction process is possible, but usually this value is lower.
Experimental measurements on WPA have indicated that a redox
potential of 0.35 V is sufficient to oxidize all uranium to U(VI). More
than 90% of uranium from phosphate rock is dissolved in WPA and
about 95% is in the hexavalent state [19].
Chemical analysis of clarified (aged) WPA of 26 to 28% P2O5
obtained from various phosphates determined for several years is
given in Table 1. The values in the table are in agreement with
literature data [2,20,21], but for the Florida phosphates, there are
various types with different concentrations.
The main component of uranium decaying products is 226Ra [22].
This element is carried in gypsum CaSO42H2O, due to the chemical
resemblance of Ca and Ra. Literature data have shown that some
226Ra is left in WPA [23]. Measurements on clarified, aged WPA
samples kept in sealed bags for 40 days indicate that this element is
absent. Literature data mention the existence in WPA of 1 to 2%
210Po initially in the phosphate rock, the remaining 98% being carried
on gypsum [24,25].
Gypsum radioactivity is important and is due primarily to 226Ra. In
the case of gypsum from sedimentary rock, its radioactivity is 25 to 45
pCi/g. Gypsum is deposited in large amounts near the plant site and is
a source of radioactive pollution, not only by the dust spread by wind,
but also by 222Rn gas evolved continuously.
Based on the literature references above, it was established that WPA
contains a small amount of radioactivity, the only component being
uranium predominantly in the
Table 1 Uranium Content in Technical
Phosphoric Acid
Uranium
Sample Phosphate origin (g/L)
1 Jordan 0.070.12
2 Morocco 0.120.16
3 Israel 0.070.14
4 Tunis 0.020.06
5 Florida 0.080.17
6 Togo 0.100.12
7 Egypt 0.080.11
8 Kola 0.010.02
 Page 241
hexavalent state. The existence of UO2 2+ species is evident, but the
complex charged and uncharged species with H2PO4 have also been
determined [19].
WPA must be concentrated, usually to 47 to 54% P2O5, before being
used in fertilizer manufacture. Uranium follows an increasing
concentration tendency, with phosphoric acid concentration to 0.17 to
0.26 g/L U (Jordan, Morocco, Israel, and Florida). Phosphoric acid
obtained from Tunisian phosphates has a lower uranium content, 0.04
to 0.06 g/L.
Volcanic phosphates processed to WPA, like the Kola type, have a
lower, almost negligible uranium content of 0.01 to 0.02 g/L. Thorium
is also present in small quantities (R. Dumitrescu, personal
communication, 1984).
In the experimental data of Table 2, radioactivities of phosphates and
phosphogypsum resulting from the dihydrate process are listed.
Phosphoric acid in concentrated form is obtained directly, on a smaller
scale, by the hemihydrate process. This process takes place at a higher
temperature, with more concentrated sulfuric acid, and in a reducing
medium. The phosphoric acid with higher concentration in H2PO4 has
a stronger complexing power on Fe(III). Smaller quantities of Fe(II)
have more efficient reducing power. The reducing medium of attack
maintains the tetravalent state of uranium in insoluble form, which is
carried on hemihydrate in greater than 70% concentration. Uranium in
this acid is usually 0.01 to 0.04 g/L, while the hemihydrate carries
0.06 to 0.1 g/kg, comparable to phosphogypsum, where the uranium
content is 0.015 to 0.03 g/kg [26].
Tetravalent uranium insolubilized and deposited on hemihydrate [24]
may be recovered partially by oxidative leaching. It is likely to be in
the form of UF4, which is difficult to solubilize even by an oxidative
effect. Most 226Ra is carried on hemihydrate.
III
Radioactivity of Fertilizers
Experimental results allow us to reach some conclusions on fertilizer
radioactivity. With the exception of single superphosphate, most
fertilizers resulting from the sulfuric attack of phosphate rock are
obtained from phosphoric acid. The usual products are triple
superphosphate (TSP), monoammonium phosphate (MAP), and
diammonium phosphate (DAP).
It was established that the only radioactive component of the WPA is
uranium. Therefore, a TSP fertilizer prepared from concentrated WPA
will contain 0.01 to 0.018%
Table 2 Radioactivity of Natural
Phosphates and Gypsum
Phosphate origin 226Ra(pCi/g)
Morocco 43
Jordan 32
Florida (central) 46
Israel 25
Kola (volcanic) 5
Morocco gypsum 30
Jordan gypsum 20
Kola gypsum 3
Concrete, bricks 0.761.7
 Page 242
U. At the same time, extra radioactivity due to uranium and radium in
the rock used in the reaction is also present. In general, TSP has a
radioactivity of 15 to 40 pCi/g 226Ra and 30 to 80 pCi/g 238U. A DAP
fertilizer has 25 to 30 pCi/g 238U and 0.5 to 2 pCi/g 226Ra. The SSP
will carry all the radioactivity in the rock.
TSP from Kola phosphate has very low radioactivity, 1 to 2 pCi/g Ra;
close to surrounding materials, such as bricks and concrete, it is 0.7 to
2 pCi/g. This is also true for DAP from Kola. In the case of phosphate
fertilizers obtained from phosphoric acid resulting from the
hemihydrate process, there is radioactivity from the rock in reaction
with phosphoric acid.
IV
Radioactivity Elimination from WPA
The types of fertilizers mentioned in this work are produced in large
quantities. Eliminating radioactivity from fertilizer itself is practically
impossible and very expensive. At present, concentrated fertilizers are
preferred and produced from WPA. The liquid phase of WPA, which
seems to be the most suitable form, permits removal of radioactivity.
It was well established that by the dihydrate process, the only
radioactive component in WPA is uranium. Uranium elimination from
WPA is desirable to obtain a nonradioactive fertilizer, but the uranium
is also valuable for nuclear purposes. The high price of uranium at the
end of the 1970s was responsible for developing technologies to
recover uranium from phosphates, a nonnegligible source of this
element. At the same time, large amounts of natural phosphates were
processed to obtain WPA and concentrated fertilizers. Recovery of
uranium from WPA was the main field of research and various
technologies were applied on an industrial scale. At present, the main
purpose might be the elimination of radioactivity for environmental
reasons. The economics of uranium recovery from phosphoric acid is
marginal, but removing radioactivity from fertilizers for the beneficial
effect on the environment may be the main incentive, together with
conservation of energy (otherwise lost), cost savings, and so on.
Experimental studies based on precipitation methods or ion-exchange
resins failed due to the low concentration of uranium and strong
complexing power of H2PO4 ions. Liquid membrane methods have
not yet shown results comparable with those of solvent extraction. At
present, the only techniques used on an industrial scale are based on
solvent extraction principles.
A
Solvent Extraction
Experimental studies carried out by Hurst et al. [21,2730] have
established that only a few extractants, mostly in the class of acidic
organophosphorus esters, are efficient for uranium extraction from
phosphoric acid media. Within this class, di(2-ethylhexyl)phosphate
(DEPA) possesses qualities that other members lack. This extractant is
efficient, very stable, easy to strip, and reasonable in cost, among
other benefits.
Other extractants of this class were proposed and even applied on an
industrial scale: alkylpyrophosphoric and phenylphosphoric acids
[i.e., octylpyrophosphoric acid (OPPA)] were used by Dow, and
octylphenylphosphoric acids (OPAPs) were used by ORNL. OPPA
was efficient for U(IV) extraction from WPA, but this extractant
easily hydrolyzes to mono-and dialkyl phosphates in acid media [24].
However, dialkyl phosphate may be a good extractant. In a mixture
with mono ester, it has an antagonistic effect. Similar behavior is seen
in the case of OPAP; although of greater stability, it also hydrolyzes
 Page 243
[28] in contact with an acid medium. Both extractants [28,30] are
difficult to strip; therefore, elements such as U, Y, and Fe may
accumulate in the organic phase. All these deficiencies led to the
predominant use of DEPA in all uranium recovery plants.
Based on literature data [31], it was established that DEPA in the
presence of a synergist of the neutral type, such as tri-n-butyl
phosphate (TBP) or tri-n-octylphosphine oxide (TOPO), gives
improved extraction [20,28]. The best effect is shown by DEPA +
TOPO (ORNL process) in kerosene. Therefore, all uranium recovery
plants in operation, mostly in the United States, use this extractant.
However, the base extractant is DEPA. The extraction efficiency is
dependent on its concentration, distribution coefficient D increasing
with DEPA concentration and decreasing significantly with H3PO4
concentration. Uranium, which must be in the hexavalent state, is
almost nonextractable from concentrated phosphoric acid (cWPA)
[31]. Temperature also has a negative effect on extraction.
The presence of a synergist may improve the value of D. In Fig. 1 the
introduction of TBP has an unsignificant effect [31]. A concentration
of DEPA higher than 1 M is needed. By contrast, TOPO has a very
pronounced effect, as shown in Fig. 2 [32]. The concentration of
TOPO synergic reagent is inscribed on each curve. In Hurst's ORNL
work it was established that the best concentration ratio of extractant
components is 0.5 M DEPA + 0.125 M TOPO [21].
Before being used in the extraction process, WPA must be [27]
clarified and organic components dissolved or colloidal matter
eliminated. Organic components in WPA are a source of trouble,
causing the accumulation of ''cruds" at the extractantWPA interface.
Crud formation is prevented by previous treatment of WPA, usually
with active carbon.
The extraction equipment used is a mixer-settler extractor, provided
with various improvements. After processing in a final separator,
where the two phases are separated, WPA is sent to posttreatment
equipment to remove small organic droplets entrained by the aqueous
phase. Uranium extracted in the organic phase is sent to the stripping
stage

Figure 1
Distribution coefficient as a
function of DEPA concentration
for U(VI) and U(IV) and 4.3 M
H3PO4. Numbers next to lines
refer to molality of TBP.
 Page 244

Figure 2
Uranium(VI) extraction from 4.3
M H3PO4 solutions with DEPA and
0.0 M, 0.05 M, 0.1 M, 0.3 M, 0.4 M,
0.6 M TOPO in kerosene.
in two or three countercurrent steps in a mixersettler type of extractor.
In contact with the stripping agent prepared from concentrated WPA
in the presence of Fe(II), uranium is transformed in a nonextractable
U(IV) species that is transferred to the aqueous phase. Due to the
aqueous/organic ratio, the strip solution will contain 10 to 20 g/L U, a
much higher concentration than that of the initial WPA.
Hurst has found that elimination of O2 dissolved in the organic phase
is a necessary step before stripping [29]. Otherwise, Fe(II) is quickly
oxidized, leading to excess consumption and other troubles. Oxygen
elimination is carried out by sparging nitrogen gas into the organic
phase.
To obtain a product of higher purity, a second extractionstripping
cycle is required. U(IV) in strip solution is diluted to 5 to 6 M H3PO4
and oxidized to U(VI).
The extractant used is the same mixture but the concentrations are
different0.3 M DEPA + 0.175 M TOPOassuring a saturation with
uranium of the organic phase and maintaining impurities in the
aqueous phase [33,34]. The stripping stage of the second cycle is
based on an ammonium carbonate reagent, which causes simultaneous
precipitation of an ammonium uranyl tricarbonate. This cycle needs
considerable attention and control.
In Figure 3 is a flowsheet of the process as used on an industrial scale
in the United States, Belgium, and Taiwan, based on a two-cycle
method. An alternative process using the same extractants, either
DEPA at higher concentration or DEPA + TOPO in the ratio noted
above, but in a one-cycle method has also been used. The same
reducing strip reagent is used, or one with similar properties may lead
to U(IV) transfer in the aqueous phase, where its precipitation is
possible with a selective reagent. In the latter method, the investment
and operation costs are lower and a useful green cake product is
obtained.
Hurst and Crouse [28] have proposed the use of OPAP in a two-cycle
process. In a similar way the WPA must be clarified and organic
components removed. In the first cycle, OPAP is used for uranium
extraction, which must be in the tetravelent state. For this purpose Fe
or Fe(II) is introduced in WPA, an important shortcoming of the
process. The strip process is carried out with 10 M H3PO4 containing
an oxidant. In this case the
 Page 245

Figure 3
Flowsheet for uranium recovery from phosphoric acid.
nonextractable species is U(VI). The strip solution may be diluted to 5
to 6 M H3PO4. In the second cycle, uranium is extracted from a more
concentrated solution in DEPA + TOPO solvent. The organic extract
is scrubbed with water and then stripped with ammonium carbonate
under conditions in which ammonium uranyl tricarbonate (AUT) is
precipitated directly. As already mentioned, OPAP extractant was
dropped due to its instability and difficult stripping conditions when
elements such as U, Y, and Fe are retained.
The OPPA [33] process first used in the early 1950s was neglected
until 1979, when it was used in a modified form at the Gardinier plant.
OPPA is even more unstable than OPAP but has the lowest cost of all
the extractants mentioned here. Uranium must be in the tetravalent
state, the extractable species, for which purpose Fe or Fe(II) must be
introduced in WPA. OPPA is prepared on site and has to be fed in the
extractant circuit continuously to offset its loss. The stripping reagent
was 15% HF, which led to UF4 · H2O precipitation. This process has
the advantages of operating on the one-cycle principle and yielding
UF4, which is a valuable component for nuclear purposes. OPPA has
many disadvantages, especially instability, the need for advanced
WPA cooling, and the requirement for reduction with Fe. The
advantage of its low cost was overridden by these disadvantages,
however, and the plant was shut down.
In all cases, WPA raffinate from the solvent extraction circuit contains
the entrained organic phase. After a separation process in a settler, the
raffinate still contains 200 to 500 ppm extractant, which must be
recovered. The process uses coalescing and air flotation to effect the
separation.
The few plants operating at present are based on the DEPA + TOPO
process. A general view is given in Table 3.
Many studies are being carried out to develop new processes as well
as to improve the existing ones. One new method developed in France
uses a "super" TOPO extractant [18], which has an improved
extraction effect but is cost prohibitive and at current marginal
uranium market prices is not competitive with DEPA + TOPO. A new
process developed by Uraphos Chemie and tested in a pilot plant uses
an amine as the extractant for uranium removal from WPA [35,36]. It
is important to note that all processes involving solvent extraction
require previous treatment with WPA to remove both solids and
dissolved organic matter. The same is valid in this case, but a more
thorough treatment is necessary. Some anions, such as Fand , have
low tolerance levels and Fe is introduced to
 Page 246
Table 3 Uranium Plants that Recover from WPA
Capacity
(tons
Country Name, location Method P2O5/yr) StartClose
U.S. URC-Grace, Bartow, Fla. OPAP 330,000 1976 1980
Farmland, Pierce, Fla. DEPA + 450,000 1978 1981
TOPO
Uncle Sam, Convent, La. DEPA + 820,000 1978
TOPO
Gardinier, Tampa, Fla. OPPA 450,000 1979 1982
IMC, New Wales, Fla. DEPA + 1500,000 1980
TOPO
CFI, Plant City, Fla. DEPA + 750,000 1980
TOPO
CFI, Bartow, Fla. DEPA + 600,000 1980
TOPO
Sunshine Bridge, Donaldsonville, DEPA + 480,000 1981
La. TOPO
BelgiumUmipray SA, Puurs DEPA + 130,000 1980
TOPO
Canada ES Western Co., Calgary OPAP 110,000 1980 1981
DEPA + 1983 1987
TOPO
Taiwan China Phosphate, Lung Tan DEPA + 33,000 1981
TOPO
Source: Ref. 18.

ensure 0.22 V (i.e., uranium must be in the tetravalent state). The

extractant used in this work is 2% trilauryl amine. To obtain uranium
extraction, it was required to introduce into the WPA a long-chain
polyphosphate complexing agent for U(IV). Stripping solution was
prepared from sodium carbonate, and finally, uranium was precipitated
with natrium hydroxide as diuranate. In the second purification cycle,
uranium diuranate was dissolved in sulfuric acid solution. This
aqueous sulfuric solution is contacted with the same extractant
(aminic) sulfuric medium, assuring due complexation. The same
carbonate stripping solution is used, and finally, pure diuranate is
precipitated.
B
Solid Ion Exchangers
Solid ion exchangers of more selective power have become attractive
in the last few years [37]. Solid resins can be used to treat WPA with a
lower uranium content. Nevertheless, their poor capacity in WPA and
other disadvantages limit their use. Some special resins were
impregnated with DEPA + TOPO or with aminophosphonic functional
groups such as DES-467.
C
Liquid Membrane Process
There are two types of liquid membranes:
1. Thin films supported by a solid; the liquid containing an extraction
reagent is impregnated in a solid polymer sheet (membrane). Uranium
is extracted by the reagent and transferred through the membrane to
the stripping agent. The extraction reagent is recirculated.
2. Emulsion membranes (liquid surfactant membranes) consist of a
stabilized dispersion of aqueous phase in organic phase. Thus the
stripping reagent is stabilized by a surfactant into the extraction
reagent. The organic phase is DEPA + TOPO and the stripping agent is
5 M H3PO4 with Fe(II), which together form a stable emulsion. This is
dispersed into WPA. Uranium is transferred to extractant, then to the
strip
 Page 247
medium. The process takes place in the mixersettler extractor with
slow stirring. Emulsion is finally separated from WPA and broken to
yield the extractant, which is recirculated, and uranium containing
strip solution. The latter is processed as usual in a second
extractionstripping cycle.
D
Hemihydrate Process
In the case of the hemihydrate process [18], 70 to 80% of the uranium
is carried on the solid phase due to reductive medium. Uranium is
partially recovered from the solid phase. Hemihydrate is converted to
gypsum, and uranium dissolved in the diluted WPA is recovered.
In the dihydrate process, in the presence of a strong Fe reducing agent,
the rock dissolution can be conducted so as to obtain most of the
uranium as U(IV), which is included in phosphogypsum. After solid-
phase separation, an acid leach in the presence of an oxidant may
recover uranium partially. As already mentioned, this leaching is
difficult.
E
Froth Flotation
Research was carried out on WPA based on components without
surface activity which were floated adding a collector (surface-active
reagent) [18]. The floated products were collected as solids on the
liquid surface [38].
F
Phosphoric Acid Extraction
There are solvent extraction processes applied to phosphoric acid
recovery from technical phosphoric acid in view of replacing that of
higher purity obtained in a thermal process. In this case uranium is
usually distributed between the organic and the aqueous phases and an
eventual selective stripping is needed for uranium recovery from
organic phase. This process is not efficient and is not important.
V
Fertilizers Produced from WPA after Uranium Removal
Gamma spectrometry of WPA samples kept 40 days in sealed bags has
clearly demonstrated the lack of radioactivity due to radium and its
decaying products. Therefore, the only radioactive component in WPA
treated by solvent extraction and eliminated is uranium. Some of this
element is still left in the raffinate [39,40]. The following are also
measurements using gamma spectrometry on WPA samples or other
methods: neutron activation analysis on similar samples, including
delayed neutron method, x-ray fluorescence, solvent extraction, and
spectrophotocolorimetry, have established that the uranium content in
the raffinate is usually 0.01 to 0.02 g/L. In concentrated WPA this
value is slightly higher, at 0.015 to 0.025 g/L U.
An eventual TSP fertilizer will contain only insignificant quantities of
uranium from CWPA, but uranium and radium from phosphate rock
used in the final reaction will be included in the product. The reaction

explains the existence of the extra uranium and radium.

 Page 248
In the case of DAP, only uranium from cWPA is present, and its
radioactivity is negligible after the extraction process. DAP is
obtained according to the reaction

It is evident that radium is absent in this fertilizer. Its radioactivity is

almost similar to that of natural materials: soil, construction, and so
on. By removing uranium from WPA there are fertilizers of
sedimentary phosphate origin with radioactivity similar to that of
volcanic Kola phosphates.
At present, there are no regulations concerning radioactivity in natural
phosphates destined for fertilizer production or for direct application.
There is no restriction concerning their use; some fertilizers have
much higher radioactivity than that of surrounding materials. It must
be remembered that fertilizers are spread on agricultural lands
continuously and to date, have not been shown to cause environmental
problems. Uranium is a heavy metal and in the presence of Zn, Cd,
Mn, and/or Ni might have a collective effect [39]. This is the chemical
aspect of the problem. Studies of uranium uptake [41] from the soil by
crops generally consider this effect negligible, but a definite
conclusion has not yet been found. Radium is important from both the
chemical and radioactive points of view.
In the process of uranium elimination from phosphoric acid other
elements are also removed, such as rare earths, especially the yttrium
group. Their role on crops is not known, but their elimination may be
beneficial.
VI
Choice of Industrial Process:
Cost and Equipment
In a metallurgical plant project, the ore that is processed is important
from both the technological and economic points of view. The
uranium recovery plant is not directly involved with mining problems.
Phosphate ores are used in phosphoric acid plants, which are
dependent on their characteristics. The mining cost is also supported
by the WPA plant. The uranium plant project is dependent on WPA
plant capacity and location. Actual recovery technologies are efficient
and applied only to the dihydrate process, which is by far the most
important process for WPA manufacture. Plant capacity is also an
important factor, due to the actual marginal uranium market because
only plants with large capacities have resisted this phenomenon. The
characteristics of WPA treated for uranium elimination play a decisive
role in the process.
Phosphoric acid may have a fluctuating chemical composition but the
main component, P2O5, does not fluctuate much (26 to 28%). Other
components with an influence on the uranium oxidation state or
complexation include Fe(II), Fe(III), F, and SO42 . As already
mentioned, the Fe(II) content is usually very low and uranium will be
present in the hexavalent state.
Another characteristic strongly influencing both capital and operation
costs of the uranium technology is the type of WPA. There are two
types with respect to dissolved organic matter: brown acid and green
acid. In brown acids, the dissolved organic matter or collodial may
reach 2 g/L, which is an important shortcoming that causes difficulties
or even makes the extraction process impossible, due to the
accumulation of large quantities of crud. Green acid, with a lower
organic content (100 ppm), does not present such problems. Organic
matter is extracted, then released as crud, which accumulates at the
solventaqueous interface, making extraction more and more difficult.
Organic matter
 Page 249
is only partially eliminated in the initial stage of acid clarification,
when solids (phosphogypsum) must also be removed from WPA.
Various flocculants and adsorption clays are used at the clarification
stage. Settled solids carry part of the organic matter of colloidal
origin. Brown acid needs a more advanced organic removal than that
needed for green acid. After the clarification stage, the WPA is treated
with activated carbon in several adsorption towers, where organic
matter is eliminated. In the case of brown acid plant design, one must
allow for frequent carbon regeneration with NaOH solution, which is
an expensive operation. A green acid uranium removal plant is 70% as
complex as a brown acid plant and the cost of operation is much
lower. In the last case, adsorption towers are not required.
In the initial stage, filtered WPA is cooled in a flash cooler to 35 to
40°C, because the extraction process is affected at higher
temperatures. A cooled acid is favorable for better elimination of
dissolved solids. The WPA type, brown or green, imposes a specific
technology. After purification, the brown acid becomes a green acid
and all remaining unit operations are the same. The degree of color is
usually measured photocolorimetrically at 408 nm and is considered
green when optical density is lower than 0.2 to 0.3. An absolute
determination of organic matter is not required. Other techniques
proposed for organic matter removal (e.g., chemical oxidation,
organic solvent scrubbing) were not efficient. Brown acids of high
optical density are difficult and expensive to process for the current
uranium market price. The continued operation at these facilities is
usually based on long-term contracts for uranium higher than the
current market price. Brown acids have a higher uranium content.
The next important factor that must be taken into account is the
uranium content of WPA in the case of uranium recovery for nuclear
purposes. WPA of Tunisian origin or other acid with uranium below
0.07 g/L is processed at higher cost. The higher the uranium content in
WPA, the lower the operation cost. In this case one combines the
advantages of uranium recovery with reduction of the radioactivity of
the final fertilizer product.
The type of extractant selected is determined by the technology
applied. It has been established that only DEPA and DEPA + TOPO
are efficient, selective, and stable. OPPA and OPAP were dropped for
two main shortcomings: extractant instability and difficult stripping.
Therefore, the only stable extractant is DEPA, eventually in the
presence of TOPO as synergist.
For technology based on DEPA on an industrial scale, the most
suitable stripping reagent is phosphoric acid containing Fe(II). There
are two alternatives to consider: a one-cycle process and a two-cycle
process.
Elimination of radioactivity from WPA must be combined with
uranium economic recovery, and this process is related to the uranium
concentration required for nuclear purposes. In a one-cycle system,
lower investment and operation costs are involved, but the resulting
product should undergo further purification. In the two-cycle process,
a higher-grade product is obtained, but investment and operation costs
are higher.
WPA raffinate posttreatment is also taken into consideration in view
of recovering the entrained extractant. The fine entrained extractant
droplets are partially separated in a settler, then the process is
completed in coalescing equipment or air flotation.
All these considerations will affect the investment and efficiency of
the process. Plant capacity, uranium content in WPA, and the type of
acid are the main factors on which investment and operation costs are
strongly dependent. Finally, the operation costs will be reflected in
reagent consumption, utilities (electricity, steam, water), and labor.
Utilities depend on eventual uranium recovery plant integration with
the sulfuricphosphoric acid
 Page 250
plants. Electricity and steam are produced by the sulfuric acid plant
and in this case the operation costs will be lower. Labor is a
significant percentage of operating costs and depends on plant
location.
For a one-cycle plant processing 100,000 tons per year P2O5, the
capital cost is approximately $10 million. The capital cost breakdown
is approximately as follows (1989):
Acid clarification and organic matter 3040%
elimination
Extractionstripping process 4050%
Acid posttreatment 57%
Uranium concentrate 5%
Operating costs (chemical reagents, $25 to
utilities, labor, etc.) are estimated at 30/kg U.

Uranium recovery plants must be installed near the WPA plants and
use the acid produced there. The equipment required for acid
conditioning consists of:
A flash cooler made of special steel resistant to WPA
Clarifierthickener systems, rubber lined
Filters
Activated carbon adsorption towers for organic matter removal from
WPA, rubber lined
Extractionstripping equipment
The main equipment is a four-to five-step countercurrent horizontal
extractor which is an improved mixersettler version of higher
efficiency. The stripping equipment is slightly different, but two steps
are required. Extractionstripping equipment is usually from ebonite-
lined steel. Similar materials are in use for the rest of the tanks,
vessels, and so on. Ebonite-lined steel gives good results. The same is
true of posttreatment equipment. While ebonite-lined tubes are used
for pumps, valves, and so on, other materials are similar to those used
in a WPA plant.
It is noted that WPA pretreatment represents an important share of
capital cost, but this is essential to obtain a clean, green acid easy to
concentrate, store, or transport, which are not negligible factors in the
WPA industry. Systems for preparation of reagents used in the process
are provided with automation and control equipment. Filtration (or
centrifuge) systems and drying of concentrate equipment are also
required.
VII
General Considerations
All economic considerations cited in the literature are based on
uranium recovery for nuclear purposes. Uranium removal related to
avoiding environmental radioactive pollution, discussed in this
chapter, might introduce the concept of evaluating the economic
aspects of the problem, which may gain increased attention in the
future. At present, the fertilizer manufacturer is interested in economic
conditions for uranium recovery, which are not favorable due to high
production costs and low uranium prices.
In large companies operating costs have been reduced, but uranium
content in WPA is an important factor. Uranium production from WPA
will survive, due to long-term contracts for the sale of uranium. It is
important to note that some plants operate with marginal profits
because of the unfavorable uranium market. Therefore, it is not
possible to extend the economic efficiency of the process to fertilizer.
 Page 251

References
1. J. R. Strutt, On the distribution of radium in the earth's crust,
Proceedings of the Royal Society of London, Series A, Vol. 78, 1906,
p. 150.
2. Z. S. Altschuler, The bearing of geochemistry on the recovery of
uranium and rare earths in phosphorites, IMPHOS Proceedings,
Boston, 1980, pp. 605623.
3. Z. S. Altschuler, R. S. Clarke, Jr., and E. J. Young, (1958),
Geochemistry of uranium in apatite and phosphorite, U.S. Geological
Survey Professional Paper 314-D, 1958, p. 45.
4. F. Habashi, (1962) Correlation between the uranium content of
marine phosphates and other rock constituents, Economic Geology,
Vol. 57, 1962, p. 1081.
5. G. Carabbi and F. Mazzi, The possibility of a partial substitution of
calcium with uranium in apatite lattice, Studie Ricerchedella
Divisione Geomineraria, Vol. 1, No. 2, 1959, p. 793.
6. R. A. Gulbrandsen, Chemical composition of phosphorites of the
Phosphoria Formation, Geochimica et Cosmochimica Acta, Vol. 30,
1966, p. 769.
7. M. Slanski, Uranium deposits in Africa: geology and exploration,
IAEA Report AG-109/14, 1978.
8. Y. Nathan and Y. Shiloni, Exploration of uranium in phosphorites: a
new study on uranium in Israel phosphorites, Proceedings of the IAEA
Symposium, Vienna, 1976, pp. 645655.
9. V. M. Gavshin, V. A. Bobrov, and L. S. Zorkina, Quantitative
relations between uranium and phosphorus in phosphatic sedimentary
rocks, Lithology and Mineral Deposits, Vol. 6, 1974, p. 118.
10. J. D. Love, Uraniferous phosphatic lake beds of Eocene age in
intermountain basins of Wyoming and Utah, U.S. Geological Survey
Professional Paper 474-E, 1964.
11. J. B. Cathcart, Uranium in phosphate rock, U.S. Geological Survey
Professional Paper 988-A, 1978.
12. F. T. Bunus, Photochemical reduction of UO2 + +, Chimia
Actinidelor, Stiintifica, Bucharest, 1981, p. 177.
13. B. G. Harvey, Radioactive decay, in Introduction to Nuclear
Physics and Chemistry, Prentice Hall, Englewood Cliffs, N.J., 1969,
p. 27.
14. C. F. Baes, Jr., The reduction of U(VI) by iron Fe(II) in
phosphoric acid solution, Journal of Physical Chemistry, Vol. 60,
1956, p. 803.
15. F. T. Bunus, V. Domocos, and D. Bold, (1978) Chemical analysis
of UF4 obtained in a wet process, Journal of Radioanalytical
Chemistry, Vol. 49, No. 1, 1978, p. 21.
16. F. T. Bunus, V. Domocos, R. Bulaceanu, P. Dumitrescu, and G.
Popescu, Uranium determination in phosphoric acid solutions,
Journal of Radioanalytical Chemistry, Vol. 33, 1976, p. 251.
17. M. Cher and N. Davidson, (1955) The kinetics of the oxygenation
of ferrous iron in phosphoric acid solution, Journal of the American
Chemical Society, Vol. 77, 1955, p. 793.
18. F. J. Hurst, Recovery of uranium from phosphates: current status
and trends, Proceedings on the Recovery of Uranium from Phosphoric
Acid, IAEA, Vienna, 1989, TECDOC 533, pp. 915.
19. B. J. Thamer, Spectrophotometric and solvent extraction studies of
uranyl phosphate complexes, Journal of the American Chemical
Society, Vol. 79, 1957, p. 4298.
20. G. Cordero, L. G. Jodra, J. L. Otero, and J. M. Josa, Recuperacion
del uranio del ácido fosfórico, Energia Nuclear, Vol. 21, 1977, p. 297.
21. F. J. Hurst, D. J. Crouse, and K. B. Brown, Recovery of uranium
from wet process phosphoric acid, Industrial Engineering Chemistry,
Process Design and Development, Vol. 11, No. 1, 1972, p. 122.
22. G. Zaddach, Gamma Spektren, Katalog von Ge(Li)-Spectrum,
Kernforschungsanlage Julich 225, 1973.
23. F. Habashi, The recovery of uranium from phosphate rock:
progress and problems, Uranium, IMPHOS Proceedings, Boston,
1980, pp. 629661.
24. T. Botella and P. Gasos, Recovery of uranium from phosphoric
acid: an overview, Proceedings on the Recovery of Uranium from
Phosphoric Acid, IAEA, Vienna, 1989, TECDOC 533, pp. 1733.
 Page 252
25. R. H. De Voto and D. N. Stevens, Uraniferous Phosphates:
Resources and Technology and Economics of Uranium Recovery from
Phosphate Resources, USA and Free World, U.S. DOE Report GJBX
110, Earth Sciences Inc., 1979.
26. F. J. Hurst and W. D. Arnold, A discussion of uranium control in
phosphogypsum, Hydrometallurgy, Vol. 9, 1982, p. 69.
27. D. W. Leyshon and R. T. Schneider, Phosphoric acid clarification:
practice in USA, Phosphorus and Potassium, No. 104, 1979, p. 32.
28. F. J. Hurst and D. J. Crouse, By product recovery of uranium from
WPA, Proceedings of the International Symposium of the American
Chemical Society on Actinide Recovery from Waste and Low Grade
Sources, New York, August 2425, 1981, pp. 201223.
29. F. J. Hurst and F. A. Posey, Long term potential of uranium
recovery from H3PO4, SEP Technical Manual, American Institute of
Chemical Engineers, New York, 1982, p. 184.
30. F. J. Hurst and D. J. Crouse, Recovery of uranium from WPA by
extraction with octylphenyl phosphoric acid, Industrial Engineering
Chemistry, Process Design and Development, Vol. 13, 1974, p. 286.
31. F. T. Bunus, Determination of low level of uranium in solutions
obtained by acid attack of phosphate rock, Talanta, Vol. 24, 1977, p.
117.
32. F. T. Bunus, C. V. Domocos, and P. Dumitrescu, Synergic
extraction of uranium from phosphate solutions with di(2-
ethylhexyl)phosphoric acid and tri-n-octylphosphine oxide, Journal of
Inorganic and Nuclear Chemistry, Vol. 40, 1978, p. 117.
33. W. L. MacCready, J. A. Wethington, Jr., and F. J. Hurst, Uranium
extraction from Florida phosphates, Nuclear Technology, Vol. 53,
1981, p. 344.
34. W. W. Berry, Recovery of uranium from phosphoric acid,
Chemical Engineering Progress, February 1981, p. 76.
35. D. Rohe, Uran aus phosphorsaure: längerfristig Chancen,
Chemische Industrie Vol. XXXVII, 1985, p. 76.
36. U. M. Boin, (1985) The new uraphos process, Proceedings on
Advances in Uranium Ore Processing and Recovery from Non-
conventional Resources, IAEA, Vienna, 1985, p. 231.
37. Y. Volkman, (1989) Recovery of uranium from phosphoric acid by
ion exchange, Proceedings on the Recovery of uranium from
phosphoric acid, IAEA, Vienna, 1989, TECDOC 533, p. 59.
38. S. Huwyler, Uranium Recovery from Phosphates and Phosphoric
Acid, Eidgenoessisches Institut fur Reaktorforschung, Wurenligen
EIR-C-1, April 3, 1985.
39. E. F. Dillard, A. W. Frasier, T. C. Woodis, and F. P. Achorn,
Precipitated impurities in 18460 fertilizer prepared from WPA,
Journal of Agricultural and Food Chemistry, Vol. 30, 1982, p. 382.
40. Z. T. Wakefield, Distribution of Cd and Selected Heavy Metals in
Phosphate Fertilizer Process, Bulletin Y-159 TVA, National Fertilizer
Development Center, Muscle Shoals, Ala., 1982.
41. J. J. Mortvedt, D. A. Mays, and G. Osborn, (1981) Uptake by
wheat of cadmium and other heavy metal contaminants in phosphate
fertilizers, Journal of Environmental Quality, Vol. 10, 1981, p. 193.
 Page 253

21
Phosphoric Acid by Wet Process:
Heavy Metals in Phosphoric Acid and Their Removal
Marinela Man*
Institute of Chemical Research
Bucharest, Romania
I
Toxic Potential of Heavy Metals
The toxic potential of heavy metals present in food and water has long
been known. Most of these metals accumulate in tissues and their
removal from the body takes a long time. It was concluded
unanimously that food carried most of the heavy metals taken into the
body from the soil by plants. Fertilizers are one of the factors involved
in soil pollution by heavy metals, but their contribution is smaller than
that of other activities, such as the nonferrous industry, fossil fuel
burning, industries using heavy metals or their salts, and waste
disposal.
The literature reports on the results of studies that have approached
trends of greater significance, such as:
Determination of heavy metal quantity needed by the human body and
plants, as well as of the optimum permissible levels
Evaluation of biotoxic potential, including manifestations in the cases
of deficiency or overdosage
Establishment of sources causing environmental pollution, especially
of the soil, with determination of their buildup
Almost all toxic metals are to be found in phosphate rocks, and thus
implicitly in fertilizers. The most frequent are Cd, Ni, Pb, Zn, Cu, V,
Mn, Mo, and Hg. Most of these metals are necessary to life, but each
has an optimum permissible threshold which when exceeded leads to
biotoxic phenomena.
Like mercury and lead, but unlike zinc, manganese, molybdenum, and
other elements, cadmium plays no part in the known vital biochemical
functions of plants or animals. It is one of the elements most toxic to
the living body as it builds up, especially in the liver and kidneys. As
cadmium can be removed from the body only with great difficulty, its
half-time amounting to 40 days and about 20 years in the liver and
kidneys,
*Retired.
 Page 254
it is a cumulative toxic substance. Its toxicity lies in its affinity for the
reactive enzyme groups, which contain nitrogen and sulfur. The main
toxic effects are neuropathological symptoms and kidney
insufficiency. Misbalance of vital minerals such as phosphorus and
calcium has also been found, which favors the occurrence of Itai-Itai
disease. For these reasons, in 1972 the World Health Organization
recommended a maximum daily dosage of 70 µg.
II
Heavy Metals in Fertilizers
Although phosphate fertilizers have long been known to contain
significant cadmium concentrations, due to cadmium's presence in
phosphates, it is only over the past few years that soil pollution by this
element has received particular attention. Depending on the origin of
the ore, cadmium may be present in a phosphate rock in one of the
following forms:
1. Contained in the crystalline lattice of apatite, as a substitute for the
Ca2+ ion
2. In the organic component of the phosphate ore
3. Bound to calcite, accompanying the phosphate ore
The content of fertilizer in cadmium and other heavy metals depends
on the nature of the phosphate rock, on the manufacturing process
(decomposition by sulfuric or nitric acids), and on the concentration in
phosphorus of the end product.
According to Mortvedt and Osborn [2], Florida phosphate rocks
contain between 3 and 15 ppm Cd, whereas the rocks in the western
United States go beyond 130 ppm, and Australian rocks range from 4
to 109 ppm. Table 1 shows the estimated cadmium levels of the rocks
in current use in industrial plants.
The table shows that the Kola apatite in Russia [the Commonwealth
of Independent States (CIS)] has an extremely low cadmium content
compared to the Senegal, Togo, and Nauru rocks with the highest
levels. Thus the cadmium input into agricultural lands through the
application of phosphorus fertilizers will be affected by the rock
origin. This is better illustrated by the data in Table 2, showing the
annual dosages of P2O5 used in countries of the European Economic
Community (EEC). These values vary considerably
Table 1 Approximate Cadmium
Contents of Commercial Phosphate
Rocks
Cadmium content
Country of origin (ppm)
CIS, Kola 1
U.S., Florida 615
U.S., Idaho 40340
U.S., North Carolina 36
Morocco 875
Israel 1530
Tunisia 2535
Togo 4280
Senegal 6690
Mauru 90
Source: Ref. 3.
 Page 255
and reflect not only the differences in the cadmium quantities in the
phosphate rock but also the fertilizer grades used and the differences
between application rates. Even so, the weighted mean cadmium
content of the phosphate rock entering the EEC would be reduced
from 26 ppm to only 20 ppm.
The Dutch and Danish governments have also proposed that the
maximum permissible cadmium content of EEC phosphate fertilizers
be set at the 15 mg/kg P2O5 by the mid-1990s. Later, Germany
followed these countries' proposals by passing severe laws against
pollution by noxious metals. It is estimated that installing equipment
to remove cadmium in fertilizer plants would increase product costs
by about 20%.
In the wet phosphoric acid process, heavy metals contained in the
phosphate rock get redistributed between phosphoric acid and waste
phosphogypsum. Von Plessen and Schimmel [3] mention that on
processing Morocco or Togo phosphate with sulfuric acid, about 80%
of the cadmium is to be found in the acid, the balance in the
phosphogypsum.
In his studies on phosphates from various zones of the United States,
Wakefield [5] has established that with central Florida grades, as
much as two-thirds of the cadmium content will be in the phosphoric
acid, the other third in the phosphogypsum. Slightly higher
proportions will be found, as in phosphogypsum from other U.S. areas
(Table 3 and Fig. 1).
III
Techniques to Remove Cadmium from Phosphoric Acid
The literature refers only to cadmium removal methods, ignoring
other heavy metals from phosphate rocks or the solution resulting
from acid decomposition, from wet filter grade acid or NP solution,
where phosphates are decomposed with nitric acid, respectively. The
process of cadmium separation from the end product is not known and
does not appear obtainable.
A
Cadmium Removal from Phosphates
1
Calcination
DorrOliver [6] has developed a process for removing cadmium by
fluidized-bed calcination of phosphate at 1100 to 1150°C for 30 min,
under reducing conditions. This
Table 2 Consumption of Phosphorus
Fertilizers and Cadmium Input in the EEC
Rate of consumption (kg Cd
Country P2O5/ha) input
Italy 29.8 1.6
Netherlands 57 4.6
U.K. 64 6.5
Germany 96 4.6
Belgium 144 9.4
France 76 5.4
Denmark 46 2.4
Irish 124 6.6
Republic
Source: Ref. 4.
Table 3 Distribution of Toxic Metals in Acidulation, Central Florida
Pounds per ton of rock feed
Component 28% WPAa Phosphogypsumb Rock input Outputc Wash water
P2O5 571.8999 47.1817 602.0000 619.0815
Ca 7.1487 789.5078 656.0000 796.6565
Cd 0.0063 0.0031 0.0100 0.0095
Cu 0.0147 0.044 0.0180 0.0588
Zn 0.082 0.0598 0.122 0.1418
Mn 0.3743 0.0629 0.416 0.4372
Ni 0.0442 0.0472 0.06 0.0913
Pb 0.0042 0.0189 0.018 0.0231
Cr 0.0988 0.0031 0.1040 0.102
V 0.2103 0.0031 0.238 0.2134
Al 10.7231 2.8309 12.8000 13.5540
Mg 6.0975 0.6291 5.0000 6.7266
Fe 17.2411 5.0327 14.4000 22.2738
aAssumed filter grade acid to contain 95% of input P2O5.
bAssumed 5.5 tons of phosphogypsum/ton of P2O5 in filter-grade acid.
cAssumed output to be the sum of acid content plus gypsum content.

technique is used in Nauru, where phosphate calcination is also necessary to

reduce the content in carbonate and organics. Thus the concentration in cadmium
and organic contaminants diminishes simultaneously to less than 20 ppm Cd and
0.01% organics.
FLS Miljo Environmental Management [1] (B. T. Anderson, personal
communication, July 1989) has improved the operating technique by ensuring
intrinsic destruction of the organic matter to reduce the cadmium content. The
process, shown in Fig. 2, integrated in the direct in-mine processing of
phosphates, permits net reduction of fuel
 Figure 1
Cadmium distribution in acidulation, central Florida region. (From Ref. 5.)
 Page 257

Figure 2
FLS process for cadmium removal from phosphate rock. (From Ref. 1.)
requirement for 900 to 1100°C calcination. The phosphate rock,
preheated in a set of cyclones, is then calcined in fluidized-bed
reactors in an oxygen-free ambience.
Pilot-scale experiments on Taiba rock (Senegal) have succeeded in
removing over 80% cadmium from the rock. Negev Phosphates [7]
uses a process similar to cadmium removal by calcination in a number
of cyclone furnaces at 850 to 1000°C with a dwell time of 15 s.
Cadmium removal efficiencies are in the range 50 to 80%. By this
process the calcite matrix around the apatite granule gets destroyed at
high temperatures and all the gas-entrained dust is retained by
cyclones.
To lower the calcination temperature (i.e., the energy requirement)
and to increase the amount of cadmium removed by as full burning of
Cd-containing organics as possible, Hoechst A. G. [3,8], furthering
this technique, has studied a chlorinating calcination. In this case the
temperature goes down to 700°C while cadmium removal efficiency
goes up to 90%.
During the process other metals, such as Fe and Hg, are also removed.
Phosphates will not sinter, thus maintaining their reactivity. As much
as 2% remnant chlorine must
 Page 258
be separated by water washing, so that all phosphates can be
processed by acid decomposition. Experiments on Togo phosphate
rock have led to full cadmium removal.
2
Solubilization with Acid Solutions
A lower cadmium content can be achieved by solubilizing its
components into a mineral acid liquor, able to dissolve only small
amounts of P2O5. Uhde GmbH Hoechst A. G. [3,9] has determined
that calcium ions reduce the solubility of calcium phosphates in
mineral acid. Therefore, they have chosen to use a weakly acid nitric
acid/calcium nitrate solution as a cadmium dissolving agent. As
proven by the data in Fig. 3, cadmium is separated so much better, as
the nitric acid/calcium nitrate solution is more dilute. Similarly, Fig. 4
shows that the pH level of the extraction solution has to be higher to
dissolve small amounts of P2O5. It should be noted that not all the
forms in which cadmium is present in phosphates are soluble in the
nitric acid/calcium nitrate solution. Thus with the Morocco phosphate
rock, 50% of the cadmium dissolves, whereas with the Florida rock,
the percentage is far lower.
The phosphorite thus purified is used for processing with nitric acid,
whereas the extraction solution, because of the cadmium presence,
cannot be discharged, as it requires pretreatment for cadmium
removal.
B
Acid Purification Process
1
Cadmium Removal by Precipitation
Like other heavy metals, such as Pb and Cu, present in phosphoric
acid, cadmium can be removed by precipitation to sulfides. CF Kalk,
Köln, Germany [10], owns a process by which cadmium is
precipitated by bubbling a hydrogen sulfide stream into minimum
45% strong phosphoric acid, at 0 to 100°C and 1 to 50 atm, for 1 to
360 min. The

Figure 3
Relation between cadmium removal efficiency
and pH of reaction mixture. (From Refs. 3 and 9.)
 Page 259

Figure 4
Relation between P2O5 losses and pH
of reaction mixture. (From Refs. 3 and 9.)
cadmium content diminishes from 65 ppm to 13 ppm. Instead of
hydrogen sulfide, a soluble sulfide, such as sodium, potassium,
calcium, iron, ammonium, and so on, can be used, in which case the
working pressure is attained with an inert gas.
An alternative of this method is described in a Japanese patent [11],
where water-soluble metal sulfide is first homogenized with calcium
hydroxide slurry, followed by its dosing in phosphoric acid. The best
results are achievable only when the Ca2+ to SO42 molar ratio is in
the range 1 to 1.5 and that of S2 to heavy metals is above 10.
Application of this process is limited by the initial phosphoric acid
grade.
A Swedish Company, Boliden Aktiebolag [12], has improved this
precipitation to sulfide, by predesulfurization of phosphoric acid, for
which purpose a calcium or barium salt is used. After sulfate
filtration, desulfurized phosphoric acid is partly neutralized with
sodium hydroxide at a Me/P ratio in the range 0.05 to 0.3 (preferably
0.1 to 0.2), where Me stands for the number of Na+ equivalents in the
alkaline solution, and P for phosphorus atoms contained in the
phosphoric acid. Further, soluble sulfide is added and it precipitates
other toxic elements, such as Cu2+, Zn2+, and As3+. In practice,
excess sulfide as high as 0.2 to 1 g of S2/L acid is necessary.
The process permits high purification not only of cadmium, which is
lowered from 15 to 20 ppm to less than 1 ppm, but also of copper and
arsenic. A number of purification methods are also known to use
organic sulfur compounds for precipitation of heavy metals.
Hoechst A. G. [1,1316] has established that good results can be
accomplished by using the esters of the following acids:
dithiophosphoric (RO)2PSSH, dithiophosphonic
 Page 260

Figure 5
Impurities purification process with
(C8H17O)2PS2H. (From Refs. 3 and 9.)
(RO)RPSSH, or dithiophosphinic R2PSSH, the last being used to
precipitate cadmium from the NP solution resulting from the nitric
attack of rock phosphate. The complex formed can be separated as
such or can be fixed with an adsorbant and then filtered.
From the esters of dialkyldithiophosphoric acid, Hoechst A. G. has
established that diethyldithiophosphoric or
di(ethylhexyl)dithiophosphoric has a particular affinity for Cd, Cu,
Pb, Hg, and As, these being likely to be extracted practically from up
to 80% P2O5 phosphoric acid, as shown in Fig. 5. An adsorbant of
high specific surface area such as aluminum silicate or a porous
polystyrene resin can be used. The organic 20 to 60% binder and 40 to
80% adsorbant can be added together or separately. A simplified
flowsheet of the Hoechst Adex process is shown in Fig. 6.
 Figure 6
Cadmium removal from phosphoric acid
by the Hoechst Adex process. (From Refs. 3 and 9.)
 Page 261
2
Cadmium Removal by Flotation
Methods based on ion flotation consist principally of the formation of
foam, by means of available surfactants, and in the removal of these metals
from the foam by means of specific flotation reagents: collectors or air
floated.
Uranium Pechiney, together with the Centre de Recherches sur la
Valorisation des Minerais (CRVM) [1719], have developed a procedure for
recovering cadmium from 28 to 30% P2O5 phosphoric acid, which, as
shown in Fig. 7, covers a four-stage process: (1) pretreatment of the crude
phosphoric acid, (2) complexing of the cadmium ion, (3) flotation and
removal of precipitate, and (4) deoiling filtration of the cadmium-free
phosphoric acid.
The first process stage includes three major operations: cooling phosphoric
acid down to about 60°C, necessary for the formation of the cadmium
complex; followed by the removal of suspended material using a flocculant
to increase the settling speed. The clarified acid is then passed over iron
scraps to reduce the Fe3+ present in the acid to Fe2+, a necessity for
enhancing the efficiency of the organic complexant, which consists of a
collector from the group of dithiophosphoric esters or their alkaline salt.
This precipitation stage must be conducted under a nitrogen blanket to
prevent the acid from reoxidizing. The foam formed during flotation can be
separated by filtration in a filtering bed composed of phosphogypsum,
which permits full removal of any complexant traces or of its degradation
products. The cadmium removal efficiency attains 90 to 95%, which
permits the content to go from 70 ppm down to 5 ppm.
3
Ion-Exchange Processes
Progressive lowering of the grade of industrially processed phosphates and
implicitly of that of phosphoric acid has convinced PECO of Lakeland,
Florida [1,20], to develop a

Figure 7
Basic outline of Uranium Pechiney cadmium removal process. (From Refs. 17 and 19.)
 Page 262
process based on ion exchange for removing sludge-forming cationic
impurities. Ion exchange against an acidic (H+ form) resin is well
known as an effective method of removing magnesium and other
impurities from phosphoric acid. The major disadvantage is the use of
spent dilute sulfuric acid used for stripping magnesium at resin
regeneration and its conversion to the H+ form. Magnesium can be
precipitated as insoluble double salt MgSO4·H2SO4, provided that the
concentration of regeneration acid is raised to 75% by blending it with
93% sulfuric acid.
The procedure, shown in Figs. 8 and 9, involves two distinct stages:
that of fixed-bed ion exchange and of processing the sulfuric acid
solution from the resin regeneration. The process efficiency is
relatively high, as the cadmium content of 30% P2O5 is reduced to 2
ppm; in addition to removal of 100% residual calcium, 74%
magnesium, and 28% iron, but only 4% of aluminum. For resin
washing, pond water can be used, which is then used in the
phosphoric acid manufacturing section to wash phosphogypsum. After
purification by this process, the acid strength decreases from 30.4% to
29.4% P2O5, requiring additional low pressure (145 kg/ton of P2O5)
to restore it to its original strength. Reagent costs are approximately
$5 per ton of P2O5.
4
LiquidLiquid Extraction Processes
From all available techniques for the removal of cadmium from
phosphoric acid, liquidliquid extraction seems to be best on an
industrial scale. This technique, also used for uranium extraction,
creates the same problem:
1. Need to clarify and settle the filter phosphoric acid prior to
extraction
2. Low distribution coefficient
3. Handling rather a large amount of solvent, in which the extractant is
dissolved
Generally, these procedures use an extractant dissolved in a solvent,
with continuous extractors, usually of the mixersettler type, as basic
equipment. Cadmium is then

Figure 8
PECO purification process: ion-exchange area. (From Refs. 1 and 20.)
 Page 263

Figure 9
PECO purification process: regenerating treatment area. (From Refs. 1 and 20.)
reextracted from the organic phase while the solvent and extractant
are regenerated and recirculated.
The literature mentions several procedures based on liquidliquid
extraction, the differences between these methods lying principally in
the extractant nature and less in that of the solvent. From the available
extractants the best results were obtained with amines or their
quaternary ammonium salts as well as with some organic sulfur
compounds. Next, the most representative processes specific to each
alternative are discussed briefly.
CF Budenheim Rudolf A. Oetker [19,2123] uses as cadmium
extractants tertiary aliphatic amines with C6C18 atoms, proven to be
most efficient; even better results have been obtained with amine
chloride salts or trioethylamine chlorohydrate, triisononyl amine, or
tridodecylamine dissolved in kerosene or other hydrocarbon, such as
xylol, or a mixture thereof. Cadmium is removed in the organic phase
as complex ions (R3NH)2+ (CdCl4 )2. The weighted ratio of the
organic phase to the aqueous phase is 1:3. The cadmium-rich organic
phase can be separated from the phosphoric acid and processed further
to be reextracted and to have the extractant and solvent recycled. For
reextraction an aqueous solution of sodium sulfate or sodium
carbonate at a controlled pH is used [8,9]. For instance, a solution
containing 5% sulfate is used to remove cadmium from the organic
phase. Chemical analysis of the phosphoric acid before and after
extraction is given in Table 4.
The same company [24,25] has also patented another class of
cadmium extractants: substances pertaining to phosphenes, phosphane
salts, or phosphonium compounds (e.g., 2% tridodecylphosphane
dissolved in a kerosenedecanol mixture or 2%
tridodecylmethylphosphonium chloride dissolved in the same solvent
mixture). Reextraction of
 Page 264
Table 4 Analysis of Phosphoric
Acid from Togo Rock Before and
After Extraction with Tertiary
Amines
H3PO3 (%)
Component Impure Purified
P2O5 52.6 52.6
H2SO4 4.4 4.4
HF 0.7 0.7
HCl 0.008 0.005
Fe 1.1 1.1
Al 0.5 0.5
Cr 0.2 0.2
Mg 0.09 0.09
Cu 0.01 0.01
Cd 60 ppm 1 ppm
Source: Ref. 22.

cadmium from the organic phase can be done with a 4.1% Na2SO4
solution (organic phase/aqueous phase ratio of 8:1), followed by the
precipitation of cadmium with sodium carbonate. The cadmium
content is lowered from 60 ppm to 1 ppm. A similar procedure, using
the same classes of solvents, has been developed by J.R. Simplot Co.
[26]: diisobutyl dithiophosphene or organophosphene sulfides such as
tri-n-propyl phosphene sulfide, dicyclohexyl phosphene sulfide,
triisobutyl phosphene sulfide, or dicyclohexyl dithiophosphinate. In
this case the polyvalent metals present in the acid must be previously
reduced to their minimum valence. The extractant is dissolved in a
solvent, preferably kerosene; however, other solvents, such as decane,
dodecane, toluene, p-xylon, or ethylbenzene, should be considered.
The basic characteristic of the process involves the use of liquidliquid
extraction for cadmium removal from the phosphoric acid combined
with ion exchange for recovering cadmium from the organic phase.
Also, depending on the extractant concentration, one can achieve
selective separation of heavy metals. Thus a low concentration in
organophosphenes favors cadmium extraction, whereas incremental
concentrations lead to successive separation of molybdenum and zinc
and then of nickel.
The main process stages as seen on the flowsheet in Fig. 10 consist of
contact in an extractor of wet phosphoric acid with the extractant (i.e.,
a thioorganophosphene) dissolved in kerosene. Cadmium separates as
a precipitate suspension in the organic phase. The raffinate, composed
of cadmium-free phosphoric acid, is directed to the next extraction
stage to have other heavy metals extracted after previously increasing
the extractant concentration of the organic phase.
The cadmium-rich organic phase is processed further to separate
cadmium and solvent extractant. This operation is carried on by two-
stage stripping with an ammonium hydroxide solution to keep
cadmium suspended. The cadmium-containing aqueous solution is
flowed over an ion-exchange column (e.g., Dowex 50 resin) to retain
the cadmium, which is eluded therefrom using a mineral acid. The
solution thus rendered free of cadmium is directed to the extractant
recovery stage, along with the organic phase from the stripping
process, where it is treated with sulfuric acid to convert it to acidic
 Page 265

Figure 10
J.R. Simplot process for recovering metal values
by ion flotation. (From Ref. 26.)
form slightly soluble in solvent. The results obtained by the process
described, using diisobutyl dithiophosphinate, are given in Table 5.
In addition to the foregoing process for removing cadmium and other
heavy metals from wet phosphoric acid by precipitationadsorption
with an organic sulfur compound, Hoechst A. G. has developed an
alternative that uses liquidliquid extraction, the extractant being the
same (i.e., the esters of dithiophosphoric acid) [3,27,28]. Di(2-
Table 5 Reduction of Heavy Metal Content in
28% P2O5 Phosphoric Acid
Extractant addition Acid analysis
Incremental Cumulative Cd Zn Mo Ni
(g added) (g added) (ppm)(ppm)(ppm)(ppm)
0 0 36.4 929 15.4 60
0.065 0.065 1.6 915 15.4 54.4
2.435 2.5 0.2 613 4.6 43.1
1.3 3.8 0.2 260 3.6 36.5
1.3 5.1 0.2 84 3.2 27.3
Source: Ref. 26.
Table 6 Removal of Heavy Metals from Wet Phosphoric Acid by LiquidLiquid Extraction,
with Various Dithiophosphoric Esters
Extraction Distribution coefficient,
(ppm) Ea
Example A/O R in (RO2)PSSH Phasea Cd Cu As Zn Ni Cd Cu Zn
1 2:12-Ethylhexyl a 11 15 4
b <0.5 <0.5 1 >57 >80
c 43 60 nd
2 20:12-Ethylhexyl a 21 36 10
b <1 <1 5 >433 >744
c 840 1500 nd
3 50:12-Ethylhexyl a 21 36 10 400 50
b 3 <1 5 390 50 300 >1813 1.37
c 1700 3400 470 1000 22
4 10:1Dodecyl a 21 36 10
b <1 <1 6 >213 >374
c 410 720 nd
5 5:1Octadecyl a 21 36 10
b 14 10 6 2.5 13
c 66 240 nd
6 20:12-Ethylhexyl a 6 8 2.5
b <0.5 <0.5 1 >235 >328
c 200 280 nd
7 50:12-Ethylhexyl a nd nd 1 nd nd
b <1 <1 0.3 540 60 >933 >1933 0.53
c 1400 2900 nd 430 70
Source: Refs. 27 and 28.
aa, Acid before extraction; b, acid after extraction; c, extract phase.
 Page 267
ethylhexyl)thiophosphoric acid or dodecyldithiophosphoric acid have
proved to be the most effective extractants. Heavy metal complexing
with these reagents runs according to the following reaction:

The 10% solution in kerosene serves as an extractant and the weighted

ratio of the phosphoric acid to the organic phase is 50:1, preferably
20:1.
The operation is effected in a single extraction stage even at a
phosphoric acid strength of 80%. Reextraction of heavy metals is
conducted in several stages: first with a dilute solution of 10% sodium
hydroxide, then with 36% hydrochloric acid. In the acid reextraction
stage the dialkylthiophosphoric ester present partly as a salt is
reconverted quantitatively to acid, being recycled to the extraction.
The alkaline and acid aqueous phases resulting from the reextraction
stages are combined and further neutralized with a base to halve the
heavy metals precipitated. Table 6 shows representative data on the
removal of heavy metals by the process discussed above.
References
1. Cadmium in phosphates: one part of a wider environmental
problem, Phosphorus and Potassium, No. 162, JulyAugust 1989, pp.
2330.
2. J. J. Mortvedt and G. Osborn, Studies on the chemical form of
cadmium contaminants in phosphate fertilizers, Soil Science, Vol. 134,
1981, pp. 185192.
3. H. von Plessen, and G. Schimmel, Verminderung des cadmium
Gehaltes von Rohphesphaten and Mineraldüngern, Chemie Ingenieur
Technik, Vol. 59, 1987, pp. 772778.
4. M. Hutton, The environmental implication of cadmium in
phosphate fertilizers, Phosphorus and Potassium, No. 123,
JanuaryFebruary 1983, pp. 3336.
5. Z. T. Wakefield, Distribution of Cadmium and Selected Heavy
Metals in Phosphate Fertilizer Processing, TVA Bulletin Y-159,
Tennessee Valley Authority, Muscle Shoals, Ala., 1980.
6. Dorr-Oliver, Inc., U.S. Patent PS 4,017,585, 1974.
7. Negev Phosphates, Ger. Offen. DE-OS 3,413,664, 1983.
8. Hoechst A. G., Eur. Patent OS 99,096, 1982.
9. Uhde GmbH, Hoechst A. G., Ger. Offen. DE 3,204,963, 1983.
10. CF Kalk GmbH, Ger. Offen. DE 2,422,902, 1975.
11. Jpn. Patent 53,075-196, 1978.
12. Boliden Aktiebolag, PCT/SE 80/00/00134, 1980.
13. Hoechst A. G., Ger. Offen. DE 3,202,658, 1983.
14. Hoechst A. G., Eur. Patent OS 91,043, 1983.
15. Hoechst A. G., Ger. Offen. DE 3,227,202, 1984.
16. Hoechst A. G., U.S. Patent 4,466,948, 1984.
17. Removing cadmium from phosphoric acid, Phosphorus and
Potassium, No. 137, MayJune 1985, pp. 2627.
18. Uranium Pechiney, U.S. Patent 4,634,580, 1987.
19. Removing impurities to improve phosphoric acid quality: acid
treatment, Phosphorus and Potassium, No. 147, JanuaryFebruary
1987, pp. 3032.
20. L. E. Bonffard and S. B. Morris, Removal of cationic impurities
from wet-process phosphoric acid, AIChE Joint Melting, Clearwater,
Fla., May 1988.
 Page 268
21. CF Budenheim Rudolf A. Oetker, Ger. Offen. DE 3,218,599,
1983.
22. CF Budenheim Rudolf A. Oetker, Ger. Offen. DE 3,327,394,
1985.
23. CF Budenheim Rudolf A. Oetker, Ger. Offen. DE 3,341,073,
1985.
24. CF Budenheim Rudolf A. Oetker, Ger. Offen. DE 3,342,211,
1985.
26. J. R. Simplot Company, U.S. Patent 4,511,541, 1988.
27. Hoechst, A. G., Ger. Offen. DE 3,127,900, 1983.
28. Hoechst, A. G., U.S. Patent 4,503,016, 1985.
 Page 269

22
Granular Triple Superphosphate and Single
Superphosphate
Henryk Jan Górecki
Technical University
Wroclaw

*, Poland

I
Technological and Environmental Aspects of Superphosphate
Production
About 35% of the phosphate rocks mined today are processed to
single or triple superphosphates. Although an increase in the world
production of fertilizer based on the wet-process phosphoric acid
process is expected, the superphosphates are still important fertilizer
products for agriculture [1].
Superphosphates containing monocalcium phosphate as a main
component are manufactured by acidulation of the phosphate raw
material by sulfuric acid, phosphoric acid, or a mixture of the two.
Until the early 1960s in the United States, enriched single (normal)
superphosphate (SSP) and triple superphosphate (TSP) belonged to
the basic phosphate fertilizers. Despite considerable progress in the
manufacture of superphosphates during the 19501960 period,
production of SSP has been declining and TSP production has
increased three times from 1960 to 1980, to a level of 780,000 tons of
available P2O5, and this level is actually constant. As an export
product, concentrated superphosphate has been exceeded by
diammonium phosphate [2]. Although the ''golden era" of
superphosphates is over, in many countries these products are still
used in agriculture. The new role for SSP production is the utilization
of waste sulfuric acid from other branches of the chemical industry
and metallurgy (e.g., copper, titanium oxide). Many superphosphate
plants, especially SSP plants in Europe, are located inside the town
agglomeration and are older than the wet-process phosphoric acid
plants. These facts and technological aspects of phosphate rock
processing cause many environmental problems, as a result of fluorine
and particulate matter emission from acidulation reactors, during
curing (aging) and in technologies that use recycling systems from
granulators, dryers, and coolers.
The intention of this chapter is to give a survey of pollution control
problems relating to the complexities of superphosphate production,
including preparation of raw materials, manufacture, storage, and
transportation. Local environmental problems related to
 Page 270
superphosphate production impinge on organization of raw materials
delivery, raw materials preparation (drying, grinding), technology,
treatment of the effluent, operation, and maintenance. Many of these
problems are similar to the control of emission in wet-process
phosphoric acid production, which is discussed in detail in Chapter
17, including the design issues related to absorption devices and
absorption systems.
A
Chemistry of Superphosphate Production Processes
Single superphosphate is made by reacting exothermically sulfuric
acid and a finely ground phosphate rock in a den or continuous reactor
[39]. During simultaneous acidulation and settling, gaseous effluents
containing CO2, water vapor, and fluorine-containing gases are
evolved. The process occurs according to the reaction

where y represents unreacted CaF2 from the fluoroapatite.

Calcium sulfate hydrate formed by the reaction is the main inert
component of this low-grade fertilizer product. The mineral admixture
of the fluoroapatite as nepheline and other aluminum silicates is
decomposed in the process:

Silica gel reacts with evolving HF:

Part of the evolving HF reacts with SiF4 in a liquid and gaseous form
according to the reactions
The HF evolved can be considered as an intermediate reactant in the
aqueous and gaseous forms with silica, forming a gaseous mixture
containing both SiF4 and HF and CO2 evolved in the process of
magnesium and calcium carbonate decomposition by sulfuric acid.
The properties of the final product depend on admixture content
which reacts with sulfuric acid and by-products according to the
reactions
 Page 271
During superphosphate production the following reactions are
possible:

which lead to lower concentrations of soluble P2O5 in water.

A similar mechanism of phosphate rock conversion takes place during
triple superphosphate production, where the principal process occurs
according to the reaction

In this process HF formed reacts with aluminum silicate, forming

SiF4. The evolving of SiF4 is also an effect of the reaction of CaF2
contained in phosphate rock and present in a fresh reactive mass of
"free" phosphoric acid, according to the reaction [6]

In this process the wet-process phosphoric acid used as a raw material

contains fluorine, whose concentration depends on the technological
system. Acid of 48 to 55% P2O5 concentration contains 0.9 to 0.4% F
and is used in den and continuous den process, and 28 to 42% P2O5
containing 1.9 to 1.1% is used in many slurry-granulation processes
[3,4,9,10].
Superphosphate processes are based technologically on the summary
reaction (1), and (17) depends on the previous rapid surface reaction
of fluoroapatite with sulfuric acid to phosphoric acid and CaSO4 ·
0.5H2O and in the next stage of reaction between fluoroapatite and
phosphoric acid. This stage lasts for several days. The reaction
product as colloidal monocalcium phosphate in a liquid phase
containing 15 to 16% H2O is characterized by tiksotropic properties.
During further processing, the liquid mass, in solidifying forms
crystalline Ca(H2PO4)2 · H2O. Processing of phosphate raw material
to superphosphate product may be realized in many ways.
B
Technological Superphosphate Production

1
Single Superphosphate
The manufacture of SSP involves:
1. Size reduction of phosphate rocks (90% < 100 mesh)
2. Acidulation in special mixers [1,3,58,1015] (cone TVA, pan,
cylindrical, TVA sigma-blade turbine mixers)
3. Solidification in a den by batch mixing (Wenk, Beskov, Milih,
Svenska den) [in most working plants solidification is done using a
continuous dening process at > 100°C and a retention time of 0.5 to 4
h (conveyor belt, Broadfield mixer and den device, TVA belt,
Nordergren den, Moritz-Standaert den)] [1,3,58,11,13,1517]
4. Final curing (completion of the reaction and hardening during 4 to
6 weeks of storage) [1,6,7]
5. Finishing (size reduction to homogeneous form bagging)
Recently, various methods have been developed, primarily continuous
methods using special solutions and mixer and den construction
[5,9,15,16].
An example of continuous methods of nongranular SSP production is
presented in Fig. 1. It is also possible to granulate nongranular product
(SSP) cured on a pile over a 4-week period in a typical recycle
system, including grinding, screening, and grinding
 Page 272
oversized particles. A similar system is used for granulation of cured
run-of-pile TSP. The points of emission of gases and dust are
indicated in Fig. 1.
2
Triple Superphosphate
The manufacture of TSP highly concentrated phosphoric acid
containing 40 to 46% P2O5 is realized in many systems and forms:
1. Nongranular (a method similar to SSP production)
2. Granulation of cured run-of-pile TSP (ROP TSP)
3. Direct granulation (GTSP) methods comprising (a) a slurry
granulation process and (b) an ex-den granulation process
The principles behind these GTSP processes and sources of emission
fluorine and dust are illustrated by Figs. 2 and 3 [57,1017].
The basic differences among these systems are:
1. The slurry processes known as Dorr-Oliver and Jacobs Dorrco use
low-analysis wet-process phosphoric acid of 38 to 42% P2O5
concentration as a raw material, and the resulting slurry is mixed
continuously with a large quantity of recycled dried fines in a blunger
(pug mill mixer) or a rotary drum, where slurry coats out the granule
surface.
2. The ex-den process is based on the direct granulation of a plastic
ex-den mass, drying, screening, and conveying to storage.
3. In the Simplot process GTSP is produced by treating limestone with
wet-process phosphoric acid in a direct granulation system under
conditions of low fluorine emission [1820].
4. Granular TSP is stored for a much shorter curing period (2 to 6
days) than that of nongranular TSP.
5. Cured nongranular TSP is granulated after curing periods of 3 to 6
weeks.

Figure 1
Manufacture of granular SSP (run of pile).
 Page 273

Figure 2
Slurry process of GTSP production.
3
Enriched Superphosphate
Enriched superphosphate is a term incorporating all grades of
superphosphates containing less than 46% and more than 20% P2O5.
This product is manufactured using a mixture of sulfuric and
phosphoric acid in a system technologically similar to that used for
TSP production [1,2,5].
4
Partially Acidulated Phosphate Rock
Partially acidulated phosphate rock (PAPR) is a mixture of
superphosphate, CaHPO4 · 2H2O and undecomposed phosphate rock.
PAPR contains CaHPO4 · 2H2O in a form
 Figure 3
Ex-den process of granular triple, single, or enriched superphosphate production.
 Page 274
insoluble in water but available for plants. This product is made in an
SSP or TSP plant using a lower proportion of H2SO4 to phosphate
rock [1,21,22].
C
Source of the Gaseous Pollution in SSP and GTSP Production
The primary source of pollution during superphosphate production are
gases evolved during chemical processes involving the phosphate raw
material and strong acids and particulate matter, such as fine
phosphate rock and dust from the final product, crushed silica. The
source of fluorine is a phosphate rock and wet-process phosphoric
acid in TSP production. A large percentage of introduced fluorine
remains in the final product, probably in the form of fluorosilicates,
partially as CaF2 and fluoro-undecomposed fluoroapatite. The amount
of fluorine evolved depends on the active silica content in phosphate
rock, the concentration of H2SO4 (SSP) and H3PO4 in TSP
production, and technological line construction. The fluorides in the
gas streams are in both the gaseous and particulate states.
The active areas of evolution of gaseous fluorine compounds are
indicated as emission points in Figs. 1, 2, and 3. Some unit stages emit
particulate only, some emit gases only, but most of the processes used
evolve a combination of fumes, mist particulate matter, carbon
dioxide droplets of liquids, and water vapor. The important sources of
SSP production are the acidulation mixer, the den (either closed or
continuous), the cutter, conveyors, transfer dumping and reloading
points, curing storage building, and bagging devices.
The main pollutants during the manufacture of granular TSP are
evolved during reaction in a mixing vessel or cone; on a continuous
slurry belt or curing belt; on the cutter; at the point of transfer of the
hot mass during granulation; during drying, cooling, or screening; in
the conveying system; during storage and bagging; and during bulk
blending. The amount and composition of the pollutants are specific
for various technological systems (e.g., granulation of run-of-pile
superphosphate, ex-den granulation, slurry granulation processes).
Digestion devices and dens must be enclosed and connected to a fume
exhaust scrubbing system, and at the other points where pollutants are
emitted, special eaves are mounted and connected with a scrubber or
collection duct spray system [1,6,20,23,24].
In SSP and TSP manufacture an individual separation system is used
for the main operation, although collective systems for the complete
technology are known. The effectiveness of the fluorine abatement
system depends on proper design, scrubber efficiency, effective
collection, the ability of the scrubbing media, the breakdown
emissions during suffer damage and occasional cleaning and dredging
silica deposits. The controlled emission is valid simultaneously for
fluorine compound recovery and for good ventilation of the work area
to allow employers to operate, especially in storage buildings, where
fluorine continues to be evolved from superphosphate during curing
and storage.
In addition to the primary contaminants developed during TSP and
SSP manufacture (fluoride compound particulates), secondary
pollutants arise as a result of reactions occurring in the scrubbing
system. As the result of this effect, plugging occurs in the pipe system
and the composition of the scrubbing media is changed. Additional
pollutants are also present when waste sulfuric acid containing
nitrogen or sulfite compounds is used, and SO2 and NOx are present
in the gaseous pollutants.
The same effect is observed during the absorption of vent gases from
dryers heated
 Page 275
by contaminated fuel [23]. Although the amounts are small, additional
fluorine emission takes place during transportation in bulk and during
bulk blending. In special, rather rare conditions when phosphate rock
is calcined in the range 650 to 1000°C, fluorine in the form of SiF4 is
evolved in rather high concentrations (between 20 and 70 g F/m3).
Fluorine is also evolved, principally in the form of HF, in the process
of thermal defluorinization of phosphate rock alone or in a mixture
with soda and furnace grade of phosphoric acid in the range 1200 to
1350°C. This process is used for feed phosphate production using
apatite as a raw material [1,24].
II
Emissions and Fluorine Gas Recovery
A
Fluorine Balance in SSP and TSP Manufacture
The amount and form of fluorine-bearing gases evolved during the
reaction between phosphate rock and sulfuric acid (SSP) or
phosphoric acid depends on the activity of the phosphate rock, the size
distribution of the rock particle, content of silicates and active silica,
the F/Si ratio in the phosphate rock composition of acids used, thermal
conditions, and the retention time in the operational units. According
to the data reported [1,3,4,6,10], in the acidulation step, about 25 to
50% of primary fluorine contained in the phosphate rock is evolved in
the mixersettling den, in the handling system including belt conveyors
and elevator, and during curing. About 50 to 70% of fluorine is
contained in the final product (0.7 to 1.8% F) in both solid (two-
thirds) and liquid (one-third) form [7].
Industrial tests carried out under similar conditions determined that
during the acidulation of rock from Florida, about 21% of fluorine
was evolved to the gaseous phase; Moroccan rock, about 25%; rock
from Tennessee, 35.5%; apatite from Kola, 37%; and rock from
Karatau, 40% [3]. During granulation and drying of run-of-pile SSP,
about 10 to 17% of fluorine contained in cured SSP is released in a
gaseous form containing 0.3 to 0.5 g F/m3 and 8 to 10 g dust/m3.
During these processes additional emission of fluoride occurs as a
result of the reaction

Under these temperature conditions HF is evolved simultaneously and

an average SiF4/HF molar ratio of about 1:2 is observed [1,3,8,10,25].
During SSP or TSP curing, conversion is continuous and about 1 to
3% of fluorine contained in SSP is evolved in the form of vent gases
[6,8].
The main fluorine-bearing gases collected from the mixer and den
contain about 15 to 70 g F/m3; those collected from drying and
cooling are 10 to 30 times lower but contain particulate matter (from
drying, 8 to 10 g/m3; from cooling, 30 to 70 g/m3) [8,10,26]. The
distribution of fluorine during TSP and enriched superphosphate
manufacture depends on the technological system and the
concentration of wet-process phosphoric acid used.
In the den method only 10% of the fluorine contained in the raw
material (rock and acid) is evolved in the form of SiF4. This gaseous
stream contains about 0.4 to 0.8 g dust/m3. In the continuous method
using slurry technology in the gaseous phase, about 50 to 60% of the
fluorine is evolved as a mixture of HF and SiF4 with a molar ratio of
2:10. The gases exhausted during pulverizing and drying contain 2.8
to 5.0 g F/m3 and 10 to 20 g dust/m3, and those from cooling contain
0.05 to 0.10 g F/m3 and 30 to 70 g dust/m3 [8,10,25]. In Table 6 of
Chapter 23 the content and amount of gases evolved during
processing operations are presented.
 Page 276
In Russian plants, reduction of fluorine emission during handling and
storage of SSP and TSP is achieved by the addition of sodium sulfate
to cured SSP or to the TSP granulation operation (20 kg/ton of
product). This procedure make possible a five-to sixfold decrease in
the emission of fluorine [10]. TSP conditioning and partial
neutralization of free phosphoric acid by ammonia [27] decrease the
amount of fluoride evolved and lead to better handling properties for
TSP in a product containing higher concentrations of magnesium.
In the Simplot process [20] of GTSP production, one source of
fluorine emission is eliminated, because lime is used instead of
phosphate rock. The average emission is estimated to be 0.08 kg F per
ton of P2O5 fed and the only gas cleaning necessary is a scrubber on
the dryer and various handling devices (screens, elevator, baghouses).
This process appears to reduce the absorption system. Curing is not
necessary, of course. However, the total fluoride emission is at the
same level, if the amount of fluorine evolved during the additional
stage of the phosphoric acid process is added to the Simplot process
emission.
The average distribution of fluorine in SSP and TSP production
according to data from Russian plants is presented in Fig. 4 [3,4,8,10].
In most plants in Kazakhstan, Ukraine, and Russia, special raw
materials systems are used. As a rule, lower-grade phosphate rock
such as Kara-tau, Kingisep, Toolsk, Egorevsk, Polpinsk, and
Verhokamsk is acidulated by wet-process phosphoric acid obtained
from high-grade phosphate rocks such as Kola apatite [2830]. In this
system the balance of fluorides is modified.
B
Absorption of Effluent Containing Fluorine
The fluorine released in process operations effluent and vent gases is
mainly in the form of SiF4, partially in HF form, when low-silica-
containing phosphate rock is processed and in the form of solid
particulate matter. These gases are usually scrubbed with water, a
dilute concentration of H2SiF6, or an alkaline solution containing soda
or lime milk or sodium sulfate [25]. Processes using an NH4F solution
[31] and an ammonia solution [32] as absorption media are also used.
The gaseous effluents contain more than one contaminant, and the
absorption medium is also a crude solution, particularly when pond
water or recycled liquids are used [1,10,20,31,33,34].
Fluorine recovery by means of scrubbing leads to the following
problems:
1. Pollutants can condense or react with the scrubbing media, with the
precipitation or formation of new compounds in liquid form.
2. Absorption of SiF4 is related to silica precipitation. This precipitate,
together with the other solids precipitated, has a tendency to plug
pipes and vessels.
3. The contamination of pond water or accumulation of recycled
liquid will limit the absorption ability of the scrubbing media. This
effect is caused by the vapor pressure of contaminants and the
temperature of the liquids used.
4. A higher scrubbing efficiency is achieved using alkaline media, but
many additional problems connected with the precipitation of
crystalline CaF2 stalactites and plugging arise.
5. The particulate matter contained in gases forms a matrix
encapsulating solids in a hard, massive layer of deposits, particularly
in narrow openings and within the system where liquids change
direction.
 Page 277

Figure 4
Fluorine balance in the manufacture of SSP and TSP.
 Page 278
In all cases, scrubbing media that utilize pond water are less efficient
than those that use well water [33].
Fluorine absorption efficiency depends on the absorption ability of the
liquid medium determined by the equilibrium in the system
H2SiF6SiF4HFH2OH3PO4. Equilibrium data convenient for design
purposes are available [25,3437], although, in fact, the exact
mechanism of SiF4 and HF absorption is more complicated. The basic
reactions occur according to reactions (6) and (7), but the process
becomes clearer if the following processes are considered [10]:

Contamination of the scrubbing liquids by calcium ion or lime milk

changes the form of existing fluorosilicic complexes toward species
whose Si/F ratio is lower, when the pH of the solution is increased
according to the transition [25]

The following process occurs when lime milk is used as an absorbing

medium:

The most important factor for efficient fluorine absorption is the

proper temperature range, which can keep the silica precipitated in an
easy-to-wash form. Too high temperatures cause silica to crystallize
on the scrubber surface, making it difficult to remove [24].
Plugging is one of the most important factors in limiting proper
operation of an abatement system. Plugging can be minimized by
better control of the temperature conditions so as to obtain deposits in
the form of soft fiber and glutinous mud, easy to remove. A second
way to avoid maintenance problems is by correct selection of
scrubbing devices and proper construction of the scrubbing system.
The most important development in this area is the use of a cyclonic
separator in the flushing system, together with a venturi and combined
venturicyclone scrubber, and foam scrubbers. Efficient dust removal
using dry cyclones also improves operating conditions [26,38].
C
Equipment for Scrubbing Fluoride-Containing Gases in SSP and TSP
Manufacture
The efficiency of any fluorine abatement system depends on proper
selection of the scrubbing devices and the absorption ability of the
scrubbing medium. Scrubber design and method of determination of a
number of transfer units are presented in Section I in Chapter 21 and
in many comprehensive papers (e.g., Ref. 39). The principal
technological problems associated with a gaseous stream evolved in
SSP and TSP production that has a high particulate-matter content
include large differences in the composition of gaseous pollutants
from various operations and the variable temperatures of the gaseous
stream. The basic considerations in pollution abatement system design
are whether they can meet the technological and environmental
requirements, whether they are convenient for operation and cleaning,
and of course, the investment and operating costs involved.
 Page 279
During the past decade many absorption devices in old SSP plants
have been modified or replaced by the newer scrubbers, which reduce
the fluorine content in gaseous streams by over 96%. The primary
factors involved in the construction of SSP and TSP fluorine
abatement systems are:
1. The efficiency of scrubbing in terms of prevalent environmental
standards
2. The properties and fluorine content of the gaseous stream inlet
3. The properties involved in multiple contamination
4. The solids formation resulting from absorption
5. The low-pressure-drop requirement (energy consumption)
6. The corrosive properties of the liquids involved
7. The quality of final by-product fluorosilicic acid solution or the
method of disposal or utilization of the other exhausted wastes (by the
lime method)
The following devices are used in the scrubbing process in SSP and
TSP manufacture [1,6,8,10,34,38,4045].

Mechanical absorber
Device used in SSP production which consists of an absorption
chamber with a turning roll with fixed plates that while rotating are
partially immersed in absorbing liquids. From this chamber the gases
(30,000 to 50,000 m3/h) are introduced to the spray chamber that
follows.

Mechanical atomizer
An immersed rotating propeller shaft in this device forms mist and
small droplets of absorption liquids. The absorbers are convenient for
gases contaminated by dust, and an efficient deentrainer is necessary
to remove droplets from outlet gases.

Surface absorber
A device in which the fluorides are absorbed on liquid layers flowing
downstream on plates and pipe (cooled) surfaces, with countercurrent
contact of gases an indirect cooling source.
 Page 280

Bubbling scrubber
A device in which a contact surface is formed by a gas stream by
using a constant layer of liquids with floating nozzles, bubbling with
calpack, and introduction of bubbles by contact with a rotating target.

Bubbling Doyle scrubber

A very convenient device to use in GSTP production for absorbing
gases with a high particulate-matter content. Installed in a three-stage
system in series, absorbs 99.5 to 99.7% of the fluorides.

Packed tower
A device with a stack mounted on the top and a deentrainment section.
Polyethylene packing is preferred to polypropylene because its surface
is smoother and thus more slippery. It can be used as the final stage in
a three-stage system including two spray towers. Installing this
scrubber as a first stage would cause a problem of silica deposition.

Spray chamber
This is a large void tower with multiple high-pressure sprays flowing
countercurrent to the gas flow. The device is used for the first-stage
treatment of gases containing a high percentage of pollutants. These
chambers are mounted in series because of low unit efficiency (80 to
90%).
 Page 281

Water impact scrubber

A device equipped with two large full-cone spray nozzles. This
system enables the formation of a very fine water mist in the air duct
with a large surface area for efficient fluoride and dust absorption.
The stream of very fine droplets must be introduced to an effective
entrainer separator.

Venturi scrubber
A device for use with high-speed gases (20 to 40 m/s), constructed
with a reduced-taper pipe throat and a diffuser connected with the
chamber. A scrubbing medium is introduced by a vaporizer or reducer
in the form of dispersed droplets that absorb fluorides. These devices
tolerate a high solids content (granulation units in SSP and TSP) and
operate in series.
 Venturicyclonic scrubber
A device consisting of a venturi followed by a cyclonic spray
scrubber. The absorbing medium is atomized in the throat area. Part of
the absorbing medium is collected in the elbow. Gas flows from the
throat and is introduced tangentially to the wet cyclone. Dust buildup
is avoided by use of a flooded elbow.
 Page 282

Venturicyclone coaxial
This device is a modification of a venturi scrubber. The scrubber is
built as an integrated hybrid mounted concentrically with a cyclonic
section. During operation the liquid does not change direction, and the
gas-droplet system in the bottom section achieves separation of the
entrained liquids. The scrubber is used in GTSP abatement systems.

Aeromixer scrubber
This is another device similar to a venturi scrubber. These absorbers
are installed in series. The first stage is a dry cyclone for dust
removing. The system must be equipped with reentrainment following
the removal of gases from the battery. Disadvantages include very low
efficiency (about 80%) and problems with turbine maintenance,
although 94 to 97% efficiency is achieved with a three-stage battery.

Spray cross-flow packed-bed scrubber

This device is an integrally built multistage system for absorbing a
low-particulate gas stream which flows horizontally through both a
rain of spray and washing packing. In the first section the particulate
matter is removed. This device can purify gas to the desired level of
environmental standards.
 Page 283

Cross-flow scrubber
This device requires a smaller amount of scrubbing medium but is
subject to greater plugging than is true of vertical scrubbers, although
without cleaning problems.

Foam scrubber
A device in which a gasliquid contact surface is formed while the gas
passing upstream through the liquid layer covers the grating, causing a
dynamic layer of foam to be formed as an effective medium for
efficient mass transfer. The special cone-shaped stabilizer stabilizes
the foam. Particulate matter is removed in the first stage. Foam
scrubbers with various modifications are used more and more
frequently in GTSP plants.
In addition to the scrubbers described for use in GSSP and GTSP
manufacture, use of cyclonic spray scrubbers; impingement scrubbers,
which are subject to plugging in the orifice and target areas; and
induced spray wet dynamic void-type spray scrubbers has been
reported. The technical characteristics of process equipment,
including pressure drop, liquid requirement, number of transfer units,
and solids recovery efficiency, are presented in Chapter 17 and in the
literature [6,8,10,43]. In SSP and TSP production, void spray towers,
venturi scrubbers with wetted elbows, and foam scrubbers have been
used predominantly. Scrubbers that have a low pressure drop, are easy
to clean of deposited silica, and are of simple construction are
optimal. As a rule, packed towers are used as the last section in the
abatement system. As a last section, efficient deentrainment
(demister) devices are used, such as cyclonic, packed, and
impingement grids, targets, fiber beds, and monofilament packing
(e.g., Kon-tane, B-Gon materials) [44,45].
D
Pollution Abatement Systems Used in SSP and TSP Production
Removing gaseous pollutants from vent effluent in SSP, GSSP, TSP,
and GTSP production is more complicated than simple absorption.
Each fertilizer process has its own requirements and the most effective
abatement system varies for each special set of conditions,
characterized by the form and properties of the effluent, local
environmental standards, and limits on the scrubbing media. During
the past decade many new systems have been developed not only to
remove pollutants according to rising environmental requirements but
also to produce high-quality fluorine-bearing concentrated solutions
as raw material for salable fluorine compound production.
 Page 284
The effectiveness of a fluorine recovery system is determined by
proper selection of the scrubbers, correct abatement line composition,
and proper inlet fluorine concentration. The design of the pollution
abatement system is controlled by the following criteria: reduction of
contaminants to below the level required, minimization of the
plugging problem, efficient separation of liquid droplets and mist,
effective precipitated silica removal, nominal maintenance and
investment costs of the exhausted gases and for wastewater disposal,
and a minimum surface for objectional buildup. Achieving these goals
depends on proper determination of the number of scrubbing stages
and amount of liquid separation, minimization of pressure drop, the
main aspects of the overall power requirements, and determination of
the time and efficiency of the gas/dispersed water contact. To provide
effective abatement of pollutants requires simplicity of disassembly of
the packing case system for periodic cleaning of deposited silica by
flushing with and tumbling in water. Based on considerable
experience in the operation of various systems for simultaneous
fluorine and dust removal in scrubbing media where many chemical
processes take place, scrubbers are preferred that have a low pressure
drop and whose induction fans consume very little energy, with
simplicity of design geometry for the gas and liquid ducts. Inasmuch
as fluorine-removal efficiencies of about 99% are required, about 4.6
to 6.0 transfer units must be provided by the system to provide
effective distribution of water droplets and water spirals at the wall,
thus preventing silica buildup. A greater number of transfer units
(NTU) is required: as many as 8 NTU when pond water is used as the
scrubbing medium. According to Teller [26], 95% fluorine recovery
can be provided by an integral system composed of cyclonics,
venturis, or spray scrubbers, with minor variations in system
performance achievable with modifications in geometry. Higher
efficiency is possible using a combined system comprised of a
primary scrubber, in which the bulk of the fluorine is absorbed, and a
tail-gas scrubber, in which the residual fluorine is removed. In this
part of the system concomitant particulate matter and SiO2
precipitates are absent, thus improving conditions for penetration of
the solute gas into the liquid.
In SSP, TSP, GSSP, and GTSP plants, an abatement system must be
installed to permit the removal of contaminants from various process
streams (mixing vessel, settling den, elevators, vent gases from the
granulator, dryer, cooler, transportation system, curing, and storage)
containing different levels of dust and fluorine driven below
atmospheric pressure by means of exhaust fans. The abatement system
can serve more than one process because the granulation section may
be used for TSP, DAP, NPK, and mixed fertilizer manufacture.
The traditional scrubbing system is formed essentially of scrubbers in
series, in which a circulating solution of fluorosilicic acid is sprayed,
with a tail-gas scrubber in the form of a packed tower followed by an
entrainment separator, fan, and stack discharge of gas to the
atmosphere.
An absorption section is presented schematically in Fig. 5. This
system was used widely in Europe in the early 1960s. Gases from the
den stage are introduced to the cyclone by fan for dust removal and to
an absorber equipped with mechanical atomizers or a mechanical
absorber with rolling, partially immersed plates. This absorber is used
to remove the bulk of the fluorine in fluorosilicic acid solution in a
multisectional device. The gas goes from the scrubber to the washing
spray tower, into the deentrainment separator, and through the exhaust
stack to the atmosphere.
The simplified version of this system is a single-stage concrete or
rubber-lined multisectional spray chamber in which gas is absorbed in
a recirculating solution of
 Page 285

Figure 5
Fluorine absorption system in SSP plant.
fluorosilicic acid. In older factories, 70 to 80% fluorine recovery
efficiency was achieved. The fluorosilicic acid in the suspended silica
is neutralized by lime and disposed of. When fluorosilicic acid is
produced as a salable by-product or processed to fluorine compounds,
the system must be provided with a silica-removing line. Construction
of such a system is shown in Fig. 6. In this system spray towers are
used in series with a common circulatory system for liquids. In one
installation known to the author (WZPN, Wroclaw

*), six spray chambers in line are used. The silica is thickened in the
settling tank and separated in the centrifuge, or silica sludge is
recirculated to the mixerden stage.

In many SSP plants horizontal scrubbers with multisections, each

consisting of a chamber with its own circulation circuit [34], are used.
Inside each chamber a special wire mesh pad is mounted and
nonclogging nozzles are used. The setup of this system is shown in
Fig. 7.
For improved absorption, various types of wet cyclones in series
followed by packed towers are used. For conservation of space and
economy, and to avoid plugging,

Figure 6
Fluorine absorption system with silica-removing line.
 Page 286

Figure 7
Multisectional absorption system used in SSP plant.
venturicyclonic, venturicoaxial, and venturiaeromixing scrubbers in
series are used. These systems operate with an efficiency of 95 to
98%. Most plants have three stages in series. During the last decade
systems have been modified using foam scrubbers and efficient spray
cross-flow packed-bed scrubbers [4,10,26,45] containing an inside
deentrainment device and adapted to solids removal in the first
section. Retrofitting in older plants involves using the newer types of
scrubbers.
In nongranular SSP or TSP production or run-of-pile granular
superphosphate production, the product must be cured. From the time
SSP or TSP has been cut from the den and transported to the curing
pile, and during curing, fluorides are evolved. In the newer plants the
storage buildings are sealed and fluoride vapors are induced to exit
through the roof collector by a fan rated at 200,000 to 300,000 m3/h.
In a plant producing 1000 tons of TSP per day, about 2500 kg F/day is
evolved in this stage [41]. According to Teller [45], from a pile of
20,000 to 60,000 tons of TSP, about 0.1 to 0.3 kg of fluorine per day
per ton of P2O5 is evolved, and fluorine concentrations of 40 to 80
mg/kg on the surface and 6000 to 9000 mg/kg in the center of the pile
have been detected. During natural off-gas ventilation of the building,
concentrations of 0.1 to 0.5 g F/m3 have been found. In a modern
curing facility the vent irrigation gases are scrubbed in a separate
scrubbing system provided with a Doyle scrubber, void towers, or a
compact spray cross-flow packed-bed scrubber with an efficiency of
97 to 98%. In most plants the gases from the primary scrubber system
are sent to a tail-gas scrubber connected with other stages of the
abatement system.
The final cured product (SSP or TSP) is dug from the pile, crushed,
screened, and shipped in bulk or granulated form. During granulation
and conditioning by coating the surface with ground phosphate rocks,
dolomite, chalk, or limestone, the gases evolved contain 0.3 to 0.5 g
F/m3 and about 8 to 10 g/m3 of dust.
After dry cyclonic dust removal the gases are scrubbed. The gases
from the dryer and cooler are rid of dust by a dry multicyclone battery
heated indirectly by steam to avoid water vapor condensation. From
this stage the gases are combined with vent gases from screens and
elevators of the recycled system and are introduced to a washing
tower and then to an absorption system in which a 1% suspension of
lime is often used in the last stage for more efficient fluorine removal.
During GTSP production using an ex-den granulation process,
fluoride evolution begins at the mixer cone and continues along the
settling belts and along elevators and transfer points. All these
emission sources are hooded and the gases collected are scrubbed in a
system such as that shown in Fig. 8.
 Page 287

Figure 8
Three-stage venturicyclonic system used in TSP plants.
In continuous slurry methods of GTSP production, granulation and
drying are important sources of fluorine emission. The absorption
system inlet is equipped with various types of dry dust-removal
devices, such as multicyclones, bag (sack) filter batteries,
electrofilters, and mechanical settling chambers. In these devices dust
is partially removed and fine particles are removed in the first stage of
the scrubbing system. The layer of dust deposited is removed by an
automatic hammer or chain device. In GTSP manufacture,
independent absorption sections are used predominantly, collecting
gases in streams from reaction, from granulation (ex-den), from the
dryer and cooler, and from storage. In many plants exhaust gases from
the scrubber section (void towers, Doyle scrubber nozzle scrubbers,
venturicyclonic, venturihybrid) are introduced together to the tail-gas
scrubber, comprised of packed tower and deentrainment separators.
When a foam scrubber or spray cross-flow packed-bed scrubber is
used, the gases are exhausted by fan through the stack to the
atmosphere. According to Osag [42], such treatment results in a
significant improvement in absorption; only 0.01 to 0.15 kg F per ton
of P2O5 is uncontrolled.
The choice of the system used to deal with gaseous pollutants of
various concentrations of fluorine and dust also has an effect on the
power requirements and investment costs. It is possible to blend all
gaseous streams and to scrub effluent from individual operations
separately. In practice, in the first integral scrubbing stage absorption
of the streams with the highest fluorine concentration occurs and the
other gases of low fluorine concentration are added to the off-gases
from this scrubber section. Using typical blending procedures, 1.5 to
2.0 times more power is consumed and the original investment costs
are about 1.3 to 1.5 times higher.
Of course, the power requirements depend on the type of scrubber
used. According to many reports and to practical experience, to
achieve the same number of transfer units as those with a packed
tower, venturi scrubbers require about 7 to 10 times as much power
and spray scrubbers, 1.2 to 1.5 times. This problem must be solved in
the design process.
Effective emission control depends not only on correct design of the
absorption section, but also on daily operation and maintenance and
proper cleaning of the ductwork piping pumps and fans. Periodically,
according to the maintenance schedule (every 7 to
 Page 288
10 days), removal of any buildup is desirable. When the packing is
flushed with a strong stream of water once a week during cleaning,
the packing must be replaced within 7 to 9 months. An important
factor in efficient operation is the provision of open geometry at the
first gasliquid contact area, for minimization of plugging.
The properties of the liquid and gaseous phases require the use of
special materials to withstand the highly corrosive conditions. The
duct systems are made of glass-reinforced polyester with a Dynel
coating of the glass fiber interior. Scrubber interiors are of rubber-
lined steel or 316 stainless steel (40 to 50% higher unit cost). The
stacks and fan housings are of rubber-lined steel. The nonclogging
nozzles and scrubber interior spray heads are of fiber cast and glass-
reinforced epoxy pipe.
Retrofitting modifications for an older system, especially one in a
poor state of repair, tend toward more efficient combined scrubber use
and the use of efficient absorbing media in the last tail-gas scrubbing
operation. In Germany, use of an absorption section containing a
scrubber tower with grating plates or bubble-cup plates has enabled
concentrations of 0.03 g F/m3 and solutions up to 30% H2SiF6 to be
produced. In Japan, concentrations of 0.013 g F/m3 in exhausted gases
from SSP plants have been achieved when gaseous pollutants are
scrubbed by two units arranged in series in which 2% NaOH solution
is used as an absorption medium and suspensions of Na2SiF6 are
produced as a by-product. In several plants in the United States a
suspension of NaF is used and by-product Na2SiF6 is removed. A
high level of efficiency is achieved because of the low partial pressure
of SiF4 over this medium.
III
Particulate Matter Recovery
A
Reasons for and Consequences of Dust Emission in SSP and TSP
Production
Control, recovery, and return of particulate matter from SSP and
GTSP processes are as or more important than control and recovery of
fluorides. The dust is emitted from practically each stage of
production, and dust losses can create economic and environmental
problems for the surrounding neighborhood. Dust emission happens
during loading of the phosphate rocks (ships, railway boxcar, trucks),
in interfactory transportation, and during handling and grinding. The
particulates arise from various stages, mainly drying, cooling, and
granulation, as well as from most size-reduction and transportation
devices, such as conveyors, screens, and particle-size reducters, and
from reloading points. The dust can also emanate from inefficient dust
collectors, from incidental spillages, and during bagging and bulk
blending with other fertilizer components.
The dust from SSP and GTSP plants comprises fine phosphate rock
and superphosphate particulates containing from 1.5 to 4% fluorine,
mainly in an insoluble in water (apatite) form and a form only
partially soluble in water (SSP and TSP). This emission does not
create a significant problem for human beings, plants, and animals in
areas surrounding the manufacturing facilities. The particulate matter
originating from SSP and TSP production is more troublesome for
plant maintenance operation and working conditions. Dust elimination
or minimization can improve operation of the drive mechanism, lead
to a better state of repair for electric motors, switches, and
connections, and lower energy consumption because dust necessitates
that additional energy be used in rotation and transportation devices.
Improvements in fertilizer unit operation, proper design of systems
and equipment,
 Page 289
and efficient control systems are installed to provide improved dust
control, recovery, and dust return. In most SSP and TSP operating
plants the contaminants are removed from all exhausted gases before
discharge to the atmosphere, and significant economic savings and
lowered levels of atmospheric pollution have been achieved. Modern
improvements eliminate losses by using a closed-circuit effluent
control system.
Dust emission has now been reduced successfully using three
methods:
1. Installation of an effective dust control and recovery system
2. Installation of special equipment, such as dust seals and a closed
transportation system
3. Changing the material properties, changing the grain size
distribution, and the addition of agents that lower dust formation
(depressants)
B
Reduction of Dust Emission in SSP and TSP Production
A discussion of the main sources of dust emission and technical
suggestions for dust abatement follow.
1
Preparation of Phosphate Raw Material (Drying, Grinding)
In SSP and TSP production only dry size reduction of the phosphate
rocks takes place. The phosphate rocks must be reduced to particles
fine enough for 90% to pass through a 100-mesh screen and 80% to
pass through an 80-mesh screen to provide a higher specific surface
than in wet-process phosphoric acid production, which creates more
dust problems.
The moisture tolerance level of phosphate rocks for various types of
mills (ball mills, <2 to 3% H2O; pendulum, ring, and rod mills, <4 to
5% H2O) requires rock drying prior to grinding to avoid a pasty form
of ground materials and problems during pneumatic rock
transportation. Phosphate rocks are dried in rotary or fluid-bed dryers
and discharged from dryers onto a conveyor belt at a temperature of
100 to 140°C. The exhausted gases contain up to 35 g dust/m3. In this
operation about 7.5 to 15 kg of fine phosphate is emitted from the
dryer of a 100-ton/h-capacity drying unit, even when a good dust
control system is used [6].
Similar amounts of dust are emitted during the calcination of
phosphate rocks [26] and during thermal defluorination of apatites or
during the manufacture of feed phosphates containing sodium and
calcium by thermal processing of phosphate rocks with a mixture of
soda and furnace-grade phosphoric acid. In a gaseous stream moving
at 195,000 m3/h from the kiln, 2 to 5 g HF/m3, 0.03 g SiF4/m3, and
0.2 to 1.5 g dust/m3 are present during the production of 10 tons of
feed phosphate per hour. The dust is collected using a settling
chamber, baghouses with cloth filters, and cyclones. In feed phosphate
production, the settling chambers and cyclones are followed by a wet-
scrubbing process using pond water and lime milk suspension as an
absorption medium (Feed Phosphate Factory Works, Bonarka,
personal communication, 1992). Elimination of phosphate rock drying
is possible if wet grinding is used, and simultaneously, many dust
problems are avoided, although such a system may operate only for
wet-process technology (WPA). In TSP-WPA plants, only dry
grinding is used, because two different systems are not practical in a
single factory [46].
The primary purpose of particle-size reduction is to form a larger
surface area of phosphate rock for easier, more effective
decomposition and a shortened curing period. Grinding and pneumatic
transportation can cause particulate emission. To illustrate the
construction of this stage, a sample SSP grinding system is shown in
Fig. 9. This system is part of a 270,000-ton SSP/year plant, including
a type 41.51 cylindrical ball mill of
 Page 290
capacity 30 tons of phosphate rock per hour and a pneumatic system
for ground phosphate (80% < 200 mesh, 90% < 100 mesh)
transportation and dust collection. The dust control system consists of
a battery of cyclones, two-stage pulsate bag filter sections, and a
central dust collection duct that enables transport of the ground
phosphate and simultaneously removes dust from the exhausted air.
This system is used for grinding the dry phosphate rock (moisture <
3%) of 90% < 3 mesh with 50,000 m3 air/h flow through the grinding
section. The air leaving the cyclone battery, which contains 100 to 300
g dust/m3, is filtered by a bag filter compact section to a level of 60 to
80 mg dust/m3
. The exhausted air from this stage is about 5 to 10% of the total flow
(SSP Works, WZPN, Wroclaw

*, personal communication, 1992).

In SSP and TSP plants, various mills and particulate matter circulation
systems are used. For optimization of the power requirements, various
types of ground rock classification methods are used: cyclonic
centrifugal, cone connected with settling chambers, and screens.
Application of these devices enable operation in a closed-loop circuit.
When a pendulum or roller ring mill is used, internal classification of
the ground rock particles by centrifuging makes it possible to recycle
the coarse particles back to the mill.
The dust emanating from the grinding stage is from local points in
which material changes the devices (inlets, discharge, shaft conveyor,
elevator assembly) when the ducts are not quite airtight. The ductwork
ought be as short and direct as possible to ensure adequate air velocity
(20 to 26 m/s). Aluminum ducts are easier to keep clean than ducts
made from other materials [47]. Improvement in the grinding system
depends on using closed-loop circuits, providing the shaft assembly
with fresh-swept interfaces, and

Figure 9
Grinding system with dust collection and pneumatic transportation.
 Page 291
lowering emission at material unloading points. The most important
influence on dust formation is retention time in the mill. During too
long a grinding process, a large amount of dust is formed, but too-
short retention times lead to problems of oversized particles.
2
Acidulation Step
The vent gases from the mixer and den (digestion) contain less dust
than effluents from other points and particulate matter estimated to be
in the range 1 to 5 µm [43]. This stream is scrubbed in the void tower.
In some TSP plants the reaction gases are scrubbed in the first stage
by wet-process phosphoric acid returned to the acidulation step (TSP
Works, GZPN Gdansk

*, personal communication, 1992). A more airtight, closed system of

efficient hoods in the mixerden is used.

3
Granulation
Particulate matter in this stage is generated as a result of attrition and
recycling of the fine particles. Dust emission from the granulation of
ex-den or run-of-pile SSP or TSP depends primarily on the type of
granulator used (drum, blunger, spherodizer, pan), construction of the
inlet and discharge assembly, and the impurity content of the
granulated recycled material. The gaseous effluents from this stage
contain about 5 to 50 g dust/m3. In many installations gases from this
operation are blended with gases from the dryer and cooler.
The modern granulator and reactorgranulator are equipped with
labyrinth-type seals enclosing the feed end of the granulator or in the
exhaust hood covering the discharge element of the granulator
[49,50]. Decreasing the dust emission is possible in this stage and in
the following operation if harder granules are formed and if dust
depressants are introduced into the granulated mass.
4
Drying and Cooling
The largest amounts of dust are emitted in the drying and cooling
operations. The dust concentration from these stages depends mainly
on the properties of the solids and less on the equipment used [51],
although the type of discharge chute is important. Gases passing
through the dryer and cooler at a velocity of 1 to 2 m/s cause a high
dust content in the gaseous pollutants, as much as 500 g dust/m3
[6,26,46]. The particulate matter is removed from these streams in
cyclones or multicyclones. To prevent condensation of the water
vapor, these units are insulated by polyurethane foam or other material
or even heated externally by steam to maintain the temperature 10 to
20°C above the dew point. To increase the efficiency of cyclones, an
additional airstream is adjusted by means of a damper [49,50] to the
gas dust before the cyclones. The cyclone stage is followed by a wet-
scrubbing system, and in many plants the liquid suspension obtained
is recycled to the granulation stage.
5
Recycling System, Including Transportation, Screening, and Size
Reduction
Dust formation in a granulation recycle system is similar to that of
other fertilizer processes and Achorn's suggestions [49,50] on dust
reduction in this operation are also valid for GSSP and TSP plants. All
elements of this system, especially crushers (hammer, cage, chain, rod
mills), screens, and transfer and terminal points of the belt conveyors,
are sources of the dust. In a crusher operating with rotating elements,
a fan-type effect is generated [4951]. To lower dust emission a bypass
system is used to recycle the induced air in the mill.
 Page 292
Dust losses from the screen are reduced by using lids on screens and
efficient hoods and flexible connections with ducts. Vibration
generates many problems. In many plants the size reduction and
classifying operations are installed in separate rooms. Solid floors are
used instead of gratings to control particulate spillages. Significant
plant dust reduction is possible when modern elevators and conveyors
(screws, cleated grabs, buckets, belts with hoods) are used. The
minimal straight path and proper duct angle and bin location have an
influence on the total dust problem.
6
Curing and Storage
From the curing pile for SSP or TSP, dust is formed when a power
shovel is used to lift the material and during transportation to the
granulation stage or bagging machine when it is distributed in
nongranular form. To avoid caking, the product is disintegrated in a
cage mill to pass a 6-mesh screen and anticaking agents such as
ground limestone and dolomite are added, or special conditioning
agents are introduced in the granulated mass or run-of-pile material:
for example, 0.5 wt % 10-weight motor oil or liquid fertilizer (10340)
[49,50]. Addition of these depressants decreases considerably the dust
formation tendency, preventing material from sticking, decreasing
fluorine emission, and improving the size and shape of granules.
Some producers add a small amount of ammonia or ammonium salts
or even sodium sulfate to the product, to decrease fluoride emission
[10].
During curing or storage the dust is scrubbed together with fluorine in
a separate system. Void towers or venturi scrubbers are often used or
the dust is collected in baghouses.
7
Handling and Transportation of Granular and Nongranular SSP and
TSP
Transportation over relatively long and sophisticated routes provides
many local sources of particulate emissons, primarily from unloading
points, screen chutes, open belt conveyors, and elevator boots. The
fine dust particles have a tendency to adhere to belts, which causes
slowing at certain transportation points. The hygroscopic properties of
the dust create many problems during cleaning and cause spillage.
Collection pans are installed beneath these points.
Effective reduction of dust emission is achieved by enclosing the belt
conveyor or replacing the belt with a cleated drag transporter or dust-
covered screw conveyor. The dust-high cyclone closure by molded
rubber connection with conveyors, installing the flat covers on an
elevator or conveyor discharge, modification of the centrifugal
discharge elevator by controlling the speed and discharge of material,
and dosing the boot and elevator by flexible slot covers or vane-type
seals for bucket elevators are useful for reduction of dust
concentration to below 2 to 3 mg dust/m3. Often, dust concentrations
above 500 mg dust/m3 are evident. There is often considerable dust
concentration (700 to 1000 mg/m3) where chutes from screens and
cyclones are connected with the recycle system. Dust reduction is
necessary here.
8
Bagging, Blending, and Application
Superphosphates are stored in bulk on piles or in bins. Dust is emitted
when the product is removed from the pile with a front-end loader,
during unloading of a bucket elevator, during introduction into a
bagging machine hopper, from discharge chutes during unloading
from bottoms of elevators, and during unloading to trucks or when
applied by farmers.
 Page 293
New types of conveyors are presently used for dust and spillage
reduction, such as closed conveyors, scaled elevators, and flexible
sleeves as discharge chute. Special rubber and canvas flats and
curtains around unloaded transport trucks also reduce dust emission.
Cleated drag conveyors installed at angles eliminate the use of bucket
elevators.
When blending granular superphosphates with other components in
rotary mixers, the dust is considerably reduced by spraying with mixer
oil, water, or liquid fertilizer (10340) as a dust depressant. Dust in the
air surrounding the bag machine may be reduced many times if this
procedure is used. Typically, about 15 to 30 mg dust/m3 is detected
near a blending mixer, about 300 mg of dust near a bagging machine,
and 25 to 70 mg/m3 near a discharge chute during operation without a
dust depressant.
Better sizing of granules, lower fines content, use of dust depressants,
and sealing the transportation system make possible considerable
limitation of dust emission, although the use of a dust control device
is necessary. In blending plants of capacity 50 tons of fertilizer per
hour, a baghouse collector of size 300 m3 air/min is needed.
Properties of granular SSP and TSP such as size of granules, size
distribution of granules (dust content), and presence of dust
depressants are important for the dust problem during application by
farmers. Undesirable dusting effects are avoided or minimized by
using an applicator with a slotted boon with auger and by using a
device by which granular fertilizer is applied simultaneously with
liquid fertilizer.
C
Systems and Equipment for Dust Reduction in SSP and GTSP Plants
Systems for collection of suspended fine solids operate using:
1. Dry systems, which remove the particulate matter in dry form using
dry electroseparators, cyclones or multicyclones, settling chambers, or
baghouses
2. Wet methods, which remove dust in suspension simultaneously
with fluorides using wet electroseparators, hydrocyclones, wet
scrubbers (void towers, venturi-impingement multisectional systems,
spray cross-flow packed-bed scrubber)
All these methods operate as recovery and return systems controlling
all dust-emanating sources. Dry systems remove only particulate
matter, although the new Teller method proposes the use of baghouses
with continuous injection of a dry additive for removing fluorides
using adsorption [52].
A dry system operating in an electrostatic field is used only for
collecting the dust from drying and calcination of phosphate rocks.
The corrosive atmosphere caused by the presence of fluoride limits its
application. Wet electrostatic precipitators effective for scrubbing fine
particles smaller than 2 µm are also used in phosphate rock dryers
[53].
In most of dust-emanating sources, high-efficiency cyclones or
multicyclones are installed. These devices remove most of the larger
particles of rock or final product. The principal operating problems
are moisture from the gaseous stream and hygroscopic properties of
the solids. There is external heating of the cyclone surface or
insulation is used to prevent plugging. The layer of plugged deposits
is removed by mechanical pulse equipment or in GTSP plants by
using hanging chains inside the cyclone which when continuously
induced are effective surface cleaning units (TSP Works, GZPN
Gdansk

*, personal communication, 1992).

Fabric filters are the most efficient particulate removal equipment.
These devices
 Page 294
collect dust with an efficiency as high as 99.5%. The construction of
these collectors depends on the methods of removing particulate
matter from bags and the filter fabric used. Mechanical shakers,
reverse flow devices, ring jets, and pulse jets are the main removal
systems. For choice of the proper fabric, the acid resistance, flex
abrasion, and melting temperature have to be considered. In TSP and
SSP baghouses, acrylic, Dynel, polypropylene, Filtron, Creslan,
Orlon, and wool cloth is used [54]. A battery of cloth bags is mounted
as a group on independent frames or groups of bags mounted in a
compartment. These devices are installed on top of holding bins with
dust pickup points and are used as the final cleaning stage before the
stack. The baghouses can collect dust without disturbing the water
balance but require a large amount of space in special rooms and need
to avoid the dew point. The baghouses are characterized by a low
pressure drop and low energy requirements.
Maintenance and operating problems also depend on the proper ratio
of air to cloth and continuous control of the pressure drop throughout
the system [55]. For dryer and cooler dust control, baghouses and
ductwork from the dryer-cooler must be installed with polyurethane
foam layers to avoid water vapor condensation.
The wet scrubbers are used primarily as a secondary cleaning
operation for gases emitted simultaneously from suspended fine
particles and mist and fumes. Wet-cleaning devices are effective in the
particle-size range from 1 to 2 µm. These particles are not removed in
cyclones. The suspension of dust collected is returned to the
granulation stage or acidulation. The cyclones are followed by a wet
scrubber resistant to plugging, such as a void tower, bubbling
scrubber, or venturi scrubber. The optimum spray particles for
particulate collection are in the range 200 to 800 µm, but for fluoride
absorption any size less than 50 µm is more effective. Gases removed
by the wet-scrubbing system contain less than 30 mg dust/m3.
Centrifugal settling of coarser particles is achieved in hydraulic
cyclones. The thickened slurry is returned to the process.
At present, many complex system for fluorine and dust recovery are
being used as SCPB scrubbers in which a cross-flow packed-bed
section operates as a final entrainment separator. The problems with
effective return of diluted suspension cause to operate in the closed-
circuit scrubbing system to avoid of holding the ponds.
Most technological operations are equipped with separate collection
systems. Also, for transfer or unloading points of solids, the screens
and size-reduction recycle operation are connected with an integral
dedusting system. For dust reduction in plants, a central cleaning
system is installed in which dust and spillage are sucked out through a
pipe network by means of a vacuum cleaner.
IV
Environmental Regulations for SSP and TSP Production
In many developed and developing countries, special regulations are
similar in form to U.S. Clean Air Act Amendments [56], establishing
emission limits and methods required for testing and reporting
fluorine and dust emission for SSP and TSP. These national
regulations form the overall tolerance limits, but local authorities
establish the specific limits for existing and new plants. This
procedure is similar to the U.S. New Source Performance Standard
(NSPS) [20]. Evaluation of the fluorine and dust regulations of the
emission limit follows the development in SSP and TSP technology of
abatement equipment and systems. For example, NSPS tolerance
limits established a value ''for
 Page 295
plants manufactured or replaced after May 1, 1964 of up to 0.182 kg F
per ton of P2O5 manufactured in any 24-hr period." Before this date,
this limit was higher (up to 0.272 kg F/ton of P2O5) for existing plants
[56]. Similar rules pertain in India, where for existing plants,
according to Indian Standard Institution regulations, the tolerance
limit is 0.5 kg F/ton of SSP and 0.3 kg F/ton of TSP and for a new
plant, 0.1 kg F/ton of SSP and 0.075 kg F/ton of TSP. The limit for
particulate matter emission through stacks is established as 500 mg
dust/m3 for all SSP plants and 4 kg dust/ton of TSP product for all
TSP plants [57]. Recent escalation of restrictions is a result of
technological developments and national acceptation of an
environmental protection policy. For example, in India, a local
pollution control board established the following limits for Godwari
Fertilizer and Chemicals: 250 mg of suspension particulate matter/m3
and 44 mg F/m3 in 1982, but the revised 1985 limits were more
restrictive: suspended particulates, 150 mg/m3, and fluorine, 10 mg
F/m3 [58].
To meet escalating standards most TSP and SSP plants must recover
fluorine as H2SiF6 or remove (by lime neutralization) 30 to 40% of
the total fluorine contained in rocks. According to various data, about
15 to 40 kg of fluorine per ton of P2O5 processed may be recovered as
useful fluorine-bearing liquids or suspensions. Modern abatement
systems should yield fluorine and dust recovery efficiencies higher
than 95%. Many abatement systems used in new or reconstructed
plants have reduced plant emission by 99.5%. Because of the
significant development of scrubbing operations in most SSP and TSP
plants and by sealing the emanating devices, emissions of gaseous
fluorine have been reduced to 0.1 kg of fluorides (as F) per ton of
P2O5 produced in the form of SSP or TSP and dust in the range 0.3 to
1.0 kg/ton of product.
Significant developments in traditional technologies such as SSP and
TSP are limited, and during the last two decades only the Simplot
GTSP process (TSP from lime and phosphoric acid) has been
identified as new. Also, the main principles of fluorine and dust
emission control using wet scrubbing systems are identified as BACT
(best available control technology), although many new modifications
and a closed-loop system were used. These facts meant that the actual
requirements for NSPS compliance in U.S. TSP plants are on the
constant level: 0.091 kg F/ton of equivalent P2O5 feed and for GTSP
storage facilities, 0.00022 kg dust per hour per ton of equivalent P2O5
stored. According to the National Ambient Air Standard, the dust
emission tolerance limit defined by the U.S. Environmental Protection
Agency (EPA) as the primary standard (protection of public health) is
0.075 mg/m3 for the annual arithmetic mean and 0.260 mg/m3 for the
maximum 24-h concentration [56].
All new and existing plants must meet overall national or individual
national requirements similar to or even more stringent than the
American NSPS [59]. The emission tolerance limit is established as
follows: in Germany, exhausted SSP gases must not exceed 5 mg
F/m3 and 50 to 150 mg dust/m3; in Belgium, 670 mg dust/m3; in the
Netherlands, 10 to 175 kg HF/day from one plant; in the United
Kingdom, 324 mg F/m3 and 460 mg dust/m3 from granulation plants;
in Norway, 0.3 to 1.6 kg HF/h from an individual plant; in Brazil, 0.1
kg F/ton P2O5 fed and 75 mg dust/m3. Many national regulations are
applied (Spain, Sweden, Poland) to each plant separately, taking into
consideration the emission balance, capacity, and environment
characteristics.
The level of emission is controlled using a periodic or continuous
standardized procedure and using reporting formats for comparison
with local regulations. Examples of measurement and analysis are
presented in TVA publications [60,61].
 Page 296

References
1. IFDC Fertilizer Manual, IFDC/UNIDO, Muscle Shoals, Ala.,
1984.
2. L. B. Nelson, Phosphatic materials, in History of U.S. Fertilizer
Industry, Tennessee Valley Authority, Muscle Shoals, Ala., 1990, pp.
371383.
3. W. A. Zajcev, Manufacture of Fluoride Compounds During
Fertilizer Production, Chimia, Moscow, 1982.
4. W. F. Karmysov, Chemical Phosphate Rock Processing, Chimia,
Moscow, 1983, pp. 276299.
5. U.S. Department of Agriculture and TVA, Superphosphate: Its
History, Chemistry and Manufacture, U.S. Government Printing
Office, Washington, D.C., 1964.
6. A. V. Slack, Phosphoric Acid, Marcel Dekker, New York, 1968, pp.
727778.
7. J. Schroeder, Superphosphate, in Technology of Phosphate
Compound, WNT, Warsaw, 1958, Vol. V, pp. 235325.
8. I. P. Muchlenov and O. C. Kovalev, Absorption of and Dust
Collection in Fertilizer Plants, Chimia, Moscow, 1987, pp. 7205.
9. F. C. Nielsen, Manual of Fertilizer Processing, Marcel Dekker,
New York, 1987.
10. S. D. Zviencikc and A. A. Brodski, Technology of Phosphate and
Multicomponent Fertilizers, Chimia, Moscow, 1987, pp. 306375.
11. S. Leszczynski

*, Developments in the Phosphate Fertilizer Industry, WNT, Warsaw,

1967.

12. M. E. Pozin, Technology of Mineral Salts, Chimia, Saint

Petersburg, 1970.
13. E. J. Fox, M. Stinson, and G. Tarbutton, Superphosphate,
Tennessee Valley Authority, Muscle Shoals, Ala., 1964.
14. M. E. Pozin and P. J. Ziniuk, Developments in the den method of
TSP manufacture, Chimie Prom. (Russian), Vol. 9, 1979, pp. 3739.
15. G. H. Wesenberg, Concentrated superphosphate: manufacturing
process, in Manual of Fertilizer Processing, F. Nielsen, ed., Marcel
Dekker, New York, 1987, p. 494.
16. A. S. Jackson, Production of single superphosphate with a TVA
cone mixer and bet den, in Manual of Fertilizer Processing, F.
Nielsen, ed., Marcel Dekker, New York, 1987, pp. 125138.
17. W. J. Sacket, Jr., and W. J. Sacket, Sr., The Suckett super-flow
process, in Manual of Fertilizer Processing, F. Nielsen, ed., Marcel
Dekker, New York, 1987, pp. 139145.
18. L. W. Bierman and J. R. Simplot, LETS process for triple
superphosphate, Chemical Engineering Progress, Vol. 149, 1982, pp.
8184.
19. J. R. Simplot, Direct granulation process for triple
superphosphate, U.S. Patent 4,101,637, 1978.
20. W. Herrings, Review of New Sources Performance Standards for
Phosphate Fertilizer Industry, 3-79-038, U.S. Environmental
Protection Agency, Research Triangle Park, N.C., 1979.
21. Sulfuric acid-based partially acidulated phosphate rock, Technical
Bulletin IFDC-T-31, International Fertilizer Development Center,
Muscle Shoals, Ala., 1986.
22. J. Hagin, Partially Acidulated Phosphate Rock: A Review,
Technion Israel Institute of Technology, Haifa, Israel, Vol. VI, 1985.
23. S. Kumar, Air pollution in superphosphate fertilizer industry in
India, Indian Chemical Manufacture, Vol. 30, 1979, pp. 15.
24. L. J. Hill, Removing gaseous pollutants from phosphate fertilizer
operation, Proceedings of the Fertilizer Institute Environmental
Symposium, New Orleans, 1976, pp. 421436.
25. L. N. Archipova, V. H. Senin, T. G. Repienkova, I. M. Malceva,
and B. I. Szramban, Purification of gaseous pollutants evolving in
continuous TSP production, Chimie Prom. (Russian), Vol. 7, 1975, pp.
510511.
26. A. J. Teller, Control of gaseous fluoride emission. Chemical
Engineering Progress, Vol. 63, No. 3, 1967, pp. 7579.
27. V. F. Karmysov, L. V. Vasilieva, L. J. Pierminova, and I. E.
Abramov, Granular TSP containing ammonium ion from Karatau
phosphate rock, Chimie Prom. (Russian), Vol. 9, 1974, pp. 669671.
 Page 297
28. T. I. Zaviertiaeva, Decomposition of phosphate rocks by wet-
process phosphoric acid, Chimie Prom. (Russian), Vol. 10, 1978, pp.
3638.
29. E. E. Aasamia, Studies on continuous TSP process using various
phosphate rocks. Chimie Prom. (Russian), Vol. 10, 1973, pp. 3639.
30. T. G. Riepienkova, L. P. Buslakova, I. F. Chripunov, and V. A.
Kopylev, TSP production by den-method, Chimie Prom. (Russian),
Vol. 7, 1975, pp. 509510.
31. S. R. Pandey, Disposal of wastes in phosphatic fertilizer plants,
Technology (India), Vol. 3, No. 4, 1966, pp. 6162.
32. J. C. Barber and T. D. Farr, Fluoride recovery from phosphorus
production, Chemical Engineering Progress, Vol. 66, 1970, pp. 5562.
33. J. M. Craig, Fluoride removal from wet-process phosphoric acid,
Environmental Sciences, Dissertation Abstract International, Vol. 31,
1971, p. 7.
34. V. V. Rao, Control of fluorine emissions in SSP plants, Fertilizer
News (India), Vol. 33, No. 5, 1988, pp. 4952.
35. B. I. Samban, L. D. Pavluhina, N. D. Afonova, and A. P. Gorskaja,
Equilibrium vaporliquid in system H2SiF6HFH2O, Zhurnal
Prikladnoi Khimii (Russian), Vol. 52, 1979, pp. 23442347.
36. G. I. Senomova, W. A. Kopylov, and M. E. Pozin, Evolving
fluorine-bearing gases from system H3PO4H2SiF6H2O, Zhurnal
Prikladnoi Khimii (Russian), Vol. 6, 1975, p. 1371.
37. W. A. Kopylov, G. I. Sienotova, and M. E. Pozin, Study on
equilibrium partial pressure HF, SiF4 and H2O in the system
H3PO4H2SiF6H2O, Zhurnal Prikladnoi Khimii (Russian), Vol. 11,
1976, pp. 24552458.
38. R. S. Mayze and P. O. Stone, Alternatives in handling fluorine
wastes from phosphate fertilizer plants, Proceedings of the 7th
Australian Conference on Chemical Engineering, 1979, pp. 175178.
39. G. Alloyer, Advances in gas scrubbing: the recovery of fluorine in
fertilizer manufacture, Phosphorus and Potassium, No. 110, 1980, pp.
2730.
40. A. O. Hansen and R. J. Danos, Scrubbers for phosphoric acid
plants, Chemical Engineering Progress, Vol. 40, 1982, pp. 8590.
41. R. C. Specht and R. R. Calaceto, Gaseous fluoride emissions from
stationary sources, Chemical Engineering Progress, Vol. 63, No. 3,
1967, pp. 7880.
42. T. I. Osag, J. A. Smith, F. L. Bunyard, and G. B. Crane, Fluoride
emission control costs, Chemical Engineering Progress, Vol. 10,
1976, pp. 3136.
43. H. O. Grant, Pollution control in a phosphoric acid plant,
Chemical Engineering Progress, Vol. 60, No. 1, 1964, pp. 5355.
44. J. Siden, An efficient scrubber, Phosphorus and Potassium, No.
175, 1991, pp. 3235.
45. A. J. Teller, Scrubbing of gaseous effluents, in Phosphoric Acid,
A. V. Slack, New York, 1968, pp. 741763.
46. P. Becker, Phosphate rock grinding, in Phosphate and Phosphoric
Acid, P. Becker, ed., Marcel Dekker, New York, 1987, pp. 195277.
47. R. D. Young, H. G. Graham, R. S. Meline, and J. R. Gahan,
Effluent control practices in granulation plants, Proceedings of the
International Superphosphate and Compound Manufacturers'
Association Conference, Sandefjord, Norway, 1970, pp. 126.
48. British Sulphur Co., The need for granular fertilizers, Phosphorus
and Potassium, No. 136 (supplement), 1985, pp. 327.
49. F. P. Achorn and H. L. Kimbrough, Fertilizer dust problems: major
source and available controls, Proceedings of the Fertilizer Institute
Dust and Noise Workshop, St. Louis, Mo., 1978, pp. 116.
50. F. P. Achorn and H. L. Kimbrough, Dust problems in plants,
warehouses and in the field, Proceedings of the Western Fertilizer and
Pesticide Safety School, Fresno, Calif., 1975, pp. 132.
51. E. Aalto and P. Suppanen, Dust control in NPK production,
Proceedings of the ISMA Conference, Prague, 1974, pp. 19.119.18.
 Page 298
52. A. J. Teller and L. E. Lombardi, Scrubbing fluoride emissions,
Proceedings of the Environmental Symposium, The Fertilizer Institute,
New Orleans, 1976, pp. 437446.
53. E. Bakke, Wet electrostatic precipitators on phosphate rock dryers,
Proceedings of the Environmental Symposium, The Fertilizer Institute,
New Orleans, 1976, pp. 401409.
54. R. A. Conway and R. D. Ross, Air pollution considerations in
industrial plant operations, in Handbook of Industrial Waste Disposal,
Van Nostrand Reinhold, New York, 1980, pp. 484549.
55. H. G. Power, Case history: use of bag dust collectors to control
particulate emission from dryers and coolers in NPK plants,
Proceedings of the Environmental Symposium, The Fertilizer Institute,
New Orleans, 1976, pp. 383399.
56. J. Baretincic, Environmental regulations, in Manual of Fertilizer
Processing, F. Nielsen, ed., Marcel Dekker, New York, 1987, pp.
473486.
57. M. R. Agarwal, Control of gaseous emissions in fertilizer industry,
Fertilizer News (India), Vol. 23, No. 10, 1978, pp. 1520.
58. B. H. Babu and V. C. Rao, Pollution and environmental control in
GFFL, Chemical Engineering World, Vol. 222, No. 11, 1987, pp.
7981.
59. K. T. Johnson, Pollution control in fertilizer industry: world
experience, Proceedings of the FAI-IFDC Fertilizer Seminar on
Trends in Consumption and Production, Delhi, 1977, pp. 214227.
60. E. D. Myers, Practical processes for determining losses from
granulation plants, Proceedings of the 32nd Annual Meeting of the
Fertilizer Industry Round Table, Atlanta, Ga., 1982, p. Z-147.
61. E. D. Myers, Particulate Sampling Guidelines and Reporting
Format, Alabama Air Pollution Control Commission, Montgomery,
Ala., 1972, pp. 133.
 Page 299

23
Utilization of Fluorine from Phosphate Fertilizer Plants
Henryk Jan Górecki
Technical University
Wroclaw

*, Poland

I
Introduction
Recovery and economic utilization of fluorine compounds from the
phosphate industry are important problems in the chemical industry,
economy, medicine, and agriculture. Recovery of fluorine-containing
compounds is theoretically and practically possible in all existing
technologies in phosphate raw material chemical processing. The
amount and quality of fluorine-containing by-products or salable
fluorine chemicals indicate the technical level of a fertilizer plant and
simultaneously, qualify or define fertilizer technologies according to
environmental and sanitary standards.
Development of the scrubbing system, new methods of purification of
scrubbing liquids, and recent developments in waste fluorine by-
product processing are destined to grow in importance in economic
fluorine recovery. Until the 1950s fertilizer plants were relatively
small and widely scattered, and practically no attempt was made for
sanitary control of fluorine emission. The only control was the use of
long chimneys to disperse gaseous fluorine effluents.
Before the mid-1960s no attempts were made for recovery, although
examples of a scrubbing water system are known. With the rapid
increase in fertilizer production and use in the 1980s when larger and
more efficient plants where built, and simultaneously with the scarcity
of fluorspar and expensive cost of fluorine compound production and
constant demand for fluorine compounds, fluorine recovery is an
important part of the fertilizer industry [1].
Describing this problem we must realize that:
1. Because of its high level of reactivity, fluorine compound emission
to the atmosphere must be drastically reduced [2,3].
2. Agricultural research and farm experience indicate the unfavorable
effects of fluorine
 Page 300
on plant vegetation and farm product properties, and the lowered
availability of phosphate fertilizer applied [4].
3. A high fluorine content is unwanted in feed phosphate used in
animal husbandry [5].
4. Production of fluorine-containing wastes in liquid and solid form
may cause disposal problems [6].
5. Chemical processing of phosphate rocks and all fluorine-containing
solutions can cause sanitary problems in terms of occupational
diseases [79].
6. Fluorine compounds in the sewage or drained sewage from
industrial piles can cause bioaccumulation of fluorides and the effects
of calcium metabolism in aquarium flora and fauna [10].
A
Motivation for Fluorine Recovery from Fertilizer Manufacturing
Facilities
Following the arguments presented earlier, the high efficiency of
fluorine recovery from phosphate raw materials is motivated by
various environmental factors and the following economic arguments:
1. Worldwide fluorine deposits available for exploitation are much
scarcer than phosphate rock deposits.
2. Recent and future fluorine recovery efficiency and the high
capacity of fluorine-containing waste solutions from fertilizer plants
are qualifying this route as being an equally valuable source to that
based on fluorine ores.
3. Possibility for financial stability for fluorine compound consumes
(e.g., the aluminum industry), because fertilizer products are in
constant agriculture demand.
4. Fluorine recovery and utilization technologies are more competitive
because of new developments in absorption systems, specific
recycling methods, efficient purification processes, and complex
utilization processing of waste solutions.
5. Natural reserves of fluorine ores are located in several countries,
and therefore delivery of fluorine products for industrial consumers
may be competitive.
Phosphate rocks of sedimentary and igneous origin contain in addition
to calcium phosphate, 10 to 15 major impurities and many trace
elements, mainly heavy metals and rare earth elements. During
chemical processing of phosphate rocks, recovery of some
components is possible, mainly during wet-process phosphoric acid
production or nitric acid acidulation of phosphate rocks. Theoretically,
rare earth elements (lanthanides) from apatites (Kola peninsula) and
scandium and titanium compounds from specific sedimentary deposits
may be recovered, but in practice, only fluorine and uranium are
recovered. The uranium content in sedimentary rock is very low,
about 100 to 120 ppm in high-grade ores (Florida, Morocco), but
thanks to convenient uranium transition (80 to 85%) to the liquid
phase composing phosphoric acid solution during the acidulation
process, it is easy to recover uranium compounds by efficient
extractive methods. The uranium recovery process started in Florida
in the 1950s, based on the process of clarification of wet-process
phosphoric acid and multistage countercurrent extraction of uranium
by a kerosene solution in phosphorusorganic liquid exchangers. Many
modifications of these recovery methods have been used industrially
in the United States, Japan, France, Canada, Morocco, and Iraq. The
development of these technologies is limited, and production in
existing facilities is unsettled partially because of the unstable
uranium market (demand and uranium prices) and rather high
recovery costs, depending on the investment costs for an additional
 Page 301
recovery line, chemical agent consumption, and special pretreatment
of crude wet-process phosphoric acid [2].
Fluorine recovery and utilization are free of these limitations and
simultaneously, all fertilizer producers must increase fluorine
emission control. The demand for fluorine compounds is stable thanks
to multiple uses for fluorine compounds.
B
Phosphate Rocks as a Secondary and Perspective Source of Fluorine
Fluorine belongs to elements widely distributed in nature. According
to many reports [1116], the fluorine content of continental rocks in the
earth's crust is estimated to be 0.02 to 0.09%, with an average of
0.065%. Finger [11] postulates its abundance as greater than that of
chlorine, copper, and lead, and it is contained in more than 100
fluorine minerals in fluorspar, cryolite, and phosphate rock, the
commercial sources of this element. World reserves and amounts of
fluorine compounds mined are presented in Table 1 [12,15,17].
Fluorspar and cryolite are classified as fluorine-rich minerals, with
phosphate rocks as low-content minerals although of importance
commercially. Many other minerals, such as topaz, bastenaesite,
lepidolite, amblygonite, lake evaporates, and villiamitte sellaite,
cannot be considered as potential industrial sources of fluorine
[11,1820]. Deposits of natural cryolites are rare and poor. The only
commercial mining, at Iviqtut in Greenland, is sufficint for short-run
operation [11,17]. Taking into account the relation of reserves to the
annual production of fluorite as a perspective source of fluorine must
be discussed. Even at 12 million short tons of 15 to 35% CaF2
containing more than 35% fluorite, these reserves are sufficient for the
short term. The United States, former Communist countries, Japan,
and Germany are the primary consumers of fluorspar, but two-thirds
of world production is located in seven countries (Mexico, the United
States, Spain, South Africa, China, Mongolia, and Thailand). In 1990
world fluorite consumption was about 9 million metric tons (average
45% CaF2).
The ratio of fluorine content in produced phosphate rock (about 5.8
million tons of fluorine) to fluorine content in fluorspar produced was
3:1. According to many geological estimations, fluorine reserves in
the form of fluoroapatite mineral are 60 to 100 times higher than in
fluorites. Current U.S. Bureau of Mines estimates indicate that the
reserves of natural fluorspar will be depleted toward the end of the
twentieth century, because total world production in this period is
estimated as actual reserves [12,14].
The geological situation and technological possibilities of fluorine
recovery from fertilizer processes must be considered in development
of fluorine industry. Fluorine
Table 1 World Fluorine Reserves and Amounts of Minerals Mined
Ore
Reserves 1990 reserves
Main Concentration (million production (years of
Mineral compound (%F) tons) (million tons) operation)
Fluorspar CaF2 2040 250.0 9000 2030
Cryolite Na3AlF6 2035 0.7 0.0466a 23
Phosphate Ca10F2(PO4)6 24 117,000 157,900 700800
rocks
Source: Data from Refs. 11, 12, 15, 17, 21, and 22.
a1982 [17].
 Page 302
Table 2 Fluorine Recovered in Fertilizer
Technologiesa
Fertilizer 197519801990
Single superphosphate 38.3 18.5 6.0
Triple superphosphate
Den method 11.9 7.6 2.7
Den continuous 3.1 10.0 6.5
Continuous 17.3 9.0 11.3
Recycled (granulated) 6.6 22.8
DAP and MAP based on wet- 11.2 19.6 18.1
process phosphoric acid
Nitrophosphate 11.5 22.6 26.5
Defluorinated phosphate rocks 7.5 6.1 6.1
Source: Ref. 15.
aTotal annual recovery in all fertilizer plants =
1000.

recovery as a professional part of fertilizer technology is a relatively

young art [1,2], but because of the scarcity of natural fluorite deposits,
the high cost of exploration and transportation, and more exact limits
of the fluorine compounds emitted, most world fluorine production
will have its source in the fertilizer industry. In 1977, about 7% of the
fluorine produced in the United States was obtained from fluorosilicic
acid recovered in the phosphate industry [15].
At present about 25% of the fluorine in phosphate rock is recovered
and used in fluorine compound manufacture. The last (1980) available
estimation of the fluorine share in production streams showed that
fertilizer products contain 60.7%; uncontrolled emitted gases, 0.54%;
neutralized sewages, 23.2%; and fluorine compounds produced,
15.6% [15]. This distribution does not include the fluorine content in
phosphogypsum.
The amount of the fluorine recovered depends on the route of
phosphate raw material processing. Taking into account
physicochemical data and technological solutions, about 1.2 million
tons of fluorine could be recovered annually from the phosphate rock
processed by the world fertilizer industry.
In Table 2, fluorine compound production is presented as a share of
fertilizer technology in the period 19751990. Actually, phosphate
rocks are the major secondary source of fluoride processed as salable
products, but in the future this source may become the primary source.
II
Principal Routes of Fluorine-Containing Mineral Processing
The industrial demand for a fluorine compound is the result of
multidirectional application of this element and its compounds
directly, as by-products, or as raw materials in various branches of
industry. The mineral fluorites and phosphate rocks are the basic
mineral source for fluorine compound production. Using this raw
material, many fluorine derivatives or commercial products are
manufactured. The processing system and principal applications are
presented schematically in Fig. 1. This system proves that there are
many
 Page 303

Figure 1
Schematic construction of fluorine-containing mineral processing system.
 Page 304
ways to alter the consumption percentages of natural fluorspar and by-
products from a phosphate plant. The production capacities
throughout this processing system will be changed as fluorine
compound consumption is reduced in the metallurgical industry,
including aluminum production. Simultaneously, increasing fluorine
derivative consumption in the chemical industry, primarily
hydrofluoric acid, is compensating for the decrease in demand of
traditional users such as the aluminum and steel industries.
In Table 3, specifications for the most important inorganic fluorine
derivatives and their application are presented. Table 4 provides
similar data for organic products containing fluorine. These two lists
show many and varied modern applications. The principal part of
fluorine production as fluorspar (natural and synthetic) and
hydrofluoric acid as basic raw material for many products, such as
aluminum fluoride and synthetic cryolite, fluorocarbons,
fluoropolymers, and fluorosilicic acid and its salts, is produced and
consumed in large amounts [2325].
In the period 19501960, the steel industry ranked first, consuming
more than 50% of the fluorine compounds produced [11]. In the next
decade this consumption decreased to about 30% and consumption of
fluorine products by the chemical industry rose to about 60%.
Fluorine material consumption in the glass, enamel, and ceramic
industry was stable at 6 to 9%. The increase in aluminum production,
in particular the expansion in fluorocarbon and fluoropolymer
production that took place in the 1970s and 1980s, increased the
demand for hydrofluoric acid as a primary starting material.
About 50% of the hydrofluoric acid produced at present is used for
AlF3, Na3AlF6, and NaF salts consumed in the manufacture of
aluminum, and nearly 40% is used in organic derivative production.
About 7% HF is consumed for special metallurgical purposes. The
remainder of the hydrofluoric acid produced is used in the
manufacture of catalysts and special inorganic and organic derivatives
for agriculture, the nuclear industry, electronic, and special materials
manufacture [11,17,20].
With the increasing demand for special inorganic and organic
products, constant increases in hydrofluoric acid consumption are to
be expected. The chemical industry is the largest user of fluorspar for
the manufacture of hydrofluoric acid, and with the rapid increase in
demand for this product will use large amounts of the waste
fluorosilicic acid solution produced in phosphate plants as an
alternative source of hydrofluoric acid, or directly for commercial
products.
Hydrofluoric acid is now and will be in the future the key source
material in the production of organofluoric compounds. These
compounds are a part of the actively developing organic technology
region. Many new compounds and polymers have been reported and
new applications are being discovered. Technological progress in this
field is more rapid than knowledge regarding properties, usage, and
environmental concerns. In 1930 the first fluorocarbons were formed,
and in 1934 the first fluoropolymers (PCTFE) were discovered. Then
came the development in 1938 of polytetrafluorethylene (PTFE), in
1948 of epoxyfluoropolymers, and in 1962 of phenoxypolymers. In
1980 more than 360 million tons of fluorocarbons were produced. Of
this total, refrigerants account for 46%, foam blowing agents for 20%,
solvents for 16%, fluoropolymers for 7%, and aerosol propellants for
less than 1% [26]. This distribution of manufactured products was
changed completely after discoveries by U.S. university scientists that
fluorocarbons used as aerosol propellants have been destroying the
ozone layer in the upper atmosphere, which protects the earth
naturally against the intensity of ultraviolet rays. Restrictions in 1978
on the use of fluorocarbons by the U.S. Environmental Protection
Agency (EPA), followed by wide international acceptation of these
restrictions, caused very rapid limitation on the
Table 3 Primary Inorganic Fluorine Compounds
Product Fluorine raw material Applications

CaF2 Fluorspar Opal, clear or opaque glass, colored

H2SiF6 glass, low-melting glass as flux and
Na2SiF6 opacifiers, prefabricated metal building
HF mineral, wool binding, brick staining,
building material, cement modification,
hydrofluoric acid, calcium cyanamide,
crystalline fluorspar, special optical
lenses, steel flux, welding additives

AlF3 HF Flux and electrolyt

H2SiF6 Flux and opacifiere
(NH4)3AlF6 Welding flux

Na3AlF6 Cryolite (natural) Electrolyte

HF
AlF3 Flux and opacifier
H2SiF6 Welding flux, casting agent
Insecticide

HF CaF2 By-product in the manufacture of

AnhydrousH2SiF6 inorganic fluorides, by-product in
organic synthesis catalyst (alkylation)

Solution By-product in the manufacture Inorganic chemical industry, distillery

of acid fluorides and fluorides, and brewery, cleaning enamels, glass
antiseptic agents; cleaning of industry, special users
colored metal, flux modifier,
frosting, etching, polishing,
removing silica from graphite,
acidizing oil wells, removing
 effloresence from bricks

(table continued on next page)

(Continued)
Table 3 Primary Inorganic Fluorine Compounds
Product Fluorine raw Applications User (industry
material
H2SiF6 Scrubbing Fluoridation, disinfectant for copper and brass Municipal water
liquids vessels, preservative, agent for casting light disinfection, wood
SiF4 metal, main by-product for inorganic fluoride protection, metallur
derivatives chemical industry
Na2SiF6 H2SiF6 Insecticide, fungicide, laundry soap, flux and Agriculture, househ
SiF4 opacifier, casting of light metals, toothpaste chemistry, glass and
Precipitate from ingredient enamel, metallurgy,
acidulation hygiene
K2SiF6 H2SiF6 Opacifier, light metal flux, synthetic mica Enamel, metallurgy,
ceramics
Precipitate
(wet-
phosphoric
acid, nitric
phosphate)
NaF H2SiF6 Raw material in the manufacture of AlF3, Chemical industry,
Na2SiF6 Na3AlF6 metallurgy, aluminu
HF Steel flux, agent for metal refining, additive to industry, agriculture
cryolite fungicide, wood preservative, municipal water
insecticide, fluoridation
Special applications (lower production)
UF6 F2 Separation and concentration of 235U from the Nuclear technology,
UF4 HF more abundant 238U, semiproduct UF4 (green uranium recovery fr
salt) manufacture, salting out UF4 from organic wet-process phosph
extracts acid
F2 KHF2 SF6 as high-voltage insulation and x-ray tube Electronics, nuclear
HF construction, UF4, UF6 manufacture, special technology, special
fluoride synthesis: CoF3, AgF2, MnF2, ClF3, application, fluoriza
and ClF6 (catalyst)
(table continued on next page)
(table continued from previous page)
Product Fluorine raw Applications User (industry)
material
ZnSiF6 H2SiF6 Wood preservative, laundry sour, Special chemicals.
MgSiF6 SiF4 concrete
(NH4)2SiF6 hardener, mothproofing agent
BaSiF6 H2SiF6 Insecticide, opacifing agent Agriculture, enamel and
CaSiF6 SiF4
PbSiF6 H2SiF6 Electrolytic lead plating Metallurgy
SiF4
AlSiF6 H2SiF6 Fluidization of concrete Building material
ZnSiF6 SiF4 Hardening of concrete
NH4F HF Tarnish glass preservative agent, Electric and glass brewe
absorption distillery
agent in HF production yeast, chemical technolo
glass industry
NH4F · H2SiF6 Frosting and etching agent
4H2O HF In elemental fluorine production, Chemical industry, wood
KF preservative, protection, glass
KHF2 tarnish agent
BaF2 HF White pigment Enamel
SbF3 HF Mordant, temper Typography, paint
SbF5 HF Fluorescence agent Organic chemistry
CoF2 NaF Catalyst Organic technology
HF
CrF3
CrF3 HF Mordant Typography
·4H2O HF Impregnation and preservation Wood and building mate
ZnF2 · HF Together with fluorosulfonic acid as Polymers, petroleum
4H2O F2 catalyst Rocket technology
HF Component of fuel
BF3
ClF3
HBF4 H2SiF6 Cleaning and pickling Metallurgy
HF
NaBF4 HF Oxidation inhibitor Casting, aluminum indus
H2SiF6 Fire-retardant, preservative wood construction
PF3, PF5, HF Raw materials for ''nerve gas" Military, organic synthes
POF3, FSF3 manufacture for
biologically active organophosphorus
fluorine compounds
Source: Data from Refs. 2, 6, 11, 13 to 16, 20, and 24.
 Page 308
Table 4 Organic Derivatives Containing Fluorine
Product Synthesis Applications
Fluorochloroalkanes HF reaction with As aerosol propellants (restricted
Fluorochlorocarbons chlorinated alkenes application), refrigerants (Freon 11,
CCl2F2, using catalysts such as 12, 22, 113, 114)
CHCl2F, CHC13F, CCl2F2 SbF3, AgF, NaF
Fluorochloroalkenes Fluorination using HF As semiproduct in many organic
and F2 synthesis
Fluorochloroarenes Solvents, semiproduct in organic
Fluorochlorination synthesis
using HF, Cl2, and
FeCl3, AlCl3 as
catalyst
Fluoroalcohols Telemerization of Dyestuffs, plastics, pharmaceuticals,
tetrafluoroethylene, surface coating, thermally stable
introducing fluoroalkyllubricants, solvents
group into
organic molecule
Trifluoromethyl aromatic Benzenoid compound Biologically active compounds
compounds stabilization
Fluorochloroketones Decarboxylation of Solvents, semiproduct in organic
fluoroketoacids synthesis
Fluorochloroacids Exchange one or Pesticides, very strong acid in
several H atoms in organic synthesis
alkyl or aryl group
Fluoropolymers Polymerization of Materials of high chemical
(e.g., hexafluoropropene) fluoroalkenes (e.g., resistance, good dielectric
tetrafluoroethylene); properties, low friction coefficient
co-polymerization
(e.g.,
hexafluoropropene)
Copolymers, fluororesins, According to polymer PVDF, PTFE, PTFCE polymers for
elastomers [e.g., Teflon technology using electrical insulation mechanical
(U.S.), Ftorplast 4 (USSR), epoxy seal, frying pans, coatings, gaskets,
Fluan (U.K.)], fluorocrystalline lining of chemical vessels, pipes,
epoxyfluoropolymers, polypropylenephenoxy electronic special materials
phenoxypolymers resins
Source: Data from Refs. 11, 20, and 26 to 28.
 Page 309
manufacture of these products. In 1974, of these restrictions aerosol
perpellants had been the principal use of the fluorocarbons produced.
Now and in the near future, continuing progress and development in
fluoroplastics production can be expected. These products are
characterized by very useful properties such as high chemical and
mechanical resistance, because these materials are nonflammable,
nonoxidizing, acid-proof, and alkaline-proof, and they are
distinguished by mechanical and electrical resistance, low coefficient
of friction, low permeability, low dielectric strength, and low moisture
absorption. These products are in common use in the chemical
industry for lining, pipes, gaskets, and seals, and are used in frying-
pan construction. Many fluorine organic materials are used for
electronic and electric purposes (e.g., insulation, construction, cables,
switches), in the agriculture and food industries (as lining, packing,
glands, gaskets), as modern materials in the automotive industry, and
in varied military and space applications.
In recent years interest and progress in the manufacture and special
uses of fluorine have increased. New processes for the production of
hydraulic fluids, special fire extinguishers, unusual plastics and
elastomers, paint and dyes, special oils, lubricants, greases, and
coolants, and agents for galvanic techniques and electroplating have
been developed. In agriculture there are new types of pesticides and
repellents, and in pharmacy and medicine, new forms of synthetic
blood, diuretics, steroids, tranquilizers, and anesthetics [11,17,26].
Many new processes and products have been proposed and developed
in inorganic and organic fluorine-containing technology. Description
of these processes is beyond the scope of this chapter.
III
Fertilizer Industry as a Competitive Source of Fluorine Compounds
During the processing of phosphate rocks using strong mineral acids,
fluoride compounds are converted into fluorosilicic acid, which is
partially evolved in various stages of fertilizer production. Details on
the technology and distribution of fluorine compounds in process
streams, including process parameters and quality (grade) of raw
materials, are presented in Chapters 22 and 23 and Part V.
A
Fluorine By-Products from Fertilizer Plants
Fluorine compounds are removed from wet-process phosphoric acid
plants and triple and single superphosphates in the form of 15 to 25%
fluorosilicic acid. Recovery is also possible [2,6,2935] from wet-
process phosphoric acid obtained from low-grade phosphate ores
containing more than 1.5% MgO using alkaline salts [15,30]. In this
process precipitation of Na2SiF6 takes place.
Fluorine compounds in the same form are recovered in the nitric
phosphate process. The fluorosilicates are precipitated from the
acidulation solution, and from a solution of calcium nitrate by adding
sodium salts. In addition to fluorosilicic acid and sodium
fluorosilicate, fluorine compounds are removed from effluent gases in
thermal defluorination processes and superphosphate plants in the
form of a mixture of CaF2 and CaCO3 using lime milk as the
absorption agent [32,33].
Specifications for fluorine compounds in various stages of fertilizer
technology are presented in Table 5. Of course, some of these
compounds are negligible for recovery
 Page 310
Table 5 Forms of Fluorine Compounds Developed During Phosphate Rock Processing
Compound Technology Occurrence
SiF4 Wet-process phosphoric Reactor vapors, cooling, condensate, filtration
acid (WPA) vapors, evaporation
HF Single superphosphate Acidulation, aging, granulation, drying, cooling
(SSP), SSP (granulated)
Triple superphosphate Acidulation, aging, granulation, drying, cooling
(TSP)-den, TSP
(granulated)
Partially acidulated Acidulation, granulation, cooling
phosphate rock (PAPR)
Na2SiF6 Nitric phosphate (NP) Removing fluorine compound by Na2CO3
addition to Ca(NO3)2 solution
WPA (low-gradehigh-Mg Removing fluorine compounds by sodium salt
phosphate rock) addition to crude WPA
WPA dihydrate Precipitate partially in reactors during slurry
cooling scaling in tanks, pipes, flash cooler
K2SiF6 WPA During shipping or storage high-K+-containing
acid (rock or incidentally, admixture with NPK)
CaF2 WPA Precipitates during evaporation crude acid to 40%
P2O5
Defluorinated feed Sludge obtained during scrubbing by lime milk
phosphate

(table continued on next page)

 Page 311
(table continued from previous page)
Compound Technology Occurrence
WPA Removing impurities from sewage by lime or
phosphate chalk treatment
fertilizers
Ca5F(PO4)3 WPA, TSP, Partially undecomposed phosphate rock (in all
SSP, PAPR, stages)
NP
CaSiF6 · 2H2O NP Precipitates from nitric acid acidulation
solution
AlFHPO4 ·2H2O WPA Precipitates during cooling of WPA
NaF WPA Component of scrubbing liquids during
absorption by alkaline solution
NH4F WPA Component of the absorbing liquids in recovery
fluorine process by ammoniatic solution
HSO3F WPA, Form occurring during distillation fluorine
connected compounds from H2SO4 (ca. > 80%)
with HF
KNaSiF6 WPA Constituent of the scaling precipitate in pipes
and flash cooler
CaSO4CaSiF6CaAlF6(Na)WPA Salt occurs under the filter cloth, scaling the
· 12H2O filter pipe system
CaSO4CaSiF6CaAlF6(x) ·
10H2O
MgAlF6Na · 6H2O WPA Salt precipitate during concentration wet-
process phosphoric acid from 30% to 50%
P2O5
H2PO3F WPA Acid exists in the concentrated solution of
phosphoric acid
Source: Data from Refs. 2, 6, 29, and 35, and Feed Phosphate Works, Bonarka, personal
communication, 1992.
 Page 312
purposes, but in reality they present technological difficulties
[2,6,30,36] (Feed Phosphate Works, Bonarka, personal
communication, 1992).
B
Fluorosilicic Acid:
The Most Important Fluorine By-Product from the Fertilizer Industry
The main fluorine by-product from the fertilizer industry is in the
form of a crude aqueous solution of fluorosilicic acid. This solution is
produced mainly in the larger wet-process phosphoric acid and
superphosphate plants and is converted in the form of salable products
in utilization plants connected with fertilizer plants (Na2SiF6, AlF3,
NaF, Na3AlF6, HF) in large amounts. The waste diluted solutions
(e.g., sulfuric acid) are recycled to the fertilizer plants. In the smaller
plants fluorine compounds are usually removed in the form of calcium
fluoride sludge that is formed in the process, using lime milk in the
absorption process. Utilization of this process is for sanitary purposes
only. The recovery process is not profitable on this scale.
Sometimes a system of collection of by-product fluorosilicic acid
from several plants for transport to a central refining and utilization
unit is used. This system can also operate profitably for smaller
fertilizer plants located in built-up regions, where there are the most
stringent environmental restrictions and highest penalties for waste
disposal [37].
Possible fluorine recovery efficiencies for various technologies are
presented in Table 6. For comparison, nitrophosphate and thermal
feed phosphate data are also reported [30,37]. In this table the
concentration of fluorine compounds in effluent gas and the flow of
these gases are reported. Sometimes, alkaline solutions (NaOH,
NH4OH, Na2CO3) are used to increase the purity of the exhausted
gases [15,30].
The fluorosilicic acid solution obtained in scrubbing systems is
partially purified using clarification and filtration methods for
removing solids components (Fe, Al, Mg salts, phosphate compounds,
SiO2). Figure 2 illustrates the possible routes of much solution
processing to obtain commercial products or by-products. The
schematic utilization system presented is simplified; it is also possible
to convert Na2SiF6 precipitate acting NH3 to obtain NH4F and
NH4HF2 and from these solutions to product CaF2 or MgF2. In
addition, K2SiF6 may be converted to KF,NH4F, and KHF2
[15,30,32,33].
C
Qualitative Requirements for Consumers of Fluorine By-Product
Solutions
From an economic standpoint, it is important to recover fluorine
compounds with high efficiency and simultaneously, to produce high-
purity fluorine by-product. Technological data, including fluorine
recovery and P2O5 concentration based on information from plants
where Kola apatite is processed, are presented in Table 7.
Of course, the composition of the received fluorine-containing
solution depends on the fertilizer technology and the construction of
the absorption system. Ranges of the most important impurities of
fluorosilicic acid are presented in Table 8. These are very wide, and
their lower points correspond with solutions that were first purified or
aged and then filtered on a semicontinuous belt filter. The by-products
in the form of sodium fluorosilicate obtained in the nitrophosphate
process or from wet-process phosphoric acid as precipitate using
sodium salts contain relatively high levels of impurities. This
precipitate contains from 10 to 30% other salts. Calcium sulfate is the
main constituent of these impurities.
Table 6 Fluorine Recovery in Fertilizer Technologies
Gaseous
streams
Technology Raw material Source of emission (1000 m3/h)
SSP Igneous Den 710
Sedimentary Den 710
Granular SSP SSP (igneous) Rotary drum 2030
from pile
SSP (sedimentary) Rotary drum 2030
from pile
Granular TSP Apatite WPA Den 710
TSP from pile Rotary drum dryer 1020
Crude phosphoric Blunger, rotary drum dryer 50150
acid and apatite
Wet-process Apatite Reactor 2050
phosphoric acid, Vacuum flash cooler 1015
dihydrate method Filter and washing stage 1015
Vacuum evaporator 10
Vacuum evaporator with 30
blowing
Defluorinated Apatite (Kowdor) Rotary kiln 5060
phosphate, Sedimentary Cyclone furance 3040
Ca3(PO4)2 (Karatau)
Thermal feed Igneous (Kola) Rotary kiln 80100
phosphate,
CaNaPO4
Nitric phosphate Apatite Removing CaSiF6 from acidulation solutions and Na
reacting Na2CO3 with calcium nitrate solution
Source: Data from Refs. 28, 30, and 36, and Feed Phosphite Works, Bonarka, personal commu
 Figure 2
Possible routes of by-product fluorosilicic acid
utilization for the main large-tonnage salable products.
Table 7 Recovery of Fluorine and P2O5 Content in Fluorosilicic Acid Scrubbing Solution
(10% H2SiF6)
P2O5 concentration Recovery of F
in 10% (kg F/1000 kg
Technology H2SiF6 solution (%) P2O
Wet-process phosphoric acid (dihydrate) 0.020.05 4852
Wet-process phosphoric acid (hemihydrate) 0.020.05 3035
Single-superphosphate apatite (Kola) 0.05 2930
Phosphorite (Karatau) 0.05 3435
Triple-superphosphate (continuous) method 0.60.8 40
Wet-process phosphoric acid with high Mg Na2SiF6 precipitate 45
concentration (Karatau)
Nitrophosphate (apatite) Na2SiF6 precipitate 45
Source: Ref. 23.
 Page 315
Table 8 Composition of Crude Fluorosilicic Acid Solutions Produced in
Fertilizer Plants
Wet-process Single-
phosphoric acid superphosphate Granulated
Component Unit plant plant TSP plant
H2SiF6 % 1525 1525 1525
P2O5 ppm 1001000 5004000 20005000
CaO ppm 50500 10005000 5005000
Fe2O3 ppm 50200 50300 100500
A12O3 ppm 3050 50150 100300
SO4 ppm 1001500 10002500 3001000
Cl ppm 1001500 1001500 1001500

Waste fluorosilicic acid may be used without additional treatment within

established limits. The quality requirements [15] for the final product are
described next.
CaF2 (synthetic fluorspar) CaF2 max. SiO2 S max. P max.
max.
metallurgical grade min. 85% 5% 0.3%
ceramic grade 8595% 2.5% 0.12% 0.1%
acid grade (for HF production and min. 97% 1% 0.05% 0.1%
moisture min. 1%)
Na2AlF6 (synthetic cryolite) SiO2 max. Fe2O3 SO4 max. P2O5
max. max.
metallurgical grade I 0.5% 0.06% 0.5% 0.05%
metallurgical grade II 0.9% 0.08% 1.0% 0.20%
ceramic and enamel grade 1.5% 0.10% 1.0% 0.6%
AlF3 (aluminum fluoride) SiO2 + SO4 P2O5
Fe2O3 max. max. max.
metallurgical grade I 0.3% 0.5% 0.05%
metallurgical grade II 0.4% 0.7% 0.10%
ceramic and enamel grade 0.5% 1.0% 0.5%
NaF (sodium fluoride) Na2CO3 Na2SO4 undiss. %
 max. max. max.
special grade 0.2% 0.3% 2.0%
second grade 4.0% 2.5%
HF (hydrofluoric acid) HF min. H2O SO2 max. SO4 H2SiF
max. max. max.
special grade 99.95% 0.030% 0.007% 0.005% 0.01%
second grade 99.87% 0.060% 0.015% 0.010% 0.03%

D
Purification of Fluorosilicic Acid Produced in Scrubbing Systems of
Fertilizer Plants
Fluorine-containing gaseous effluents from the reaction section of a wet-
process phosphoric acid process, an evaporation section, and in particular,
from other technological stages in SSP and TSP manufacture contain
droplets of phosphoric acid and small particulates of phosphate rocks,
TSP or SSP, silica, and iron magnesium and aluminum salts. Scrubbing
liquids of 15 to 25% H2SiF6 concentration in which components of
gaseous effluents were absorbed may be utilized for salable products such
as Na2SiF6, AlF3, Na3AlF6, and synthetic fluorspar-CaF2. Several
methods can be used for contamination reduction to the level called for
by technological standards for the raw materials of these technologies.
 Page 316
The primary methods used depend on the following processes and
equipment:
Specific entertainers
Dust collector
Multistage absorption system, used as a preliminary absorption step in
concentrated H2SiF6 solution droplets of acid (H3PO4 and H2SO4)
Chemical precipitation
Clarification and fine filtration of by-product solution
One of the most popular systems, called the English system [3941],
depends on removing phosphate compounds by scrubbing the vapors
with an aqueous solution that absorbs the impurities but minimally
absorbs the fluorine compounds. This process occurs at a temperature
that assures water vapor condensation.
In this stage, concentration of H2SiF6 is higher than 30% H2SiF6, and
about 5 to 7% H3PO4 concentration is achieved in a sealed box
container. This solution is removed from the system and transferred to
the acidulation step. The basis for this system is shown in Fig. 3.
Vapors leaving the evaporating vessel (e.g., phosphoric wet-acid
evaporation stage) via the outlet contain mainly SiF4, HF, water, and
phosphoric acid droplets. Most systems, such as the Hydro-Supra
solution [41], are equipped with special entrainment separators to
remove large droplets of phosphoric acid. These entrainers are located
 Figure 3
Absorption system with preliminary absorptionpurification
scrubber for phosphoric acid removal. (From Ref. 40.)
 Page 317
between the separator vessel and the first scrubber. A similar effect
can be achieved using the sulfuric acid solution in the first absorption
stage similar to the way it is used in hydrofluoric acid manufacture.
According to an Allied Co. patent [42], tetrafluoride gas is evolved
from concentrated sulfuric acid in a vertical column of special
construction. The gases are introduced to the middle part of the
column, and a solution of lower H2SO4 concentration than that in the
feed sulfuric acid containing impurities is recycled or processed in
some other way. The SiF4 evolved is free of phosphates and other
impurities.
Important and effective methods of removing impurities from
fluorine-containing effluents depend on passing a gaseous stream
through a fluidized packed impingement separator in which the
droplets are removed by contact will ball packing in the separator
(Fig. 4). The exhaust gaseous stream is droplet-free. The phosphoric
acid solution is removed by spraying the packing with a H2SiF6
solution of 40 to 70% concentration. The preliminary purified gas
stream may be scrubbed by a solution of H2SiF6, alkali metal
fluoride, or alkaline earth metal fluoride. This system makes it
possible to produce a fluorosilicic acid solution containing less than
150 ppm phosphoric acid. When the first absorption step is omitted,
the fluorosilicic acid will be contaminated by phosphoric acid at a
concentration of 0.1 to 0.5% P2O5.
A modification of this system by Mitchell Ltd. [44] depends on the
introduction into the boiling solution of a concentrated phosphoric
acid aqueous solution of the fluorosilicic acid. As a result,
stoichiometric quantities of 2HF and SiF4 are desorbed. According to
this procedure, the vapors obtained are next scrubbed with a solution
of H2SiF6 of concentration higher than 30% and absorbed in the next
absorption stage, in which a solution of less than 30% H2SiF6 is used.
The vapors from the boiling tank first pass through a deentrainer and
then to the first cleaning scrubber, where concentrated H2SiF6 is used
to remove phosphoric acid fine droplets without fluorine absorption
and without

Figure 4
Fluidized packed impingement system for phosphoric
acid droplet removal from fluorine-containing gases. (From Ref. 43.)
 Page 318
water condensation. This system also assures a low silica content in
the by-product fluorosilicic solution.
An important and essential development in this field is the result of
new processes proposed by Parish [45,46]. These methods are based
on properties of the boiling temperature of the system
HFH2SiF6H2SO4H2O. In the Parish systems [39] for concentrating
fluorosilicic acid, concentrated sulfuric acid is added. The addition of
silica causes SiF desorbtion from the dehydrated fluorosilicic solution.
In this process the concentration of H2SO4 is lowered by 10 to 20%.
The basis of these methods is shown in Fig. 5.
Other techniques for the removal of phosphoric acid droplets and
particulates in the manufacture of superphosphate have been
proposed. One of them [48] comprises diluting the gaseous pollutant
8-to 12-fold by volume of air per volume of waste gases, separating
the droplets and dust by passing into a cyclone, and subsequently
eluting the resulting waste gases with water to form fluorosilicic acid.
This method utilizes the fact that condensation can be avoided very
simply by diluting the waste gases before they enter the cyclone with
air to the point that under no circumstances does the temperature fall
below the dew point.
Other ways of removing the impurities from fluorosilicic crude acid
consist of chemical treatment of the solution [47]. The favorable effect
achieved when removing AlPO precipitate occurs by addition of
aluminum salts as AlCl3·6H2O, Al(NO3)3·9H2O, Al(SO4)3·13.5H2O,
A12(SO4)3(NH4)2SO4·2H2O, and (NH4)3AlF6. The purified solution
contains P2O5 below 0.04 to 0.001% [48].
Fractional crystallization of (NH4)2SiF6 from H2SiF6 solution using
NH3 makes it possible to remove pure, phosphate-free ammonium
fluoride crystals [49]. In another proposed method, purification of the
sodium fluorosilicate precipitate by special repulping techniques with
a sodium salt solution leads to a pure product [50].

Figure 5
Parish system used for concentration of fluorosilicic acid. (From Ref. 39.)
 Page 319

IV
Principal Process for Converting Fluorine-Containing By-Products
into Salable Products
Although many processes for fluorine by-product production from
fertilizer manufacture have been developed, only a few methods have
been realized on a commercial scale for large volume production. The
H2SiF6 for direct use, Na2SiF6 and K2SiF6, HF, NaF, AlF3, Na3AlF6,
NH4Cl, and synthetic fluorspar are presently manufactured in close
technological connection with fertilizer plants. Some details of the
manufacturing methods used to produce these products from fluorine
by-products of the fertilizer industry are presented below.
A
Direct Use
Crude fluorosilicic acid and its sodium salts have many uses based on
their bactericidal and fungicidal effects. These products are used as
fungicidals, steriliants, disinfectants, and wood preservatives.
Fluorosilicic acid is used in brewing and distilling and for many
purposes in building material production (e.g., as a concrete
hardener). Small amounts of purified fluorosilicic acid and its sodium
salts may be used for the fluoridation of municipal water, with up to 1
ppm F concentration in drinking water. Although the applications
noted are wide ranging, they utilize only very small amounts of the
fluorosilicic acid produced. The largest part is used as a raw material
for fluorosilicic salts and hydrofluoric acid production.
B
Recycled Acidulation Raw Material
Beginning in the 1960s, an H2SiF6 utilization test was used for using
this solution to phosphate rock acidulation in the Soviet Union. These
tests attempted to improve the wet-process phosphoric acid obtained
from phosphate rock containing magnesium (e.g., Karatau and
Kingisep) and to design a waste-free process without phosphogypsum
[51]. Processing of raw material containing a high level of magnesium
in the manufacture of triple superphosphates leads to a low-quality
product that has a tendency to form lumps. This effect is removed
when H2SiF6 is introduced to the reaction slurry, causing
MgSiF6·2H2O precipitation. In this process an MgO/P2O5 ratio of
0.02 is reached when a fluorosilicic acid solution of 24.7% H2SiF6
concentration is used to acidulate the phosphate rocks. The acid
obtained is suitable for TSP manufacture. As a result of the laboratory
tests, a process of H2SiF6 utilization for phosphate raw material
acidulation was presented [52] based on the reaction

The HF evolved reacts with SiO2 and SiF4 and is absorbed. The
reaction slurry is filtered off and soluble CaSiF6 reacts according to

and filter cake containing CaSiF6·2H2O is thermically decomposed:

 Page 320

Fluorine-containing gases are absorbed and recycled to the acidation

process. The interesting idea behind this complex process has never
been used in practice.
In Saint Petersburg (NIUNIF) a method for the manufacture of wet-
process phosphoric acid has been proposed. The LTL Lensovieta [23]
process for ammophosphate manufacture according to the reaction

and using a regeneration process according to the reaction

and together with the Merseburg [6] conversion process,

CaSO4·2H2O using CO2 and NH3 to CaCO3 forms an interesting idea
for the utilization of H2SiF6. This is another example of a theoretical
process that has never been used, as is the method [53] elaborated at
Kiev Technical University based on the process

As the technological result of these multidirectional investigations, the

reaction

may be used in the defluorination of by-product liquid containing

ammonium silicofluorates.
An objective of the Montedison SpA modification single-
superphosphate process was changed to eliminate fluorine compound
production as a by-product, simultaneously restricting sulfuric acid
consumption. This process depends on recycling to the acidulation
step the adsorbed H2SiF6 solution, which together with H2SO4 is used
for decomposition of the phosphate rocks. This process leads to 5%
sulfuric acid consumption, avoids waste fluorosilicic acid processing,
and decreases fluorine gas concentration in the working area. This
closed system requires increasing sulfuric acid concentration by about
8%, higher efficiency in the scrubbing system, and of course leads to
higher fluoride concentration in the superphosphate fertilizer
produced. The F/P2O5 ratio in the final product is similar to the value
of this ratio in the phosphate rock. The content of the available P2O5
is the same as in SSP produced conventionally. This process was
tested in a plant equipped with a rotary drum type of den and rubber
belt in which Land Pabble and Khouribga rock were used [5456].
Several single-superphosphate plants in India [57] use a fluorosilicic
acid recycle system equipped with two spray towers and a final
packed tower operating as a countercurrent system. The exhausted
gases from this absorption stage contain about 25 mg F/N·m3.
C
CaF2:
Synthetic Fluorspar
Recently there has been a trend toward the manufacture of synthetic
fluorspar from waste fluorosilicic acid. This route may also be used
for the production of pure silica. The pure fluorspar can be used as a
raw material in the manufacture of hydrofluoric acid as a starting
material in fluoroorganic chemistry.
Presently about half the natural fluorspar mined is processed to
hydrofluoric acid. This conventional route is expensive and high in
the consumption of sulfuric acid. In this
 Page 321
traditional process the highest grade of fluorspar is required, whose
price continues to increase sharply. These problems cause increasing
interest in fluorspar production based on the use of fluorine by-
products and wastes sludges recovered during the neutralization of
liquid effluents in the lime neutralization process.
1
Fluorspar Produced in the Pseudomorphic Replacement Process
From the diluted fluorosilicic acid obtained in the absorption of
effluent gases it is possible to produce CaF2 by a pseudomorphic
replacement process [58]. In the first stage H2SiF6 reacts with
ammonia and the NH4F solution obtained reacts with granular CaCO3
to form granular-shaped particles of synthetic fluorspar with an
efficiency of about 97%. This countercurrent [59] process was
demonstrated in a pilot plant. The results obtained proved that the
particles of calcium fluoride formed were similar in shape and size to
those of the starting granular feed. The test showed that the product
obtained had a favorable chemical analysis. The calcined product at
1000°C contained 47.5% F, 50% Ca, 0.1% P, 0.1% Si, 0.7% A12O3,
and 0.9% Fe2O3 [58]. This process may be used for CaF2 production
using hydrofluoric acid solution or as a sewage treatment process
[60].
2
Fluorspar Produced in the Lime Adsorption Method
The many small fertilizer plants and, in particular, feed phosphate
production based on the thermal defluorination method prefer to
neutralize fluorine gases with limestone or milk of an acid. In these
processes a precipitant sludge of CaF2 and SiO2 is produced. Under
special process parameters and composition of the absorption liquids,
silica can be precipitated first. This effect can be achieved if
ammonium salt solution is used as the absorbing medium [61]. The
fluorosilicic acid reacts with ammonia to form NH4F, which is then
converted to CaF2. The silica in this process is precipitated first. In the
other process, silica remains in metastable form and the precipitate
can be separated by filtration or centrifugation.
The direct absorption of fluorine gases in the thermal defluorination
process is a two-stage system in which in the first stage, diluted HF is
absorbed, and in the second stage, lime milk is used. A by-product
containing 75% CaF2, 22% CaCO3, 1% SiO2, 1% CaF(PO4)3, and
1% Ca(OH)2 is produced. This product is utilized successfully in the
manufacture of special building materials [37]. The problem of
separating silica from a fluoride solution or suspension is the difficult
part of these methods. Modifications depending on clarification or
speed setting fluorspar using polyacrylamide flocculants have been
proposed [62].
3
Fluorspar Produced in the Sewage Treatment Process
In many sewage treatment processes in the fertilizer industry and in
many methods used to recycle pond water, fluorspar is removed by
using CaCO3 or CaCO3. This process occurs according to the reaction

The solids are settled and water is returned to the fertilizer processes
after thickening of the solid matter, which is washed in a
countercurrent system. The solids are dried and briquetted and can be
used as lower-grade fluorspar containing about 75% CaF2, 4% P2O5,
and 1.0% SO4. The Agrico Chemical method [63] utilizes as an
additional process the stage of sulfate precipitation in the form of
CaSO4·2H2O and alkaline washing of the solid matter precipitated.
 Page 322
D
HF:
Hydrofluoric Acid
Pure hydrofluoric acid in aqueous solution or in the form of
anhydrous acid is a material proposed for use in many organic
syntheses, mainly in the production of fluorocarbons and
fluoroorganic polymers and special fluorine derivatives. Admittedly,
the traditional route of HF production depends on the mineral
fluorspar, containing more than 97% CaF2, with decomposition by
concentrated sulfuric acid. But recently, the waste fluorosilicic acid
from fertilizer plants is being used more and more frequently as a
competitive process. Many different procedures have been proposed,
but only a few of these have been put into commercial operation.
Problems with impurities, undefined physicochemical equilibrium and
kinetic data, corrosive conditions, and with separation and utilization
of the waste silica limit practical use of this process [17,64].
The primary patented and developed processes that comprise the
routes of HF production from waste H2SiF6 are described next.
1
Combined System of Waste Fluorosilicic Acid Decomposition by
Concentrated Sulfuric Acid
Methods have been developed to utilize the properties of the system
H2SO4H2OHFH2SiF6SiF4 for separation of HF as a pure product.
Many modifications of these methods have been developed, which
differ in their approach to SiF4 recycling and the use of diluted
sulfuric acid.
In the preliminary step, solid particles are removed by clarification or
filtration and in special modifications, inorganic agents are added for
the precipitation of phosphate compounds. In the first stage, SiF4 is
evolved by treating fluorosilicic acid with concentrated sulfuric acid.
In the second stage, hydrofluoric acid is removed as substantially dry
vapor by distillation of the liquid phase from the first stage containing
H2SO4 and HF. The basis of this method is illustrated in Fig. 6. This
process is modified by the introduction of concentrated sulfuric acid
to the weaker acid from the first stage. The process is related to
fertilizer technology as shown in Fig. 7.
The essential differences among the methods developed depends on
the processing of sulfuric acid containing HF [66,67]. To increase the
amount of HF evolved from sulfuric acid, an inert condensable sweep
gas is selected from the group consisting of paraffinic hydrocarbons,
steam, and SO3 [65] and from blowing sweep gases such as air,
oxygen, nitrogen, and argon [68,69]. In this process, nonreactive
stripping agents such as hexane are used for better HF separation [66].
It is proposed that the tetrasilica fluoride removed in the first stage be
recycled to the SiF4 absorption stage [7072]. Fluorosilicic acid
decomposition and simultaneous desorption of fluorine gases are
endothemic processes that consume a large amount of heat energy.
The introduction of concentrated fluorosilicic acid (40 to 50%
H2SiF6) at this stage improves the heat and water balance. The
decreased amount of water in the recycling streams is convenient for
process distillation and the reversible reaction

depending on temperature and the water content in the solution

[72,73].
The hydrated silica precipitate formed in this process is filtered off or
centrifuged, and the solution is recycled to the first stage. It is possible
to produce a special grade of silica in the process in which H2SiF6 is
used for HF production. During distillation of the feed solution any
excess of H2O and HCl is removed without formation of SiO2 until
 Page 323

Figure 6
Separation of hydrofluoric acid and silicon tetrafluoride. (From Ref. 65.)
anazeotropic solution containing about 36% H2SiF6, 10% HF, and
54% H2O is reached. This azeotrope is suitable for making fumed
SiO2 [74,75]. This solution is burned by the combustion of fuel and
oxygen to make fumed, commercially desirable SiO2 and HF.
Modification of the fluorosilicic acid decomposition process in a
solution of strong sulfuric acid is used for pure SiF production, based
on the reaction
 Figure 7
Hydrofluoric acid production from waste fluorosilicic acid. (From Ref. 66.)
 Page 324

The continuous manufacture of SiF4 is possible when a H2SiF6

solution is fed to an intermediate point of the vertical column and
sulfuric acid of concentration higher than 80% H2SO4 is fed to the top
of the column. Fluorosilicic acid is fed to silica in suspended and
dissolved form [42].
2
Acidulation of Synthetic Fluorspar
Acidulation of synthetic fluorspar can be realized in a manner similar
to processes that use natural fluorite as the raw material. The U.S.
Bureau of Mines has proposed a process in which HF and a small
amount of cryolite are produced [76]. Most of the technical problems
in synthetic fluorspar production involve the removal of silica and
other impurities.
3
Pyrohydrolithic Reaction
The method involving pyrohydrolithic reaction is based on a
hydrothermal process in which synthetic fluorspar reacts with water
vapor and silica according to the reaction

4
Fluorosilicic Acid Thermal Decomposition
The process developed by Buss Ltd. (Swiss) depends on H2SiF6
thermal decomposition and pure hydrogen fluoride separation using a
solvent absorption method to recover HF from the silica by-product
[72,76]. The improved solubility of HF in an organic solvent such as
polyether can be used in the process of evaporation of fluorosilicic
acid [33].
5
Thermal Decomposition of Salted-Out Fluoroaluminum Salts from
Crude Wet-Process Phosphoric Acid
In the precipitated salt Falphite, which is the calcium-free compound
AlFHPO4·2H2O, (FeAl)HPO4F·2H2O is thermically decomposed to
pure HF and water vapor and aluminum phosphate. This salt can be
precipitated from impure wet-process phosphoric acid containing a
large quantity of aluminum during aging at an elevated temperature
[77].
6
Decomposition of Ammonium Fluoride
Ammonium fluoride solution obtained from fluorosilicic acid and
ammonia is decomposed by sulfuric acid in the processes

This Kemira Oy process [17,78] is utilized in the manufacture of

hydrofluoric acid containing 0.1 to 0.2% H2O. The by-product
NH4HSO4 and H2SO4 solution containing about 0.1% HF is recycled
to the fertilizer process.
The hydrogen ammonium salt NH4HF2 may be oxidized to produce
HF [79] according to the reaction

The other proposed process [80,81] utilizes a precipitate containing

Na2SiF6 as raw material. In the process
 Page 325

a concentrated solution of NH4F with big crystals of NaF and silica is

obtained [81]. In another process, K2SiF6 is used as the raw material
[80]. These processes [7981] are possible only in theory.
7
Decomposition of Sodium Fluoride
The properties of the system NaFNaHF2H2O are utilized in HF
manufacture. According to Imperial Chemical Industries Ltd. [82], HF
contained in gaseous streams may be absorbed on dry sodium
fluoride, forming sodium bifluoride. This salt is next decomposed
thermally for NaF regeneration and evolving of pure, dry hydrofluoric
acid. The ICI process uses a special absorption system consisting of a
suspension of NaF in high-boiling inert organic liquids. A modified
process [83] of HF production from waste fluorosilicic acid in the first
stage comprises evaporation, followed by combustion of these vapors
to obtain HF, H2O, and fumed SiO2. After removal of the SiO2, the
gas is passed through a fluidized bed containing subdivided NaF. The
NaHF2 formed is decomposed to NaF, which is recycled and HF is
produced.
E
Na3AlF6:
Synthetic Cryolite
The natural cryolite resources have been depleted. Synthetic cryolite
of a high degree of purity produced from waste fluorosilicic acid is
suitable for use in the aluminum industry [33,8486]. Several routes to
synthetic cryolite production in an integrated fluorine recovery system
are in practical use.
1
TVA Process for Ammonium Cryolite and Ammonium Fluoride
This process consists of an integrated absorption system in a fertilizer
plant in which fluorine compounds are absorbed in scrubbers in which
ammonium solution is used as the scrubbing liquid [84]. After silica
precipitation in the second stage by action of gaseous ammonia and
separation of this precipitate,

part of the solution is recycled to the scrubbing system and sulfuric

acid and aluminum sulfate are added to a second part of the solution.
The process occurs according to the reaction

The precipitated (NH4)2SiF6 is dried and may be used in this form or

converted to Na3AlF6 or AlF by addition of Al(OH)3 and heated in
the kiln, recycling ammonia to the absorption stage.
A Romanian modification [17] of this process includes preliminary
removal of phosphate compounds and involves a hydrolysis stage in
which ammonia is used. The hydrated silica that is precipitated is
removed, the NH4F solution is partially recycled to the absorption
process and to the filtrate product of Al(OH)3, and an NaOH reaction
is added. Gaseous ammonia is recycled to the process of P2O5
removal, and the Na3AlF6 final product is filtered off and dried.
 Page 326
2
Chemie Linz Process, Including an Indirect NaF Production Stage
The Chemie Linz process comprises the reaction fluorosilicic acid
with soda according to the reaction

Part of the H2SiF6 reacts with Al(OH)3:

and in the last stage,

The product precipitated is filtered off using drum filters and is

calcinated [33,39,85,86]. In a Romanian modification of this process
[17,85], after silica has been removed by centrifuging, the final
product is produced by adding alumina, Al2O3, soda, and carbon
dioxide to the sodium fluoride solution.
3
Integrated System for Simultaneous Manufacture of Synthetic
Cryolite and Other Fluorides Using H2SiF6 Neutralization by NH3
One of the most interesting methods of fluorine compound recovery
was developed at the Dungemittel Kombinat VEB Stickstoffwerk
Piesterliz. This practical method makes possible the production of
Na2SiF6, NaF, and Na3AlF6 [85,87]. In the process NH4F formed by
the neutralization of H2SiF6 as with NH3 is converted to cryolite by
reaction with NaOH and then with AlF3.
4
Integrated Complex System for Manufacture of Cryolite
Simultaneously with Other Fluorine Compounds, Based on Alkaline
Fluorosilicates Precipitated from Wet-Process Phosphoric Acid
A method developed by Fertilizer & Chemicals Travancore Ltd.
(FACT) is capable of producing Na2SiF6, K2SiF6, Na3AlF6, AlF3,
NaF, and hydrated silica as commercially salable products. This
technology comprises the ammoniation of Na2SiF6 precipitate to yield
a mixture of NaF and NH4F. After the addition of aluminum salt to
this mixture, NaF precipitate is separated and AlF3 containing filtrate
is processed to Na3AlF6 by treatment with sodium chloride. The
sodium fluorosilicate used as the starting material is precipitated from
crude wet-process phosphoric acid by treatment with either trisodium
phosphate or potassium carbonate.
5
Kaiser Aluminum Process Using Indirect Na2SiF6 Production
A process for producing cryolite [85,88] comprises a reaction between
H2SiF6 or NaSiF previously prepared and a solution of NaCO
according to the reactions

After silica removal, the final product is obtained in the reaction

 Page 327
6
Cryolite from Hydrofluoric Acid
In Russian chemical works, synthetic cryolite production is based on
hydrofluoric acid. This method uses the reaction

Precipitated cryolite is removed and NaF (as impurity)-containing

solution is transferred to the cryolite according to the reaction

This method can be used for utilization of an aqueous HF solution

obtained from waste fluorine compound in fertilizer manufacture
[15,23].
F
Aluminum Fluoride
Aluminum fluoride and cryolite are used together as a flux in the
aluminum industry. Traditionally, AlF3 is obtained from hydrogen
fluoride and aluminum hydroxide. Of course, as a raw material HF is
obtained from fluorine solutions produced in fertilizer manufacture,
but only low levels of P2O5 and SO4 are required. Many other
methods of AlF3 production have also been developed and used.
1
Using H2SiF6 Waste Solutions and Al(OH)3
In this process [33,39] fluorosilicic acid reacts with aluminum
hydroxide at the boiling points with the reaction

This process depends on the possibility of metastable silicic acid

removal before trihydrate aluminum precipitation. This process occurs
slowly in a batch crystallyzer. The separated crystals are calcined at
550°C, with AlF3 obtained as the final product. This is a difficult
process because there is a danger that the hydrolytical fluoride will
decompose to Al2O3 and HF. This process has been used at the
Osterreichische Stickstoffwerke in Linz and other works [89]. For
better energy and material consumption efficiency, this process has
been modified by the Romanian firms Icechim and Iprochim. In this
process the filtered-off mother liquid containing AlF3 is treated by
Al(OH)3, and cryolite is produced after adding soda to the reaction
mixture [17].
A similar method of processing H2SiF6 has been proposed [90],
although another aluminum salt (e.g., AlCl3) is used as a raw material:

The next stages are the same as those in the OSW Linz process or
similar processes using sophisticated systems for metastable silica
removal and processing of AlF3·3H2O, and connection with fertilizer
manufacture [91].
Research has also been carried out on a process for monohydrate
aluminum fluoride production [92]. The essential feature of this
process is removal of the phosphate compounds in the form of FePO4
and (FeF2)2SiF6 and crystallization of AlF3·H2O under autoclave
conditions.
 Page 328
2
AlF3 Production by Ammonia Treatment of Fluorosilicic Acid
This method of AlF3 production is based on the basic principle of the
TVA process: absorption of fluorine-containing gases by an ammonia
solution, with indirect production of ammonium cryolite [38,84],
leading to the following sequence of reactions:

G
Fluorosilicates:
Na2SiF6 and K2SiF6
Fluorosilicates of alkali and alkaline earth metals are characterized by
low solubility. Therefore, most methods utilize the following
processes:
1. Methods of fluorosilicic acid neutralization by hydroxides, oxides,
or carbonates or reaction with chlorides or sulfates.
2. Alkaline absorption of fluorine-containing gases. This method is
inconvenient because it leads to the production of weak solutions and
the precipitation of salts in pipes and tanks.
3. Defluorination of wet-process phosphoric acid by sodium
phosphate or another alkaline and further processing of fluorosilicates
[9399].
4. Precipitation of Na2SiF6 and K2SiF6 directly from crude
phosphoric acid solution produced by a wet method from low-grade
phosphate raw material with a high magnesium content
[15,23,85,100].
5. Precipitation of fluorosilicates during nitrophosphate processing of
phosphate rock directly from acidulation solution or by addition of
sodium salts from a calcium nitrate solution [23,35,101].
The principal problems with these methods deal with purification of
the salts obtained, removal of calcium sulfate and other admixtures,
and crystallization of these salts in a filterable form. Many
modifications of these methods are used in practice by FACT, Fisons
Ltd., Wellman Incandescent Furnace (UK), and the Chemical Works
Police (Poland). The possibility of direct use of BaSiF6 and MgSiF6 in
a scrubbing system absorbing effluent fluorine gases has been
reported [85].
H
Fluorides:
NaF, KF, and NH4F
Most methods developed for fluoride production are components of
technological lines in the manufacture of synthetic cryolite and
aluminum fluoride. Fluorides may be produced by:
1. Neutralization of waste fluorosilicic acid with oxides, hydroxides,
and basic salts [84,85,102].
2. Hydrolysis of fluorosilicates by hydroxides, oxides, and carbonates
[33,85,87].
3. Reaction of fluorspar with soda. This process may be used for
utilization of waste synthetic fluorspar in the reaction
 Page 329
For higher yields in NaF manufacture, a 70 to 75% CaF2 content in
the raw material is desirable, and thermal agglomeration increases the
efficiency of the process. For better NaF extraction, amorphous silica
should be used rather than quartz. This additive makes it possible to
obtain dry agglomerate in a rotary kiln without unfavorable kiln rings.
The aqueous NaF solution extracted using NaAlO2 can, in turn, be
used in the manufacture of synthetic cryolite [15,23].
4. Reaction of hydrogen fluoride with ammonia, soda, and alkaline
carbonates [15,23].
V
Ecological Aspects of the Utilization of Fluorine-Containing
Materials
Recovery of fluorine from fertilizer plants and utilization of the by-
product for the manufacture of commercial salable fluoride products
is an integral part of modern fertilizer manufacturing. Monitoring
systems now report on gaseous effluent concentration, making it
possible to avoid health hazards to human beings and animals and
harmful effects to plant life. Indian data [103] from plants that do not
recover fluoride compounds show that about 0.6 to 0.7 kg of fluorine
is emitted during production of 1 metric ton of single superphosphate,
which is about 8 to 10 times higher than in plants with a fluorine
recovery system.
Ecological aspects connected with fluorine compound manufacture
and use are multidirectional and partially unpredictable
environmentally. An example of such an unexpected effect is the
destruction of the ozone layer in the upper atmosphere by
chlorofluorocarbons. The fluorine form in the by-products and wastes
produced by the fertilizer industry is not as dangerous, because
ecological equilibrium in nature causes a low level of fluorine in
water and soil in the soluble (available) form. The influence of the
fertilizer industry must be considered in the work area and
environment near the factory [3]. The total environmental fluorine
from the fertilizer industry is but a small amount of the total world
emission. According to one report 115], in the United States 39% of
soluble fluorine compounds were emitted by the steel industry, 16%
from coal-based energy power stations, 13% from the fertilizer
industry, 9.8% from aluminum production, 5.8% in brick, title, and
ceramic manufacture, and 5.3% from HF production. In Canada, 56%
of fluorine emissions came from the aluminum industry, about 17%
from the fertilizer industry, and 15.5% from the steel industry.
The worst conditions for workers exist in the aluminum industry,
despite continuing attempts to control fluorine emission. During coal
combustion 34 to 72% of the fluorine contained in coal (0.0025 to
0.039% F) is passed to the fumes and gaseous effluents and causes
more harmful environmental effects than do emissions from the
fertilizer industry, particularly in countries whose energy systems are
based on bituminous coal or brown coal (Eastern Europe).
Near large fertilizer factories fluorine concentrations are 10 to 100
times higher than in other areas (0.00017 to 0.00023 mg F/m, to a
maximum of 0.005 mgF/m) [104]. The fluorine compounds emitted
are scrubbed by rains and mist and thus are constantly being removed
from the atmosphere. Plant leaves are damaged by fluorine
concentrations higher than 0.003%. Extensive studies have been made
of the health of workers in fertilizer plants exposed for 25-year
periods. No indication of fluorosis was found in schoolchildren
residing in an area near a fertilizer plant. It has been reported by the
U.S. Department of Commerce that gross pathology changes found in
aluminum industry workers were rarely encountered among fertilizer
workers [3].
 Page 330
Medical reports describing the effects connected with long-term
exposure of workers to fluorine compounds in fertilizer plants showed
that fluorine concentrations were generally less than 10 mg F/m (in
cryolite plants, 30 to 40 mg F/m). Chronic effects are observed, such
as changes in or damage to the skeleton in the form of permanent
crippling of skeleton rigidity, stiffness, and rheumatism [8]. Fluorine
uptake into the lung by occasional emissions leads to irritation, but
constant exposure causes fluorspar formation in the form of symptoms
of mild osteosclerosis. According to Donnel [7], men who have
ingested 20 to 80 mg of fluorine daily for 10 to 20 years in plants
processing fluorine-containing materials even up to 13,000 ppm
fluorine content in bones have been indicated. These changes are
observed more frequently among aluminum industry workers than in
the fertilizer industry. Periodic x-rays detection and urinalysis can be
used to monitor illness.
Acute effects may occur accidentally, with such symptoms as nausea,
vomiting, diarrhea, and acute abdominal pains, when inorganic
soluble fluorides are taken orally. Sometimes, indigestion of only 150
mg of NaF, for example, can cause serious illness, although it is easy
to eliminate all effects by intravenous and intramuscular calcium
therapy [9].
Workers react to various fluorine concentrations in air as follows
[103]:
2 to 4 mg No immediate effects
F/m
8 mg Many persons experience discomfort
F/m
24 mg All persons complain and object to staying in the
F/m environment
48 mg Brief exposures cause irritation of eyes, nose, pharynx,
F/m and upper chest
96 mg Tolerable for no more than 1 minute; all the effects above
F/m intensified; smarting of skin

The most serious problems are experienced by those using

concentrated fluorides or hydrogen fluoride. Moisture in the skin will
cause hydrolysis to HF, causing eye or skin burns, although pain is not
felt immediately and damage is unexpected. Use of a previously
prepared paste of magnesium oxide in glycerol is good medicine for
accidental contact with these agents [6].
The last Soviet medical research pertaining to the production of the
phosphate fertilizers and fluorine recovery plants located in these
works detected no cases of occupational fluorosis. The total fluorine
load was calculated in workers of many professions, and in the
fertilizer workers at an annual intake of 2 to 4 g of fluorine,
corresponding to a mean shift concentration of 1 to 2 mg/m.
VI
Future Fluorine Recovery from Phosphate Rock
Fluorine recovered from phosphate rock is presently competitive
economically with fluorine mined in the form of fluorspar. The
increase in fluorine recovery efficiency is in accord with increasing
environmental responsibility. This effect may be achieved by
technological modifications of the principal fertilizer production lines,
more efficient scrubbing systems, recycling of liquid streams, and
better connection of utilization units where commercially available
fluorine products are manufactured using an absorption unit. Full
utilization of fluorine compounds recovered depends on the form and
purity of the final products. The higher quality of these products can
be obtained by special technical and technological advances in
scrubbing systems and the use of a preliminary purification
 Page 331
step for cleaning the liquids absorbed. A second, much more common
process relies on a two-stage technology in which in the first stage,
pure hydrogen fluoride is produced and high-purity fluorine
compounds are manufactured. Another important utilization trend
involves the production of synthetic fluorspar for direct use instead of
natural fluorspar in steel metallurgy or decomposed by sulfuric acid to
obtain pure hydrogen fluoride.
The future of fluorine recovery processes also depends on the demand
for AlF3 and Na3AlF6 in the aluminum industry. A lowered fluorine
consumption of 1 ton of aluminum produced has been observed. In the
United States, Japan, and Western Europe new technological
modifications are being used which feature recycling and a more
efficient absorption system.
Fluorine compounds are the basic flux agents in the electrolytic
reduction process in aluminum smelters. Flux losses are made up in
the form of 0.03 to 0.04 ton of AlF or 0.03 to 0.05 ton of Na3AlF6 for
every ton of aluminum produced. These unit consumption standards
are tending to decrease as a result of higher cell efficiency and
increased reclamation of cryolite from pot linings and the introduction
of Li2CO3 to the flux. This method has made it possible to reduce
fluorine emissions some 25 to 50% and to lower energy consumption.
The combination of the methods discussed above has increased
fluorine consumption 2-to 2.5-fold. Pilot-plant demonstrations by
Alcoa (the Aluminum Corporation of America) have shown that using
AlCl3 as flux can also decrease AlF3 and Na3AlF6 consumption.
Nevertheless, the demand for fluorine flux is constant because of
expansion in the aluminum industry. A similar tendency is observed in
the steel industry, which is looking for a safe fluorine substitute. In the
past decade the average consumption of fluorspar per 1 ton of steel is
below 2 kg of CaF2 per ton of steel produced. In 1970, unit
consumption was 5 to 6 kg of CaF2 per ton of steel. This effect is the
result of a new type of flux based on such natural additives as borax,
colemanite, illmenite, dolomite, manganese, and ferrous manganese.
A similar effect was achieved when fluorspar was formed into
briquettes containing 20 to 25% and other additives. The same
tendency is observed in welding technology; for example,
Sogerem/Sacme (France) proposes using a mixture of fluorspar and
selaite. Of course, the modifications described above are uninfluenced
by total fluorine compounds consumption. All these reductions in
consumption by traditional fluorine-consuming industries are
compensated by increased fluorine consumption, primarily for plastics
and the manufacture of specialty chemicals.
The largest suppliers of fluorine compounds, in association with the
aluminum industry, are interested in processing waste fluorine by-
products from fertilizer plants (e.g., Kaiser Aluminum Chemical Co.,
Alcoa, Montecatini, Showa Denko, Nihon Rinsan KK, Mitsubishi,
Hiryo KK, Central Glass, Onoda Industry, Werke OSW, Chimprom,
Comprimo NV of the Netherlands). Many producers that formerly
manufactured synthetic cryolite and aluminum fluoride based on
natural fluorspar are building plants that consume waste fluorine
compounds from fertilizer plants. It is difficult to estimate the share of
fluorine derived from phosphate rock in total fluorine production, but
based on many statistical reports from the fertilizer, aluminum, and
steel industries, it is possible to estimate that about 55% of the
fluorine now produced is recovered from phosphate plants.
References
1. L. B. Nelson, History of the U.S. fertilizer industry, in Plant
Environmental Problems, J. H. Parker, ed., Tennessee Valley
Authority, Muscle Shoals, Ala., 1990, pp. 399, 491.
 Page 332
2. P. Becker, Phosphates and Phosphoric Acid, Fertilizer Science and
Technology Series, Marcel Dekker, New York, 1983.
3. R. C. Specht and R. R. Calaceto, Gaseous fluoride emissions from
stationary sources, Chemical Engineering Progress, Vol. 63, 1967, p.
78.
4. G. H. Collins, Commercial Fertilizers: Their Sources and Uses,
McGraw-Hill, New York, 1955, p. 215.
5. E. Sepehri-Nik, The manufacture of calcium phosphate,
Phosphorus and Potassium, No. 149, 1987, pp. 2829.
6. A. V. Slack, ed., Phosphoric Acid, Marcel Dekker, New York,
1968, pp. 765778.
7. T. A. O. Donnel, Biological activity of fluorides, in The Chemistry
of Fluorine, Pergamon Press, Melbourne Fla., 1975, p. 1028.
8. B. C. Saunders, The physiological action of organic compounds
containing fluorine, Chemistry of Fluorine, Vol. 6, 1965, pp. 183210.
9. R. D. Klebanov and A. T. Sidenko, On occupational fluorosis-
related hazards in phosphorus-containing manufacture, Gigina Truda,
Professional Zabol (Russian), Vol. 9, 1989, pp. 1719.
10. D. A. Wright and A. W. Davison, The accumulation of fluoride by
marine and in terrestial animals, Environmental Pollution, Vol. 8,
1975, p. 1.
11. G. C. Finger, Fluorine Resources and Fluorine Utilization,
Butterworth, London, 1968, pp. 3556.
12. Bureau of Mines, Phosphate Rock Availability, World Bureau of
Mines Information Circular IC 8989, 1984.
13. A. Morawinski, Inorganic Chemical Mineral Resources and Their
Chemical Utilization, Wydawnictwo Geologiczne, Warsaw, 1975.
14. Fluorine, in Encyclopedia of Chemical Technology, ed. Kirk and
Othmer, J. Wiley & Sons Vol. 10, 3rd ed. 1980, p. 630.
15. W. A. Zaicew, Manufacture of Fluoride Compounds During
Fertilizer Production, Chimia, Moscow, 1982, p. 248.
16. S. Pawlikowski, Fluorine World Mineral Resources Ba, B, F, Sr,
Wydawnictwo Geologiczne, Warsaw, 1979, p. 197.
17. N. V. Dimcev, N. N. Popovici, and P. Petreanu, Phosphate rocks
used in the fertilizer industry, Iprochim, Bucharest.
18. F. J. Kopytov and A. A. Kaplan, Fluorine deposits, resources and
production, Proceedings of the First Conference on Fluorspar
Production, Moscow, 1974, pp. 46.
19. A. A. Ivanova, Prognosis of the Regional Resources in Geological
Research, Mingeo, Moscow, 1973, pp. 162167.
20. A. Bolewski, H. Gruszczyk, and S. Pawlikowski, Processing of
the fluorine raw materials, in World Mineral Resources Ba, B, F, Sr,
Wydawnictwo Geologiczne, Warsaw, 1979, p. 290.
21. B. Bak and B. Radwanek, Phosphorus P2O5 World Mineral
Resources, Wyd. Geol., Warsaw, 1987, p. 20.
22. Phosphate rock statistical supplement, Phosphorus and Potassium,
No. 176, 1991, p. 41.
23. W. F. Karmysov, Recovery and utilization of fluorine compounds,
in Chemical Phosphate Rocks Processing, Chimia, Moscow, 1983, pp.
276296.
24. R. Schmutzler, Fluorides of phosphorus, Advances in Fluorine
Chemistry, Vol. 5, Butterworth, London, 1965, pp. 3171.
25. U. Glabisz, H. Gabryel, A. Kacalski, and B. Grzmil, Utilization of
fluorine by-products from fertilizer plants (Polish), Przemysl
Chemiczny, Vol. 68, 1989, p. 20.
26. G. T. Austin, Shreve's Chemical Process Industry, McGraw-Hill,
New York, 1984, pp. 352355, 668679.
27. R. Filler, Trifluoromethyl aromatic compounds, in Advances in
Fluorine Chemistry, Vol. 6, Butterworth, London, 1968 pp. 142.
28. S. Kumar and S. R. Palit, Fluoro-alcohols, in Advances in
Fluorine Chemistry, Vol. 6, Butterworth, London, 1968, pp. 6979.
29. Fertilizer Manual, International Fertilizer Development Center,
Muscle Shoals, Ala., 1979.
 Page 333
30. S. D. Zviencikc and A. A. Brodski, Technology of Phosphate and
Multicomponent Fertilizers, Chimia, Moscow, 1987, pp. 306375.
31. I. P. Muchlenow and O. C. Kovalev, Absorption and Dust
Collection in Fertilizer Plants, Chimia, Moscow, 1987, pp. 7205.
32. J. Bidder and J. A. Hallsworth, Two processes recover
fluorosilicic acid in useful concentration, Phosphorus and Potassium,
No. 73, 1974, pp. 3539.
33. H. J. Denziger, H. J. Konig, and G. E. W. Kruger, Fluorine
recovery in the fertilizer industry: a review, Phosphorus and
Potassium, No. 103, 1979, pp. 3339.
34. British Sulphur Corp., Advances in gas scrubbing: the recovery of
fluorine in fertilizer manufacture, Phosphorus and Potassium, No.
110, 1980, pp. 2730.
35. F. Nielsen, ed., Manual of Fertilizer Processing, Marcel Dekker,
New York, 1987.
36. N. F. Chrupow, W. A. Kopylew, and T. G. Riepkova,
Defluorination of the wet-process phosphoric acid, Zhurnal
Prikladnoi Khimii. (Russian), Vol. 7, 1977, p. 1617.
37. A. W. Frazier and Y. K. Kim, Redistribution of Impurities in Wet-
Process Phosphoric Acid, Bulletin TVA INFDC, Y-202, 1988, pp.
124.
38. H. L. Antoskina, Utilization of the wastes from phosphate plants,
Chimie Prom. (Russian), Vol. 4, 1979, pp. 5558.
39. J. Massone, Stand und Entwicklung der industriellen Fluorchemie,
Chemiker Zeitung, Vol. 96, 1972, pp. 6575.
40. M. English, (Mitchell Ltd.), Purification of vapours containing
fluoride compounds, Brit. Patent 1,179,366, 1970.
41. British Sulfur Co., Phosphoric acid plant revamps, Phosphorus
and Potassium, No. 170, 1990, pp. 3032.
42. J. S. Tawar (Allied Co.), Continuous production of silicon
tetrafluoride gas in a vertical column, U.S. Patent 4,470,959, 1984.
43. H. O. Crawford (Fisons Ltd.), Improvements in or relating to the
treatment of gases, Brit. Patent 1,453,830, 1976.
44. M. English (Mitchell Ltd.), Purification of fluorosilicic acid, Brit.
Patent 1,243,802, 1971.
45. W. P. Parish (Wellman), Concentrating fluorosilicic acid. U.S.
Patent 3,645,679, 1972.
46. W. P. Parish (Wellman), Concentrating fluorosilicic acid, U.S.
Patent 3,645,680, 1972.
47. H. Goller, Method of producing fluorosilicic acid of low
phosphorus content from waste gases, U.S. Patent 3640 685, 1972.
48. G. E. Kidde, Removal of phosphate compounds from fluoride
containing solution, U.S. Patent 3,870,786, 1975.
49. D. H. Reeve (IMC Co.), Preparation of fluorine compounds, U.S.
Patent 2 945 745 1960.
50. Central Glass Co., Process for purification of sodium
silicofluoride, U.S. Patent 1,246,020, 1968.
51. S. Szpunt, Manufacture of wet-process phosphoric acid by sulfuric
acid decomposition process using phosphate rock with magnesium,
Chimie Prom. (Russian), Vol. 41, 1965, p. 674.
52. S. Szpunt, Fluorosilicic acid for phosphate rock processing,
Chimie Prom. (Russian), Vol. 37, 1961, p. 714.
53. S. Szpunt, Manufacture of phosphate fertilizers, Chimie Prom.
(Russian), Vol. 10, 1955, p. 674.
54. R. Monaldi, Process for recyling MgSiF6 solution recovered by
gas washing in the superphosphate den, in New Developments in
Phosphate Fertilizer Technology, Elsevier, Amsterdam, 1977, pp. 110.
55. M. C. Griem and J. Rust, High efficiency fluoride removal for
recycling in the fertilizer industry, Technical Literature Ceilcote
GmOH, Beisbeheim, Germany, 1983.
56. A. G. Cherleston, Mass and energy balances in a fluoride scrubber
when the fluoride is returned to SSP, Fertilizer Research Netherlands,
Vol. 12, 1984, pp. 6784.
57. V. V. Rao, Control of fluorine emissions in SSP plants, Fertiliser
News, Vol. 33, 1988, pp. 4952.
 Page 334
58. W. C. Warneke, Recovery of granular fluorides from phosphate
rock by pseudomorphic replacement, Proceedings of the 2nd
International Congress on Phosphorus Compounds, Boston, 1980, pp.
819825.
59. C. A. Hollingsworth, Counter-current process for passing a solid
particulate reactant to a liquid reactant, U.S. Patent 4,098,873, 1978.
60. W. C. Warneke, Direct production of coarse particle inorganic
fluorides, U.S. Patent 4,120,940, 1978.
61. R. Pandey, Disposal of wastes in phosphatic fertilizer plant,
Technology, Vol. 3, 1965, pp. 6162.
62. T. A. Graceva, A. D. Dmitrievskij, and B. I. Szrambin,
Introduction of floculants to a fluorine-containing solution,
Udobrienia Doklady (Russian), Vol. 42, 1975, p. 38.
63. British Sulphur Corp, CaF2 manufacture from waste fluorine
values, Phosphorus and Potassium, No. 107, 1980, pp. 3638.
64. N. B. Vidienov, I. N. Gryncarov, and E. Pelovski, Chemical
Utilization of the Inorganic Chemistry (Bulgarian), ZKOZHOOPM,
Sofia, Bulgaria, 1989, pp. 137145.
65. J. E. Barker and J. W. Markey (Tennessee Corp., N.Y.), Separation
of hydrogen fluoride and silicon tetrafluoride, U.S. Patent 3,218,129,
1965.
66. A. C. Mohr and R. P. Obrecht (Stauffer Chemical Co.), Process for
recovering strong HF from phosphate rock digestion processes, U.S.
Patent 3,257,167, 1966.
67. L. C. Oakley and T. T. Houston (Tennessee Corp., N.Y.), Process
of producing hydrogen fluoride as a dry gas from clear fluorosilicic
acidcontaining solution, U.S. Patent 3,218,124, 1965.
68. T. T. Houston (Tennessee Corp., N.Y.), Process of producing
hydrogen fluoride in a two-stage procedure and effecting rapid
evolution of the hydrogen fluoride by sweeping the second stage with
an inert gas, U.S. Patent 3,218,127, 1965.
69. F. J. Klemm (Tennessee Corp., N.Y.), Process of producing
hydrogen fluoride in a two-stage procedure, U.S. Patent 3,218,128,
1965.
70. B. Zawadzki, A. Buliska, R. Loski, and Z. Brzostowski,
(Chemical Works, Lubón), Method of production of anhydrous
hydrogen fluoride, U.S. Patent 4,062,930, 1977.
71. B. Zawadzki, A. Buliska, Z. Szulc, R. Loski, and Z. Brzostowski
(Chemical Works Lubón), Pol. Patent 82,438, 1977.
72. B. Zawadzki, Hydrofluoric acid from waste fluorosilicic acid
(Polish), Proceedings of the National Conference on Fluorine
Compounds, Brzózki, Poland, 1980, p. 65.
73. N. I. Miszin, W. I. Rodin, W. A. Zajcev, W. I. Komarov, and E. C.
Zitova, Study of waste-free method of HF manufacture based on
decomposition of H2SiF6 by sulfuric acid with full semi-product
utilization, Trudy Institute Moskovskii imeni Mendeleeva (Russian),
Vol. 109, 1974, pp. 2030.
74. G. L. Flemmert, Process for converting silicon and fluorine
containing waste gases into silicon dioxide and hydrogen fluoride,
hydrogen fluoride, U.S. Patent 3,969,485, 1976.
75. J. F. Leathers and D. Calvin, Method for distilling a ternary
azeotrope of fluorosilicic acid, hydrogen fluoride, water, U.S. Patent
4,008,130, 1977.
76. British Sulphur Corp., Fluorine recovery, Phosphorus and
Potassium, No. 65, 1973, p. 44.
77. E. M. Chemtob and G. L. Beer, Process for production of HF from
impure phosphoric acid, U.S. Patent 4,505,884, 1985.
78. F. Redies, B. Redies, and D. Turk (Kemira Oy), Verfahren zur
Gewinnung von Fluorwasserstoff ans einer wabrigen Loung ans
Fluorwasserstoff und Schwefelsaure, Ger. Patent 2,757,228, 1978.
79. R. N. Secord and C. R. Carpentier (Cabot Co.), Process for
recovering hydrogen fluoride from aqueous fluorosilicic acid solution,
U.S. Patent 3,128,152, 1964.
80. W. Augustyn, K. Dubik, A. Kossuth, and M. Oskedra, HF from
H2SiF6 based on the alloy KHF2NH4HF2NH4F and their
decomposition with H2SO4, Przemysl Chemiczny (Polish), Vol. 53,
1974, pp. 48100.
 Page 335
81. W. Augustyn and J. Chmiel-Pela, The method of recovery of HF
from phosphate rock based on NH4HF2 as semiproduct, Przemysl
Chemiczny (Polish), Vol. 53, 1974, p. 40.
82. British Sulphur Corp., ICI process recovers anhydrous HF from
wet gas streams, Phosphorus and Potassium, No. 68, 1973, p. 50.
83. R. G. Harting (Gulf Design Development), Recovery of hydrogen
fluoride, U.S. Patent 3,711,596, 1973.
84. J. C. Barber and T. D. Farr, Fluoride recovery from phosphorus
production, Chemical Engineering Progress, Vol. 66, 1970, pp. 5662.
85. British Sulphur Corp., Fluoride recovery by the phosphate
industry. Part II. Commercially available processes for recovering
fluorides, Phosphorus and Potassium, No. 54, 1971, pp. 2129.
86. British Sulphur Corp., Fluorine recovery by the phosphate
industry. Part I. Wider interest in conversion to salable fluorides,
Phosphorus and Potassium, Vol. 53, 1971, pp. 2325.
87. British Sulphur Corp. DDR fluorine recovery system available for
licence, Phosphorus and Potassium, Vol. 49 1970, p. 25.
88. A. C. Byrns and C. Kaiser, Al production of cryolite, U.S. Patent
2,994,582, 1961.
89. R. Richter, The production of aluminium fluoride from waste
gases of phosphoric acid plants, Proceedings of the ISMA Conference,
Prague, 1974, pp. 1720.
90. R. B. Derr, Production of aluminum fluoride, U.S. Patent
3,175,882, 1961.
91. D. C. Gernes and W. King (Kaiser Aluminum), Producing
aluminum fluoride, U.S. Patent 3,057,681, 1962.
92. M. Grobelny, R. Rozycka, and J. Chmiel, A monohydrate process
for manufacture of aluminium fluoride from fluorosilicic acid,
Phosphorus and Potassium, Vol. 100, 1979, pp. 2124.
93. W. A. Hodge (Swift), Production of sodium silicofluoride and
fluorine-free phosphoric acid, U.S. Patent 2,883,266, 1956.
94. E. N. Case (Atlantic Richfield Co.), Process for the recovery in
useful form of fluorine values from phosphate rock while producing
phosphoric acid, U.S. Patent 3,743,725, 1973.
95. W. A. Satterwhite and J. M. Carter (United States Steel Co.),
Production of sodium silicofluoride from wet process phosphoric
acid, U.S. Patent 3,554,694, 1971.
96. J. E. Barker and J. H. Robinson (Tennessee Corp.), Process for
producing fluorosilicates, U.S. Patents 3,462,242, 1969, and
3,554,694, 1971.
97. W. A. Zajcev, A. F. Gafarova, B. G. Gromov, and I. D. Spycyn,
Removing of fluorosilicates by precipitation using sodium and
potassium salts, Chimie Prom. (Russian), Vol. 7: 1974, pp. 523526.
98. A. I. Antosikov, J. M. Razvadovski, and W. A. Zajcev,
Improvements in fluorosilicates production from wet-process
phosphoric acid, Chimie Prom. (Russian), Vol 11, 1977, pp. 854859.
99. W. A. Kopylov, T. V. Tomakova, W. M. Archipova, and O. I.
Kruteva, Removing of sodium and potassium fluorosilicates, Chimie
Prom. (Russian), Vol. 1, 1979, pp. 2628.
100. M. H. Cybina, W. F. Karmysow, L. M. Archipova, and O. I.
Kruteva, Removing fluorine from wet-process phosphoric acid from
Chubsuglsk phosphate rock, Trudy NIUIF, Vol. 288, 1976, pp.
118121.
101. W. M. Bieglov, A. H. Magritova, and G. R. Rinberg, The method
of removing fluorine compound from nitric phosphate solution, Sov.
Union Patent 592,750, 1978.
102. G. E. Kidde, Ammonium fluoride process for defluorination of
phosphoric acids and production of ammonium fluorosilicate, U.S.
Patent 4,046,860, 1977.
103. M. R. Agarwal, Control of agseous emissions in fertilizer
industry, Fertilizer News, Vol. 23, 1978, pp. 1420.
104. S. Borowiec, B. Hapanowicz, and Z. Zablocki, Differentiation of
gladiolus leaves necrosis and fluorine content during the years
19771986 on the area affected by emissions from the chemical plants
Police, Agriculture Academy Proceedings, Szczecin (Polish), Vol. 130,
1987, pp. 1620.
 Page 337

24
Introduction to the Potassium Fertilizer Industry
Maurice O. Klein
International Fertilizer Development Center
Muscle Shoals, Alabama

Kenneth W. Reid
Potash Corporation of Saskatchewan Inc.
Saskatoon, Saskatchewan, Canada

Elmar L. Goldsmith
Kalium Canada, Ltd.
Regina, Saskatchewan, Canada
I
Introduction
Potassium (K) is one of the three primary nutrients essential to plant
life. Potassium is vital to protein formation and photosynthesis. It also
activates and controls the reaction rates of enzymes and increases the
tolerance of plants to diseases [1]. Potassium is the seventh most
common element on earth in combination with various compounds
because of its high reactivity and affinity for other elements. However,
potassium is normally of commercial value when found in large,
relatively concentrated deposits of potassium salts that lend
themselves to economies of scale and low-cost mining and
production. Deposits of this nature are large bodies of marine
evaporites and bodies of salty water such as certain landlocked seas,
lakes, and natural brines. Sodium chloride is a major component in all
these deposits. Potassium salts are also known as potash salts; the
word potash traditionally refers to the potassium oxide (K2O) content,
although no oxide is present.
The production of potassium salts in 1991 came from 50 mines
located in 15 countries and was about 46.5 million metric tons or
about 26 million metric tons of K2O. About 90% of this production
was potassium chloride (60 to 62% K2O) or muriate of potash (MOP),
and about 4% was potassium sulfate (50 to 53% K2O) or sulfate of
potash (SOP). About 96% of the overall production was used in
agriculture as potash fertilizers [2].
Muriate of potash is manufactured from sylvite (KCl). Sylvite is the
preferred potassium-based agromineral because (1) it has the highest
K content, almost 63%; (2) it occurs abundantly in nature; and (3) the
process to extract sylvite from hosting ores is relatively simple. As
shown in Table 1, sylvite occurs in sylvinite ores (sylvite + halite +
anhydrite + clay), in carnallite ores (sylvite + halite + carnallite +
clays), in hartsalts (sylvite + kieserite + halite + clays), and in kainite
ores. Sylvite can also be formed synthetically from naturally occurring
surface and subsurface brines that contain dissolved
 Page 338
potassium and chlorides. Sylvinite ores are by far the most commonly
utilized ores in the production of muriate of potash.
Potassium sulfate is the most common substitute for potassium
chloride when potassium fertilizers having a low chloride content are
needed to meet crop or soil requirements. Potassium sulfate is
manufactured mostly from langbeinite (K2SO4·2MgSO4), synthetic
schoenite, and potassium chloride and sulfuric acid.
In this chapter we discuss pollution and its prevention in the
manufacture of potassium chloride from sylvinite ores. The main
environmental concern is the salinization of surface and subsurface
water and of the land. We introduce briefly methods of mining and
refining of sylvinite ores. We then describe the type of effluents
produced during the manufacture of potassium chloride, how these
effluents are processed prior to their release into the environment, the
methods of waste management that have been developed to store these
effluents, and methods for effluent disposal. Also discussed are the
sources of air emissions, the type of dust removal equipment utilized
on these sources, the characteristics of the emissions, the impact of air
emissions on the environment, and the relevant legislation and
regulations for pollution prevention and control for this industry.
In developing this chapter, the potassium chloride industry in
Saskatchewan, Canada, has been taken as a case study for several
reasons: (1) the industry is one of the largest in the world; (2) the
industry encompasses the major methods of mining, beneficiation, and
other processing used globally in the production of potassium
chloride; (3) its mines and its refineries are located in agricultural
areas, and some are close to urban areas; and (4) the operations are
subject to severe and fluctuating climatic conditions. Reference is also
made to methods used by potash industries located elsewhere in the
world. As we shall see, much of the pollution control technology used
in the manufacture of potassium chloride in Saskatchewan applies
generally among all potash producers. However, the degree of
employing and tailoring this technology to meet specific needs
depends on the local geological, climatic, and sociopolitical situation.
II
Mining and Refining Methods Utilized in KCl Production
Saskatchewan potash mines are located in the south central and
southeastern portion of the province in the midst of fertile agricultural
land used for cereal crops, particularly
Table 1 Primary Potassium Deposits Utilized in the Manufacture of
Potassium Fertilizers
Deposit Minerals % Deposit location
type K2O
Sylvinite KCl + NaCl + 1040 Canada, Russia, Ukraine, U.S.,
clays U.K., France, Germany, Brazil,
Argentina
Carnallite KCl + MgCl2 1016 Spain, Congo, Ethiopia, Israel,
·6H2O + NaCl + Thailand
clays
Hartsalts KCl + MgSO4 · 1020Germany
H2O + NaCl +
clay
Kainite 4KCl + 4MgSO4 1318Italy
· 11H2O + NaCl
LangbeiniteK2SO4 + 712U.S.
2MgSO4 + NaCl
Dead Sea Brine Israel, Jordan
Great Salt Brine U.S.
Lake
 Figure 1
Simplified schematic of the conventional beneficiation process to manufacture potassium chlo
 Page 340
wheat. The underlying sylvinitic-type ores, which slope southeast, are
considered among the world's largest and richest in potassium. These
deposits contain approximately 30 to 40% KCl, 50 to 60% NaCl, and
1 to 8% insolubles, mostly in the form of clay, silt, sand, and dolomite
[3]. Most of the ores are mined by conventional means, but some are
solution mined. Ore bodies located approximately 1000 m below the
surface are mined mechanically through conventional means. Ores
located significantly beyond the 1000-m depth are solution mined. A
conventional mine lost to accidental flooding is also solution mined
[4]. Typical mechanically mined operations consist of two shafts
connected by a network of tunnels, which develop with mining
activities. One shaft serves for mine air intake, the other for mine air
exhaust. Similarly, one is the personnel shaft, and the other is the shaft
for production or hoisting of the run-of-mine ore.
Solution mining involves the injection of unsaturated brines directly
into the ore body to dissolve the KCl it containes. The impregnated
brine is brought to the surface through a closed circuit and fed into the
refinery for further processing. The insolubles and most of the NaCl
are left behind in the subterranean cavity. For conventional mining, a
typical surface refinery for the run-of-mine ore incorporates three
basic stages, as shown in Fig. 1.
1. The beneficiation stage, including ore crushing and grinding, ore
desliming, separation of the potassium chloride particles from a
mixture of KCl and NaCl (mainly by froth flotation), and tailings
preparation for disposal.
2. The product preparation stage, which includes KCl concentrate
debrining, float drying in either natural gasfired rotary or fluid-bed
dryers, screening, compaction/granulation, product conditioning, and
product storage.
3. The dissolution/crystallization stage, which includes solids
dissolution, a 24 effect vacuum-cooling crystallization circuit, and
product debrining. However, when a separate refined product is
manufactured, this stage also incorporates product drying and storage
and in some refineries compaction/granulation, screening, and product
conditioning. The crystallization requires a heat sink, which is
provided with recycled brine or water/brine from a cooling tower.
A simplified block flowsheet of a solution mining production facility
for potash is illustrated in Fig. 2. The surface refinery consists of the
following basic processing stages:
1. Precipitation of sodium chloride from solution using a quadruple-
effect evaporation circuit, separation and debrining of sodium chloride
precipitate from the evaporator's discharge, and preparation of tails for
their disposal.
2. The potassium chloride production stage, which includes a
multiple-effect vacuum cooling crystallization circuit, product
separation and debrining, drying, screening, compaction/granulation,
product conditioning, and storage.
Potassium chloride may be manufactured on a seasonal basis using
large, relatively shallow ponds and brine from epeiric water bodies or
flooded mines. The precipitate is normally dredged and fed to the
surface complex for further processing to obtain a marketable product
[5].
 Figure 2
Block flow diagram for solution mining and processing potassium chloride in Saskatchewan.

References
1. Soil Fertility and Plant Nutrition, Potash and Phosphate Institute, Atlanta,
Ga., 1990.
2. G. Barry, Potash, in Metals and Minerals Annual Review, The Mining Journal
Limited, July 1992.
3. R. T. Hart, A Review of Waste Management in the Potash Industry and
Options and Decommissioning and Abandonment of Potash Tailings Piles:
Report on CANMET Project 35024, Energy, Mines and Resources, Ottawa,
Ontario, Canada, 1985.
4. R. C. Smith, From catastrophe to opportunity (converting a flooded potash
mine to solution mining), In Proceedings of the 39th Annual Meeting of the
Fertilizer Industry Round Table, Atlanta, Ga., October 31November 2, 1989.
5. E. L. Goldsmith and M. O. Klein, Potassium chloride crystallization in
Saskatchewan, In Proceedings of 2nd International Conference on Separations
Science and Technology, M. H. I. Baird and S. Vijayan, eds., Hamilton, Ontario,
Canada, October 14, 1989.
 Page 343

25
Environmental Aspects in Mining and Production of KCl
Kenneth W. Reid
Potash Corporation of Saskatchewan Inc.
Saskatoon, Saskatchewan, Canada

Maurice O. Klein
International Fertilizer Development Center
Muscle Shoals, Alabama
I
Introduction
The effluents in a typical conventional processing operation for
potassium chloride are (1) liquid effluents, (2) solid effluents (tails),
and (3) air emissions. The major concern with these effluents is
potential contamination of the environment with salt. To minimize the
risk of such contamination, various waste management methods for
fertilizer pollution control have been adopted or developed by the
potash industry in Saskatchewan. The methods and equipment used
for pollution control are discussed together with the characteristics of
the effluents, their impact on the environment, relevant legislation and
regulations, and some mine decommissioning plans currently being
evaluated.
II
Solid Wastes
A
Sources and Composition
Because sylvinite ores contain NaCl, KCl, MgCl2, and water
insolubles in the form of silts, sands, clays, and dolomites, and
because the market demands a product with a purity in excess of 95%
KCl, most of the non-KCl ore constituents are removed during the
beneficiation and refining process. The solid constituents removed
form the solid effluent or solid wastes produced by a conventional
mining and refining operation to manufacture potassium chloride.
These are commonly known as tails.
About 300 million metric tons (mt) of tails have been produced over
the last 30 years by the potassium chloride industry in Saskatchewan
at the rate of 1.8 to 2.5 mt of solid waste for every metric ton of
Saskatchewan potash (K2O) product, depending on the KCl content of
the run-of-mine ores. Tails make up the largest component of wastes
with run-of-mine ores; common table salt is the major constituent.
The quantity of tails increases
 Page 344
with lower-grade ore bodies. For example, in the former USSR, 75 to
80% of run-of-mine ores are processed into tailings [1].
The composition of tails varies with the composition of the ores being
mined, which also varies within the ore body. Two general types of
tails are produced in Saskatchewan, as indicated in Table 1. Tails
containing up to 13% insolubles are produced from the northern ore
body, which may contain up to 7% insolubles. Tails containing up to
2% insolubles and 3% MgCl2 are produced from the southeastern ore
body, which may contain up to 2% insolubles and some
carnallitehence the elevated presence of MgCl2. All tails contain small
amounts of KCl.
B
Description of Waste Management Area
The waste management area is the site used to store the wastes
produced during beneficiation and refining of the run-of-mine potash
ores into muriate of potash products. In Saskatchewan, the area varies
in size but may occupy up to 350 ha of land. A waste management
area is diked and contains a salt pile and a brine pond as illustrated in
Fig. 1. The salt pile has a wedge shape with sides sloping at 30 to 40°
and generally a back slope of less than 4°. Adjacent to the salt pile is
the brine pond, generally less than 3 m deep and up to 150 ha in size.
The brine pond is normally contained by dikes, which are usually less
than 4 m high but may reach 6 m in some locations.
Dikes are built from the most appropriate material available locally. A
fine sand or sandy till is the preferred material. Frequently, the only
material available is a silty till, which is prone to wave erosion and
loses cohesion upon contact with brine [2]. Dikes constructed with
silty till require constant monitoring and maintenance. Moreover, silty
till material has a high potential for pipe failure, which can be
prevented by placing a sand filter on the downstream toe of the dike.
In some cases, the dikes are protected from erosion with stone riprap
or a sacrificial berm.
A waste management site generally is located in a low-lying region,
which may have been a slough or a marsh, on a 1-to 4-m-thick deposit
of silt, sand, and gravel. This
Table 1 General Composition of Tails from
Saskatchewan Mines
Tails produced from:
Southeastern Northern
mines mines
NaCl (%) 9598 9095
KCl (%) 23 13
MgCl (%) 23 00.5
Water insolubles (%) 1.52 515
Composition of insolubles
Quartz (%) 1.52 515
K-spar (%) 7.510 12.516
Dolomite (%) 5565 28.532
Anhydrite (%) 016 03
Illite (%) 612.5 17.521
Chlorite (%) 02.5 4.55.5
Septechlorite (%) 5.56.5 16.518.5
Source: Ref. 3.
 Page 345
deposit is permeable and contains a surficial aquifer, also illustrated in
Fig. 1. Under the surficial aquifer is a relatively impermeable glacial
till ranging from 1 to 50 m in thickness. An intertill sand aquifer may
be found under the glacial till. Beneath the intertill aquifer, or the first
till formation in the absence of intertill aquifer, is a second till layer,
which may be 50 m thick and is normally highly impermeable.
Beneath the impermeable layer is a deep aquifer. This aquifer, along
with deeper aquifers, normally is under artesian pressure. Bedrock
shales form the bottom of the deep aquifers. The major environmental
problem of potash operations in Saskatchewan is to prevent brine
from escaping into the surficial and sometimes the intertill aquifer to
avoid salinization of these water sources.
Most waste management sites are in a groundwater-recharged area. As
illustrated in Fig. 2a, in a recharge area the natural flow pattern of
water is into the ground and away from the site. If not prevented,
brine will also enter the aquifer and flow away from the mine. A few
waste management sites are located on groundwater-discharge areas
where the natural flow pattern is toward the mine, as shown in Fig. 2b.
Brine flows in this setting are much easier to control because the
natural hydraulic pressures resist the downward and lateral movement
of the brine from the ground surface. Groundwater-discharge sites
contain natural salty water because in southern Saskatchewan
evaporation rates exceed the precipitation rates. These sites are always
located in sloughs and marshy areas.
The runoff from a mine site is either drained directly into the nearby
brine pond or is first drained into a collection pond and then pumped
to the brine pond. Surface runoff
 Figure 1
Generalized stratigraphy under the waste management
area of potassium chloride production in Saskatchewan.
 Page 346

Figure 2
Brine movement and groundwater flow in different groundwater
regimens: (a) groundwater recharge site; (b) groundwater discharge site.
originating off the site and flowing to the mine site is diverted around
the area or is collected and added to the brine pond.
C
Solid Waste Handling
The solid waste is slurried in the surface refinery and transported from
the refinery to the waste management area by pipeline. As indicated in
Fig. 3, a tails slurry of about 20 to 35% solids is produced by
combining (1) the solids from the tails filter, (2) the slurry from the
slimes thickener underflow, and (3) reclaimed brine from the brine
pond. The solids that discharge from the tails filter are the salt tails,
which normally contain 5 to 7% moisture content [4]. The underflow
from the slimes thickener is a mixture of water insolubles and fine salt
and is collectively known as slimes. The composition of tails slurry
may vary from refinery to refinery and at any given time. A sample
composition of tails slurry from a mine working the northern ore
bodies is shown in Table 2.
 Page 347

Figure 3
Schematic configuration of conventional waste handling
in the production of potassium chloride in Saskatchewan.
Table 2 Analysis of Constituents of the
Solid and Liquid Waste Produced by a
Potash Mine in Saskatchewan
Composition of tails slurry
Density 1.44
pH 6.40
Temperature 22°C
Solids 32%
Composition of solids in slurry
(%)
Na 35.5
K 1.3
Cl 53.8
Mg 0.8
Ca 0.7
SO4 0.8
Water insolubles 7.1
Al2O3 0.7
 Fe2O3 0.3
SiO2 2.3
Composition of liquid portion of
tails slurry (%)
KCl 7.7
NaCl 17.7
CaCl2 0.5
MgCl2 1.1
Specific gravity 1.226
 Page 348
In the waste management area the slurry is discharged onto the top of
the tails pile, as also indicated in Fig. 3. As the slurry dissipates by
gravity over the gentle declining slope of the back of the tails pile,
most of the solids in the slurry settle out; those of the coarsest size
settle on the higher slopes at 3 to 4° near the top of the pile, and those
of the finer size settle on lower slopes about 2° farther away from the
top of the pile. Finer salts and some of the insolubles settle out near
the toe of the pile, where the slope is about 0.5°. The brine is
ultimately clarified as it passes through the brine pond on its return to
the refinery through the reclaiming brine pump station.
At the top of the tails pile, the tails rapidly drain and within a few
minutes can support heavy equipment. The drained tails are bulldozed
around the face of the pile with earth-moving equipment to channel
the slurry to maximize the length of the flow path to increase the
amount of material settling near the top of the tails pile. Settled tails
are also bulldozed around to form the sides of the tails pile at an angle
of 35 to 40° and over the gentler backslope. Although piles up to 60 m
high have been built using this method, the average height of a tails
pile is closer to 35 m.
In contrast to coarse tails, slimes drain very slowly. Slimes produced
from northern ore bodies and discharged in a depression on the tails
pile settle slowly and form an almost impermeable base. Evaporation
of the brine above the slimes leaves behind a salt crust on top of the
slimes, which prevents normal rates of evaporation. In general, this
type of area cannot support foot access and requires extended periods
of drying.
D
Pile Characteristics
Piles are features of the local landscape. They vary in height from 20
to 65 m, the differences occurring because of the method of tails
stacking, which is discussed later, and the physical characteristics of
the tails. Coarser tails have a higher angle of repose and therefore
permit the construction of higher piles with a smaller base.
Conversely, finer tails have a lower angle of repose, permitting the
formation of piles with significantly reduced height but an extended
base area, which may reach 200 ha. The tails piles are either pink,
reddish, or white in appearance. Fresh and wet tails appear pink or red
because of iron compounds present in the ore. Old or dry tails appear
white because NaCl that is present in the pile as brine is pulled to the
surface and precipitates as water from solution evaporates. The area at
the toe of the pile, which is the main settling location for slimes,
resembles a mud flat. The area becomes fully accessible when
covered with salt as the tails pile grows and advances.
The salt in a tails pile falls within the size range of fine gravel and
coarse sand [6]. The insolubles are of medium to high plasticity and
contain silt and clay-size particles. Fine materials are moved off the
pile either by erosion forces or into the underlying salt as leaching
rainwater falls on the tails pile. On the basis of in situ filtration
studies, the surface permeability of a tails pile ranges from 1.5 × 105
to 4 × 105 M/s, a range similar to the permeability of sand. Because of
this permeability, rain penetrates the surface of the pile. However, the
pile temporarily stores the penetrating rain and releases it gradually
within 14 days, with 80% of the water discharging during the first 48
h [7]. The in situ density of the pile is about 1.5 mt/m3 and may
increase to 1.95 mt/m3 near the base of the pile. Evidence suggests
that a dense layer of NaCl precipitate forms at the pile's base and
prevents the brine from moving into the soil under the tails pile [8].
 Page 349
E
Pile Stability
Most potash operators attempt to build steeper and taller piles.
However, there were concerns that piles may contain potential failure
planes if built on layers of slimes containing 20 to 50% clay [3].
These concerns of pile stability led to a series of experimental studies
on the expected strength of a slurry tails pile under short-term loading.
Results from laboratory experiments showed that solid tails
conglomerate is a relatively strong material having an internal friction
of 40° while its strength improves in the presence of up to 20%
slimes. Pure slimes also form a reasonably strong material exhibiting
an internal friction of about 30°. Much of the strength of potash tails
and subsequently of a tails pile is due to the cementation that occurs
between the salt particles. In fact, strength tests on samples taken from
a tails pile have measured an angle of internal friction greater than
45°, which is a characteristic of very strong granular soils. The
strength component caused by interlocking and crushing is considered
substantial in these types of soils [9].
The stability of potash tails piles has also been examined
parametrically [5] and with field studies [10]. These studies have
shown that tails piles creep near the top but not enough to cause
serious concerns about failure. In addition, computer modeling studies
using the limit equilibrium analysis have been performed on a pile of
strong tails material over relatively soft material such as glacial
sediments. These studies indicated that a tails pile built on a weak
foundation of soil could fail only if the pore-water pressure,
particularly below the toe of the pile, exceeded 25% of the height of
the pile [11]. On the basis of these findings and conditions in
Saskatchewan, pile stability does not appear to be a significant
concern even for piles exceeding 200 m in height.
F
Advantages and Disadvantages of Conventional Waste Solids
Handling
The major advantages of the conventional technique for waste solids
handling is that this technique is a relatively trouble-free operation
and provides an effective material handling system. Some operating
problems occur because of the occasional plugging of the tails
discharge pipe, in which case it requires disassembly for cleaning, and
during winter shutdowns when brine left in the tails pipe because of
poor drainage freezes, thus hindering production startup efforts.
Occasionally, unsaturated brine is released from the surface refinery
because of process upsets. When discharged on top of the tails pile,
this brine dissolves a hole through the tails pile and generally exits the
lower part of the pile in an undesired location. Similarly, rain falling
on the tails pile can also create these holes. The holes must be filled
with tails salt to maintain a proper tails slurry path of flow.
The requirement for expanding the waste management area every few
years is the major disadvantage and concern with the conventional
tails handling technique. As the pond fills with slimes, this storage
capacity must be replaced to provide an adequate operating reservoir
for reclaimed brine and to allow for a runoff from a maximum
probable storm falling on the waste management area. Moreover, even
the coarsest tails settle require relatively flat slopes to deposit. The
combined effect of relatively flat slopes and gradual filling of brine
ponds and their replacement results in an expanding waste
management area having a large surface-to-volume ratio. As this ratio
increases, the potential environmental damage per unit salt stored also
increases. Moreover, a larger waste management area has a larger
perimeter that must be protected against brine escape. In addition,
government agencies regulating potash mines such as the one in
Saskatchewan
 Page 350
are becoming increasingly reluctant to approve further expansions of
existing waste management areas.
G
New Approaches to Solid Waste Handling
Several potash operators have taken different approaches in
attempting to increase the life of the brine ponds by retaining the
maximum amount of tails on the tails pile. The most common
approach is to build salt barriers on the tails pile to create a small
settling area not only to trap material but also to increase the flow path
substantially and thus increase the opportunity for the tails to settle
out from the flowing slurry. Some operators have created actual
settling ponds on the pile. These ponds have been only partially
successful because the dikes, particularly when built from salt, break
frequently when in contact with undersaturated brine.
Some potash operators have experimented with portable cyclones to
debrine the slurry on top of the tails pile. The denser material from the
cyclone underflow has a higher angle of repose, and steeper piles can
be built. This concept of debrined tails handling has been practiced in
the past in Saskatchewan by utilizing a centrifuge located in the
refinery and a conveyor between the refinery and the tails pile. This
method has been abandoned mostly because the conveyor could not
be operated in 40°C temperatures, which are common with
Saskatchewan winters.
The Rocanville Division of the Potash Corporation of Saskatchewan
Inc. has taken a modified approach to solid waste handling to
maximize the use of its existing waste management area and to reduce
the potential contamination with salt of a nearby stream and
surrounding soils. This stream is a source of water to a municipality,
and the soils in the vicinity of the mine are highly permeable. The
Rocanville approach is illustrated in Fig. 4. The underflow of the large
slimes thickener is pumped directly to a separate slimes settling area
without being combined with salt tails. In turn, the salt tails are
slurried

Figure 4
Tails handling at PCS Inc., Rocanville Division, Saskatchewan.
 Page 351
to about 20% solids and pumped at about 500 L/s to the top of the 60-
m-high tails pile to two of three self-propelled cyclone/stackers and
debrined to about 15% moisture content. Each of the three units has
13 cyclones, each 30.5 cm in diameter. A short drag chain discharges
the underflow tails over the edge of the tails pile, where tails stand at
an angle of about 30°. The cyclone overflow is directed to the brine
pond. With this method, less brine percolates through and to the base
of the tails pile and the hydraulic pressure causing the brine to
percolate into the subsoil strata is also reduced. Moreover, since
cycloned tails will stand at a steeper angle than tails precipitated from
a slurry, space utilization is also improved because piles with a steeper
side angle can be built. Not only can more tails be stored over a given
area of land, but the area of soil through which the brine can penetrate
per volume of tailings is reduced [4].
III
Liquid Wastes
A
Sources
Liquid waste is produced mostly from fresh water used in the
beneficiation and refining process and from precipitation falling on
the mine site on or close to the waste management area. Liquid waste
is contained in the brine pond before being reclaimed for tails slurry,
for makeup brine in the refining process, for heat sinks in vacuum
cooling crystallizers in some cases, or for disposal. About 1 m3 of
fresh water is used for every metric ton of potash (K2O) produced
[12]. About 80% of this water enters the brine pond as part of the
brine in tails slurry [13]. The brine in the pond is generally saturated
with NaCl.
Liquid waste also enters the brine pond from seepage collection
ditches, drains, and wells. These collect groundwater, which may
contain brine escaping from the perimeter of the waste management
area. Although some of the liquid is brine, most is natural
groundwater.
B
Liquid Waste Disposal
Because of process requirements, particularly process bleeds to keep
certain brine constituents or entrained solids in the process brine
below certain levels, excess brine is produced and disposed of. The
method of disposal depends on the location of the mine, availability of
rivers and large bodies of water, and regulations. In New Brunswick,
Canada, excess brine is controlled discharged into the Bay of Fundy
through a 32-km-long pipeline [14]. In the German and French potash
industry, excess brine is also disposed of in bodies of moving waters.
In Saskatchewan, however, excess brine must be discarded by means
of injection into deep wells because regulations do not allow the
discharge of potash brines into any body of surface water. Nor do they
allow for brines to enter any aquifers that contain usable water or any
aquifer that connects with an aquifer with usable water.
The amount of brine to be disposed of is based in part on the
prevailing climatic conditions and regulations. For example, Fig. 5
illustrates the water balance for a potash waste management area in
Saskatchewan of 180 ha storing 3.6 million mt/yr of solid waste.
Although evaporation is significant because of the semiarid climate,
the amount of brine injected is determined to comply with (1) the
environmental regulations, which require that the brine pond always
be capable of containing the runoff from a maximum probable storm,
defined as 0.3 m of water in 24 h; and (2) the dike's free board
standard, defined in the CANMET Pit Slope Manual [15].
 Page 352

Figure 5
Approximate annual water balance (m3) of a Saskatchewan potash waste
management area of 180 ha storing 3.6 million mt/yr of solid waste from
a Saskatchewan potassium chloride manufacturing complex.
The excess brine is injected into deep aquifers located beneath all
potassium chloride mines in Saskatchewan. The stratigraphic sections
of sedimentary geology in Saskatchewan are illustrated in Fig. 6. The
northern mines use the sandstones of the Deadwood Formation. This
formation is up to 250 m thick under these mines and is about 1300 m
below ground level. The Deadwood Formation is at the bottom of a
basin that is closed off on the sides without having a defined discharge
area. Impermeable shale isolates the Deadwood Formation from the
mining horizon and the near-surface aquifers. The natural water in this
formation contains about 35% NaCl and is generally similar in
composition to waste potash brines. The formation has been absorbing
brine from five injection wells, one well per mine site, for almost 20
years without any evidence of pressure buildup [16]. The rate of
injection into deep wells of the Deadwood Formation ranges from 30
to 65 L/s brine with pressures from 2400 to 6200 kPa.
In the southeastern part of the province, the Deadwood Formation
becomes too thin to receive injected brine. Instead, mines in this area
use the Interlake Formation, which is 100 m thick, highly fractured,
and made of limestones and dolomite; its natural formation water
contains about 20% NaCl. The Interlake Formation is believed to
discharge about 300 km to the east near Lake Winnipeg, located in
Manitoba [16]. The Interlake Formation is less receptive than the
Deadwood Formation to injected brine, but it absorbs brine without
excessive pressure buildup at the well head. Because of the lesser
receptivity, mines using the Interlake Formation have multiple
injection wells. On the basis of two locations that use the Interlake
Formation for brine disposal, considerable local variations exist in the
ability to accept brine. Typical injection rates range from 15 to 44 L/s
with pressures between about 2700 and 5500 kPa. The majority of the
injected brine appears to be taken up by a few highly permeable
fractured zones. To increase its capacity to receive brine, the
formation may be perforated at the deep-well interface.
The deep injection wells are constructed using conventional oil field
technology. A
 Page 353

Figure 6
Idealized stratigraphic sections of sedimentary geology in Saskatchewan.
number of disposal configurations have been used, but the most
productive appears to be the open-hole type illustrated in Fig. 7. In
this configuration, a hole is drilled from the ground level down to and
through the disposal formation. The casing installed from the ground
level to the top of the formation is cemented to the adjoining rock.
The injection pipe is then inserted through the casing, down to the top
of the disposal formation. A packer is also placed down the hole at the
top of the receiving formation to seal off the space or annulus between
the two pipes. This space is filled up with diesel fuel, which is under
pressure. The diesel fuel serves as a corrosion inhibitor and as a leak
detector. A leak of diesel fuel or its failure to hold pressure indicates a
leak into the annulus,
 Page 354

Figure 7
Idealized brine disposal well.
usually because of packer failure. Cathodic protection is required on
each well. An annual pressure-falloff test is also required on each
well. About 80% of the pressure dissipates within a day of shutting
down an injection well [16].
A well's performance can be enhanced with fresh water, which dilutes
the injected brine. At one well, the injection rate has increased from
about 91 L/s to almost 152 L/s at the same pressure after a prolonged
rainy period, which diluted the brine in the pond. About 5 to 30% of
the waste salt now produced by potash operators in Saskatchewan is
being dissolved by rain and disposed of by deep injection.
Major operating problems of the deep wells have been packer failures
and leaks in the wall of the injection pipe. These types of failure
require the removal of the injection string to replace the failed part. In
addition, a well that has been shut down and allowed
 Page 355
to drain back by induced pressure in the formation, instead of being pumped
back to the surface, can become obstructed by debris being forced back to
the well. Moreover, the back pressure can unseat the packer and move it
upward in the annulus.
C
Liquid Waste Containment
No brine is deliberately allowed to escape from the waste management area.
Any runoff produced by precipitation falling on the mine site is collected
and directed to the brine ponds. Surface flow originating off the site but
flowing toward the mine is diverted around the facilities or is collected and
added to the brine pond. The only surface flow of brine from the waste
management area is brine seeping through dikes. However, the seepage is
collected with subsurface seepage collection systems.
The major environmental problem to resolve is the containment of brine
escaping the waste management area into the aquifers located under this
area. This brine is usually confined to the shallow surficial aquifer. Various
techniques are utilized in an attempt to contain the brine seepage. The main
technique is the use of dikes as illustrated in Fig. 8a. Dikes that contain an
impermeable core (or key) greatly reduce brine seepage. However, unless
these dikes are extremely impermeable because of thickness or material of
construction, they cannot completely stop seepage of brine. A high degree of
quality control during construction of these types of dikes is thus required if
they are to perform effectively.
 Figure 8
Brine containment techniques employed by Saskatchewan's potassium chloride industry.
 Page 356
Polyethylene liners have been tried with little success. Attempts to use
30-and 60-mil high-density polyethylene liners have failed because of
the severe weather elements in Saskatchewan. It appears that exposed
synthetic liners are not suitable for extreme variations in temperatures
(50 to +40°C), bright sunshine, and high winds, all typical of
Saskatchewan.
Because almost all dikes leak at least a little, subseepage backup
systems used to control brine seepage have been installed by each
producer around their waste management area. One such system
consists of an impermeable cutoff wall cutting through an aquifer to
halt seepage under and through the dikes. This system is illustrated in
Fig. 8b. Like dikes, however, cutoff walls cannot completely stop
leakage, only retard it; the amount of leakage is dependent on the
hydraulic head against the wall and its permeability. Moreover, when
clay is used in the construction of the impermeable barrier, the
stresses acting on the barrier must be confined. Clays in contact with
potash waste brines may shrink, thus decreasing or destroying the
effectiveness of a cutoff when built from clay material [2].
Another system for collecting subseepage is the open-drain collection
system, in which brine moving in the aquifer enters an open ditch.
From the ditch, the brine is pumped back to the brine pond, as
illustrated in Fig. 8c. The aquifer in this system has to be relatively
shallow. Otherwise, a very large ground cut would be required to build
a ditch with stable slopes. In addition, the level of the brine in the
ditch must be kept low to prevent this brine from flowing across the
ditch and reentering the aquifer. Unlike a dike or a cutoff wall, a
properly constructed and operated drain can stop seepage nearly
completely. Where conditions permit, the open-drain system is the
most cost-effective seepage containment technique for the potassium
chloride industry of Saskatchewan.
Where the aquifer is too deep for an open-drain application, a buried
drain may be employed to collect brine seepage, as illustrated in Fig.
8d. A buried-drain collection system operates on the same principle as
the open-drain system except that the brine is collected in a perforated
drainpipe. The cut is backfilled with a permeable material to (1)
enable brine to pass to the pipe and (2) eliminate concerns about the
slope stability of the drain's excavation. This method is becoming a
standard practice in the Saskatchewan potash industry.
In a few cases, some brine seeped into the intertill aquifer. Because
this aquifer is too deep to permit the construction of a drain or cutoff
walls, containment wells have been used as illustrated in Fig. 8e. The
containment wells are located around the waste management area. The
seeping brine is intercepted by creating intercepting depressions in the
groundwater level to draw brine into the well. From the well the brine
is pumped back into the brine pond. The pumps must be spaced
adequately to prevent avenues for brine to escape between wells.
Because the wells collect water from all directions, a containment
well produces large volumes of diluted brines.
A unique system for seepage containment in the intertill aquifer is
illustrated in Fig. 9. In this system the intertill aquifer is under artesian
pressure with a piezometric level higher than the bottom of the
overlying surficial aquifer. An open drain is constructed to the bottom
of the overlying surficial aquifer and gravity flow wells are installed
in the intertill aquifer. The overflow from these wells is at the level of
the bottom of the open drain. Thus brine moving in the surficial
aquifer is collected in the open drain. Brine from the intertill aquifer,
which flows by artesian pressure up through the well, overflows into
the open drain and mixes with the surficial aquifer drainage in the
interceptor drain. The mix is pumped back to the brine pond.
Surface escape of brine is uncommon. However, some spills have
occurred because
 Page 357

Figure 9
Schematic of brine containment system at PCS Rocanville.
of breaks in brine or slurry pipes but have been contained because of
the mine site's pattern of drainage to the waste management area.
Some spills have required emergency cleanup action. Cleanup
procedures normally call for the removal of any collectible brine or
salt and the contaminated soil into the waste management area. The
area in which the spill occurred may be flushed with clean water to
dilute the remaining salt below levels that can cause environmental
damage.
D
Monitoring Waste Brine Movement
The degree of monitoring for seepage of the waters around the waste
management area of a potassium chloride manufacturing complex
depends largely on the regulation under which the industry operates.
In Saskatchewan, potassium chloride producers are required to
monitor groundwater and potentially affected surface water around the
waste management area as a condition for operating licenses at the
rate of twice a year for levels and once a year for constituent and
conductivity determinations. The level of the groundwater indicates
flow pattern. The chloride concentrations in groundwater indicate
potash brines concentrations because chloride ions are not attenuated
by soils and they flow freely with groundwater. In contrast, the major
cations of potash brines, namely Na and K, are attenuated by a soil's
cation exchange and cannot serve as reliable indicators. Furthermore,
in Saskatchewan's groundwater, the chloride ion is a rare anion
relative to sulfate and bicarbonate ions.
In an analysis of groundwater data, the background level of chloride
must first be established for both the area and the aquifer of interest.
The observed chloride
 Page 358
concentration in a monitored well can normally be related to
background concentration and flow pattern to establish the degree of
potash brine contamination and the performance of the brine seepage
containment system. At times it may be difficult to determine whether
a chloride observation is natural or from brine contamination; in such
cases the ratio of chloride to sulfate or chloride to total dissolved
solids usually can establish whether the source of the chloride
concentration is natural. In Saskatchewan, chloride concentrations
should not exceed about 25% of the sulfate levels. This is equivalent
to about 5% of the total dissolved solids. An exception would be
found in a naturally saline area where the sulfate ion concentration has
exceeded the solubility of calcium sulfate product, hence leaving
behind a solution in which the chloride-to-sulfate ratio is uncommonly
high.
Wells 5 to 10 cm in diameter are installed to access the aquifer of
interest. Larger wells allow easier access for monitoring and sampling
equipment, and therefore wells with a smaller diameter have been
superseded by those with a larger diameter. Both the screen and the
riser of a well are constructed from polychlorine vinyl to prevent
corrosion and permit downhole geophysical equipment. The annulus
around the screen is sand packed. The annulus above it is sealed with
bentonite. Cement grout seals the annulus from the bentonite to the
surface.
Various types of equipment have been used for drilling a monitoring
well, depending on the stratigraphy involved and the need to collect
soil samples. Such equipment includes cable tools, top-and bottom-
drive rotaries, augers, hollow-steam augers, and air percussion drills.
The wells around the waste management area, which supply water to
the nearby farms, are also monitored. These wells are located at least
1 mile from the waste management area because the normal practice
of the potassium chloride industry in Saskatchewan is to acquire all
lands within 1 mile of the waste management area. Since the inception
of the monitoring program, there has been little evidence to suggest
that any of these wells have been affected by escaping potassium
brine. Data from these wells are often used for public relations and for
protecting the potash operators against legal action by the surrounding
farmers. Conversely, the water in these wells may be contaminated
from livestock use.
The monitoring program also includes the sampling of surface water
from springs, streams, and ponds located near a potash operation. In
analyzing the data from these samples, attention is always given to ion
ratios rather than to absolute ion concentrations. In general, data from
these sources are of marginal use because they are highly dependent
on both season and weather conditions.
After almost 20 years of monitoring the groundwater around potash
operations in Saskatchewan, monitoring systems show that brine has
moved at most a few hundred meters from the brine ponds. Brine
movement beyond these points has been prevented by containment
facilities. With rare exceptions, brine contamination is restricted to the
surficial aquifer. Moreover, all areas of brine contamination are on
lands owned by the potassium chloride producers.
E
Use of Geophysical Methods to Monitor Movement of Brine
Establishing an effective network of monitoring wells requires a
laborious and costly trial-and-error process to determine the
appropriate location for the monitoring wells. Because a brine is
highly conductive, the use of electromagnetic resistivity surveys is
being studied by the Saskatchewan potassium chloride industry as a
substitute and/or
 Page 359
supplement to groundwater sampling and analysis. Electromagnetic
resistivity geophysical surveys do not require direct contact with the
aquifer being monitored, and a large area can be surveyed relatively
inexpensively. The testing equipment involved sends a pulse of
electricity into the earth and measures the resistivity of the soil on the
basis of time and intensity of the received pulse. The spacing between
the receiver and the transmitter determines the depth at which the
resistance is measured. The use of groundwater penetrating radar and
shallow seismic surveys has also been tested. However, useful data to
determine the brine concentration in the groundwater could not be
produced.
Equipment designed to measure conductivity across a soil profile
intersected by a well or a tube installed in the ground is also being
tested and thus far has produced promising results. The profiling
instrument allows an operator to determine the location of the brine in
the soil's profile without having to take multiple samples.
References
1. A. G. Lopushnyak, Environmental protection at potassium plants of
the USSR, paper presented at Kali '91, 2nd International Potash
Technology Conference, Hamburg, Germany, May 2629, 1991.
2. S. L. Barbour and N. Yang, The influence of claybrine interaction
on the geotechnical properties of clay soils of western Canada, paper
presented at Kali '91, 2nd International Potash Technology
Conference, Hamburg, Germany, May 2629, 1991.
3. R. F. Johnson, Investigation of the Physical and Chemical
Characteristics of Potash Tailings in Saskatchewan, File 26SQ-
23440-39177, Supply and Services, Ottawa, Ontario, Canada, 1984.
4. K. W. Reid and G. A. Maki, Innovative waste management
practices at PCS Inc., Rocanville Division, paper presented at Kali
'91, 2nd International Potash Technology Conference, Hamburg,
Germany, May 2629, 1991.
5. S. R. Tkach, K. W. Reid, C. L. Hall, and L. L. Hlady, Potash
Corporation of Saskatchewan, Inc., Lanigan Division,
Decommissioning Report, Potash Corporation of Saskatchewan, Inc.,
Saskatoon, Saskatchewan, Canada, 1990.
6. D. G. Fredlung, Long term stability of potash tails piles on soft
foundations, paper presented at Kali '91, 2nd International Potash
Technology Conference, Hamburg, Germany, May 2629, 1991.
7. K. Shook and K. W. Reid, Attenuation of Storm Water Runoff from
Potash Tails Piles, Final Report, File 14SQ-23440-5-9164, Supply
and Services Canada, Ottawa, Ontario, Canada, 1986.
8. P. G. Landine, S. L. Barbour, and M. J. Reeves, Characterization of
weathering of potash tailings piles, Annual Report to the
Saskatchewan Potash Producers, Department of Civil Engineering,
University of Saskatchewan, Saskatoon, Saskatchewan, Canada,
1992.
9. Waste management practices in Saskatchewan, in Phosphorus and
Potassium, No. 180, JulyAugust 1992.
10. D. C. Goodall, D. G. Fredlung, and K. N. Lamb, Methods and
applications of a state-of-the-art geotechnical field investigations of
potash tailings near Lanigan, Saskatchewan, in Potash Technology
'83, R. M. McKercher, ed., Pergamon Press Canada, Toronto, Ontario,
Canada, 1983.
11. Z. M. Zhang, K. MacDonald, and D. G. Fredlung, Stability of
Potash Tails Piles of Extended Heights: A Report to the Potash
Corporation of Saskatchewan, Inc. , Saskatoon, Saskatchewan,
Geotechnical Engineering Group, Department of Civil Engineering,
University of Saskatchewan, Saskatoon, Saskatchewan Canada, 1991.
12. R. T. Hart, A Review of Waste Management in the Potash Industry
and Options and Decommissioning and Abandonment of Potash
Tailings Piles: Report on CANMET Project 35024, Energy Mines and
Resources, Ottawa, Ontario, Canada, 1985.
 Page 360
13. K. W. Reid, 1984, Water use in the Saskatchewan potash industry
and opportunities for water recycling/conservation, Canada Water
Resources Journal, Vol. 9, No. 1, 1984, pp. 2126.
14. J. R. B. Ross and D. C. E. Waugh, 1991, Meeting the challenge of
backfilling salt tailings at the DPPC potash mine, paper presented at
Kali '91, 2nd International Potash Technology Conference, Hamburg,
Germany, May 2629, 1991.
15. CANMET, 1977, Waste embankments, Chapter 9 in Pit Slope
Manual, Supply and Services Canada, Ottawa, Ontario, Canada, 1977.
16. G. McLaughlin and K. W. Reid, Use of Injection Wells at PCS
Mining for Disposing of Waste Potash Brine: Final Report, File
23SQ-23440-5-9165, Supply and Services Canada, Ottawa, Ontario,
Canada, 1985.
 Page 361

26
Long-Term Solid Waste Management and KCl Facilities
Decommissioning
Kenneth W. Reid
Potash Corporation of Saskatchewan Inc.
Saskatoon, Saskatchewan, Canada

Maurice O. Klein
International Fertilizer Development Center
Muscle Shoals, Alabama
I
Introduction
Most of the potassium chloride operations in the world, including
some in Saskatchewan, were built before the current high level of
regulations and public awareness of the need to protect the
environment. However, while initial design and subsequent remedial
actions taken by potassium chloride producers in Saskatchewan have
restricted the environmental damage to the loss of land immediately
under the surface refining facilities and the waste management areas,
in many cases an imbalance exists between the life of a mine and the
storage capacity of the waste management facility. Specifically, most
mines have large ore reserves, which would allow these mines to
operate for more than 100 years at current design production rates.
However, using existing waste management methods, few mines have
waste storage facilities that would last beyond 25 years without
expanding their boundaries.
Because the regulatory authority in Saskatchewan is reluctant to allow
a potash operator to expand its current waste management area
beyond current boundaries and because the same authority has been
pressing the potash industry not only to stop storing tailings on the
surface but also to produce decommissioning plans for both the potash
mines and the waste management areas, the potash producers in
Saskatchewan have been studying several alternatives. These
alternatives are based on (1) managing the waste so that further
expansions of existing waste management areas can be avoided, (2)
halting the surface storage of salt, and (3) removing the salt already in
storage or isolating the salt from the environment. A discussion of
these alternatives follows.
II
Long-Term Management Option
The long-term management option is based on the belief of potash
operators that (1) the mines could operate without expanding their
waste management area, and hence storing
 Page 362
salt on the surface should be allowed, (2) the present environmental
impact is minimal and would not increase by adding more salt to an
area that is already covered with salt, and (3) once a mine has been
exhausted, arrangements could be made to maintain the existing brine
containment facility and the injection wells. After closing the mine,
the brine pond would be drained through the injection wells and
would be kept empty except during and after a major rainstorm. With
time and adequate precipitation, the rain would dissolve the salt pile.
The resulting brine and that collected in the seepage collection system
would be injected into deep aquifers through the injection wells. This
process would continue until all the salt in the tails would be dissolved
and injected. The difficulty with this management option is that up to
2000 years may be required to dissolve a pile given the normal rate of
rainfall in Saskatchewan.
III
Enhanced Long-Term Management Option
A variation of the long-term management option is the enhanced long-
term management option. In this option the water now used for potash
production, once potash production ceases, would be used to enhance
the dissolution of the pile. Pile dissolution time in this approach would
be reduced to less than 500 years and in some cases to less than 200
years. However, two main challenges remain to be resolved with this
approach. One is the uncertainty of water availability 100 years in the
future. The other is the effect on the solubility of tails pile material
when stored over long periods. The physical and chemical processes
taking place in the pile as it weathers are extremely complex [1].
IV
Dissolution and Injection Option
A variation of the dissolution and injection principle would be to use
more water for salt dissolution, followed by more brine injection
while potassium chloride is being produced (1) to reduce or eliminate
storage of salt in the waste management area and (2) to reduce the size
of the existing pile before exhausting the mine of its ore. Although
this option may be technically feasible, it may not be economically or
politically feasible because (1) the cost of water and brine injection
may be more than the historical low profit margins for potash, (2)
there is political reluctance to allowing the use of potable water for
this purpose in an area such as Saskatchewan where a net precipitation
deficit exists, and (3) a new dimension in waste handling would be
added, together with unresolved questions on how an operating potash
tails pile could be dissolved without creating problems of safety or
tails management.
The Saskatchewan industry has searched for a source of groundwater
for this purpose in the Manville Formation (see Fig. 6 in Chapter 25).
This is an extensive sandstone formation that has been known to
contain large quantities of brackish water under high pressure.
However, during an extensive drilling program, it was found that the
hydraulic conductivity of this aquifer was too low to provide
sufficient quantities of water. No other known potential sources of
geological formation water are available so far.
V
Backfilling Option
The backfilling option consists of taking the tails and placing them in
mined-out areas as backfill. In addition to the environmental
advantage, the backfilling option may prolong the life of the mine
because it may permit secondary mining of the support pillars.
 Page 363
However, backfilling is highly dependent on the geographical
conditions of the potash beds. Large high cavities (stopes) are created
by mining folding potash ore bodies, as in Germany and eastern
Canada. These cavities are filled without major difficulties.
A typical backfilling system adopted by one of the potash operators in
New Brunswick, Canada is illustrated in Fig. 1. In this system, the
debrined tails from the drum and scroll filter are repulped with
process, underground return and with reclaim brine to about 50 to
60% solids. The slurry is gravity conveyed through a pipe 203 mm in
diameter located in the production shaft down to the mine horizon,
which is about 800 m below the surface. The slurry is then pumped
and disposed of in the disposal area, where the brine is separated,
collected, and returned to the surface. The reported fill factor and
backfill density have been 90% and 1.5 g/mL, respectively, and the
monthly
 Figure 1
Schematic of slurry backfill and brine handling systems.
 Page 364
backfilling rate above 121,000 metric tons (mt) of tailings [2]. The
size distribution of the tails handled in this system is illustrated in Fig.
2.
Unlike folding potash beds of Germany and eastern Canada, the ore
bodies in Saskatchewan are fairly flat. Moreover, most Saskatchewan
mines have a high closure rate because the salt rock is plastic, flowing
under high pressure, and therefore mined-out areas close within a few
years. As a result, the available underground space is insufficient for
the disposal of tailings currently stored on the surface. However,
wastes now being produced could be stored underground as backfill,
in the manner described earlier, shortly after the openings in the ore
body have been created. In Saskatchewan the backfill method has
been shown to be technically feasible in pilot studies, but the
implications for cost, mine safety, and efficiency remain to be
addressed.
VI
Capping of Tails Pile Option
Consideration has also been given to covering the tails pile with an
impermeable cap. On the basis of extensive laboratory testing and
small-scale field experiments, two pilot-scale caps have been placed
on a tails pile. One of the caps has been constructed of a consolidated
glacial till separated from the pile surface by a capillary break of
granular material and protected from weathering by another granular
layer on its surface. Another cap has been made from a mixture of
sand and bentonite polymerized to resist salt. The performance of both
caps has been monitored for two years, and both have performed
reasonably well [3]. However, their performance in the long run
remains to be ascertained. These caps must retain their low
permeability and resist degradation from severe weather conditions
and from stresses caused by salt. If any amount of water penetrates the
caps,

Figure 2
Grain-size distribution of DPPC halite backfill.
 Page 365
this water would dissolve the underlying salt support for the overlying
cover material, causing it to collapse into the depression formed and
thereby destroying the integrity of the heap.
Capping presents two additional potential problems that remain to be
resolved. One is the degree of difficulty in capping the steep sides and
back slopes of the tails piles. The other is the enormous amount of
capping material required to cover a pile with an eventual area of
perhaps 500 ha and a height of 100 m.
VII
Solution Mining Option
The process of decommissioning tails piles by dissolution and
injection into a flooded mine could produce potassium chloride as
well. This is possible because of the relationship between the
solubilities in water of potassium chloride and sodium chloride with
respect to temperature. When cold water is saturated with sodium
chloride and injected into a flooded mine, the solution absorbs heat
and dissolves KCl from the surrounding ore body; consequently, NaCl
is precipitated from the solution. The KCl-enriched solution is
pumped to the surface and cooled to precipitate potassium chloride
discriminately, which could be further processed to obtain a salable
product. The solution from which the potash has been precipitated
would then be resaturated with NaCl and reinjected into the flooded
mine to begin the cycle again.
The principle of solution mining is already practiced at two locations
in Saskatchewan, as indicated earlier. At one location, those cavities
in the ore body produced by solution mining could be backfilled with
salt originally brought to the surface in solution. The other operation
produces potash by dissolving it out of the walls of a flooded potash
mine. Among the factors to be determined are the dissolution rate of
the tails and the long-term effect of solution mining on the heat
transfer and temperature profile of the ore body in a flooded
conventional mine. Moreover, the ratio of NaCl removed to KCl
produced appears to be only about 0.25 [3].
VIII
Alternative Slimes Handling Tails Management Option
As indicated earlier, the presence of slimes requires shallow slopes
and therefore an expanded waste management area. Unless the slimes
can be backfilled or vegetated, in the long run the need for an
expanded waste management area remains. The Saskatchewan potash
industry has considered placing the slimes underground in a saturated
brine slurry but has been reluctant to do so because of mine safety
concerns. However, an alternative solution has been developed, which
is illustrated in Fig. 3. The proposed solution involves the mixing of
slimes in water of low salinity to remove about 99.95% of the salt
from the slimes originating in slimes thickener underflow. The salt
would be leached by countercurrent washing and pumping the washed
slimes to a pond where the seepage would be collected until the salt
content of the settled material reached acceptable levels [4]. The
desalinated material, a claylike soil able to support vegetation, would
then be abandoned. Also, because the excess effluent from the
counterwashing circuit would be undersaturated with respect to
sodium chloride, it could be used to dissolve more salt from the tails
pile before being disposed of through deep injection. This not only
addresses the slimes disposal problem but also reduces the amount of
salt stored on the surface by the amount of salt contained in the
unwashed slimes plus the amount of salt dissolved
 Page 366

Figure 3
Schematic illustration of slimes washing in a countercurrent decantation (CCD) fashion.
from the tails pile in the bleed from the washing circuit. Further, because the
slimes are handled separately from the tails slurry, a coarser salt would be
produced and transported to the tails pile either as slurry to be debrined on
the pile in cyclones or as unslurried salt tails; conveyors could be used
provided that the problems of freezing conveyors could be overcome. With
either method, a tails pile with slopes of at least 30 to 35° could be built and
could be raised to heights exceeding 100 m. A pile of this size could likely
store, within an existing waste management area, the salt wastes produced
during the operating life of a mine.
One distinct disadvantage of washing the slimes countercurrently is the
substantial water consumption, estimated at 10 mt for every metric ton of
slimes. Therefore, the lack of adequate amounts of available water for
decommissioning purposes mentioned earlier also applies in this case.
References
1. P. G. Landine, S. L. Barbour, and M. J. Reeves, Characterization of
weathering of potash tailings piles, Annual Report to the Saskatchewan
Potash Producers, Department of Civil Engineering, University of
Saskatchewan, Saskatoon, Saskatchewan, Canada, 1992.
2. J. R. B. Ross, and D. C. E. Waugh, Meeting the challenge of backfiling
salt tailings at the
 Page 367
DPPC potash mine, paper presented at Kali '91, 2nd International
Potash Technology Conference, Hamburg, Germany, May 2629, 1991.
3. M. D. Haug, and K. W. Reid, 1991, Tails pile decommissioning
options for Saskatchewan potash mines, paper presented at Kali '91,
2nd International Potash Technology Conference, Hamburg,
Germany, May 2629, 1991
4. B. Gotts, Alternative methods for the handling and storage of
potash slimes, 1991 Annual Report to the Saskatchewan Potash
Producers Environmental Committee, Saskatoon, Saskatchewan,
Canada, 1992.
 Page 369

27
Air Emissions at KCl Facilities
Maurice O. Klein
International Fertilizer Development Center
Muscle Shoals, Alabama

Kenneth W. Reid
Potash Corporation of Saskatchewan Inc.
Saskatoon, Saskatchewan, Canada

Elmar L. Goldsmith
Kalium Canada, Ltd.
Regina, Saskatchewan, Canada
I
Introduction
Potassium chloride is fairly soft, with a hardness index of 2 on the
Mohs scale and behaves in a brittle manner when subjected to impact,
grinding, shear, and abrasion. Consequently, all handling of potassium
chloride products tend to produce dust. Potassium chloride product is
not considered a ''hazardous," "dangerous," or "toxic" substance in
Canada and the United States and therefore it is not regulated under
legislation specifically covering these issues. For example, potassium
chloride is not specified in the Canadian Transportation of Dangerous
Goods Act [1], nor in the corresponding U.S. classifications for
dangerous goods. However, as indicated later in this section, various
jurisdictions at the national, state, provincial, or even municipal level
have regulations governing the types and levels of dust that may be
discharged into the atmosphere.
The principal sources for air emissions from a conventional potassium
chlorideproducing operation are product dryers, refinery air exhaust,
mine ventilation exhaust, and fugitive dust. However, the most
environmentally significant sources are the emissions from dryers in
which the final debrining of the potassium chloride product is carried
out utilizing natural gas as a heat source. The major environmental
concern in these emissions is the chemical composition of the
particulates emitted.
In Saskatchewan, particulates in air emissions from dryers contain
about 90 to 95% KCl and 3 to 10% NaCl. These emissions are also
acidic and in some cases may contain more than 5% organics. Other
than KCl and NaCl, few elements are present in potash dryer dust at
levels greater than trace amounts. As indicated in Table 1, the toxic
metals are either absent or at very low concentrations, particularly in
the emissions from dryers used on chemical-grade products.
Because the material fed into the dryers contains more than 95% KCl,
it is popularly believed that the KCl content of the dust produced
during drying should also be above 95%. This is not the case because
the surface brine on the product material entering the
 Page 370
Table 1 Particulates in Air Emissions from
Dryers
Concentration (ppm)
Fertilizer-grade Chemical-grade
Element dryer dryer
K 469,000 520,000
Cl 469,000 410,000
Na 50,000 3,000
Br 260 150
Fe 340 <20
Ca 230 <20
S <20 <20
Rb 50 50
Si 900 <100
O 20 <10
P <30 <30
(ortho)
Mn 10 <10
Pb <50 <50
Hg 1.17 1.33
(ppb)
Mg 530 <40
Ni 10 10
Al 310 <50
As 14 130
Co <10 <10
Ci <5 <5
SiO2 <1,000 <1,000
Ba 50 <50
F 500 400
V 53 46
Cd <0.5 <0.5
Ag <2 <2
H2O 120 700
Source: PCS, Inc., internal report.

dryers is saturated with NaCl, which appears to be preferentially

emitted in the dryer exhaust as water from the solution enveloping the
product evaporates.
The pH of dryer emissions ranges from less than 1 to about 4 [2]. The
acid causing the low pH is believed to be hydrochloric acid, which
appears to form in the dryer by the decomposition in steam at elevated
temperatures of MgCl2 contained in the feed to the dryers. The
magnesium content in the feed increases as the carnallitic content of
the ore body being mined increases. An example of water and pH
balance of a drying circuit in the potash industry of Saskatchewan is
illustrated in Fig. 1.
The organic material in the dryer emissions appears to originate from
the amines and oils added to the flotation stages of the overall
beneficiation and refining process. Amines and oils carried with the
product become liberated in the dryers. The organic material,
characterized as black, sticky, and oily, can build up on the inside of
air pollution control equipment and the filters of air sampling
equipment. The temperature of the gases at the dryer's exit ranges
from 175 to 230°C. The temperature at which they enter the
atmosphere depends on the type and lengths of the duct and stacks and
on the type of dust collection equipment.
 Page 371

Figure 1
Typical flows of liquids and their pH values in a potassium
chloride drying/scrubbing circuit in Saskatchewan.
Emissions are highly variable in terms of both dust load and size of
the dust particulate, as also indicated in Table 2. Dust load values as
high as 20 g/m3 have been reported [4]. Regarding the size of the dust
particulates, it is important to consider not only the variability of the
size distribution of the dust produced by a potassium chloride dryer
but also the factors influencing that distribution, because the difficulty
of removing a dust particle from a moving airstream is inversely
related to the particle size. An example of typical size distribution
range of dust emitted from a potassium chloride dryer in
Saskatchewan is illustrated in Fig. 2.
A major factor influencing the particle size of the airborne dust is the
type of dryer utilized, as evidenced in Table 2. A fluid-bed dryer
produces a much coarser dust than that produced by the more
conventional and widely used rotary-type dryer. More degradation of
potash particles occurs in a rotary dryer, resulting in production of
more fine particles, which become entrained in the exhaust gases.
Evidence suggests that the size distribution of dust particulates is also
influenced by the flame configuration of a particular rotary dryer. A
rotary dryer may either be directly fired, in which case the product
passes through the flame as it tumbles through the dryer, or the dryer
may have
 Page 372
Table 2 Dust Emissions and Particle-Size
Distribution from Selected Potash Dryers in
Saskatchewan
Dust Dust particle sizes:
Dryer emissions percent smaller than:
Dryertype (g/m3) 1 µm 2 µm 5 µm
1 Rotary 1.33 40 54 70
2 Rotary 1.37 25 40 56
3 Rotary 2.06 26 45 71
4 Fluid 9.62 7 19 47
bed
5 Rotary 8.47 22 32 48
6 Rotary 46 61 77
7 Rotary 1.19 23 41 69
8 Fluid 9.85 5 15 44
bed

a separate enclosed combustion chamber, in which case the product is

not exposed directly to the flame. It is believed that upon contacting
the flame, some damp potash particles fracture and produce a fine
dust, which becomes entrained in the exhaust gases. Similarly, rotary
dryers with less lifter and a slower turning speed cause less product
degradation and produce coarser dust.
The size distribution of the dust is also influenced by the drying
temperature [3]. This is illustrated in Fig. 3. A higher temperature may
cause more degradation of the particles being dried. Possibly because
dryer temperatures are lower with higher airflows, an inverse
relationship exists between airflows and fine dust loading [4].
The size distribution of the dust particulates is also affected by the
size of the potash product being dried, whose size in turn varies with
the size of the monocrystals of sylvite in the ore body. As expected,
dryers handling a finer product generally produce a finer dust. There
also appears to be a relationship between fine particle discharge and
the brine

Figure 2
Typical size-distribution envelope of dust emitted from rotary potash dryer.
 Page 373

Figure 3
Impact of potash dryer temperature on size of dust emitted.
content of the product entering the dryer. This is attributed to the
rapidity of crystallization of salt from the brine, which creates
relatively finer particles.
An evaluation by PCS Inc. found no direct relationship between
product feed rates to the dryers and the mass of dust exhausted.
However, there may be a relationship between feed rates and the size
distribution of the dust [5].
II
Recovery and Control of Dryer Emissions
The level of particulates released into the atmosphere is controlled by
processing the dryer exhaust in primary and secondary emission
control stages as illustrated in Fig. 4. The primary emission control
stages normally consist of using high-efficiency cyclones to separate
and remove entrained solid particulates in the larger size range.
Cyclones may be arranged either as a single cyclone or as several
cyclones joined together to form a multicyclone unit as illustrated in
Fig. 5. A typical multicyclone contains four cyclones, each 122 to 127
cm in diameter and about 335 to 365 cm in length. Similarly, a single-
unit cyclone has a diameter of about 213 cm and a length of 914 cm.
The operating pressure drop across a cyclone varies from 3.4 to 20.4
kPa. The separation efficiency ranges between 80 and 90%.
Although cyclones offer high reliability and performance efficiency,
they must be well maintained to sustain these levels. Cyclone
performance can be reduced by fugitive air entering the cyclones.
Fugitive air interferes with the hydrodynamic forces required to
induce separation of the dust particulate from the airstream. Such air
can enter through holes caused by erosion of the cyclone's body.
Usually, however, air enters from the hopper through the dust
discharge roto valve, which is employed to remove the dust to the
collection hopper and which frequently fails to seal properly.
 Page 374

Figure 4
Typical configuration of air pollution control used on dryer exhaust
in a potassium chloride manufacturing complex in Saskatchewan.
Secondary emission control stages consist of using either wet
scrubbers or electrostatic precipitators in which the finer particulates
are separated and removed from the gaseous stream before the
stream's discharge into the atmosphere. Wet scrubbers facilitate
separation of entrained particulate from gaseous streams based on
inertia. The operating principle of scrubbers is to increase the velocity
of the gaseous stream and change the direction of the flow abruptly.
Because of different inertia forces, gases change direction more
readily than particulates do, causing the entrained particulates to
separate from the gaseous stream [6]. A typical wet scrubber consists
of two stages, which may be physically separate or combined. In the
contactor stage, the gaseous stream is contacted with a liquid spray to
wet the surface of the particulates entrained in the gaseous stream.
The level of
 Page 375

Figure 5
Multiclone dust collector of a type used for primary
dust collection in the Saskatchewan potash industry.
saturation of the gaseous stream is also increased, causing its
temperature to decrease. The lower temperature is known to improve
the separation of entrained solid particulates. In the entrainment
separation stage, the liquid droplets from the contactor stage are
removed from the gaseous stream by cyclonic separation [7]. The
depleted stream is then released into the atmosphere. Although
various designs of venturi wet scrubbers are utilized in Saskatchewan,
the adjustable-throat venturi have been employed more successfully
than the fixed venturi in removing fine dust from dryer emissions
because the adjustable throat allows the pressure drop to be adjusted
easily. A commonly used design of venturi scrubbers is the flooded
elbow, which is illustrated in Fig. 6.
Another type of wet scrubber utilized is the multivane wet scrubber
shown in Fig. 7. Air with entrained dust is introduced into the bottom
of the scrubber while water or undersaturated brine is sprayed onto
two or more conical vanes in the body of the scrubber. As a result, the
fluid forms a curtain that the dust-laden air must pass through on its
way to the top air discharge point. A space is provided above the
water spray inlet and below the air discharge point to allow any
entrained solution droplets to disengage from the air. This type of wet
scrubber can reduce the dust emissions from a rotary dryer from 1.5
g/m3 at the
 Page 376

Figure 6
Variable-throat venturi scrubber of a type
used in the Saskatchewan potash industry.
air discharge from the dryer to about 0.040 g/m3 at the discharge of
the scrubber. A typical size distribution of the dust emitted from this
type of scrubber is shown in Fig. 8.
The factors affecting scrubbing efficiency include the pressure drop
across the scrubber, the amount and type of liquid used for scrubbing,
the size of the spray droplet, the location of the spraying source, and
the dust load entering the scrubber. The scrubber's efficiency is also
affected by conditions at the dryer's discharge, such as air velocity,
flights design, and proximity to the discharge end in the case of rotary
dryers.
In the scrubbing section, the scrubbing liquid is normally brine,
although water is also used. Results from experiments have shown
that the performance of the scrubber improves only marginally when
scrubbing with water; this improvement does not generally warrant
the incremental cost in using water. Exceptions are related primarily
to circuits producing higher grades of potassium chloride products
where salt contamination of the circuit is decreased by using water
instead of brine.
The point of entry of the scrubbing liquid in the scrubbing chamber
and the droplet size of the scrubbing liquid significantly affect the
performance of a wet scrubber because both affect the degree to
which the hot gases entering the scrubber become saturated. This is
particularly important for scrubbers operating at the upper limit of
their design.
 Page 377

Figure 7
Multivane wet scrubber.
Some potash operators have been unable to achieve the required
degree of dust removal because of inadequate gas saturation despite
the scrubbing liquid flow being at the maximum. The problem has
been corrected in one case by adding a prequenching chamber to
provide a longer period of contact between the scrubbing liquid and
the hot gaseous stream. In another case where space restriction
prevented the addition of a prequenching chamber, it was necessary to
accept the higher operating costs by installing high-pressure atomizing
nozzles to inject the brine upstream of the scrubber.
Scrubbers handle up to 680 R m3/s and operate at pressure gradients
in the range 15 to 17 kPa. The lower range of the pressure drop is
typically for coarser dust from the dryers. In addition, on the basis of
operating experience in Saskatchewan, venturi scrubbers that are
properly designed, installed, and operated require only routine
monitoring and maintenance. Problem areas are usually in the throat
of the venturi, where erosion is intense, requiring replacement of the
manufacturer's parts with components lined with ceramic. Other
problems are materials buildup in the throat of the venturi and
plugging of the brine injection nozzles, particularly the high-pressure
nozzles. In one case, the drain of the cyclonic separator plugged while
the overflow became sealed. The weight of accumulated brine in the
separator body eventually caused the bottom of the scrubber to rip
away.
On the basis of observations, potassium chloride refineries that use
higher amounts of heavy oil in the process produce air emissions that
contain above the average 5% organic materials. This black, viscous
material is deposited on the walls of the cyclonic separator of the wet
scrubber, and as noted, the oil entering the venturi throat may interfere
with the contact between the dust particles and the droplets of the
scrubbing liquid. Consideration has been given to adding a surfactant
with the scrubbing liquid to increase the wetting characteristics of the
particles. Attempts to use packed beds, wire mesh, or metal chevrons
as mist eliminators downstream from the wet scrubbers proved
ineffective because salt and organic materials in the gas stream rapidly
plugged the openings in the mist eliminators.
Operators with potash dryers producing a coarse dust seldom have any
difficulty
 Page 378

Figure 8
Typical size distribution of dust emitted from multivane wet scrubber.
 Page 379
achieving the emission standards. Moreover, only minimal
maintenance on their scrubbers is required because these scrubbers are
not operated at the upper end of their operating range. In contrast,
potash operators having a finer dust entrained in the dryer exhaust
have to spend considerable effort in fine tuning the air emission
control equipment and must vigorously monitor the performance to
ensure that the equipment continues to perform near the design limits
to meet the emission standards.
In contrast to wet scrubbers, electrostatic precipitators are considered
low-energy systems. Removal of suspended dust particulates from the
gaseous stream through electrostatic precipitators is possible because
the solid particulates have a surface charge and while passing through
the stationary electric field between two electrodes the particulates
deflect toward one of the electrodes [7]. The particles are then
collected and reprocessed. Because the gaseous stream that passes
through the electrostatic precipitator may be at an elevated
temperature, it may contain a significant amount of heat.
In Saskatchewan, electrostatic precipitators are used only on dryer
exhaust from products extracted from the southern potash ore body,
which contains less than 2% insolubles. Dryer dust from this ore has a
resistivity range of 107 to 1010 W·cm and can be processed in an
electrostatic precipitator. The number of electrostatic precipitators
used on dryer emissions depends on local processing requirements.
One refinery uses two electrostatic precipitators to remove the dust
from its dryer emissions prior to their release into the atmosphere.
One of the electrostatic precipitators has two sections and the other
has three. The typical operating conditions of the three-section
electrostatic precipitator are illustrated in Table 3.
Table 3 Data on Typical Operating
Conditions and Peformance of
Electrostatic Precipitator Used in the
Saskatchewan Potash Industry

Dryer exhaust temperature 218°C

Exhaust fan current 72 A
Exhaust fan damper setting 61% open
Burner gas valve 32% open
Electrostatic precipitator
transformers
Section A
Primary current 67 A
Primary voltage 340 V
Secondary voltage 37,400 V
Section B
Primary current 63 A
Primary voltage 290 V
Secondary voltage 31,900 V
Section C
Primary current 40 A
Primary voltage 225 V
Secondary voltage 31,900 V
Stack exhaust gas flow 925
dsm3/min
Stack exhaust temperature 165°C
Moisture content of exhaust 13%
gas
Emission particulate 0.14
concentration g/dsm3
Particulate emission rate 7.5 kg/h
 Page 380
The electrostatic precipitator require that it be properly designed,
installed, and maintained to allow a potash refinery to comply with the
emission standards using run-of-mine ores from the southern ore
body. Because electrostatic precipitators have more components that
can fail, they also require more attention to operating and maintaining
than do wet scrubbers. Types of failure associated with electrostatic
precipitators used for potash dust include broken electrode wires,
dislocated rod weights, failed electronic controls, dust deposits
causing poor airflows, holes in the discharge screw, and poor rapper
performance. Moreover, because poor performance of secondary dust
collection equipment does not directly affect the production of
potassium chloride, special efforts must be taken to ensure that the
precipitator receives the required attention.
III
Mine Ventilation
The second source of air emissions is mine ventilation air.
Mechanically mined potash mines, like all underground conventional
mines, require large volumes of air for ventilation to (1) remove diesel
exhaust generated by various types of mining equipment, (2) to ensure
that the environment in the mine is acceptably dust free, and (3)
control the temperature of the air in the mine. Flows of the ventilating
air range from 80 to 150 m3/s. The return air is typically discharged at
ground level from the mine exhaust fans. The exhausted airstream
contains 10 to 20 mg/m3 of potassium ore dust. In general, this air is
not treated for dust removal, but in some mines it is diverted through
old workings to allow some of the entrained dust to settle out before
the air is diverted for discharge. The dust level in the exhausted mine
air is lower when this air is exhausted through the service shaft. Mine
ventilation air exhausted through the production shaft entrains dust
produced in skipping the run-of-mine ores.
The main headframes of the mining shafts usually have baghouses or
wet collectors to remove the dust generated from transferring the run-
of-mine ore to the refinery. The volume of the air discharges from this
source ranges from 5 to 11 m3/s with a dust content of about 50
mg/m3.
IV
Refinery Ventilation
Processing the run-of-mine ores in the surface complex generates
dust, heat, and moisture, which are removed with refinery ventilation.
Dust is generated during the crushing of the run-of-mine ore and the
oversized product, in product screening, and in various dry product
handling stages. Typically, the quantity of exhaust air from the
crushing plants is about 10 m3/s. From the screening and open and
ventilated areas, the exhausted air ranges between 40 and 60 m3/s.
Some of the dust contained in these streams is removed in baghouses
and wet scrubbers to reduce the dust level to about 200 mg/m3 before
the refinery air is released into the atmosphere. Because the dust from
the above areas is hygroscopic, it causes plugging in the baghouse
system. Moreover, the heat from the screening area causes the
deterioration of the bag fabric. Broken filter bags are common; during
their repair, the heat and dust in the refinery may produce extremely
unpleasant working conditions.
Additional air ventilation is provided for the screening area in which
the product is handled at elevated temperatures. Air exhaust flows
from this area range from 25 to 50 m3/s. Because this air contains
only about 10 to 60 mg/m3, it is normally exhausted directly into the
atmosphere. The air exhaust from the product storage and loadout
areas
 Page 381
is also discharged into the atmosphere without dust removal treatment
because (1) these areas are not heated, (2) wet scrubbers are difficult
to operate during the harsh, long winter season, and (3) potassium
dust is hygroscopic.
V
Fugitive Emissions from Tails Piles
As mentioned earlier, the large piles of potassium chloride tails
associated with conventional mining and refining consist primarily of
NaCl and a significant amount of insolubles, such as clays, silts, and
sands. Despite pile heights close to 60 m and prevailing strong winds,
no noticeable fugitive emissions from these piles occur because the
materials in the pile cement together to form a solid mass that tends to
resist wind erosion.
VI
Boiler Stacks
Because electricity may be purchased externally, natural gas is used
mostly to provide steam and other heating media at each mine site.
Natural gas is burned in boilers that are considered relatively small
units. Because of their size and the use of natural gas and
economizers, concerns about boiler stack emissions have been
minimal.
VII
Air Quality Legislation
Air emissions are normally regulated by a government arm through
active or passive legislation. In Saskatchewan, air emissions from the
potassium chloride industry are regulated by the provincial
government through the Department of Environment and Public
Safety. The Air Quality Act is the major legislative instrument used
for this purpose. This act requires a permit to operate an industrial
source of air emissions. Although the potassium chloride industry
does not require a permit under the Air Quality Act, the act forbids the
burning on site of anything other than wood or paper wastes. It can
also order a stop or a change of any air emissions and it can order the
industry to provide any information on air emissions that the Minister
of the Department may require. Although this act exempts all mining
activities from these requirements, the Environmental Management
and Protection Act requires an air discharge permit for any mining
activities.
The potassium chloride industry in Saskatchewan must also comply
with the applicable regulations promulgated under the Air Quality
Act, such as the ambient air quality standards for Saskatchewan
shown in Table 4. The major regulations that affect air emissions in
the Saskatchewan potassium chloride industry are the Potash Refining
Air Emissions Regulations of Saskatchewan, which limit the dust load
of potassium chloride in dryer emissions to 570 mg/m3 [8]. Moreover,
under this act measurements of air emissions are required on each
dryer stack at least once per year under specified operating conditions.
The measurements must be based on an average of three successive
tests using a specified standard method. Although maximum
allowable emissions from sources other than dryers are not specified,
these sources of air emissions are small compared with those from
potash dryers.
VIII
Monitoring of Stack Emissions
Stack emissions are monitored using Saskatchewan method ACP/31,
which is the standard reference method for source testing
measurement of emissions of particulates from
Table 4 Ambient Air Quality Standards of Saskatchewana
Pollutant 1h 8h 24 h 30 days
Suspended 120 µg/m3 70
particulates
Settleable 2.0 mg/cm2
particulates
Soil index 1.5 COH
units
Potash 0.15 mg K/cm2 or 0.15
mg Cl/cm2
Sulfur dioxide 450 (0.17) 150 (0.06) 30 (0.01)
µg/m3 µg/m3 µg/m
Sulfation 30 mg/100 cm2
Carbon 15 (13) 6 (5)
monoxide mg/m3 mg/m3
Oxidants 160 (0.08) 50 (0.025)
(ozone) µg/m3 µg/m3
Nitrogen dioxide400 (0.2) 200 (0.1) 100 (0.5)
µg/m3 µg/m3 µg/m
Hydrogen 14 (10) 4 (2.9)
sulfide µg/m3 d µg/m3 d
aVolume units (ppm) in parentheses.
bGeometric mean.
cArithmetic mean.
dVolume units (ppb) in parentheses.
 Page 383
stationary sources. This method is similar to the standard particulate
measurement methods used by all North American jurisdictions. Pitot
tubes are used to measure velocities at various points in the stack to
define sampling points. Samples from the stack gas are taken and
analyzed to determine elemental composition. A precisely measured
gas sample is withdrawn isokinetically from the series of impingers.
The weight of the particulate matter and water collected in the train is
measured. The particulate emissions are then expressed as g/dry R m3.
In the past, attempts have also been made to measure fallout to
determine compliance with the ambient standards, which are 0.15 mg
K or Cl/cm2 per month. The limits are normally set in terms of
chloride as well because chloride is considered more toxic than
potassium to the surrounding vegetation. Fallout jars have been
located on elevated posts around each refinery. However, an excess of
natural dust material and of animal activity entering the jar has
prevented any meaningful determination of the amount of fallout from
the refinery's emission sources.
Biological and soil monitoring of air emissions have also been carried
out. These activities are discussed in the next section.
IX
Environmental Impact of Air Emissions
A
Impact on Soil
Given the concern about the refinery dust fallout and its impact on the
surrounding soil and vegetation, a number of studies have been
conducted to determine the impact of the dust fallout on the
productivity of the soil and vegetation. Results from field and
laboratory work have shown that the productivity of soils deteriorates
when the potassium content of the ion-exchange complex exceeds 20
to 25% of the cation-exchange capacity [9]. Because sandy soils have
a smaller cation-exchange capacity than clay soils, sandy soils are
more affected by potash fallout. Soils around the potassium chloride
operations in Saskatchewan are capable of absorbing about 3250 kg
K/ha without a significant impact on their productivity.
Monitoring of potassium accumulation in soils around some of the
potash operations in Saskatchewan was initiated in 1983. Findings
indicate that the potassium content of the cation exchange has been as
high as 20% in sandy soils located close to refineries, while the
chloride levels have been elevated above apparent background levels.
However, at most of the sites monitored, the cation exchange was at
much lower levels with no apparent trend at any location. In 1985 a
new air emissions regulation came into effect, reducing the allowable
emissions levels to less than 25% of the previous values. Despite
reduced emission levels, no downward trend of the salt content in the
soils has occurred thus far.
Emissions rates range between approximately 0.6 and 1.1 mg/m3 dust
per metric ton of product dried. Up to 55% of the emitted dust settles
in the vicinity of the surface complex in an area having a radius of
about 25 km [10]. The highest rate of dust depositions occurs within a
1-km radius of the emission source [11]. The average chloride
deposition on soils within a 5-km radius of the mining/refining
complex is about 0.01 to 0.05 mg/ha·month, which is below the air
quality standard for Saskatchewan. Detectable accumulations of
chloride on soils near the emission source appear to follow a circular
pattern of about 2.5 km radius; the size and shape of the area are
affected by prevailing wind conditions [9].
 Page 384
B
Impact on Vegetation
Air emissions can also affect the vegetation by falling directly on the
vegetation itself. Although foliar application of potassium dust to
plants can cause damage or death, no damage to annual crops in the
vicinity of the potassium chloride refineries has been observed in
Saskatchewan. Further, forage crops and native grasses are considered
more resistant to potassium dust than are annual crops [12].
Damage caused by dust fallout has been evident around each refinery
on the widely scattered stands of aspen (Aspen tremoloides), one of
the common woody species in Saskatchewan. When damaged, the
aspen appears stunted and clubbed because the potassium dust causes
death of the terminal buds formed the previous year. Terminal buds
are responsible for stem growth. The death of terminal buds leads to
the formation of other buds below, resulting in a broomlike
appearance, as illustrated in Fig. 9. Aspens are more susceptible to
damage in the spring, when their buds open. A temporary failure in
the air pollution control equipment at this time would be more
damaging than would one occurring after leaf fall when the trees are
affected little because the buds are dormant.
Damage to aspen increases with humidity [13]. A wet application of
potassium causes more damage than does a dry application. The
damage is also higher with higher relative humidities. Therefore, it is
believed that the damage to vegetation caused by emissions from a
dryer's stack equipped with a wet scrubber will be greater than the
damage caused by emissions from an electrostatic precipitator.
The damage zone around each refinery generally corresponds to the
prevailing winds. However, it is usually confined to a few kilometers
around the mine site. Trees are normally not killed and will return to
normal life form if the level of potassium dust fallout drops below the
critical level.

Figure 9
Impact of potash dust fallout on the form of aspen trees in Saskatchewan.
 Page 385

References
1. A Commentary on the Transportation of Dangerous Goods Act and
Regulations, Osler, Hoskin, and Harcourt, Toronto, Ontario, Canada,
1985.
2. R. A. Heal and K. W. Reid, Acid emissions from potash dryer
stacks at PCS Mining refineries, unpublished report, Environmental
Affairs Section, Potash Corporation of Saskatchewan Mining Limited,
Saskatoon, Saskatchewan, Canada, 1986.
3. H. Haugen, Saskatchewan Mining Association potash dryer survey,
unpublished report, 1983.
4. H. Haugen and McVittie, Sources and control of fugitive emissions
from Saskatchewan potash refineries, unpublished report,
Environmental Committee, Saskatchewan Mining Association,
Saskatoon, Saskatchewan, Canada, undated.
5. R. J. Schutzman, Size distribution of particulate in potash dryer
exhausts, in Potash Technology '83, R. M. McKercher, ed., Pergamon
Press Canada, Toronto, Ontario, Canada, 1983.
6. T. Baumeister, E. Avallone, and T. Baumeiter III, eds., Mark's
Standard Handbook for Mechanical Engineers, 8th ed., McGraw-Hill,
New York, 1978.
7. H. P. Perry, D. W. Green, and J. O. Maloney, eds., Perry's Chemical
Engineer's Handbook, 6th ed., McGraw-Hill, New York, 1984.
8. K. W. Reid, Canadian environmental legislation, Proceedings of the
39th Annual Meeting, Fertilizer Industry Round Table, Atlanta, Ga.,
October 31November 2, 1989.
9. D. A. Rennie, 1983, Fate and impact of potash dust on the soil
environment in Saskatchewan, in Potash Technology '83, R. M.
McKercher, ed., Pergamon Press Canada, Toronto, Ontario, Canada.
10. L. J. Townley-Smith and R. E. Redmann, Canadian Journal of
Botany, Vol. 58, 1980, pp. 26262633.
11. K. M. Pretty, R. R. Sentis, and R. L. Thomas, 1977, The Effects of
Alkali Halides in the Canadian Environment, N.R.C. Publication
15019, Saskatoon, Saskatchewan, Canada, 1977, pp. 3761.
12. R. E. Redmann, 1983, Resistance of plant species to airborne
residual salts from potash refineries, in Potash Technology '83, R. M.
McKercher, ed., Pergamon Press Canada, Toronto, Ontario, Canada.
13. R. T. Hart, 1983, The effect of potash emissions on vegetation, in
Potash Technology '83, R. M. McKercher, ed., Pergamon Press
Canada, Toronto, Ontario, Canada.
 Page 387

28
KCl Handling, Storage, and Transport
Elmar L. Goldsmith
Kalium Canada, Ltd.
Regina, Saskatchewan, Canada
I
Introduction
The main environmental consideration regarding the distribution of
the product from the manufacturer to the end user is the fugitive dust
generated during various stages of product handling, storage, and
transport. Without appropriate control measures, there can be
excessive dust formation and emissions from the various stages of
storage, handling, and transportation systems. Following are some of
the measures that have been developed in attempting to minimize dust
generation in these systems.
II
Handling
Reducing the severity of impact on the potassium chloride particles,
the amount of dust generated is decreased; therefore, product
dustiness is reduced. As shown in Fig. 1, the size degradation or
breakage of the product is directly proportional to the drop height and
to the type of impact material. For the 15-m drop height the product
suffered a 50% breakage when dropped on a metal surface and only
about 10% and 15% when dropped on a conveyor belt and a
potassium chloride impact surface, respectively. The latter situation
would occur when filling a bin or a warehouse. From the same graph
and utilizing a metal impact surface, product degradation can be
reduced from 50% to 20% by reducing the drop height from 15 m to 3
m, respectively.
Based on the aforementioned principle, the industry has adopted steps
to reduce the breakage. For example, when using belt conveyors, the
impact of conveyor housing on potassium chloride particles can be
reduced by (1) running the belt at minimum speed consistent with the
required transfer rate, (2) by physically moving farther back the
conveyor housing when room permits, and (3) by lining the housing
with an impact-absorbing material such as discarded old conveyor
belting. Moreover, vertical chutes through which the product falls
freely are kept to a minimum. When chutes are required, various
 Page 388

Figure 1
Effect of drop height and impact material on the breakage of crystalline potash.
 Page 389
chokes or baffles can be installed to keep the chute partially filled or
slow down the particles to minimize impact.
To minimize product breakage, certain conveying equipment is
avoided whenever possible. Screw and drag chain conveyors are
particularly impact generators and may significantly degrade the size
of the potassium chloride particles because of the grinding action.
Where screw conveyors must be used, it is critical to keep the
bearings of the screw in good repair. Worn bearings allow the screw to
''bounce," greatly increasing dust generation.
The single most important measure that potassium chloride producers
employ to control product dustiness is the use of dust control reagents.
These are mostly mineral oils in various forms, from common diesel
oil to proprietary dust control fluids. Producers also apply a mixture of
mineral oil and anticaking reagents such as fatty acid amine to achieve
the combined effect of preventing both caking and dustiness of the
product. At least one Saskatchewan producer applies a wax to its
products for dust control. Other reagents, such as polyglycols, have
been used by potassium chloride producers outside Saskatchewan [1].
Dust control reagents are normally applied by spraying onto the
finished product in various conveying equipment. In some cases,
particularly when a mixture of oil and anticaking reagent is added to
the product, the reagent is applied before product is transferred to the
storage at the production site. In other cases, the anticaking reagent is
applied onto the product, just before the product is loaded for
shipment into railcars or trucks.
Application rates for dust control reagents range between 0.05 and 0.3
wt %. The effect of dustiness on compacted potassium chloride
treated with wax is shown in Fig. 2. Dust reagents work by sticking
the dust particles to the larger particles. This way only a few of the
dust particles become airborne during handling.
Potassium chloride product may also be retreated with dust control
reagents at ports as product is transferred from railcars to ships. In
Vancouver, British Columbia, Canada, wharf operators often apply
antidust oil to the product. Otherwise, harbor authorities may shut
down loading facilities when airborne dust is visible, which may
violate local clean air regulations. Clean air regulations for dust are
also enforced at Belgian port facilities where fertilizers are
transloaded. The Greenpeace organization has been instrumental in
having clean air regulation enforced at these ports [2].
III
Storage
Potassium chloride products are stored at the producing mines in
warehouses of 20,000 to 200,000 metric ton capacity. These products
are also stored at off-site warehouses, which producers may lease or
own. Products are also stored at customers' sites. These customers are
normally fertilizer blenders/appliers. Off-site and customer storage
vary in capacity from very small [100 metric tons (mt) or less] to very
large (a few hundred thousand metric tons with many different
fertilizer products stored simultaneously).
Dust is the primary environmental concern in storage. Dust may be
present, particularly during filling or emptying of the storage facility.
Many off-site and customer facilities are located in or near urban
centers, and clean air regulations are enforced with respect to
contamination of the area with nuisance dust.
Steps to minimize dust problems in warehouses include minimizing
drop heights, called soft-drop systems, and additional treatment of
particularly dusty products with dust
 Page 390

Figure 2
Suppression of dust with wax.
 Page 391
control oils. Drop heights can be minimized by installing inclined
chutes from the center drop (i.e., fill points) in the warehouse as
shown in Fig. 3. The chutes can be rotated around the warehouse and
shortened as the warehouse fills. Soft-drop systems include ladderlike
devices in which the product drops in a series of small steps rather
than a single large one.
Occasionally, potassium chloride is stored outside on a concrete or
asphalt pad designed in accordance to local regulations governing
runoff of contaminated rainwater. Normally, outside potassium
chloride piles are covered with plastic or tarpaulins to minimize rain
and wind damage.
IV
Transportation
Potassium chloride products are transported in trucks (typically, 20 mt
per truck trailer), in railcars (typically, 90 mt per hopper car), and over
water (in barges or ships of various sizes). The environmental
concerns in the transportation of potassium chloride products are
limited to leakage of product from loaded trailers or railcars. Such
product leakage may damage the vegetation at the side of roads and
railways. However, this potential problem has been virtually
eliminated in Canada and the United States through rigorous
maintenance and inspection of the transport equipment.
In addition to inspection by rail and trucking companies, potassium
chloride producers also carry out their own inspection of every loaded
railcar and truck to ensure that bottom gates do not leak product. They
also check to ensure that top-loading hatches do not leak to avoid the
entry of rainwater. Water entering the trailer or railcar will ultimately
dissolve potassium chloride and will leak as a solution. Most railcar
operators even sweep spilled product off the top of the railcars after
loading to prevent product from falling or blowing from the car during
transit.
In the event of a catastrophic failure of a truck or a railcar, such as an
accident or a derailment, spilled product is cleared up by removing the
spilled material and any affected soil. In general, most of the spilled
material can be salvaged for fertilizer use. When desired, the area of
the spill can be flushed thoroughly with water to dilute the potassium
chloride to levels below where local damage to vegetation may be
done.
For transportation over water, in the event of a vessel sinking, local
salinity levels would initially increase but decrease with time because
of flushing caused by currents, tides, and the like.

Figure 3
Use of chutes to minimize product dust.
 Page 392

References
1. L. I. Skvirski, S. N. Titkov, A. M. Somenov, et al., Dust
suppression for water soluble salts, Author's certificate 1,096,272,
November 17, 1982.
2. L. I. Skvirski, A. A. Crristyakov, and Z. L. Kozel, Production of
non-dusty and non-caking potash fertilizers, paper presented at Kali
'91, 2nd International Potash Technology Conference, Hamburg,
Germany, May 2629, 1991.
 Page 393

29
MAP, DAP, and NPK Fertilizers:
Pollution Control Practices in Japan
Jumpei Ando
Chuo University
Tokyo, Japan

Ichiki Hamada
Coop Chemical Company
Tokyo, Japan
I
Overview of Pollutants and Control
MAP (monoammonium phosphate) and DAP (diammonium
phosphate) are produced from ammoniated slurry of wet-process
phosphoric acid by granulation, drying, and cooling. NPK products
are obtained by adding a potassium salt during the granulation. The
production process is called slurry granulation. Various grades of NPK
products are also manufactured by blending and granulation of
powdery MAP, DAP, or superphosphates with nitrogen compounds
such as urea and ammonium sulfate and potassium chloride or sulfate.
The process is called powder blending granulation.
Major air pollutants due to the production are dusts, ammonia,
fluorine, SOx, and NOx; major water pollutants are fluorine,
suspending solids, nitrogen, and phosphorus. In Japan, pollutants are
controlled in the following three ways: (1) standards by the central
government, (2) ordinances by the prefectures and major cities, and
(3) agreements between each factory and prefecture or city. For
example, dust is controlled by the standard below 200 or 300
mg/N·m3 (milligram per cubic meter at 0°C, 1 atm pressure)
according to the gas volume. Dust is often controlled by the
ordinances at 100 or 150 mg/N·m3, and at even lower values by
agreements.
Dusts are removed by a bag filter or a wet scrubber. A small amount
of ammonium chloride fumes is often contained in the dryer waste gas
and is often visible even when the particulate content meets the
regulations. Some local governments ask the fertilizer factory to
reduce the fumes. A mist cottrell (wet electrostatic precipitator) is
used to reduce the fumes.
Ammonia is contained in the waste gas leaving an ammoniator
granulator or neutralizer of phosphoric acid. It is usually recovered by
scrubbing with phosphoric or sulfuric acid and returned to the
granulator. Fluorine compounds are contained in the gases from the
granulator and dryer but are usually not controlled because the
concentrations are low.
 Page 394
The waste gas from a dryer contains some SOx and NOx derived from
the fuel to produce the hot air. The allowable SOx emissions (N·m3/h)
are calculated by the following equation using a k value between 1.17
and 17.5, assigned by the central government for each region:

where He is the effective stack height, in meters. The allowable

amount is calculated from the k value, stack height, gas temperature,
and so on. The allowable SOx concentration is calculated from the
amount and gas volume and usually ranges from 10 to 30 ppm in a
dryer gas that contains a large amount of air. The regulation is usually
met by the using low-sulfur fuels. When a high-sulfur fuel is used,
scrubbing may be needed to meet the regulation.
NOx is controlled by the standards at 150 or 250 ppm according to the
gas volume. In certain regions with stringent local regulations, NOx is
controlled below 100 ppm. Waste gases from ammonium
phosphatebased mixed fertilizer plants usually contains much less
NOx. Major pollutants to be removed from wastewaters of fertilizer
plants are fluorine and suspending solids. Heavy metals occur in small
amounts and give no problems. Nitrogen and phosphorus in the
wastewater are not controlled at most factories bordering the ocean.
Recently, nitrogen and phosphorus control has begun at factories
facing a bay or inland sea.
II
Removal of Air Pollutants
A
Ammonia
Ammonia is released mainly from a neutralizer of phosphoric acid and
an ammoniater granulator. With powder blending granulation, almost
no ammonia is released except when ammonia is used together with
sulfuric or phosphoric acid to promote the granulation. Ammonia in
the waste gases is recovered mainly by scrubbing with phosphoric or
sulfuric acid, and at some plants with water. The liquor leaving the
scrubber is returned to the neutralizer or granulator, eliminating the
wastewater.
B
Fluorine Compounds
Fluorine is present in the waste gases in the form of SiF4, and HF,
which undergo only limited removal by scrubbing with phosphoric or
sulfuric acid for ammonia recovery. Fluorine is removed by water
scrubbing, if necessary. The scrubber liquor contains fluorine in two
forms, HF and H2SiF6. The liquor is usually treated with lime to
precipitate the fluorine.
C
Dust
Gases from the granulator and dryer containing both dust and
ammonia are usually subjected to wet scrubbing to recover both. The
dusts in the gases from a dryer and a cooler with little ammonia are
often removed by a bag filter (fabric filter) and returned to the
granulator. The bag filter has the advantages of a removal efficiency
of over 99% and yields no wastewater. A possible problem with the
bag is clogging by dust, due mainly to moisture condensation. When
the bag is cold at startup, hot air is passed through the bag for a while
to warm it and thus prevent moisture condensation.
 Page 395
D
Ammonium Chloride Fumes
Ammonium chloride is formed by the reaction of potassium chloride
with ammonia salts in the granulator and dryer.

Reaction (1) may proceed nearly completely when finely pulverized

chloride and sulfate are used, while reactions (2) and (3) proceed to a
limited extent due to the equilibrium. DAP hardly reacts with KCl to
form ammonium chloride.
Ammonium chloride formed by those reactions tends to vaporize
above 130°C, while pure crystalline ammonium chloride vaporizes
little below 150°C to give the fumes. The fumes consist of fine
particles smaller than 1 µm and are hardly removed by wet scrubbing
(Fig. 1). A bag filter may remove many of the fumes but not enough.
A wet electrostatic precipitator (mist cottrell) is often used to remove
the fumes (Fig. 1).
The formation of ammonia chloride can be prevented by formulating
with DAP, urea, and KCl, or by replacing KCl with potassium sulfate.
Actually, a minor amount of ammonium chloride may still form
because DAP usually contains a small amount of monoammonium
phosphate, while fertilizer-grade potassium sulfate contains a little
chloride. Lowering the drying temperature reduces the vaporization of
ammonium chloride but lowers the drying efficiency. Cocurrent
drying may be preferred to countercurrent drying to reduce
vaporization.
III
Examples of Air Pollution Control
A
DAP Plant, Company A

The plant produces 500 tons/day of DAP by a spray dryer process.

Wet-process phosphoric acid is sprayed from the top of a tower and
reacted with gaseous ammonia introduced

Figure 1
Particulate removal efficiency. ESP:
electrostatic precipitator, BAG: bag filter.
 Page 396
from the lower part of the tower. Powdery DAP is produced by an
exothermic reaction. The product DAP has a crystal structure different
from that of usual DAP. The structure changes into the normal form
slowly during storage and rapidly when the product is moistened for
granulation.
The gas leaving the spray tower (about 125,000 N·m3/h) contains 700
to 1000 ppm NH3 with a considerable amount of DAP dust. A water
scrubber and a wet ESP (mist cottrell) were installed in 1975 to
recover nearly all of ammonia and dust. All of the water leaving the
scrubber containing the ammonia and DAP as well as the water
leaving the wet ESP is sprayed into the spray tower to adjust the
reaction temperature. The scrubber and wet ESP cost 300 million yen
in 1975. The recovered DAP and ammonia account for about 50
million yen yearly, providing an overall economic advantage.
B
Chiba Factory, Japan Union Fertilizer Co., Factory A
This factory produces NPK products using ammonium chloride, the
by-product of sodium carbonate production from sodium chloride,
ammonia, and carbonic acid gas. The factory is located in Ichihara
City, Chiba Prefecture, which has quite stringent regulations on
pollution control. An ammoniated slurry of phosphoric acid, mainly in
the form of monoammonium phosphate, is added with ammonium and
potassium chlorides, usually with magnesium hydroxide, and then is
granulated and cooled to give a 121614 product with 4MgO, one of
141414, and so on. The dryer has a capacity of 70 tons/h and gives
about 90,000 N·m3/h of a waste gas containing about 12 g/N·m3 of
dust with fumes of ammonium chloride.
The plant started operation in 1963 and initially treated the gas using a
multicyclone and a two-stage water scrubber and reduced the dust
below 0.1 g/N·m3 to meet the regulations of the central government,
at 0.2 g/N·m3, and that of the Chiba Prefecture, at 0.1 g/N·m3 (Fig.
2). However, the ammonium chloride fumes were visible. Since a
reduction in fumes was desired by the city, a wet ESP (mist cottrell)
was installed in

Figure 2
Waste gas cleaning system at the Chiba factory, Japan Union
Fertilizer Co. Figures show particulate content in g/N·m3.
 Page 397
1975, which has reduced the fumes sufficiently to make them
invisible. The current prices for a multicyclone, scrubber, and wet ESP
can be estimated at 30 million yen, 100 to 150 million yen, and 250
million yen, respectively.
A high-sulfur oil is used for drying, so the dryer gas contains about
100 ppm SOx. The SOx is almost completely removed in the scrubber
because the magnesium hydroxide dust and ammonia in the dryer gas
help the removal. The wash water of the wet ESP is used for
scrubbing and the liquor leaving the scrubber is used for granulation,
yielding no wastewater (Fig. 2). The total cleaning system has been
operated successfully.
Successful operation of the wet ESP is due to good maintainance as
well as the low dust load. The electrodes are washed with a water
spray three times a day for about 5 min each time. Once a year when
the total system, including the production plant, is subjected to a
scheduled shutdown for maintenance, and the electrodes are cleaned
by a jet cleaner. A rectifier, heater, and hot air fan were replaced after
10 years of operation. Electrodes and a controller panel have not yet
been replaced.
C
Factory B, Company B (Slurry Base)
This factory produces 84,000 tons/year of granular MAP and various
grades of NPK fertilizers based on ammoniated slurry of phosphoric
acid, urea, ammonium sulfate, potassium chloride, by-product
magnesium oxide, and so on, using an ammoniator granulator. Figure
3 shows a gas cleaning system of a type popular for mixed fertilizer
plants in Japan. An example of the gas volume and composition for
the production of MAP is shown in Table 1.
The dryer gas, at about 51,000 N·m3/h and 80 to 130°C, is passed
through a bag

Figure 3
Gas cleaning system (factory B).
 Page 398
Table 1 Example of Gas Composition in
Each Part of Fig. 2a
A B C
Ammonia (g/N·m3) 11.5 0.2 0.03
Fluorine (mg/N·m3) 34 3
aGas volume 18,000 N·m3/h.

filter to remove more than 99.5% of the dust. A typical example of the
composition of the treated gas and regulations is shown in Table 2.
The cooler waste gas at 50 to 60°C is subjected to dust removal by a
bag filter. The life of the bags for the dryer and cooler gases is 1 to 2
years.
D
Factory C, Company C (Slurry Base)
This factory manufactures 26,000 tons/year of products, mainly
ammonium phosphate containing magnesia, using phosphoric acid,
ammonia, and by-product magnesia. The gas leaving the ammoniator
granulator at about 42,000 N·m3/h, contains amonia with 20 g/N·m3
of dust and 10 mg/N·m3 of fluorine, and is subjected to scrubbing
with phosphoric acid (Fig. 4).
Nearly all of the ammonia and dust are recovered and returned to the
granulator. The gas is then subjected to water scrubbing. The gas
leaving the water scrubber contains 70 mg/N·m3 of dust and less than
3 ppm fluorine. The dryer gas at 47,000 N·m3/h is treated by a
cyclone and then by a venturi scrubber. A typical gas composition is
shown in Table 3.
Most of the wastewater leaving the venturi scrubber is used for
granulation. Actually, a wet ESP is placed downstream of the venturi
to remove ammonium chloride fumes when potassium chloride is
used to produce various NPK grades. The ESP has not been operated
since the plant stopped using chloride. The venturi may not be best
suited for the current products. The gas from the cooler, at 7000
N·m3/h and 70 to 70°C contains 400 mg/N·m3 of dust is treated by a
bag filter to reduce dust below 20 mg/N·m3.
E
Factory D, Company D (Powder Blending Granulation)
The factory produces various NPK grades using powdery DAP, MAP,
urea, ammonium sulfate, potassium chloride, and so on. The dryer
uses a low-sulfur oil for fuel, and gas at about 10,000 N·m3/h leaving
the dryer is treated by a bag filter to remove more than 99.5% of the
dust. The treated gas contains 3 ppm SOx, 70 mg/N·m3 of dust, and
15
Table 2 Composition of Dryer Gas After Dust Removal and
the Regulations
SOx Dust NOx Fluorine Oz
(ppm) (mg/N·m3) (ppm) (mg/N·m3) (%)
Waste gas 4 67 27 Below 1 18.5
Regulations 12a 150 150 NRb
ak-value control at k = 6.0.
bNo regulation.
 Page 399

Figure 4
Gas cleaning system (factory C).
ppm NOx with 19.5% O2. The gas from the cooler is also treated by a
separate bag filter for more than 99.5% dust removal.
F
Factory E, Company E (Powder Blending Granulation)
This factory produces various grades of NPK fertilizers using MAP,
DAP, urea, ammonium sulfate, potassium chloride, and so on. Small
amounts of sulfuric acid and ammonia are normally used to promote
granulation. The gas leaving the granulator is treated by a water
scrubber mainly for ammonia recovery (Fig. 5). The gas leaving the
dryer, 22,000 N·m3/h, is first treated by a bag filter and then
introduced into the water scrubber (Fig. 5), All of the liquor leaving
the scrubber is used for the granulation. The composition of the gas
after the scrubbing is shown in Table 4. A considerable portion of SOx
in the dryer gas is caught by the scrubber together with ammonia and
dust. The regulations for the factory are relatively relaxed and are well
met by the cleaning process.
Table 3 Example of Dryer Gas Composition
SOx Dust NOx F O2
(ppm) (mg/N·m3) (ppm) (mg/N·m3) (%)
After cleaning 11 70 30 3 19.0
Regulations 25a 150 250 NRb
ak-value control at k = 14.5.
bNo regulation.
 Page 400

Figure 5
Gas cleaning system (factory E).

IV
Water Pollution Control
A
Major Water Pollutants and Regulations
Wastewaters at MAP, DAP, and NPK plants are derived mainly from
water scrubbing of the waste gases. Waste gases containing ammonia
discharged from a phosphoric acid neutralizer and an ammoniator
granulator are normally subjected to scrubbing with phosphoric or
sulfuric acid for ammonia recovery and the scrubber liquor is fed to
the neutralizer or granulator without being discharged.
At powder blending granulation plants, the waste gas is normally
subjected to dust removal by a bag filter, yielding no wastewater.
Some fertilizer plants have water scrubbers yielding wastewaters that
require treatment. The wastewater from the fertilizer plant is often
treated together with the wastewater from a wet-process phosphoric
acid plant or other plants.
Table 5 shows the effluent standards by the central government and
examples of agreements and targets. COD (chemical oxygen demand),
BOD (biological oxygen demand), and SS (suspending solids) are
actually controlled by the agreements, which are more stringent than
the standards. SS are controlled at every factory, while the reduction
of COD and BOD are needed at factories using organic additives. The
pH of the wastewater is controlled between 5.8 and 8.6.
Table 4 Composition of Gas After Cleaning
SOx Dust NOx O2
(ppm) (mg/N·m3) (ppm) (%)
After cleaning 4 120 35 19.5
Regulations 11a 300 250
ak-value control at k = 17.5
 Page 401
Table 5 Effluent Standards and Examples
of Agreement and Target on Water
Pollutants (mg/L)
Standards Agreement Target
DAa Max. DA Max. DA Max.
COD 25 35 10 20
BOD 30 40 30 40
SS 60 80 20 50
F 15 15
N 60b 120b 30
P 8b 16b 3.4 8
aDaily average.
bFor waters going into lakes and ponds.

Effluent standards for nitrogen and phosphorus have been developed

for wastewaters going into lakes and ponds. Most of the Japanese
fertilizer plants are close to the sea and have no standards for nitrogen
and phosphorus. The desired standards for water going into the sea are
being studied by the central government and will be determined in the
near future. Currently factories facing inland seas and bays have
targets on the effluent concentrations made through discussions with
local authorities. An example of a target is shown in Table 5. Table 6
shows the effluent standard and an example of a ordinance for heavy
metals and a cyanate. Removal of these is usually unnecessary
because of their low content in wastewaters.
B
Chemistry of Fluorine and Phosphate Removal [1,2]
Slaked lime is generally used to precipitate fluorine and phosphate in
wastewater. Fluorine is present in two forms in wastewater, fluoride
(F) and fluorosilicate (SiF62 ). By lime addition, CaF2 is precipitated
from both these ions. CaF2 is slightly soluble in water to give 8 mg/L
F, which is lower than the effluent standard, 15 mg/L. However, the
filtrate of CaF2 precipates often contains larger amounts of fluorine.
The CaF2 precipitated normally consists of very fine crystals smaller
than 0.1 µm, which cannot be precipitated or removed by filtration
unless coagulated to form much larger particles. A coagulant is often
used to promote removal.
Aluminum dissolved in acidic wastewater is precipitated by lime
addition to form gelatinous aluminum hydroxide, which aids in the
coagulation of fine CaF2 particles. SiF62 ion, when reacted with lime
to precipitate CaF2 , gives gelatinous silica, which also aids in
coagulation to some extent. The fluorine in the treated water often
exceeds 15 mg/L, due primarily to the effects of various impurities.
Fluorine is not well removed
Table 6 Effluent Standards and Ordinances for
Heavy Metals and Cyanate (mg/L)
Material Hg Cd Cr6+ As Pb Zn CN
Standard 0.005 0.10 0.50 0.50 1.0 5 1.0
Ordinance 0.005 0.05 0.35 0.35 0.70 5 0.70
 Page 402
when the wastewater is neutral or alkaline (containing sodium,
ammonium, etc.) because lime is not well dissolved. For such water,
calcium chloride is useful to remove fluorine. Otherwise, the water is
acidified by hydrochloric acid prior to the addition of lime, so that the
lime dissolves sufficiently to precipitate fluorine.
At some fertilizer plants, a small amount of boron is added to a mixed
fertilizer as a micronutrient. In wastewater containing boron, fluorine
is precipitated very little by the addition of lime because soluble BF4
ion is formed which is very stable. Wastewater containing boron
should not be mixed with wastewater containing fluorine. By adding
lime to a wastewater containing both phosphate and fluorine,
fluorapatite Ca5(PO4)3F is precipitated, which is much less soluble
than CaF2. The apatite usually consists of very fine crystals smaller
than 0.02 µm, which tend to coagulate to form gelatinous material.
Theoretically, both phosphorus and fluorine can be reduced below 1
mg/L by precipitating in the form of fluorapatite. Such low
concentrations, however, are rarely attained at commercial plants due
to the effect of various inpurities that influence the coagulation
precipitation. For example, magnesium tends to increase fluorine and
phosphate concentrations in the liquid phase. The weight ratio of P to
F in fluorapatite is about 5:1. When acidic wastewater, with a lower
ratio, is added with lime, both CaF2 and fluorapatite are precipitated.
In this case the fluorine content of the treated water is influenced by
CaF2.
For water rich in fluorine, the following two-stage process may be
useful: (1) lime addition to pH 3 to 4, to precipitate most of fluorine in
the form of CaF2, which is separated; and (2) further lime addition to
pH 7 to 8, to precipitate the remaining fluoride in the form of
fluorapatite. In this way, fluorine and phosphorus may be reduced
below 10 and 5 mg/L, respectively. When the water is rich in
phosphate and lean in fluorine, a mixture or solid solution of
fluorapatite and hydroxyapatite Ca5(PO4)3OH is formed by lime
addition. In this case, both fluorine and phosphorus may be kept
below 5 mg/L at a pH of about 9.
At a pH above 9, hydroxyapatite is more stable than fluorapatite,
resulting in the release of fluorine from fluorapatite into the solution:

Optimum methods of removal should be established for each

wastewater, since both fluorine and phosphorus removal are variously
influenced by impurities, reaction conditions, and other factors.
C
Example of Wastewater Treatment:
I
At this factory, wastewaters from both a wet-process phosphoric acid
plant and a mixed fertilizer plant (Table 7) are treated using the
system shown in Fig. 6. The water from the acid plant is first added
with slaked lime in two stages to remove mainly fluorine in
Table 7 Wastewaters to Be Treated
Amount Contents (mg/L)
Source (m3/h) P2O5 F SS pH
Acid 100 11001300 500 23
plant
NPK 1 500 500 250 78
plant
 Page 403

Figure 6
Water treatment system (example I)
the form of CaF2. The slurry is treated in a thickener. The overflow
from the thickener is mixed with the water from the fertilizer plant
and added with slaked lime in two stages to remove most of fluorine
and phosphate. The slurry is then treated in another thickener.
Table 8 shows the composition of the treated water and the
regulations. The pollutants have been reduced to meet the regulations.
Phosphorus has no regulation because the wastewater goes into an
open sea but is usually reduced to about 16 mg/L, the standard for
water going into lakes and ponds. Nitrogen in the wastewater is also
not controlled but is usually in low concentrations.
D
Example of Wastewater Treatment:
II
This factory treats 1500 m3/h of the wastewater from fertilizer plants
together with about 5000 m3/h of the wastewater from chemical
factories. The flowsheet of the treatment system is shown in Fig. 7.
The compositions of the waters before and after treatment are shown
in Table 9. The pH of the wastewater before treatment ranges from 3
to 5. Usually, acid is added to lower the pH before lime addition. The
water is subjected to neutralization with slaked lime, precipitation,
thickening using a coagulant, and dewatering with a filter process.
The pH of the treated water is between 6 and 8.
About 90% of both fluorine and phosphorus are removed. Nitrogen
removal is not necessary because of the low concentrations. A portion
of the nitrogen seems to be removed with the sludge. The treatments
satisfy the standards, agreements, and targets shown in Tables 5 and 6.
Heavy metals are in low concentrations and require no treatment.
Table 8 Composition of Treated Water
Contents (mg/L)
pH F Pb As COD SS
Treated water 7.7 7.1 NDa 0.07 6.8 9.9
Standardb 5.09.0 15.0 1.0 0.50 25 30
aNot detected.
bFor waters going into sea.
 Page 404

Figure 7
Water treatment system (example II).
Table 9 Composition of
Wastewaters (mg/L)
Before After
treatment treatment
Usual Max. Usual Max.
COD 30 50 510 68
SS 500700 900 1020 50
F 5080 100 510 15
N 1020 50 1020 30
P 5070 100 48 10

References
1. J. Ando, M. Mori, et al., Removal of fluorine from wastewaters (I,
II, III), Gypsum and Lime, No. 133, 1974; No. 140, 1976; No. 146,
1977 (in Japanese).
2. J. Ando, E. Toya, and S. Matsuno, Effects of fluorine upon
coagulation precipitation of phosphate by lime addition, Gypsum and
Lime, No. 189, 1984 (in Japanese).
 Page 405

30
MAP, DAP, and NPK Fertilizers:
Pollution Control Practices in the United States and
Europe
Frank P. Achorn
SEME Inc. Chemical Process Consultants
Florence, Alabama
I
Introduction
The process for each of these products involves different types of
pollution problems. However, each process has the following
pollutants, which are common to all three processes [1,2]:
1. Dust: solid particles that are generated by handling, crushing,
grinding, and decrepitation. Dust does not tend to flocculate except
under electrostatic forces. Dust does not diffuse in air but settles by
gravity. Most fertilizer dust varies widely in particle size.
2. Fume: solid particles generated by condensation from gaseous
states. Typical fertilizer fumes are ammonium chloride and
ammonium fluoride.
3. Mist: suspended liquid drops by condensation from gaseous to
liquid state or by breaking up a liquid into a dispersed state (i.e., as in
splashing and foaming).
4. Gas: normally, formless fluid that occupies the space of an
enclosure and can be changed to liquid or solid only by the
combination of increased pressure and decreased temperature.
5. Vapor: gaseous states of a solid or liquid that can be returned to the
original state either by increasing pressure or decreasing temperature.
Typical vapors in the fertilizer industry are ammonia and fluoride
compounds.
II
Dry Collectors
In all processes for the production of MAP, DAP, and NPK, there are
larger dust particles that can easily be removed by dry cyclonic dust
collectors such as that shown in Fig. 1. In this low pressure drop,
cyclone centrifugal forces push particles in the gas stream toward the
outside walls of the collector, where they descend under the influence
of
 Page 406

Figure 1
Cyclonic dust collector for solidsgas separation.
gravity and the descending vortex of the gas stream. Experience has
shown that collection efficiency is best in these types of collectors
when the particle sizes are greater than 40 µm. These are types of dry
collectors usually used for the largest particles of dust normally
encountered in fertilizer plants. The pressure drop across these types
of collectors is usually 0.2 to 0.7 kPa.
A few companies do use higher-efficiency centrifugal collectors
which have pressure drops in the range 0.7 to 1.2 kPa and can
efficiently remove particles as low as 10 µm. One type of high-
efficiency centrifugal collector has several smaller collectors in a
single unit and is sometimes referred to as a multicyclone unit. Most
companies prefer to use a low-cost, low-pressure collector of the type
shown in Fig. 1. The exit gases from these cyclone collectors are
either filtered in fabric collectors (bag filter) or scrubbed in a wet
scrubber.
One problem with dust collectors is the devices used to steal their
discharge end. Usually, the exhaust blower is located downstream
from the collector so that its blades will not be coated with excessive
amounts of dust, which in turn will cause excessive vibrations of the
blower. With this equipment arrangement, the cyclone collectors are
operated under negative pressure. If the discharge is not well sealed,
air will be sucked into the discharge and will hinder the flow of
materials from the collector and interrupt the cyclonic action of the
collector. In turn, this will cause an overload of the downstream
 Page 407
scrubbing equipment. Several devices have worked well to seal these
collectors. One uses a motor-driven star feeder; other companies use
molded rubber discharge tubes, such as that shown in Fig. 2. Still
other companies use a simpler type of sealing device, such as the one
shown in Fig. 3.
To ensure that the dust load into the collectors is kept at a reasonable
level, the air velocity through the dryers and coolers should not be
excessive. Figure 4 shows a typical ammoniation granulation plant for
a NPK mixture. It shows a plant that has a rotary ammoniator
granulator, a rotary dryer, and a rotary cooler. These drums are usually
common to MAP and DAP plants also. Tests have shown if the gas
velocity through the dryer and cooler exceeds 2.0 m/s, the dust
loading to the cyclone collectors will be too high for economical
design of the collectors. Some companies overdesign the exhaust
system so that they can realize higher velocities and increase
efficiencies through the cyclones. When this is done, an extra fresh-air
inlet duct is installed such as that shown in Fig. 5. The fresh-air inlet
is generally used to suck dust from screening and crushing operations.
With this arrangement it is possible to have optimum velocities in the
dryer, cooler, and cyclone collectors.
Final cleaning of exit fumes is usually done either by wet scrubbers or
by passing the fumes through closely woven filter cloth (bag filters).
These collection devices are
 Figure 2
Dust-tight cyclone with closure-molded rubber seal.
 Page 408

Figure 3
Dust collector seal.
efficient, reliable units that can remove particulates from the exit
fumes of particle size 2 µm and larger to less than 0.5 µm. They may
be classified in terms of the three ways they are cleaned: intermittent,
periodic, and continual. A typical intermittent type is shown in Fig. 6.
It normally has fabric bags or envelopes and the unit must be shut
down regularly to remove the dust cake. Its use is good for plants that
can readily be shut down, such as bulk blending and other dry mix
plants. It can also be used for bagging operations.
The periodically cleaned collector, such as the one shown in Fig. 7, is
a series of intermittent compartments connected by inlet and outlet
manifolds. On a predetermined basis, one section of the collector is
removed from service by closing the air dampers, which isolates it
from the dirty airstream. Then a reverse airflow is passed through the
bags to clean them and the section is returned to service. Five timers
are usually set so that each section is cleaned at 90-min intervals. The
timers are adjustable if the cleaning interval needs to be changed. In
Fig. 7, compartments 1, 2, and 3 are in service while compartment 4 is
being cleaned. This type of exit gas filter is the one most frequently
used by the fertilizer industry.
A few companies use bag filters that are cleaned continuously during
operations. Cleaning is usually done with a reverse jet of air from
traveling blow rings that move continuously up and down the bags.
Figure 8 shows this type of collector. The blow-ring
 Page 409

Figure 4
Typical state-of-the-art ammoniation-granulation plant.
 Page 410

Figure 5
Cyclone gas velocity control.

Figure 6
Fabric dust collector cleaned by vibrator.
 Page 411

Figure 7
Fabric dust collector with multiple sections.
directs air at about 0.2 to 0.4 kPA back through the fabric and dust
falls into the hopper at the bottom of the collector. Fabricators of these
types of filters report that a much higher air-to-cloth ratio can be used
with this type of cleaning system.
Advantages of Fabric Collectors
1. Fabric collectors produce a dry product that can be recycled without
adversely affecting the water balance and thus the production rate.
2. Their efficiencies are excellent and their plumes are usually smaller
than those from wet scrubbers.
 Figure 8
Fabric dust collector with continuous air cleaning.
 Page 412
3. There are very few problems with entrainment losses.
Disadvantages of Fabric Collectors
1. Fabric collectors will not capture gases such as ammonia fluoride.
2. The inlet gases in the collector must remain above the critical
humidity of the particulate in the gas stream.
3. There must be negative pressure throughout the system.
4. The filament in the bag should be kept below its melting point.
Figure 4 shows an ammoniation granulation plant that employs a
completely dry dust collecting system. In this system the ammoniator
granulator and cooler discharge their gases into the drier at the feed
end. Within the dryer, the gases are heated so that the relative
humidity of the gases exiting the dryer are well above the critical
relative humidity (CRH) of the particulate in the gas stream. Critical
relative humidity is defined as the relative humidity at which the
particulate starts to absorb moisture from the airstream. If the gas
cools too much in the bag filter, the particulate becomes wet. This in
turn will cause wet mud to collect on the walls of the bags, which in
turn causes the fabric's pores to become plugged and will result in
insufficient airflow for plant operation. Safety devices should be
installed that will cause the burner of the dryer to be shut down when
the temperature exiting the dryer approaches the melting-point
temperature of the fabric in the bags. The fabric chosen must be one
that does not react with the chemicals in the exit gas stream.
One major problem with this type of dust collector is that outside air
leaks into part of the system after the dryer. When leaks occur, the
cool air cools the main exit gas stream so that its relative humidity is
higher than the CRH in the gas stream, which causes mud formation
in the collector. Another problem has been insufficient negative air
pressure in the rotary vessels and the dust collector. This problem can
be avoided by installing an exhaust blower of sufficient size at the gas
discharge end of the collector. When this is done, there is an even
flow of exit gases from the drums and the collector. The dry material
is usually emptied by a screw conveyor that is designed to remain full.
This keeps the discharges of the hoppers in the collector sealed from
the entry of outside cool air.
III
Wet Scrubber
In the DAP and MAP processes, gases exit from the dryer, cooler,
granulator, and preneutralizer. All contain ammonia and in some
instances, fluorine compounds which cannot be removed by dry
collectors. Figure 9 shows a typical DAP plant. This plant uses two-
stage scrubbing and a pipe reactor. However, the common practice is
to use a tank prereactor instead of the pipe reactor. The exit gases
from the prereactor and ammoniator granulator are first scrubbed in
the first-stage scrubber, and the exit gases from the first stage are
scrubbed in the second-stage scrubber. The scrubbing medium is
usually filter-grade phosphoric acid (28 to 30% P2O5), which supplies
part of the product's P2O5 (usually about 50% of the total P2O5). The
acid flows in a direction counter to the exit gas flow. Therefore, fresh
acid is added to the second stage, and partially neutralized acid from
this stage is added to the first-stage scrubber. Slurry from this scrubber
is pumped to the prereactor, where higher-strength phosphoric acid is
added along with additional ammonia and sometimes a small amount
of sulfuric acid. Slurry from the prereactor is sprayed into the
granulator, where it is further ammoniated to DAP. Some of the
operating conditions are shown in Table 1. Under these conditions
there is essentially no fluorine lost and very little ammonia lost in the
plant's stack gases. Some companies have installed tail-gas scrubbers
to remove traces
 Page 413

Figure 9
Two-stage scrubbing for manufacture of DAP.
 Page 414
of ammonia. They usually use gypsum pond water as a scrubbing
medium. This pond water contains acids that capture the traces of
ammonia. Dutch State Mines Co. conducted tests to show that if
sufficient ammonia is added either from the process (DAP) or
externally (MAP) so that a N/P mole ratio of 0.35 is maintained, there
will be essentially no loss of fluorine from the scrubber's (plant) stack
[3].
Generally, a plant used to make DAP is also used to manufacture
MAP. Single-stage scrubbing is all that is needed for MAP. There is
little or no problem with ammonia loss. However, there is a tendency
toward higher flourine losses when manufacturing MAP. The scrubber
should be operated mostly with water. The pH of the water needs to be
kept between 4 and 5 by the addition of phosphoric or sulfuric acid,
and in some cases, by the addition of ammonia. When a tank reactor is
used, it is recommended that the N/P mole ratio of the slurry in the
reactor be at least 0.9, to avoid high fluorine losses. Also, to ensure
that there is a favorable water balance in the plant, it is recommended
that the tank reactor be operated at an N/P mole ratio of 1.45 and a
slurry temperature of 118°C. With these operating conditions there are
low fluorine losses and there only a slight ammonia loss from the tank
reactor. This is because the tank reactor is operated at conditions of
high solubility. Slurries of relatively low water content (15 to 20%)
are produced that can be effectively sprayed within the granulator.
Phosphoric acid is also sprayed into the granulator to lower the N/P
mole ratio of the product from the granulator to 1.0. This is a ratio of
relatively low solubility that is conductive to low recycle of high
production rates.
Some companies have found it to be highly desirable to use a TVA
pipe cross reactor (PCR) instead of a tank reactor to manufacture
MAP [4]. When this is done, 90% or more of the ammonia used in the
process is added to the PCR. With these conditions the loss of fluorine
from the PCR is low because the PCR is operated at a relatively high
pressure and temperature (140°C). Also, at these conditions the melt
(>5.0% water) can be effectively sprayed into the granulator. There is
essentially no loss of ammonia from the PCR during the manufacture
of MAP. The scrubber is operated using the same conditions as those
described above for use of a tank reactor to manufacture MAP. Since
the PCR can operate and spray well at high temperatures and low
slurry water content, it is possible to have higher production rates
when it is used in preference to a tank reactor.
A
Types and Design of Wet Scrubbers
There are generally two categories of wet scrubbers: (1) low energy
and (2) high energy.
1
Low-Energy Scrubbers
The low-energy scrubber makes use of restricted passages through a
liquid. It has a plate or packing on which liquid is sprayed. The gas
flow is usually in a direction counter to the flow of liquids. Figure 10
shows several types of these scrubbers:
Table 1 Operating Conditions for a Wet
Scrubber
NH3
N/P Slurry loss
Mole specific Temperature (% of
Equipment ratio gravity (°C) total)
Granulator 1.95 N/Aa 94 1520
First stage 1.30 1.4 110 10
Second 0.35 1.3 107 001
stage
aN/A, not applicable.
 Page 415

Figure 10
Low-energy impingement scrubbers.
1. A wet centrifugal or cyclonic type, which uses centrifugal forces
and scrubbing fluid contact to effect collection.
2. An open spray tower, which uses sprays in an open tower.
3. A packed tower, in which gases and liquids pass countercurrent
through the packing.
4. A flooded bed tower, which uses one or more beds made of marbles
or Ping-Pong balls. This packing is placed loosely in the tower so they
can float and rotate in the gas stream.
 Page 416

Figure 11
High-energy venturi scrubbers.
These scrubbers require a minimum of energy. However, in some
instances there is insufficient contact between the particulate in the
exit gas stream and scrubbing fluid. When the particulate is too small
or concentration of gas to be absorbed in the scrubbing medium is
higher than usual, a more efficient, high-energy scrubber is used.
 Page 417
2
High-Energy Scrubbers
The scrubber most commonly used by the fertilizer industry is the
venturi scrubber. Venturi scrubbers are used for intermittent mixing of
dust and scrubbing liquid. They have a high impact velocity by the
gas on the injected liquid streams. Some typical venturi scrubbers are
shown in Fig. 11. Part A shows a dry approach venturi, in which the
scrubbing liquid is injected transversely at the converging section and
leaves this section to discharge into the disengaging chamber, where
the liquid is separated from the clean air. Part B a wet approach
venturi, in which a presaturator is used to wet and quench the gas
prior to passing it through the venturi section. Either this or the dry
approach venturi may be equipped with an adjustable throat
mechanism which enables the venturi to handle a wide range of gas
flows without sacrificing efficiency. Part C shows an ejector venturi,
used by some companies in the fertilizer industry.
Figure 12 shows a three-phase injector venturi which is used in a
superphosphate plant [5]. Tests show that this scrubber recovered
99.6% of the fluorine added to it. It produces a 23 to 26%
hydrofluosilicic acid (H2SiF6) which is used by municipalities to treat
their drinking water with fluorine. In this type, venturi scrubber liquid
is added under pressure through a spray nozzle. The passage of fluid
through the spray nozzle sucks the contaminated air into the venturi.
The dust collection efficiency of venturi scrubbers varies with power
input. The higher the power input, the higher the efficiency. Tests
show that higher power input is required for the smaller particulates.
 Figure 12
Three-stages venturi scrubber for normal superphosphate plant.
 Page 418

References
1. J. L. Munson, American Air Filter Co., Air pollution-dry collectors,
Chemical Engineering, October 14, 1968, pp. 147159.
2. F. P. Achorn and H. L. Kimbrough, Fertilizer Dust Problems:
Major Sources and Available Controls, Tennessee Valley Authority,
Muscle Shoals, Ala., 19 .
3. J. T. Boont, Fluorine emission control at the new UKF-NPK plant,
Pernis, The Netherlands, ISMA Technical Conference, November
1980.
4. F. P. Achorn and J. S. Lewis, Jr., Tennessee Valley Authority,
Granular ammonium phosphate sulfate and monoamonium phosphate
using common pipe-cross type reactor, U.S. Patent 3,954,942, May 4,
1976.
5. F. P. Achorn, J. L. Balay, and D. G. Salladay, Fluorine recovery
from superphosphate plants, American Chemical Society Meeting,
Chicago, August 24, 1975, p. 99.
 Page 419

31
Nitrophosphate-Based NPK Fertilizer:
Air Pollution Control
Robert E. Nitzschmann and Johannes G. A. Reuvers
BASF Aktiengesellschaft
Ludwigshafen, Germany
I
Introduction
Most phosphate fertilizers currently produced worldwide contain
P2O5 in a highly water-soluble form. These fertilizers are often
derived from phosphoric acid. The phosphoric acid is obtained from
the reaction of rock phosphate with sulfuric acid. After initial
acidulation of rock phosphate, the unwanted solid reaction products,
mainly phosphogypsum and silica, are separated off. The resulting
dilute phosphoric acid is processed further into water-soluble
phosphate salts. Roughly 40 million metric tons of P2O5 are
consumed annually worldwide [1]. As a result, about 150 million tons
of contaminated phosphogypsum are produced and washed and must
be discharged annually [1].
Nitrophosphate technology presents the main alternative to the above-
mentioned sulfuric route. In the nitrophosphate process, rock
phosphate is digested with nitric acid. The resultant dissolving
solution contains the ionic ingredients of phosphoric acid and calcium
nitrate. The latter can be removed physically by cooling [2,3]. By
crystallizing out and separating off calcium nitrate tetrahydrate (the
Odda process), a highly soluble nitrophosphate fertilizer can be
obtained [2]. Moreover, calcium nitrate can (1) be granulated and
applied directly as fertilizer, (2) be incorporated into fertilizer
solutions, or as the main use in Europe, (3) be converted into
ammonium nitrate and lime, the principal components of calcium
ammonium nitrate (CAN) [4]. Apart from a small quantity of
insoluble sand present in the rock phosphate, the nitrophosphate route
thus produces no solid by-products that need to be discharged. Even
this sand can be washed and used in the building industry [5].
These advantagesin particular, the independence on sulfur or sulfuric
acid and the absence of solid wasteof Odda-type processes, as well as
improved product qualities, have in recent years led to major policy
decisions in favor of the nitrophosphate route. Agrolinz in Linz,
Austria, has chosen the nitrophosphate process to replace its former
 Page 420
facilities (which included an environmentally damaging sulfuric acid
plant), and BASF has decided to alter its phosphoric acid plant at
Antwerp, Belgium, through expansion of the Odda capacity at that
location by 50% [3,4].
Improved environmental standards as well as legislative requirements
have led many producers to focus in particular on those aspects of the
nitrophosphate process that involve avoiding and minimizing
pollution and waste. As a result, this technology now includes
solutions to environmental problems such as air pollution, solid waste,
and process wastewater. Solutions to these environmental challenges,
as developed and implemented by BASF in recent new plants, are
discussed in this chapter. These developments take due care of the
advanced European regulations governing emissions into air and
water and, in particular, adhere to the stringent German environmental
legislation. The nitrophosphate process as developed and practiced by
BASF has been described in detail on several occasions [27].
Licensing is normally done in cooperation with the German
engineering company Uhde GmbH.
II
Process Description
The main characteristic of the nitrophosphate route is the digestion of
rock phosphate with nitric acid. The nitric acid needed for acidulation
is normally produced on site. Ammonia serves as a feedstock for nitric
acid and is used directly for neutralization. Depending on the
capacities of the nitric acid and nitrophosphate plants, ammonia is
also preferentially manufactured on site (Fig. 1). However, ammonia
can be bought, shipped, and stored. In our process the attack of rock
phosphate by nitric acid is followed by an effective and inexpensive
inert (sand) separation step to reduce the ballast in the dissolving
solution and to minimize erosion [2,4,7].
Cooling is achieved by vaporizing liquid ammonia to obtain gaseous
NH3 and/or by electrical means. Vaporizing ammonia is
advantageous, as ammonia gas is needed in the

Figure 1
Production of ammonia, nitric acid, and NPK fertilizers.
 Page 421
conversion and neutralization sections. In some special cases
recompression of gaseous NH3 may be required. The dissolving
solution is cooled to crystallize out calcium nitrate tetrahydrate
(CNTH). The crystals are separated off and the resulting filtrate (NP
acid) forms the starting material for the production of multinutrient
fertilizers in the NPK plant (Fig. 2). Here the NP acid is ammoniated
in two stages, in some cases followed by an evaporation step.
Ammonium nitrate, potassium salts, and/or other nutrients can be
added to the slurry to obtain the product composition desired. The hot
NPK slurry is then granulated using drum granulation, dried, and
screened. The screened product is coated and cooled. Capacities of up
to 1200 tons/day of, for example, 151515, in a single train are
possible.
Rock phosphates usually show a CaO/P2O5 ratio in the range 1.40 to
1.65. Such high
 Figure 2
BASF concept of fertilizer production.
 Page 422
CaO/P2O5 ratios imply low nutrient values in the final product. The
nutrient content is, however, increased significantly by separating off
part of the calcium nitrate. The CaO/P2O5 ratios achievable in the NP
acid depend on the final crystallization temperature and to a great
extent on the quality of the cooling and of CNTH separation. With
appropriately designed cooling systems, final CaO/P2O5 values of
about 0.21 to 0.24 are obtainable. This corresponds to final fertilizer
nutrient values of about 16.716.716.7. Knowledge of the Ca(NO3)2
solubility curve, heat transfer technology, and the BASF experience
with different cooling systems allows a thorough understanding of the
relationship between cooling time, cooling energy, and the resulting
CaO/P2O5 ratio. For each specific case the most favorable fit can thus
be properly realized.
The separated and washed (to recuperate any adhering phosphate)
CNTH crystals, dissolved in dilute ammonium nitrate, are reacted
with ammonium carbonate, synthesized in turn from NH3 and CO2, in
the conversion section. As such, the carbon dioxide released from an
ammonia plant can be put to use (Fig. 1). The resultant calcium
carbonate and ammonium nitrate are preferentially utilized in the
CAN process for straight nitrogen fertilizers [4] (see Chapter 9).
CNTH, calcium carbonate, and ammonium nitrate are themselves
valued solid and/or liquid fertilizers that can be applied directly to the
soil [4].
III
Environmental Aspects:
Pollution Control in Waste Air and Wastewater
Waste products arising from the nitrophosphate process are sand,
water, and air [5]. They are either substances set free from the raw
materials, such as sand from the rock phosphate and water from the
nitric acid, or water and air used for washing and drying. The
quantities of substances set free from the starting materials cannot be
reduced. For example, all the water introduced with the nitric acid
used for digesting the rock phosphate has to be removed by the end of
the process, as the granulated fertilizers are dry products. This water
leaves the process as condensate from the ammonium nitrate
concentration unit or as vapor from the various granulation and drying
stages.
The quantities of water and air made use of in the process for washing
and drying are essentially reducible. For example, the air used for
drying in our NPK production can be recirculated until it is highly
saturated, to reduce the rate of discharge. These waste productssand,
air, and waterare all natural substances and thus can be disposed of or
discharged without restrictions if they are uncontaminated. However,
in the course of the process they take up substances that would pollute
the environment: the sand separated from the rock phosphate is
contaminated with digestion solution; the condensate from the
ammonium nitrate concentration contains ammonium and nitrate ions.
The air from the acid stages of the nitrophosphate process is
contaminated with nitrogen oxides and fluorides; air from the alkaline
stages contains ammonia and fluorides. Air from granulation and
drying carries mainly ammonia and fertilizer dust (Fig. 3).
Given these facts, it is pertinent to reduce the quantities of waste
products, in particular water and air, to the extent possible. At the
same time the contaminants in these waste products ought to be
avoided, eliminated, or minimized. The remaining waste products are
to be treated using processes by which contaminants are recovered
and recycled to the process. These processes, as developed and
realized by BASF for the treatment of waste air and wastewater from
the various sections of the nitrophosphate process, are described
below.
 Page 423

Figure 3
Contaminants in by-products of fertilizer
production (nitrophosphate process).
A
Treatment of Waste Air and Wastewater from the Acidic and Neutral
Stages of the Nitrophosphate Process:
Reducing NOx Pollution
Due to the presence of fluctuating amounts of reducing agents in the
rock phosphate, the attack of nitric acid results in the formation of
NOx gases. Our concept of treating NOx-containing waste gas from
dissolving, inert separation as well as from the crystallization and
separation of CNTH allows us to tolerate fluctuating concentrations of
NOx. Hence rock phosphates of different origin containing various
amounts of, for example, Fe2+ and organic material can be used. The
NOx-containing waste air is treated in a two-stage scrubbing section
(Fig. 4). The amount of fresh water required to obtain a good
 Page 424

Figure 4
NOx scrubbing with water: conventional wastewater quantities and pollutants.
scrubbing efficiency is reduced by operating the two scrubbing towers
in series as well as by a proper combination of the various gas streams
[5,8].
Less air originates from the digestion and inert separation stages, but
the concentration of nitrogen oxides is about 50 times as much as in
the air from the cooling stage. Air from the digestion and inert
separation stages is treated in the first scrubbing tower. Despite a very
good scrubbing efficiency, the mass concentration of nitrogen oxides
in the air emerging from the first packed tower can be as much as
2000 mg/m3, expressed as nitrogen dioxide but actually due almost
entirely to nitrogen monoxide, which is difficult to absorb. The air
from the cooling stage and from the CNTH belt filters contains
nitrogen oxides with a high proportion of nitrogen dioxide, which is
also difficult to scrub. Mixing the two airstreams before they enter the
second packed tower produces a mixture of nitrogen oxides with a
degree of oxidation between about 40 and 60%, optimal for
absorption. The mass concentration of nitrogen oxides in the air
leaving the second tower can be brought below 500 mg/m3 (200
ppm), the maximum presently allowed by German regulations.
Based on this general concept, several alternatives for treating waste
gas and wastewater have been advanced, depending on the local
situation. Next we present some examples.
Example 1. To maintain a good scrubbing efficiency, the temperature
of the fresh water fed to the second absorption tower should be kept
below about 18°C. Hence, depending on the season and the
geographical plant location, fresh water can be used directly or after
appropriate cooling (winter versus summer operation, subtropical
versus northern latitudes). This route for treating NOx-containing
waste air from the various acidic stages of the nitrophosphate process
has been realized at the fertilizer plant in Bharuch, India. During
normal operation conditions the treated air contains less than 300
mg/m3 (STP) of NOx.
 Page 425
Example 2. More than 90% of the nitrogen transferred from the waste
air to the scrubbing water originates from the digestion stage. A
prescrubbing of just these gases with dilute nitric acid in a closed
circuit allows us to reduce water pollution. Absorption of nitrogen
oxides tends to increase the acid concentration, but by returning acid
to the process and adding some fresh water, the concentration of nitric
acid is kept constant at about 30 to 40 wt %. The air leaving the
prescrubbing tower still contains NOx and is scrubbed further as
described above. This approach has been realized in the BASF Odda
plant in Antwerp.
Example 3.The next stage in reducing environmental pollution from
the acidic sections of the Odda process is to eliminate the wastewater
totally from the scrubbing towers. This is possible if one neutralizes
the nitric acid, HNO3, generated with ammonia and decomposes most
of the nitrous acid, HNO2, also formed. This is shown in Fig. 5. The
remaining dilute ammonium nitrate solution absorbs nitrogen oxides
as well as water does. The concentration of ammonium nitrate is kept
constant between 10 and 20% by returning AN solution to the process
and replacing it with fresh water or condensate. The AN solution is
preferentially discharged to the conversion section of the
nitrophosphate process. Decomposition of nitrous acid can be carried
out thermally, by steam stripping at about 100°C, or by treatment with
urea at about 70°C. After decomposing the nitrous acid, the solution is
cooled before it is returned to the scrubbing column. This version of
the BASF concept for minimizing air and water pollution is now
being operated at Linz [3]. No wastewater at all is thus discharged
from the NOx scrubbing step.
Example 4. Ultimately, the final consequence of our concept for
reducing pollution is the total elimination of wastewater from the
Odda process. This approach has
 Figure 5
NOx scrubbing with elimination of waste scrubbing water.
 Page 426
been developed by BASF but has not yet been realized. After optimal
(re)use, all process waters from the acid and neutral sections of the
Odda process are combined and treated by means of an ion-exchange
process. A concentrated AN solution is returned to the process and the
eluate discharged as highly purified water, usable, for example, as
boiler feedwater.
Example 5. The idea of treating wastewater by means of ion exchange
has been followed further for individual wastewater streams. In AN
evaporation, condensates containing up to 2 g/L of nitrogen (as
ammonium nitrate, N-NH4 +, and N-NO3 ) are produced. This hot
(about 80°C) condensate is often used for cleaning of equipment and
floors or as makeup water for scrubbing towers and the like.
Depending on the needs and the possibilities of each individual plant
location, this condensate may be freed from N-NH4 + and/or N-NO3 .
To remove only N-NH4 + a single cation-exchange unit can be used.
Optionally an anion-exchange unit to remove N-NO3 can be added,
even at a later date. This, for example, may hold for locations where
biological treatment is still available but is expected to become the
limiting factor. With ion exchange, final concentrations of NH4 + and
NO3 of nearly zero are obtained in the treated condensate. This water
can be used as boiler feedwater. Using reverse osmosis both ions are
removed to the same extent. The economically significant
concentration factor, using a two-or three-stage unit, is about 10. Final
concentrations of N-NO3 and N-NH4 + in the treated condensate of
below 50 mg/L can be maintained. Both ion exchange and reverse
osmosis have been thoroughly investigated by BASF and can be
included in future projects. A reverse-osmosis unit for treating AN-
containing condensates is likely to be built at our Antwerp site.
Cation-exchange is being considered for reducing NH4 + emissions
from our AN plant in Ludwigshafen.
B
Treatment of Waste Air and Wastewater from the Alkaline Sections:
NH3 Scrubbing
To avoid the formation of aerosoles we maintain a strict separation of
waste gases containing NOx and NH3, respectively. Thus waste gases
originating from the production and storage of NP acid, as well as
from the mixing of NP acid with ammonium nitrate and/or other
inorganic salts, are fed to the NOx scrubbing system. NH3-containing
waste gases result from the conversion section, from the ammoniation
of NP acid and subsequent granulation and drying steps and from the
CAN process. For scrubbing purposes all or part of these gases can be
combined to save investment and operating costs [4] (see Chapter 9).
To reduce the NH3 emissions from the conversion section of a
nitrophosphate plant, the waste gas is scrubbed in a packed column.
The scrubbing liquid is circulated in a closed circuit, while the pH
value is adjusted through the addition of HNO3. Part of the resulting
concentrated ammonium nitrate solution is returned to the CN
conversion section. Fig. 6 shows the parameters of a typical scrubbing
tower used for treating 20,000 m3/h NH3-laden waste air.
The basic principles underlying the absorption of ammonia are well
known [5,9]:
1. A certain minimum pH value (depending in particular on the
temperature) is to be maintained at the top of the scrubbing tower to
avoid the formation and emission of aerosols.
 Page 427

Figure 6
Scrubbing of NH3-containing waste air.
2. To avoid a breakthrough of ammonia, the AN solution must be
circulated continuously at a sufficiently high rate.
Given the quantity of waste air to be treated, the operating
temperature, and the NH3 (peak) concentrations, the dimensions of the
scrubbing tower can be set. Here BASF can build on its experience in
designing, engineering, and operating a variety of scrubbing columns.
Under normal operating conditions the air emitted from the scrubbing
tower carries far less than 50 mg/m3 (STP) ammonia and less than 5
mg/m3 (STP) fluoride. No waste process water arises from the
conversion section, as no evaporation or drying is involved.
Moreover, in the BASF process for manufacturing CAN, the filtered
calcium carbonate need not be washed with water or dried (see
Chapter 9).
C
Treatment of Waste Air from the Ammoniation and Evaporation of NP
Acid
Exhaust air from the two ammoniation stages and, in some cases,
from further evaporation of the ammoniated NP slurry contains NH3
and fluorides. It is combined with waste air from the subsequent step,
in which potassium and/or other salts are added. The temperature of
the combined airstream is about 95°C. This hot, nearly saturated air is
scrubbed with nitric acid solution to remove ammonia, which is
returned to the process as ammonium nitrate. The scrubbing tower is
operated as described in Sec. III.B. Due to the high temperature of the
waste air, ammonia is not removed quite so effectively, but its
concentration in the discharged air is less than 100 mg/m3 (STP), well
below the expected legislative limit. However, to accommodate peak
concentrations of NH3 and to minimize the risk of formation of
ammonium nitrate aerosols, the flow of the circulating solution as
well as the pH value at the top of the scrubbing tower are increased.
For example,
 Page 428
producing 1200 tons/day of 151515, this section may require a flow of
about 800 m3/h for scrubbing 20,000 m3/h (STP) waste air, depending
on the prevailing NH3 peak concentrations.
During ammoniation and evaporation the NP slurry is heated to about
115 to 130°C. Volatile fluor-containing compounds may be freed from
the liquid phase and emitted with the waste air. Their concentration in
the waste air depends, on, among other factors, the F content of the
rock phosphates and the final pH value of the slurry. Up to 100 mg/m3
(STP) of fluor-containing compounds, calculated as HF, can be
present. The concentration of gaseous F-containing compounds in the
waste air leaving the NH3 scrubbing tower is, however, less than 20
mg/m3 (STP), expressed and calculated as HF. If required, the waste
air can also be scrubbed with plain water in a once-through fashion to
bring this F value down to less than about 5 mg/m3 (STP). A further
integral treatment of waste air and wastewater from this process
section is described below.
D
Treatment of Waste Air from Granulation and Drying
Granulation is achieved by spraying hot, ammoniated NPK slurry onto
a granule bed in a rotating drum. After this granule-forming step, the
granulation loop normally includes the following processing
operations: drying, screening, crushing of oversized material,
recycling of oversized and undersized material as well as part of the
product fraction, and finally, conditioning and cooling of the
withdrawn on-size NPK product [10]. The water vapor freed from the
slurry during spraying is carried away by a sufficient quantity of air,
which is drawn cocurrently through the granulating drum. The water
evaporates partially due to the sensible heat of the slurry as well as to
the heat of solidification released during the granule-forming process.
In the drying drum hot air is used to dry the NPK product further.
Usually, the water content of the final product is distinctly less than
1.0 wt %. The large quantities of air, about 90,000 m3/h (STP) for the
production of 50 metric tons per hour of 151515, not only carry water
vapor but also become laden with NH3, dust, and minor amounts of
volatile F compounds and NOx. Part of the hot exhaust air is
recirculated to the drying drum after dust removal in cyclones. The
waste air so discharged from the granulating and drying drums is first
treated in a battery of cyclones. The cyclone separators are of no
special design. The waste air so subjected to dust removal carries on
average less than about 50 to 75 mg/m3 (STP) of dust after passing
the cyclones.
In some cases a venturi scrubber is involved to remove dust particles.
Venturi scrubbers show a rather high dust separation efficiency and
are low in construction and maintenance costs. This high dust
collection efficiency is obtained at the expense of a medium-sized
flow (0.5 to 5 L water/m3 waste air) of the circulating solution
coupled with a high pressure drop (30 to 200 mbar) and hence
substantial operating costs [12]. Waste air from which dust has been
removed by means of a venturi carries less than 30 mg/m3 (STP) dust.
However, due to the short contact time between waste air and washing
liquid, venturi scrubbers show a low efficiency for NH3 absorption.
The amount of ammonia contained in the waste air from granulation
and drying depends on, among other factors, the type of NPK
fertilizer produced, the operating temperatures, and in particular, the
final pH value of the ammoniated slurry. Maintaining a final pH value
of about 5.0 and under normal operating conditions, the average
ammonia content is less than 100 mg/m3 (STP). Scrubbing is not
always required.
There remains, however, a certain incentive for further treatment of
this waste air
 Page 429
and for recuperation of the remaining quantities of nutrients (NPK
dust, NH3). This is substantiated by the following aspects:
1. As described in Sec. III.C and due to the higher saturation and
higher temperature of the waste air, ammonia is not removed quite as
effectively from the waste air from the ammoniation and evaporation
of NP acid.
2. In some cases an additional absorption tower is provided to reduce
the F content of the waste air of Sec. III.C.
3. The reduction of emissions from all stages of the NPK process in a
single step is very practicable and economically favorable.
4. The thermal energy contained in the waste air can be used to
concentrate the wash water before it is returned to the process. It is, of
course, advantageous to perform scrubbing and concentrating in the
same equipment.
These and other aspects have led BASF to develop a scrubbing system
that combines the removal of NH3, F, and dust with the recuperation
and concentration of nutrients. This development has been advanced
to the extent that we now prepare the construction of such a scrubbing
unit.
This single-step waste gas treatment system is shown schematically in
Fig. 7. The nearly saturated waste air from the neutralization and
evaporation stage (Sec. III.C) and the waste air from the granulation
and drying steps are fed to the packed column by means
 Figure 7
Single-step waste gas treatment with
nutrient recuperation and concentration.
 Page 430
of separate gas distributors. The dimensions of the scrubbing tower
are determined primarily by the gas load and by the desired reduction
in the NH3 and F concentrations. The pH value of the circulating wash
liquid is adjusted through the addition of HNO3. With such a column,
dust can also be separated [13].
To recuperate ammonia from the waste air, nitric acid needs to be
added (Sec. III.B); to remove F-containing compounds, neutral or
basic conditions are normally practiced. To absorb both types of
compounds, the pH value needs to be increased. Normally, phosphates
arising from the NPK dust are present in the circulating suspension
and show no negative effects on the absorption [13].
The scrubbing system yields excellent results: both NH3 and dust are
less than 50 mg/m3 (STP); F is less than 5 mg/m3 (STP). There is,
moreover, an additional advantage: the thermal energy content of the
waste air leads to the evaporation of water from the circulating
suspension. Various polluted waters can be used as makeup water. The
concentration of the circulating suspension is kept constant at about
35 to 45 wt % by diverting a small constant flow back to the process,
either directly or after further concentration. Problems related to
unwanted deposits, materials of construction, and the like could be
solved.
The foregoing treatment system produces no wastewater from the
ammoniation, evaporation, granulation, and drying stages. With this
scrubbing concept nutrients are recovered and low-temperature energy
is used.
E
Treatment of Waste Air from Conditioning and Cooling
After granulating and drying the NPK product is screened. The on-
size hot product may then first be conditioned and subsequently
cooled, or be cooled first. External conditioning of the NPK granules
is usually achieved in a rotary drum. Coating can be done with fine
inorganic powders such as talcum, with nonionic organic sealants
such as polyethylene wax, or with hydrophobic surfactants. For
example, polyethylene wax is applied to the hot product, whereas fatty
amines can be applied after cooling.
Cooling of the granulated and conditioned but still hot on-size NPK
product can be achieved by using rotary drums and/or fluidized-bed
coolers. A fluidized-bed cooler, comprising, for example, two or three
cooling compartments, is easy to operate and may cool the final
product from about 95 to 105°C down to about 35 to 50°C. Ambient
air is drawn in, this air being preconditioned under certain climatic
conditions to control its relative humidity and its temperature. Waste
air from the latter stages need notbut maybe treated in cyclones. This
waste air contains less than 20 mg/m3 (STP) of dust, minor amounts
of NH3 [less than 20 mg/m3 (STP)], F [less than 5 mg/m3 (STP)], and
NOx [less than 30 mg/m3 (STP)]. This waste air need not be treated.
IV
Product Safety
The nitrophosphate products made by BASF are not prone to self-
sustaining thermal decomposition (''cigar burning" [14]). To achieve
this we have, where necessary, adapted the compositions of our
products, thereby (1) maintaining the final nutrient values, and (2)
increasing the water solubility of the P2O5 component. Due to these
modifications, many possible environmental risks and hazards
involved in the handling and use of NPK fertilizers could be reduced.
 Page 431

V
Summary
The nitrophosphate route to compound NP(K) fertilizers offers a
variety of environmental as well as economic advantages. The
nitrophosphate process is independent of sulfur and, in particular,
produces no gypsum. It is flexible with respect to raw materials as
well as final NP(K) products. The modern process includes preventive
measures to avoid and minimize pollution and waste. Depending on
the local situation and the existing or expected environmental
legislation, a selection from various flexible alternatives can be made.
The methods described above for treatment of waste air and
wastewater from all the diverse sections of the nitrophosphate process
allow BASF to provide technical solutions for meeting all
environmental requirements despite ever-more-stringent legislation.
References
1. J. J. Schultz, D. I. Gregory, and O. P. Engelstad, Phosphate
Fertilizers and the Environment, a Discussion Paper, International
Fertilizer Development Center, Muscle Shoals, Ala., March 2, 1992.
2. L. Diehl, K. F. Kummer, and H. Oertel, Nitrophosphates with
variable water solubility, paper read before the Fertilizer Society of
London, London, April 17, 1986.
3. Agrolinz's new high-solubility nitrophosphate plant, Phosphorus
and Potassium, No. 166, MarchApril 1990, pp. 2124.
4. J. G. A. Reuvers, L. Diehl, and R. E. Nitzschmann, The BASF
nitrophosphate concept: experience and environmental aspects related
to three recently commissioned NP(K)-plants, proceedings of an
International Workshop on Phosphate Fertilizers and the
Environment, Tampa, Fla., March 26, 1992, pp. 241250.
5. L. Diehl, Nitrophosphate and fertiliser plants: the recycling concept
for minimising pollution of air and water, IFA Technical Conference,
Edmonton, Alberta, Canada, September 1216, 1988, TA/88/12/122.
6. A. H. Durocher, L. Diehl, and H. -J. Eisen, The BASF ODDA-NPK
process, IFA Technical Conference, Paris, 1985.
7. Fertilizers by the nitrophosphate route, Phosphorus and Potassium,
No. 155, MayJune 1988, pp. 2833.
8. P. Reinhard, L. Diehl, and R. Wunder, Verfahren zum Entfernen
von NOx aus den bei der Herstellung von Duengemitteln anfallenden
Abgasen, Eur. Patent 351,613.
9. L. Diehl and R. Wunder, Prozebintegrierte Abgasreinigung bei der
Duengemittelherstellung nach dem Nitrophosphat-Prozess, VDI
Berichte, No. 730, 1989, pp. 395415.
10. Ullman's Encyclopedia of Industrial Chemistry, 5th ed. Vol. A 10,
VCH Verlagsgesellschaft, Weinheim, Germany, 1987, pp. 323431.
11. R. Wunder, R. E. Nitzschmann, and B. Meyer, Verfahren zur
Entfernung von Ammoniak aus Abgasen, Ger. Offen. 4,000,540, Eur.
Patent 440,932.
12. K. Schaber, Gas scrubbing plants for acid and caustic exhaust
gases with consideration of the new TA-Luft, Chemie-Ingenieur-
Technik, Vol. 59, No. 5, 1987, pp. 376383.
13. K. Grimmer, R. E. Nitzschmann, E. Minx, B. Purucker, M. Jaeger,
R. Wunder, B. Meyer, and G. Kling. Verfahren zur
Prozebwasserkonzentrierung und Abgasreinigung, Ger. Offen.
4,209,498.
14. B. Purucker, Safety requirements in fertiliser plants, IFA Technical
Conference, Edmonton, Alberta, Canada, September 1216, 1988,
TA/88/15/117.
 Page 433

32
Nitrophosphate-Based NPK Fertilizer:
Wastewater Treatment
Paula-Gabriela Petreanu
Ipran-Iprochim
Bucharest, Romania
I
Introduction
The term nitrophosphate is used to designate various complex
fertilizer grades, usually obtained when phosphate rock is digested
with nitric acid. Sometimes, authors or technical publications enlarge
the scope of the term to include digestion with an acidic mixture in
which at least one component is nitric acid (the others being sulfuric
acid, phosphoric acid, or even carbon dioxide). This meaning of the
term is usually to be found in older publications [13]. Today, this term
is generally used when nitric acid is the only agent for phosphate rock
digestion. The process is rather complex, including many distinct
steps, a wide variety of reaction media, and a wide range of working
temperatures (for this type of process), so that multiple environmental
pollution problems are generated, whose solution requires very
sophisticated engineering and designing work. In such a plant the
resulting liquid effluents are of complex chemical composition,
including almost all the harmful components common to the fertilizer
industry, such as NOx , NH4 +, P2O5, F, SO4 2, SiO2, and sometimes
even Clor various types of insoluble matter.
Another important aspect to be considered are the large dimensions of
problems generated by such an industry. As is well known, most such
plants have large capacities, so that the pollution problems to be
solved are of considerable significance; the resulting polluted effluent
quantity is important, and if not properly treated, may have a rather
impressive polluting effect on the entire environmental area.
In this chapter we discuss the problems connected with environmental
protection against the polluting effects of the liquid effluents
discharged by a nitrophosphate plant. To simplify the discussion we
assume that nitric acid is the only phosphate rock digestion agent used
in the plant, that problems generated are typical for this field, and that
the problems would be the same if acid mixture were used.
 Page 434

II
Processes Used in the Nitrophosphate Industry
Although the most important quantities of complex fertilizers are
manufactured by the "sulfuric route," where the phosphate rock is
digested in sulfuric acid in order to obtain phosphoric acid as an
intermediate product, the main alternative that can be adopted is
offered by the nitrophosphate route, where nitric acid is used as a
phosphate rock digestion agent. It is assumed that about 10 to 15% of
the present world quantity of complex fertilizer is obtained by this
route (not including processes using acid mixtures) [4,5]. This
percentage has stayed the same until lately; discussions related to
adopting such a process type involve primarily comparative economic
aspects of the two alternatives, their respective operating flexibility,
the grades that may be obtained, and so on.
The up-to-date processes used in the nitrophosphate industry have
many specific steps in common. The main differences between the
processes, also used for various process classifications, are given by
the method of calcium separation. The primary chemical reaction of
phosphate rock digestion when using nitric acid is the same for all
processes and may generally be expressed as follows:

where x is the HNO3 excess. Some of the HF (maximum 3 to 5%) is

released from the liquid as a gaseous component, but part of it may
react in the solution as follows:

and is to be found in the final product.

As is well known, calcium nitrate is a soluble chemical component
and due to the fact that in the agricultural use of fertilizers, products
with a very high water solubility of P2O5 are required, the calcium ion
has to be removed from the mother liquor. At present, two principal
process routes are used for this purpose: (1) calcium nitrate
[Ca(NO3)2·4H2O] crystallization by refrigeration, and (2) calcium
sulfate (CaSO4·2H2O) precipitation by the addition of sulfate ions
(especially as ammonium sulfate). Obviously, combined processes
have been developed as well. Processes using calcium nitrate
crystallization through refrigeration have been developed starting
without exception from the Odda process, introduced in the 1930s in
Norway and represented today by several well-known processes, such
as Norsk-Hydro, BASF-Odda, Hoechst-Uhde, Kamka-Nitro, SCHZ-
Kaltenbach, and so on. These differ with respect to some operating
parameters, as well as equipment characteristics. The most
representative may be considered the Norsk-Hydro process.
Processes using calcium removal as insoluble calcium sulfate (usually
accompanied by sulfate ion recycling in the process) have been
developed by various producers, some processes being promoted on
an industrial scale, such as DSM-Stamicarbon, Power-Gas, Chemico,
TWA, and Foster-Wheeler. The most representative, with some
industrial-scale applications as well, may be the Stamicarbon process.
The differences among the various processes is represented by some
of the working conditions as well as the equipment used in the plant.
In Figs. 1 and 2 are shown block diagrams of two types of processes.
As regards process performance, the crystallization route offers
greater flexibility with respect to fertilizer grades that can be obtained
(a part
 Page 435

Figure 1
Nitrophosphate process block diagram: calcium nitrate separation by crystallization.
 Figure 2
Nitrophosphate process block diagram: calcium nitrate separation by sulfuric precipitation.
 Page 437
of the nitrogen being obtained as a simple nitrogen fertilizer) as a
separate final product, such as calcium nitrate or ammonium nitrate
(respectively nitrochalk).
As regards pollution problems, the calcium nitrate crystallization
processes have some advantages, due to the fact that the solutions
produced in the plant are of higher concentration, so that any possible
effluent amounts that result are also smaller in amount and more
concentrated. Further, they can be recirculated into the process more
easily and are less costly to treat. In any case, fewer process steps are
required, and the vessel's working volumes are minimized.
Development of this type of process is more advanced, making
possible step-by-step introduction of new and important
improvements in the process as well as in the polluting-abatement
systems. A deeper understanding of the physicochemical phenomena,
research work related to equipment construction, and not least, the
experience obtained operating such large-capacity plants for long
periods have made possible really important progress.
In nitrophosphate plants, the granulation step is performed by melts
prilling or pan-granulation methods. Pan granulation is used only in
smaller-capacity plants. In some plants, granulation in pug mills or
spherodizers has been adopted, the product being dried with hot gases.
Sometimes, this alternative is mandatory, due primarily to the
phosphate rock quality. That is, too many impurities in the rock render
impossible concentration of the mother liquor to the melt stage. In
such conditions, a specific washingcleaning gas system should be
adopted for the granulationdrying stage. These types of problems are
discussed in Chapters 29 and 30, so no additional details will be given
here.
III
Pollutants
The various polluting components possible in liquid effluents from a
nitrophosphate plant include various forms of nitrogen, such as NOx
and NH4 +, P2O5, F, Cl, SO4 2, SiO2, insoluble material, and other
substances. These components are carried over in the liquid effluents
in various ways and from different sources, such as:
1. Scrubbing of gases evolved in various process steps. Suction of
these gases from vessels is necessary to maintain normal working
conditions for personnel inside the plant building. Other gas sources
are vacuum filtration operations and a vacuum evaporation step.
2. Vapor condensation in direct-contact water-cooled condensors in
various process steps requiring evaporation, vacuum filtration, and so
on.
3. Washing of acid-insoluble waste solids.
4. Leakage from equipment by overflow, emptying operations, or
sealing inadequacies.
5. Equipment washing, as well as washing of piping, ducts, and so on
(routine or incidental).
6. Product drying (when the granulationdrying route is used).
7. Floor and channel washing.
8. Leakage out of waste piles or ponds.
9. Washing in areas where dry, solid fertilizers are handled.
Admissible pollutants quantities in effluents released by a fertilizer
plant are strictly limited by law in many countries, but they have
different values depending on plant location. This type of problem is
discussed in detail in Chapters 2 and 3 and is not
 Page 438
treated here. It is important to note that in terms of environmental
protection, in this type of fabrication the effluent amounts released
(the liquid flows and pollutant content) must be minimized as much as
possible. This is the major goal of all specialists teams working in this
industry.
IV
Liquid Effluent Pollution Control
A
Gaseous Effluent Problems
Most of the contaminated water in a nitrophosphate plant is generated
by contact with various gases released from the process (see also
Chapter 31). Practically all process steps in nitrophosphate plants
generate gases containing higher or lower amounts of chemical
contaminants. For this reason, all the equipment is connected to a
suction system, and depending on the composition and/or
characteristics of the contaminants, they are directed to scrubbing,
condensing, dry separation, or other operations. For a better
understanding of the cleaning process, gases can be divided into
several classes, depending on their composition. Figure 3 is a diagram
of the gas categories in a nitrophosphate plant according to their
characteristics.
Gas streams may be treated in several ways, the most important being
(1) collection of gases in a very few quality categories and treatment
of the resulting streams in a single large scrubbing system, and (2)
selective collection of gases in many quality categories and separate
and thorough treatment of each stream. In large-scale practice the
second alternative is more advantageous, because even if the
investment costs are greater and plant operation involves more
controls, the pollution-abatement effects are greatly
 Figure 3
Nitrophosphate process: typical released gases classification. Gas stream
categories: 1, fairly concentrated F and NOx containing gases; 2, diluted
NOx containing gases; 3, diluted NH3 containing gases; 4, concentrated
NH3 containing, almost water vapors saturated gases; 5, NH3 and water
vapors containing gases; 6, NH3 and ammonium nitrate containing water
vapors; 7, F, NOx, NH3 containing water vapors; 8, NH3, water vapors
and fertilizer dust containing gases; 9, fertilizer dust low containing
gases; 10, dry gases with fertilizer and/or coating agent dust.
 Page 439
improved. More efficient control and treatment by scrubbing of the
effluents has the advantage of permitting a high level of effluent
recycling back to the process, and consequently, a reduction of
environmental polluting effects. In addition, there is better overall
economic efficiency, as a portion of the useful components contained
in the effluents are recovered in the final products. We illustrate this
concept with a specific example.
Example.
In an Odda-type nitrophosphate plant, the gases released in the
digestion step (see Fig. 3, type 1) have the following composition
(gravimetric and approximate):
F 4%
NOx (expressed as NO2) 8%
NO3 1%
CO2 8%
Silica 2%
Water vapor 10%
Air 67%

The gases released in the Ca(NO3)2·4H2O crystallization area (see

Fig. 3, type 2), which are also acidic in character, have the following
composition (gravimetric and approximate):
F 0.03%
NOx (expressed as 0.07%
NO2)
NO3 0.03%
Si 0.001%
Water vapors 1.3%
Air 98.57%
The streams ratio is about 1:30. It is obvious that despite the fact that
the two streams have the same acidic character and almost the same
chemical components, it is much more convenient, and perhaps even
mandatory, that the two streams be treated separately, one of them
(type 1) having a low flow rate and a rather high concentration of
pollutants, while the other has a high flow rate and a very low
pollutant content. Even more narrow separation between gas streams
may be recommended if the main stream has a high flow rate (type 2)
while some secondary streams have such a low content of harmful
components that no washing is necessary before they are released into
the atmosphere.
It is, however, extremely important that gas streams of totally different
composition and characteristics be collected separately (and treated
separately). Three or four principal gas categories can be
distinguished (see Fig. 3):
1. Wet gases, acidic in character (types 1 and 2)
2. Wet gases, alkaline in character (types 3, 4, 5, and 8)
3. Water vapors, mainly alkaline in character (types 6 and 7)
4. Dust containing dry gases (types 9 and 10)
In-plant operations have demonstrated that the best solution for the
fourth category is to treat it completely separate from the rest.
Sometimes even complete separation of the work area is
recommended to avoid water or water vapor penetration in the area, to
keep liquid effluents from developing. For this reason, all purification
systems are of dry type (cyclone, bag dust separators, etc.) and the use
of water to wash equipment or the work area is totally forbidden, thus
no additional liquid effluent streams can be generated.
 Page 440
If such streams are present, they are very difficult to manage, as their
composition is seldom constant over time. Also, chlorine may be
present as a pollutant, arising from the use of potash (for ternary
grades). This type of effluent cannot be treated efficiently and cannot
be returned to the process, as the chlorine ions are highly corrosive on
the stainless steels used in the process equipment, piping, ducts,
valves, and so on.
B
Liquid Effluent Treatment
In a nitrophosphate plant, most liquid effluents arise from two major
sources: (1) scrubbing of pollutant-containing gases, and (2) direct-
contact condensation of water vapors released when solutions are
evaporated or in vacuum filtration of slurries. For example, it may be
shown that for a plant of typical capacity (about 2800 metric tons/day
of complex fertilizer) using a Ca(NO3)2·4H2O crystallization process,
the total water flow required is about 4700 m3/h, of which about 2200
m3/h is conventional cooling water for indirect heat exchangers,
recirculated back to the plant after cooling in forced-air cooling
towers; about 2300 m3/h is gas scrubbing water and vapors from
direct condensation of water; and about 200 to 300 m3/h is water used
for other purposes (as process water, cake washing on centrifuges or
filters, accidental washings, etc.).
In an open-circuit system, the entire 2300 m3/h of the contaminated
water would be emitted, so that all pollutants in the water would be
emitted, so that all pollutants in the water will leave with the water or
will have to be retained before the water is released, this being more
difficult technically and extremely costly. It is more efficient to
recirculate this water back to the plant after proper cooling. In older
systems the contaminated recirculated water, together with important
amounts of solid waste [in fact, all CaCO3 resulting from the
conversion of Ca(NO3)2 or CaSO4], is sent to high-volume settling
ponds that occupy a huge land area. Generally, the cooling takes place
under natural conditions. In fact, in addition to the important
drawbacks mentioned above, the system is highly inefficient as to
cooling when the plant is located in temperate continental climatic
zones, and is very difficult to monitor. This makes the plant's the
water balance impossible to control, and thus the system becomes
totally inefficient.
To avoid these difficulties, in plants built in Romania in the 1970s, a
new system was developed based on the following principles [6,7]:
1. The closed water stream was divided into several separate streams
(based on the pollutants contained).
2. Separate collection and recirculation of each stream was instituted.
3. Purges were given almost total treatment and were used
extensively: as process water and as wash water for equipment,
piping, and floors.
The typical system adopted, shown in Fig. 4, consists of a
contaminated recirculated water stream that is divided into two
separate streams:
1. Acidic stream: consists primarily of water used to scrub and
condense gas containing for F and NOx.
2. Alkaline stream: consists primarily of water used to scrub and
condense water vapor containing ammonia.
The flow ratio between the two streams is about 1:5.
The treatment of the two major streams is completely separate and is
carried out as best for each.
 Page 441

Figure 4
Contaminated water balance in a nitrophosphate plant (Norsk-
Hydro process) block diagram. a, Alkaline water flow; b, acidic
water flow; c, water to/from accidental washing; d, fresh water
(eventually condensate); e, purge flow; f, water vapors and carried
over droplets; g, water vapors from purge evaporation unit; h,
concentrated ammonium nitrate solution (back to the plant).
1. The acid water is collected and after an amount of calcium
carbonate (a by-product of the plant) sufficient for fluorine-
component blocking has been added, the stream is pumped to a pond
large enough to allow fluorine-compound precipitation and solids
settling, usually for 5 to 7 days. After settling, the clear upper layer is
pumped back to the nitrophosphate plant. A purge is taken out of this
stream so that its chemical composition can be kept constant. The
average chemical composition of this stream was fixed as: NH4NO3 +
Ca(NO3)2 = 5 + 6% and F = 0.01 + 0.05%. The purge is sent to a
purge treatment station, where the pH is corrected, and then the
solution is evaporated. The concentrated solution, containing mainly
55 + 60% ammonium nitrate, is then pumped back to the
nitrophosphate plant, usually at the stage of mother liquor
neutralization, so that the useful components appear in the final
products.
2. The alkaline water stream is collected and dispensed to forced-draft
cooling towers built of special corrosion-proof materials. The water is
cooled down 20 to 22°C, part of the flow rate being entrained out of
the cooling towers, in the airflow, as vapor or droplets. A rather small
purge may be taken out for nitrophosphate plant equipment, piping, or
duct washing. The water balance is monitored by adding water
condensate to the cooling towers (slightly contaminated condensate
may be used as well). Nitric acid is used to control the water stream
pH. Therefore, if the plant operated properly according to its design,
no liquid effluent will be expelled. Recently, additional improvements
have been implemented, such as diminishing NOx release in the
digestion gas stream by feeding urea to the digestion tanks. Smaller
amounts of noxious components are carried over into the water
stream, thus also minimizing the purge to be treated.
The harmful effects produced by the liquid effluent released by a
nitrophosphate plant have also been of concern to other fertilizer
producers, among them Norsk-Hydro [8]. The company has
developed and is using in its newest plant, in Porgsgrunn, Norway, a
 Page 442
selective treatment system for effluents whose concept is somewhat
different from that outlined above. Its main features are:
1. The amount of NOx released in the digestion phase is reduced by
adding urea in the digesters.
2. Chlorine-containing water is eliminated by a doublecounter change
with the P2O5-containing stream. The second change is treated to
eliminate F and P2O5 by adding NH3 and Ca(OH)2, with further
solids settling (polyelectrolyte is used as well). The settled solids are
sent to a centrifuge.
3. Washing of equipment, piping, floors, and so on, is performed using
a closed-loop water system.
4. The ammonia content in gas streams is either blocked by nitric acid
or is released by condensate stripping and returned to the plant.
Other considerations have been reported by Agro-Linz Austria for its
newest plant, built based on the BASF-Odda process [4]. Water
resulting from scrubbing of the digestion phase which released NOx-
containing gas, is used as wash water for the calcium nitrate crystals
and therefore is returned to the process.
It is noted that the NOx-containing gases and the wash water
developed in this process are very difficult to use or to treat. That is
why some fertilizer manufacturers have tried to destroy the nitrogen
oxides using conventional methods such as those proved successful in
other industries where such materials are released, such as
thermopower stations and nitric acid manufacture. But the gases
released by a nitrophosphate plant have very particular properties that
must be considered when treating them for NOx abatement: the gases
are cold and water saturated (60 to 80°C), and they contain fluorine.
This makes use of the best studied processes almost impossible.
As the problem of disposal and treatment of liquid effluents from a
nitrophosphate plant is very sophisticated and difficult, the first
concern should be to reduce their quantities as much as possible. In
practice the most efficient way to do this proved to be as follows.
1. Select the most suitable type of process equipment for the particular
plant. Pay particular attention to:
a. Gas scrubbers. These must have a high level of efficiency, using
nozzles and special fittings to ensure a good flow of liquid and
perfect gasliquid contact, to avoid solids deposits or clogging.
Many types of scrubbers are used, such as venturi type, with
Intalox or metallic rings, with floating Doyle-type balls, and so on.
The most important consideration is to select the most adequate
construction type for every specific operating condition, to
facilitate proper long-term operation.
b. Pumps. A great number of heavy-duty pumps are used in a
nitrophosphate plant, so proper selection is critical to avoid
significant leakage. Experience has demonstrated that the best
devices for this purpose are centrifugal pumps with an auxiliary
impeller and stuffing box sealing. Leakage is minimal for this type
of pump.
2. Select equipment and piping that will provide the best collecting
systems, as well as for equipment overflows. In Romanian plant
operating practice, a good solution has been found to be the use of
underground sumps, consisting of:
a. A stainless steel tank with agitator and an overflow device
leading to a nearby sump built in concrete and protected with
acidproof bricks. In the first stainless steel tank,
 Page 443
plant equipment overflows, as well as emptyings from tanks,
vessels, and pipes, are collected and routed to every building level
through piping. The liquids collected are analyzed as to chemical
composition and pumped by a submerged pump back to the plant at
the best location in the process, depending on the analysis.
b. An acidproof concrete sump, also provided with agitation, near
the first one, where the contaminated sewage from the floor traps
and sloped channels is collected. Into the sump also go the
incidental wash water, incidental equipment leakage, floor
washings, as well as any other incidental leakage. Chemical
analysis is performed and the liquid is pumped accordingly, by a
submerged pump, either to a certain spot in the plant or to the pond
water loop. The sumps are interconnected, so that if need be their
capacities can be joined in case large effluent quantities used to be
collected.
3. Select the most suitable construction materials for equipment,
piping, ducts, and so on, in the plant. It is well known that the working
media in a nitrophosphate plant are highly corrosive. The stainless
steels most commonly used are AISI 304, AISI 304L, AISI 316, AISI
316L (L for low-carbon steel used for welded pieces), and W.1.4136
or W.1.4435 (for pumps). In addition, the materials used for concrete
marks, acidproof floor linings, and floor traps, and well as coating
materials, have to be selected to resist to the heavy-duty working
conditions.
4. Periodic routine washing procedures should be implemented in the
plant area where such operations are needed. Such a system usually
consists of collecting vessels, pumps, piping connections, and so on.
The wash water is used in a closed circuit, and only after a certain
number of recirculation cycles have been completed can the
contaminated water be recycled back to the plant, if at all. The system
is then refilled with fresh water (usually, purges out of the general
recirculation water loops are used).
A special, but important problem is that of pond construction and
operation. The pond is a part of the collecting and treatment system
for liquid effluents expelled from a nitrophosphate plant. The
following matters are of importance:
1. An effort should be made to diminish pond size as much as
possible, as ponds can be a principal pollution hazard. Problems with
ponds can pose extremely dangerous environmental hazards,
particularly with regard to groundwater but also for the neighboring
land area. This goal is met by avoiding hydraulic transport of solid
wastewater slurry. Mechanical solids conveyance is used, with storage
in stockpiles located in specially prepared areas (see also Chapter 34).
This minimizes the solids storage area [7], the pond itself being
strictly limited by solids settling considerations, as the effluents
always contain small amounts of solids. The clear upper layer is
pumped back to the nitrophosphate plant.
2. Ponds should be situated in areas of low permeable soil (clay or
marl). Even under such conditions, all pond areas including the
perimeter sloped planes, must be provided with protective plastic
layers or a block of special concrete, major concern being paid to
sealing all open spaces. After settling takes place, the clear upper
layer is usually collected in smaller auxiliary basins, either by
variable-height dams or reverse underground pipes. The clear liquid is
then pumped by submerged pumps made of special stainless steel.
When the pond capacity is exhausted by solid waste (usually, after 10
or 12 years if the plant has been operated correctly), the pond is
abandoned and another one built. The older one is then emptied and
the useful components are recycled; the pond area may be covered
with soil and used for agriculture. If it is decided to empty the pond
and reuse it for its original purpose, it must be examined very
carefully for tightness of the acidproof insulation and any repairwork
done before the pond is reused.
 Page 444
3. Ponds should be located as close as possible to the plant. This
avoids long lengths of piping to and from the pond, which keeps
operating costs to a minimum and reduces the risk of leakage of
harmful substances.
4. The piping should be built of material that is highly resistant to the
fluids being pumped and should have good mechanical strength. It is
recommended that piping be located above ground, so that any
leakage may be seen and repaired as rapidly as possible.
5. A drain should encircle the pond, and there should be an adequate
number of check wells from which the leaked liquid is frequently
collected and analyzed, so as to detect any damage to the pond
structure and remedy it immediately.
It is important that the entire pond system be considered an integral
part of the nitrophosphate plant and be designed, built, and operated
with the greatest care, as any deficiency in the system could lead to
shutdown of the entire plant with consequent production losses.
C
Economics
It is obvious that it is very costly to handle of the liquid effluents
problem in a nitrophosphate plant, as environment-protection factors
are generally involved. The necessary expenses are related to the
implementation and operation of such sophisticated systems as those
described in Secs. IV.A and IV.B. The costs for efficient, well-
designed systems may be as much as 15 to 25% of the total
investment cost (based on the entire assembly cost of a nitrophosphate
plant); or 5 to 10% of the cost of operation. All costs must be justified
by the economics of the final product (i.e., the fertilizers). To
minimize these costs, several factors should be considered:
1. Keeping the freshwater makeup to the process as low as possible,
and replacing most of it with recirculated contaminated water
2. Returning to the process the maximum quantities of the nutrients
contained in liquid effluents
3. Developing markets for a variety of by-products
4. Cooperation with other plants in the industrial park or locale where
the nitrophosphate plant is located in the use or treatment of liquid
effluents
D
Factors to Consider When Locating a Nitrophosphate Plant
Nitrophosphate plants are great water consumers. In addition, these
plants are often located in industrial parks or areas where plants
manufacturing products of the following type are located: ammonia,
nitric acid, calcium nitrate or ammonium nitrate (nitrochalk) (calcium
nitrate crystallization process), and ammonium sulfate sources (for
processes using sulfate ion for calcium precipitation; perhaps for
caprolactame production, with ammonium sulfate as a by-product).
All such plants would be large freshwater consumers, even if provided
with substantial water cooling and recirculation stations. They also
often use a water source to assist in the dispatch of contaminated
water resulted from the manufacturing processes. In the past, such
industrial units were often located near rivers or the sea, where at least
the effluent problem could be managed more easily. But as
environmental protection regulations and restrictions become more
strict, liquid effluents should be handled so as to minimize harmful
effects under all circumstances. Rarely, despite great care to monitor
such a sophisticated problem, some incidental effluent contamination
may occur. In such a case, a retention basin should be avoidable where
 Page 445
contaminated effluent can be collected and analyzed, then returned to
the plant or emitted in small quantities to keep any pollution within
allowable limits. It bears emphasis that even when great effort has
gone into monitoring the effluent problem, some pollution effects are
difficult, if not impossible, to avoid. Because of this problem, and
because nitrophosphate plants need a great deal of land in addition to
that required for the plant itself, it is highly desirable to locate the
plant in largely uninhabited areas. Climatic and geographical
conditions are also important when choosing the plant location. The
temperature regime, dominant wind direction, conditions of
atmosphere dispersion, and atmospheric humidity are all factors that
cannot be neglected. It should also be noted that when an industrial
complex is located in close proximity to an important agricultural
area, it is sometimes possible to arrange to have farmers to utilize
some quantity of by-product (or waste) material, or even to arrange
for the use of substantial amounts of high-nutrient-containing liquid
effluents to be spread on the fields as liquid fertilizers, after treatment.
V
Waste Calcium Carbonate:
Transport, Storage, and Utilization
In nitrophosphate fabrication (Sec. II) the most common solid waste is
calcium carbonate, which has a certain content of ammonium nitrate
(when the calcium nitrate crystallization process is used) or of
ammonium sulfate (when the calcium sulfate precipitation process is
used).
A
Product Characteristics
Product quality differs depending on the process used. When a well-
designed and properly operated calcium nitrate conversion process is
used, the solid product specifications following filtration on rotating
filters may be expressed as follows:
Humidity 1520%
H2O
Nitrogen content (as 0.71% N
AN)
Granule size 40200 µm

For solids separation from the slurry, other types of equipment may be
used as well, such as belt filters and centrifuges, but the product
specifications are not very different from those noted above. When
calcium sulfate conversion is used, the separated solids have a higher
humidity and smaller granule size.
B
Possible Utilization
In older plants, calcium carbonate was considered a waste and was
stored in stockpiles or ponds. Later, research was undertaken to find a
use for this product, such as an agriculture amendment for acidic soils
[6], an additive to nitrochalk manufacture [6,9], and an additive to
certain complex fertilizer grades when an inert material is required to
lower the nutrient content (e.g., 20200, 151515). In principle, when
using calcium carbonate in nitrochalk manufacture or as an additive to
complex fertilizers, the product must be dried. This operation can be
performed in flash driers. Some fertilizer producers claim that
processes have been developed which make it possible to use the
product without drying (e.g., BASF-Odda process used in the
Agrolinz plant) [8]. In some
 Page 446
nitrochalk processes, the material has proved to be highly reactive and
has caused problems. This has led to process modifications [6] so that
the product to eliminate problems.
C
Transport and Storage
When considered as a waste, calcium carbonate can be transported
and stored in two main ways: (1) by hydraulic transport and pond
storage, or (2) by mechanical conveying and stockpile storage.
Usually, hydraulic transport is associated with cooling and
recirculation of contaminated water in high-capacity ponds (see also
Sec. IV.B) and includes solids reslurrying of the entire quantity of
water exiting the plant in a specially designed sump, the slurry
pumping (solids content is 1 to 2%) to a high-capacity settlingcooling
pond. The area necessary for an up-to-date average-capacity
nitrophosphate plant (about 2800 tons/yr fertilizers) could be as large
as 30 to 50 ha (for a 10-year storage period). The clear layer that
results after setting is collected in a separate small basin and is
pumped back to the plant. The calcium carbonate settles on the pond
bottom and cannot be recovered until the pond completes its cycle of
operation. In addition, there is a high level of pollution hazard (above
and underground) over the entire area, very high investment costs, and
poor cooling ability in temperate climatic zones.
Mechanical conveying and stockpile storage, if well designed and
properly operated, avoid most of the disadvantages noted above. This
method consists of belt-conveyor (usually) transport of the solid
product, which has good transport properties, being a rather dry,
consistent material that does not adhere to the belt at a repose angle of
30 to 40°. A rubber belt can be used with good results. All-metallic-
component conveyors are not recommended, as the nitrogen
component is corrosive on carbon steel parts.
Storage is in specially prepared stockpiles 1 to 2 ha in size. The land
area must be compacted and covered with a concrete layer and tarmac.
A stockpile feeding system has to be provided; either a mobile
discharge device on a belt conveyor or a central tower of great height
(25 to 30 m) may be used with good results. The material settles in
accordance with its repose-angle characteristics up to heights of 25 to
30 m. Sloped ditches are provided all around the pile to collect
rainwater leading to the sump. The water collected (having some
solids in it) may be pumped to the purge treating station (see Chapter
9). Generally speaking, experience has shown that rainwater
penetrates the solid surface layer for only 10 to 15 cm, and that no
danger of pile crushing exists. The product can be removed from the
pile using mechanical equipment such as a bucket loader, and
transport to the client is by railroad car or by truck. To access the pile
area, roads or lanes must be provided.
References
1. Phosphate FertilizersProperties and Processes: A Study of
Technological, Economic and Agronomic Considerations, Technical
Bulletin 8, 2nd ed., The Sulphur Institute, October 1966, pp. 2529.
2. C. H. Davis, R. S. Meline, and H. G. Graham, Jr., TVA mixed acid
nitric phosphate process, Chemical Engineering Progress, (Vol. 64,
No. 5, May 1968) pp. 7582.
3. S. Swanstrom, The Kemira NPK process: a design for flexibility
with environmental cleanliness, Phosphorous and Potassium, No.
145, SeptemberOctober 1986, pp. 3436.
4. S. Swanstrom, Agrolinz's new high-solubility nitrophosphate plant,
Phosphorus and Potassium, No. 166, MarchApril 1990, pp. 2124.
 Page 447
5. S. Swanstrom, Fertilizers by the nitrophosphate route, Phosphorus
and Potassium, No. 155, MayJune 1988, pp. 2833.
6. N. Popovici, Fertilizer industryenvironment pollution source:
technical solutions and technological advances made in Romania to
control environmental pollution effects arisen from fertilizer plants,
UNIDO Expert Group Meeting, Helsinki, August 1974.
7. Romanian Patent 89,808 1977.
8. E. Assum, Modern nitrophosphate technology based on the Norsk-
Hydro process, paper presented at the Fertilizer (Planning and
Development), India Ltd., Norsk Hydro Seminar, Delhi, December 2,
1981.
9. E. Aasum, Agrolinz's new high-solubility nitrophosphate plant
(designed for maximum environmental protection), Phosphorous and
Potassium, No. 166, MarchApril 1990.
 Page 449

33
Nitrophosphate-Based NPK Fertilizer:
Radioactive Components and Their Removal
Florin T. Bunus
Institute of Atomic Physics,
Bucharest, Romania
I
Radioactivity of Phosphates Used in the Nitric Acid Process
In Chapter 20, a description of natural phosphate deposits used in the
process of fertilizer production through sulfuric acid attack noted the
importance of radioactivity in the final product. The radioactivity is
due to two main components: uranium and radium.
Although dominant, sulfuric acid attack is not the only way of
obtaining phosphate fertilizer. At present there are many plants in the
world that process phosphate rock by nitric acid attack, the second
method of importance in producing chemical fertilizer. The same
sedimentary or volcanic phosphates as those discussed in Chapter 20
are used during the manufacture nitrophosphate fertilizers.
Because organic matter is not tolerated in the process of nitric attack,
calcined phosphates are used instead. This does not significantly
change the uranium content of the rock, but it might change the
chemical state of uranium due to the reductive medium. This last
factor is not very important during solubilization in nitric acid. This
reagent is more efficient, and because of the permanent presence of
traces of HNO2, it is an oxidant for Fe(II) and U(IV).
Calcination of phosphate may affect the radioactive equilibrium, but
no important influence on radium content was noted. Radioactive
RaRn equilibrium, even if disturbed by Rn removal, is reinstated with
time. Therefore, the principal radioactive components, U and Ra, are
practically unaffected before rock solubilization.
As mentioned previously, radioactive equilibrium is sometimes
disturbed in natural phosphate deposits because of natural factors.
Chemical analyses of uranium in several natural calcined phosphates
destined for nitric acid attack are given in Table 1.
 Page 450
Table 1 Chemical Analysis of
Calcined Phosphates
Phosphate %U
origin
Morocco 0.0130.015
Israel 0.0080.013
Florida 0.0100.016
Kola 0.001 (resulted from
peninsula flotation)

II
Phosphate Rock Solubilization by Nitric Acid Attack
In the sulfuric acid process the rock is digested and sulfur is thrown
away as gypsum mostly contaminated with Ra. In the nitric acid
attack of phosphate rock, a phosphonitric solution is obtained [1,2]. At
the same time, Ca(NO3)2 may be separated by crystallization or
CaSO4·2H2O by precipitation in the presence of (NH4)2SO4.
Uranium in the phosphate rock is dissolved in a phosphonitric
solution. It must be noted that the digesting process is carried out with
an excess of HNO3. The same time with uranium, its decay products
are found in phosphonitric solution. Due to the catalytic influence of
HNO3, uranium is always found in solution in the hexavalent state.
In the event that part of the Ca(NO3)2·4H2O is separated by
crystallization, only a small quantity of radium is carried on solid
crystal; therefore, its transformation with (NH4)2SO4 to NH4NO3 and
gypsum has no influence on the radioactivity that is left in the solid
phase. The same thing is valid in the case of conversion with
(NH4)2CO3 to CaCO3 and NH4NO3. The solid phase carries the low
radioactivity content of Ca(NO3)2·4H2O. Therefore, in the nitric acid
attack of phosphate rock, practically all the uranium and radium are
left in phosphonitric solution even after separation of
Ca(NO3)2·4H2O or CaSO4·2H2O. Determination of uranium from
phosphonitric solutions of various sources of phosphates is given in
Table 2.
A typical chemical analysis of a phosphonitric solution used in the
determination of uranium was 25% P2O5 and 10% HNO3, with a
density of 1500 kg/m3. The Norsk-Hydro process was used on an
industrial scale, and most of the Ca(NO3)2·4H2O was separated by
crystallization (cooling). It has already been mentioned that almost all
226Ra was dissolved in phosphonitric solution and a negligible
quantity was carried in Ca(NO3)2·4H2O. Awadalla and Habashi [3]
dissolved a 100-g rock sample containing 0.012% U in
Table 2 Uranium
Determination in Phosphonitric
Solutions of Various Sources
Phosphate Phosphonitric
origin solution (g/L U)
Morocco 0.120.14
Israel 0.100.13
Florida 0.100.16
Kola 0.010.02
peninsula
 Page 451
equilibrium with decay products. Thus the radium content of the
phosphate rock is 3.96 × 109 g Ra. The solution obtained was kept in
sealed bags for 40 days and Ge-Li gamma spectrometry
measurements gave the radium content. This behavior clearly
demonstrates the complete solubility of Ra in the phosphate rock.
However, the authors [3] found that by cooling and crystallization of
Ca(NO3)2·4H2O, about 10% of the Ra was carried in the solid phase.
Uranium in phosphonitric solution is not carried by Ca(NO3)2·4H2O.
226Ra content in phosphonitric solutions is dependent on the initial
rock dissolved and the degree of URa equilibrium, and is usually 20 to
50 pCi/g. Experimental determinations have shown that both U and
Ra from phosphonitric solutions are responsible for aqueous-phase
radioactivity. In the case of phosphoric acid (WPA), only uranium was
present in the aqueous phase. When Kola rock of volcanic origin is
used, the radioactivity of phosphonitric solution is low, approximately
2 to 5 pCi/g. Therefore, all fertilizers of nitrophosphate type (NP or
NPK) obtained by nitric acid attack of the rock will be radioactive.
Both uranium and radium of the natural phosphate are found in the
final product.
Because of the strong complexing power of H2PO4 and NO3 ,
uranium present in phosphonitric solutions is likely to form various
complex species when in the hexavalent state. Ionic species such as
UO2 2+ are also present because extractions with DEPA are possible
and explain this behavior. The assumption of neutral species existence
with NO3 must also be taken into consideration due to uranium
behavior in an extraction process with TBP or TOPO.
III
Radioactivity of Fertilizers of Nitrophosphate Type
In the process of nitrophosphate manufacture, phosphonitric solutions,
with or without separation of Ca(NO3)2·4H2O, is treated with
ammonia and is concentrated. Various methods are possible, but a
product fertilizer of NP type is finally obtained. Potassium is
sometimes added in the process; then a NPK complex fertilizer is
produced. It should be noted that nothing is separated during the
production process and all radioactivity due to uranium and radium is
also present. In the alternative of potassium introduction, its
radioactivity, however negligible, is added. Such a product will have
0.01 to 0.016%U and 25 to 40 pCi/g Ra. This value is high compared
with that of such natural materials as concrete and bricks, 0.7 to 2.0
pCi/g Ra, or Kola fertilizer, 2 to 4 pCi/g Ra.
Nitrophosphate fertilizers are continuously spread on agricultural
lands, and their radioactivity effects on crops are not yet well known
[4]. Therefore, a reduction of radioactivity was taken into
consideration in this study and is based primarily on work carried out
by the author. Uranium from fertilizer is a heavy metal and may show
some cumulative effect as a heavy metal in the soil, apart from the
fact that it is radioactive. In its turn, radium is considered a very
dangerous radioactive element either in water or air, and drastic limits
have been imposed.
However, no special regulations have yet been imposed on fertilizer
use, but it is understood that reduction of uranium and radium is
beneficial.
IV
Removal of Radioactivity from Nitrophosphate Fertilizer
The elimination of radioactivity from solid fertilizer is not possible;
this process is preferably carried out in phosphonitric solution. Until
now removal of radioactivity on
 Page 452
an industrial scale has not been reported, nor has experimental work
on a laboratory scale. However, there has been experimental work [5]
in which recovery of uranium and rare earths was attempted directly
from phosphonitric solution, and useful results have been obtained.
The purpose of this work was primarily uranium recovery as a source
of energy rather than the reduction of fertilizer radioactivity. In the
alternative of uranium elimination from phosphonitric solution, the
fertilizer produced will have lower radioactivity.
In the case of phosphonitric solutions both uranium and radium
elimination are required to attain a nonradioactive nitrophosphate
fertilizer. Therefore, two aspects of the problem are involved here
because simultaneous efficient, economical removal of radioactivity
with the same reagent is not yet feasible. Uranium and radium were
eliminated separately from the same phsophonitric solution. A
precipitation process was used to remove the radium; then a solvent
extraction process will remove and recover uranium as a valuable
energy source that would otherwise be wasted.
In work by the author it was established that by solvent extraction,
most rare earth elements present at an average concentration of 1 to 2
g/L are also extracted. It should be noted that rare earths are heavy
elements that might play a role in a nitrophosphate fertilizer when
spread on agricultural lands. Rare earth concentration in phosphonitric
solutions is much higher than in the case of phosphoric acid, where
only 25% are solubilized; therefore, emphasis on this subject is
worthwhile.
A
Radium Elimination
Awadalla and Habashi [3] have determined that on cooling
phosphonitric solutions, with Ca(NO3)2·4H2O separation by
crystallization only 10% of radium is coprecipitated. Experiments
based on CaSO4·2H2O precipitation are not efficient from
phosphonitric solutions as was true in the case of phosphoric acid
from sulfuric attack of the rock.
The best carrier for radium from acidic solution was determined to be
barium as BaSO4 [6] when microcomponent Ra is coprecipitated as
RaSO4. This is the classic method, which has a slightly different
efficiency from phosphonitric solutions. In this medium, Ba(NO3)2
and Na2SO4 or H2SO4 are added in small quantities up to 10 g/L
Ba2+. Radium is coprecipitated by BaSO4, the maximum precipitation
yield is 80%, and most of this element is removed. In similar work
given in the literature [3], a higher precipitation yield of Ra on BaSO4
has been obtained. The process was carried out in the presence of
Ca(NO3)2 and the authors [3] found that a mixed precipitate of BaSO4
+ CaSO4 was built up. In this treatment 90% of the radium was
removed from phosphonitric solutions. To attain this yield, 20 kg of
Na2SO4 and 20 kg of Ba(NO3)2 are required for 1 ton phosphate rock
(0.012% U). The consumption of Ba(NO3)2 is not negligible, but an
acceptable amount of radium removal was noted without affecting
phosphonitric solutions.
There are various ways of obtaining increased radium elimination by
increasing the pH of phosphonitric solutions with ammonia. The
process used in nitrophosphate manufacture does not add to the
precipitation yield of radium. After radium removal, the only major
radioactive component left is uranium. At present, no industrial
process is known to remove uranium and radium concurrently from
phosphonitric solution.
B
Uranium Removal from Phosphonitric Solution
In a previous work, the ORNL method used on an industrial scale [79]
based on a DEPA + TOPO synergic mixture led to uranium
elimination from technical phosphoric acid
 Page 453
obtained in the sulfuric acid attack of phosphate rock. Extraction
efficiency depends on DEPA concentration, temperature, WPA
concentration, and others factors [10,11].
There is only one experimental study in which uranium was extracted
directly from phosphonitric solution [12]. In this paper the usual
organophosphorus esters were used and proved to be efficient for
uranium extraction. It is noted in this paper that the interest was
related to uranium recovery from phosphonitric solution.
In phosphonitric solutions, U(IV) is easily oxidized to U(VI);
therefore, it is always present in hexavalent state.
Due to the strong complexing character of H2PO4 , UO2 2+ is strongly
bound to it. There is also a UO2 2+ uncomplexed species, which
behaves as such in the extraction mechanism with DEPA. In the
presence of NO3 ion, the existence of UO2(NO3)2 species is
supported by an extraction mechanism with TOPO and TBP. Other
charged species may also exist. In that paper [12] it was established
that U(VI) is not extracted significantly with TBP, due to the stronger
complexation power with H2PO4 but with a more basic extractant
such as TOPO, the extraction efficiency is higher. If uranium
extraction (distribution coefficient) in TOPO at various concentrations
in kerosene is plotted against the HNO3 concentration, Fig. 1 is
obtained. Phosphoric acid concentration was maintained constant at
3.5 M.
Uranium extraction increases with HNO3 concentration until a
maximum is attained; then a decreasing effect appears due to the
competition of HNO3 in the extraction process.
A solvation mechanism typical for UO2(NO3)2 species extraction is
evident, and in the organic phase, UO2(NO3)22·TOPO. Distribution
coefficient dependence on TOPO
 Figure 1
Dependence of distribution coefficient
on HNO3 molality in the aqueous phase for
3.5 M H3PO4. (1) 0.05 M TOPO; (2) 0.10
M TOPO; (3) 0.20 M TOPO; (4) 0.30 M TOPO.
 Page 454
molality in the organic phase is given in Fig. 2, where the slope is 2,
confirming this mechanism.
TOPO efficiency is important starting with 0.2 to 0.3 M concentration,
but this extractant is expensive for an eventual industrial process.
Another extractant described by the author [12] is DEPA in kerosene.
Uranium extraction D is given as a log-log plot against DEPA molality
for constant H3PO4 and various HNO3 concentrations (constant for
each curve) in the aqueous phase. With increasing HNO3
concentration, the value of D decreases. It is important to mention that
the U(VI) extraction yield is acceptable starting with 0.8 to 0.9 M
DEPA.
When TOPO synergist is added, the extraction effect is strongly
marked. Similar behavior is seen in the ORNL process [79] applied to
WPA resulting from sulfuric acid attack of the rock. Plotting log D
versus log TOPO molality, straight lines with unity slope are obtained
[12], as shown in Fig. 4.
In Fig. 3 in the absence of a synergist, there is also linear dependence
starting with approximately 0.6 M DEPA with slope value 2.
Dependence on H3PO4 concentration is also a straight line with slope
2 (absence of synergist). Data for the Figs. 3 and 4 confirm the
following extraction mechanism:

where (HX)2 is the chelating DEPA reagent, which is dimerized in a

liquid paraffinic hydrocarbon such as kerosene.
 Figure 2
Distribution coefficient dependence on TOPO
molality in the organic phase for 2.38 M HNO3
and 3.5 M H3PO4 in the aqueous phase.
 Page 455

Figure 3
Dependence of the distribution coefficient
on DEPA molality in the organic phase for
3.5 M H3PO4 in the aqueous phase and
(1) 2.38 M HNO3; (2) 0.79 M HNO3; (3)
0.15 M HNO3.
It is evident from Fig. 4 that D values are much higher for 0.6 M
DEPA and increase linearly with TOPO concentration. The acidity of
the aqueous phase was maintained constant in H3PO4, but three
values for HNO3 concentrations are given. Extraction yield decreases
with HNO3 concentration.
For each set of determinations, HNO3 was maintained constant. For
the alternative of replacing TOPO with TBP, the remainder of the
conditions being the same, the behavior is different. In this case the
synergistic effect, although comparable with DEPA alone, is lower.
For increasing TBP concentrations, an antagonistic effect appears.
This study makes it possible to establish the most suitable extractant
for uranium elimination from phosphonitric solutions. Experimental
data have shown that DEPA + TOPO give the best extraction yield,
similar to that with WPA.
The author has continued experimental work on technical
phosphonitric solutions obtained in an industrial fertilizer plant. The
extractants used were DEPA alone at 1 M concentration or 0.6 M
DEPA + 0.1 M TOPO. At the same time another extractant was
selected from the phenylphosphoric class. Based on previous
experiments, nonylphenyl phosphoric acid (NPPA, a mixture of mono
and diesters) has been used.
Experiments were carried out in four-step countercurrent extraction
equipment of the
 Page 456

Figure 4
Dependence of the distribution
coefficient on TOPO concentration
for 0.6 M DEPA in the organic phase.
Aqueous phase was 3.5 M H3PO4
and (1) 2.38 M HNO3; (2) 0.79 M
HNO3; (3) 0.15 M HNO3.
mixersettler type. The compound used to extract uranium from
phosphonitric solution was stripped with an appropriate reagent, and
uranium was found in the aqueous phase. The strip solution is the one
already noted, a mixture of H3PO4 containing Fe(II). Uranium is
reduced to U(IV)-nonextractible species in DEPA. The pure extractant
is recirculated to the extraction equipment. Uranium was recovered to
90%, and rare earths are obtained in a similar manner.
DEPA extractant from a phosphonitric medium was recycled many
times, each extraction being followed by an acidic stripping. DEPA
behavior was maintained constant without a change in the value of D,
which was 2.7 (in the absence of TOPO). Infrared spectra of organic
phase were left unchanged after many extractionstripping cycles.
Similar behavior is found by pH-metric titration when one
characteristic step of DEPA (diester) is kept unchanged with a number
of extractionstripping cycles. Therefore, no change in extractant
structure was detected and no hydrolysis process occurs. DEPA
extractant is very stable in both acidic media used.
For NPPA extractant in kerosene (0.1 to 0.2 M) the hydrolysis effect is
easily found in the change in the P-O (shift) absorption band, which
splits, and the appearance of an OH vibration band, indicating the
presence of phenol. The two characteristic steps in pH-metric titration
are also affected. Three or four extractionstripping cycles are enough
to detect the changes mentioned.
Another difficulty with NPPA is related to extractant strong retention
of U(IV) in
 Page 457
case of WPA and in this case of U(VI), Fe(III), and rare earths.
Stripping of these elements is a very difficult process and accumulates
with the number of cycles. The effect of this behavior is a decrease in
D with the number of cycles.
DEPA is completely stripped and is very stable, therefore is the only
extractant of practical interest. The presence of a synergistic effect due
to TOPO is reflected in a D value of 6 and removal of uranium from a
phosphonitric solution is carried out efficiently. This process, similar
to that in use for WPA, could be extended to phosphonitric solutions
with some modifications related mostly to HNO3 elimination from the
organic phase.
In another study [13] the idea was to extract phosphoric acid from
phosphonitric solutions. The extractant used was tertiary amyl
alcohol, which extracts both phosphoric acid and uranium. This
element is recovered from the alcohol phase by selective stripping,
and rare earths from aqueous raffinate by TBP. Two kinds of
fertilizers are obtained by this process; but radium will probably be
found in the aqueous phase, therefore in the low-grade fertilizer. Many
steps are involved in this process, and other data are required to draw
relevant conclusions.
As already mentioned, few data are found in the literature on this
subject; therefore, until now the only choice for eventual industrial
application must be based on DEPA as the extractant or DEPA +
TOPO. These extractants were applied on an industrial scale in the
case of WPA and are stable in phosphonitric solutions.
Starting with laboratory experimental data, a flowsheet was proposed
for radioactivity removal from phosphonitric solutions in the nitric
acid attack of the rock. There are two alternatives involved in this
flowsheet: a one-or two-cycle system. The first process requires lower
investment and operation costs, but the purity of the uranium
recovered is higher in the two-cycle system. For the purpose of
removing uranium from nitrophosphate fertilizer, the one-cycle
process is more adequate.
Because solvent extraction is used for uranium removal and the
organic phase is sensitive to solids in the aqueous phase, pretreatment
is required for phosphonitric solutions after separation of
Ca(NO3)2·4H2O. Clarification of phosphonitric solutions is carried
out in the presence of a polyacrylamide flocculant efficient at lower
temperature. Elimination of radium is possible at this stage in the
presence of Ba2+, and SO42 on the resulting BaSO4 .
Phosphonitric solutions undergo control filtration and are sent to the
extraction unit only in clarified form. Extraction takes place in a four-
step countercurrent mixersettler extractor. Uranium and rare earths are
extracted in the organic phase. Some inorganic impurities are
extracted at the same time, with small quantities of HNO3.
Phosphonitric raffinate leaving the extractor needs posttreatment to
eliminate organic extractant droplets entrained in the extraction
process. A settler is introduced for this purpose. Phosphonitric
raffinate leaving this settler still contains 100 to 500 ppm of organic
extractant fog. Organic matter is not tolerated in phosphonitric
solution, due to the danger encountered in the process of fertilizer
manufacture. Therefore, advanced removal of organic droplets is
required. Air flotation or special coalescers associated with active
carbon are involved. A very sensible, exact chromatographic method
was devised for entrained organics determination. Very stringent
conditions are imposed and 1 to 3 ppm organics are determined. Only
after this posttreatment is the phosphonitric solution sent to fertilizer
production.
Extractant leaving the extractor is sent to a scrubbing unit to eliminate
the HNO3 extracted; otherwise, uranium stripping is not possible.
Organic phase leaving the scrubbing unit, with the HNO3 eliminated,
is sent to the first stripping unit. At this stage
 Page 458
rare earths are removed in the aqueous phase. Uranium is left
unaffected in the organic phase.
The next step is the uranium stripping process and is based on U(VI)
reduction by stripping solution to U(IV). The strip solution consists of
6 to 10 M H3PO4 containing Fe(II). This medium is favorable to
U(VI) reduction:

Uranium is precipitated from strip solution as green cake, which is

filtered, washed, and dried.
In the two-cycle process, strip solution containing uranium as U(IV) is
diluted to 6 M H3PO4 and is also oxidized to transform the element in
hexavalent state. The second extractionstripping cycle is then applied.
The same extractant is used, but the stripping of organic phase is
carried out with ammonium carbonate with a precipitate of high purity
obtained. The flowsheet is given in Fig. 5.
There is an alternative to stripping both uranium and rare earths in the
first cycle as in Fig. 5. Uranium purification is then carried out in the
second extractionstripping cycle. In this case uranium and rare earths
are stripped together in the first cycle; separation takes place in the
second cycle.
V
Choice of Industrial Process, Costs, and Equipment
Actual literature data do not permit many choices concerning uranium
removal from phosphonitric solution. The only acceptable way with
industrial prospects is based on solvent extraction.
 Figure 5
Two-cycle flowsheet for uranium recovery from phosphonitric solution.
 Page 459
Several extractants have been tried, but only those from acidic
organophosphorus esters are efficient. Within this class, DEPA or a
synergistic DEPA + TOPO mixture is efficient, selective, and stable
with reasonable cost. Other extractants, such as NPPA, OPAP, or
OPPA, discussed in Chapter 20, have lower stability in phosphonitric
solution and are difficult to strip and therefore must be excluded.
DEPA extractant, the principal component, is very stable for long
contact with an acidic medium. No hydrolysis process was detected
by IR spectra, pH titration, or chromatography.
An important advantage of phosphonitric solution processing is
related to the absence of organic matter, which in the case of WPA is a
source of much trouble. Thus for brown WPA, appreciable investment
costs are required for the equipment used in organic matter removal,
as well as appreciable operation costs.
Phosphonitric solutions need a good clarification process, which is
possible only in the presence of a flocculant, and control filtration is
required for both the clarified solution and for the solids eventually
separated. The solids could be used as a fertilizer. Extraction
equipment consists of horizontal mixersettler extractors, which are
easier to control.
The difference between this process and that mentioned for WPA is a
scrubbing stage introduced for the organic phase before stripping of
uranium, to eliminate the partially extracted HNO3. Nitric acid in the
organic phase hinders uranium stripping (reduction).
Stripping stages are the same as discussed in Chapter 20, with an
emphasis on rare earth recovery. Uranium stripping is strongly
affected if traces of HNO3 are left in the organic phase. It is important
to notice that rare earth quantities are higher in this process, thus
providing an extra incentive, in addition to radioactivity removal, for
using the phosphonitric solutions.
Some experiments lead us to conclude that apart from the special
steels used for phosphonitric solutions, ebonite-lined equipment might
be used for large-scale industrial processing. The choice of a one-
cycle circuit seems to be more adequate for this purpose, when rare
earths and green cake are obtained with simple recirculation of
extractant without further treatment. The cost of such plants will not
be higher than that of a one-cycle process selected for WPA in the
case of green acid. The operation costs of this plant must be similar or
even lower if rare earths are taken into consideration.
However, extra equipment is required before or after uranium removal
for radium elimination. This could also be simplified if radium is
initially eliminated using the clarifier. The operation costs might be
affected by Ba(NO3)2 and Na2SO4 required for this process. Other
considerations already discussed in Chapter 20 are valid here also.
In this discussion the laboratory data were extrapolated on an
industrial scale, but some valuable data are taken from the WPA
process and also extrapolated here. However, pilot-plant experiments
are required before extablishing an actual industrial process. The role
of phosphonitric solution posttreatment is more important here due to
a self-ignition hazard in the final product if organics are present.
Therefore, air flotation and special coalescer equipment are used, and
when necessary, active carbon adsorption is required.
In recent years not much interest has been shown in uranium recovery
from phosphoric media in general and data do not exist for
phosphonitric solution, due to the large uranium surpluses during the
1980s. This trend, and market effects, have both changed. In terms of
the important role of radioactivity removal combined with uranium
and rare earths recovery, there have been additional incentives to
change the trends in market effects on uranium recovery.
 Page 460

VI
Fertilizer Radioactivity after Uranium and Radium Removal
It has already been mentioned that nitrophosphates resulting from
phosphonitric solutions are radioactive, containing radium and
uranium at concentrations many times higher than those of
environment materials. Spreading uranium on agricultural lands is not
only a waste of energy but a continuous source of radioactive
contamination. Radium is known to be very hazardous.
Elimination of radium and uranium from nitrophosphate fertilizer may
play a beneficial role in the environment. After such treatment, the
fertilizer obtained will have very low radioactivity, on the order of 2
to 4 pCi/g, similar to that of a fertilizer from Kola rock (of volcanic
origin). Such a fertilizer will also have a low rare earth content.
References
1. F. Habashi, Sulphur and the fertilizer industry, Arab Mining
Journal, Vol. 5, 1985, p. 74.
2. F. Habashi, Trends in fertilizer technology and its impacts on the
environment, Mater. Soc. , Vol. 9, 1985, p. 393.
3. F. T. Awadalla and F. Habashi, The removal of radium during the
production of nitrophosphate fertilizer, Radiochimica Acta, Vol. 38,
1985, p. 207.
4. D. A. Mays and J. J. Mortvedt, Crops response to soil application
of phosphogypsum, Journal of Environmental Quality, Vol. 14, 1986,
p. 78.
5. F. Bunus and R. Dumitrescu, Simultaneous extraction of rare earths
and uranium from phosphoric acid, Hydrometallurgy, Vol. 28, 1992,
p. 331.
6. G. Friedlander and J. W. Kennedy, Chemical phenomena at tracer
concentrations, in Nuclear and Radiochemistry, Wiley, New York,
1960, p 326.
7. F. J. Hurst, D. J. Crouse, and K. B. Brown, Recovery of uranium
from wet process phosphoric acid, Industrial and Engineering
Chemistry Process Design and Development, Vol. 11, 1972, p. 122.
8. G. Cordero, L. G. Jodra, J. L. Otero, and J. M. Josa, (1977)
Recuperacion del uranio del acido fosforico, Energia Nuclear, Vol.
21, 1977, p. 297.
9. F. J. Hurst, Recovery of uranium from phosphates: current status
and trends, Proceedings on the Recovery of Uranium from Phosphoric
Acid, IAEA Vienna, TECDOC 533, 1980, pp. 915.
10. F. T. Bunus, (1977). Determinations of low levels of uranium in
solutions obtained by acid attack of phosphate rock, Talanta, Vol. 24,
1977, p. 117.
11. F. T. Bunus, C. V. Domocos, and P. Dumitrescu, Synergic
extraction of uranium from phosphate solution with di(2-
ethylhexyl)phosphoric acid and tri-n-octylphosphine oxide, Journal of
Inorganic and Nuclear Chemistry, Vol. 40, 1978, p. 117.
12. F. Bunus, and P. Dumitrescu, Uranium (VI) extraction from acid
mixtures with organophosphorus esters, Hydrometallurgy, Vol. 16,
1986, p. 167.
13. F. Habashi and F. T. Awadalla, (1986) The recovery of uranium
and lanthanides during the production of nitrophosphate fertilizers
using tertiary amyl alcohol, Journal of Chemical Technology,
 Page 461

34
Storage and Transport of Solid Fertilizers
Robert E. Robinson*
Robert E. Robinson and Associates
Beaver, Pennsylvania
I
Introduction
Pollution develops in solid fertilizer storage and transport operations
from a number of causes. Regulatory as well as ethical requirements
of environmental protection, occupational health, and safety dictate
that extensive efforts be made to prevent such pollution in the first
place and to capture and handle properly pollutants that do escape.
Control and collection systems are not inexpensive but can provide a
number of benefits to production operations. As is always true in
process work, careful engineering can maximize the value obtained
and minimize cost. It is customary to focus on the more serious toxic
or hazardous sources first, then the less serious nuisance conditions.
Where possible it is usually cost-effective to prevent the release of
pollutants by incorporating relevant features into the process, storage,
and handling facilities during the original design or by adding capture
devices to act at points of generation. It is usually more difficult to
recapture material after it has been dispersed. The benefits of
pollution control include the provision of better working conditions,
improved employee health, greater safety, reduced housekeeping cost,
and reduction of loss of product inventory (shrinkage).
II
Collection Systems
A typical particulate collection system consists of a selected number
of pickup points at which dust-laden air is removed by hoods or
connections to machine housings through a system of ducts and
delivered to a central dust collector or wet scrubber by an exhaust fan.
Cleaned air is then vented safely to the atmosphere. If collected dry,
the solids are usually put back into the process at suitable points and
incorporated into products. If
*Formerly of Renneburg Division of Heyl & Patterson, Inc., Pittsburgh,
Pennsylvania
 Page 462
scrubbed, material is delivered in a slurry as scrubber effluent and
must be impounded or concentrated and processed.
III
Sources of Fugitive Pollutants
Unintentional spillage from broken bags, front-end loaders, belt
conveyors, and bucket elevators can deposit solids on roadways,
floors, platforms, and walkways. Subsequently, some of this spilled
material can be ground under truck and tractor tires in roadways and
work aisles, stirred, and become airborne fugitive dust. The best
practice for control of this type of fugitive dust is careful operation of
loading operations and regular cleaning.
Material handling equipment can generate dust when spillage occurs,
especially at loading points, discharge points, and where material is
energized mechanically, as in mixing, blending, screening, or
pulverizing operations. Dusting resulting from discharge of material
onto storage piles can be reduced by various manufacturing practices,
including adjustments to formulations and the use of coating
substances. Modern medium-and high-analysis mixed granular
fertilizers can be produced with a high percentage of smooth, hard,
dense, uniformly sized granules which inherently generate less dust in
subsequent handling. Certain raw materials and formulations enhance
production of less dusty products. The spraying of small quantities of
fuel oil or commercially available coating agents onto the cooled and
sized product can reduce dusting. Open transfer points where dust
could be released in fairly restricted spaces can be hooded and vented
by drawing appropriate volumes of dust-laden air into collection
systems.
Some equipment housings operate under conditions of positive air
pressure and tend to blow dust into the atmosphere. Housings for
screens, pulverizers, mixers, and elevators can also be vented into
collection systems. The most satisfactory method is to provide
plenums or breechings as appropriately located additions to equipment
housings, into which some of the dust-laden air is drawn. It is
desirable to operate the collection system under sufficient suction so
that negative pressure exists inside such items of equipment, if
possible.
Particulate and gaseous fugitive pollutants from fertilizers include
ammonia, ammonium chloride, nitrous and sulfur oxides, and
phosphoric and sulfuric acid mists. None of these substances are on
the list of 17 high-volume toxic industrial chemical pollutants listed in
the Pollution Prevention Act of 1990, but all are regulated by the U.S.
Environmental Protection Agency. Fluorine compounds occur in some
fertilizer materials and are quite toxic. It is important to be informed
about circumstances of release of these substances, their hazardous
effects on life and the environment, and methods of control, capture,
and safe disposal.
IV
Suggested Design of Procedure for Typical Collection System
1. Identify all pickup points to be incorporated into the collection
system. All points for a single system should be for similar types of
pollutants and be reasonably close together for the best system
economy and performance.
2. Quantify the actual design cubic feet per minute of air required to
be collected at each point for effective performance. It is advisable to
design on the high side on flows and pressure drops and to design
with a suitable margin of safety in the collector and
 Page 463
exhaust fan so that if a system change is necessary or if a design error
is discovered later, the excess capacity can overcome any unforeseen
deficiency.
3. Estimate dust particle characteristics, including particle-size
distributions, true particle densities, particle shapes, and grain
loadings of solids per standard cubic foot of air to be collected for
each collection point.
4. Prepare a preliminary equipment general arrangement sketch or
drawing of the proposed system, locating the various hoods and
pickup points, ducts, dust collector, induced-draft exhaust fan, exhaust
stack, and hood.
5. Choose an approximate average system gas velocity for efficient
particle transport to be used in the collection ducts, which considers
dust particle characteristics.
6. Calculate approximate preliminary duct areas and diameters for
each branch and trunk duct, dividing total actual cubic feet per minute
of airflow for that branch by the average system gas velocity chosen.
Round the calculated values to the nearest standard duct diameters.
7. Convert each run of duct into an equivalent run of straight duct,
summing straight lengths plus equivalent values for elbows,
transitions, and so on, to obtain equivalent total lengths.
8. Calculate trial estimates of total system pressure drops for each
branch, including allowances for pickup hood entry losses and plenum
entry losses. Where a trunk duct consolidates more than one branch
duct, the flows and pressure drops must be calculated separately for
each branch and each consolidated trunk and then combined. Ideally,
each complete branch, from pickup points to collector, should have
the same total pressure drop. If one branch has a significantly lower
pressure drop, this means less resistance to airflow in that branch, and
in operation more of the total quantity of air than intended will then
flow in that branch. In such a case the diameter of that branch should
be reduced in an effort to achieve a balance of all flows close to the
desired amounts.
9. Using the total design system flow, including any safety margins for
excess capacity, the pressure drop for the dust collector and/or
scrubber chosen and for the exhaust stack and hood should be
determined and added to the collection duct system pressure drop, for
a total system pressure drop. This pressure loss must be overcome by
the exhaust fan, which can now be sized. The fan should be chosen
with excess volumetric and pressure capacity for a design safety
factor. Where V-belt drives are used, a simple speed and inexpensive
fan speed change makes system volume flow adjustments quite simple
during startup.
10. The preliminary design should be reviewed for any imbalance of
pressures or flows. If necessary, the design should be adjusted by
changing a duct diameter or rerouting a duct to bring the entire system
into as close a good overall balance as possible. Shifting the location
of the dust collector and/or making trade-offs in the total system are
sometimes necessary. Pulling more air than necessary from a
particular pickup point is usually not a serious compromise as long as
all points receive adequate suction and flow.
11. When the preliminary design is satisfactory, the final system can
be designed in detail. The foregoing design steps should then be
carefully repeated using accurate dimensions from the final general
arrangement drawing to check all pressure drops and flows.
12. Provisions for access for inspection and cleaning of the entire
system should be provided carefully. Efficient system performance is
very dependent on system maintenance and cleanliness.
13. The possibility of excessive gas cooling and formation of
unwanted condensate
 Page 464
in the ducts, dust collector, or exhaust fan should be considered. In
some climates insulation and/or heaters are used to keep humid
mixtures of air and water vapor well above the dew point of the
mixture. A dry-bulb temperature 30 to 40°C above the dew point of
the humid air mixture is customary. Wetting of duct walls will lead to
sticking and buildup of dust, and wetting of the walls and bags in a
dry dust collector will lead to blinding of the bags, requiring frequent
cleaning of the system.
V
Recovery of Pollutants by Chemical Reactions
In many fugitive pollutant situations, the pollutants can be collected
dry and often can be reintroduced into the process, which provides a
convenient, low-cost method of disposal. Occasionally, a particular
pollutant can be recovered in a scrubbing application and the effluent
used in a process, avoiding the necessity of a collection pond or
subsequent treatment costs. These situations occur more often in
manufacturing than in storage and transport situations. An example is
the scrubbing of air containing ammonia with a weak phosphoric acid
solution and using the effluent in DAP manufacturing processes. It is
necessary that the effluent concentrations be sufficient so that not too
much water is introduced into the processes. Another example is the
scrubbing of silicon tetrafluorideladen air from phosphate acidulation
production units and storage and handling facilities with water in
multistage scrubbers with subsequent filtration of effluents and
adsorption of impurities to yield a salable fluosilicic acid by-product.
Collection of highly dispersed pollutants from remote sources is
usually not economical but may be legally necessary.
VI
Impoundment of Contaminated Rainwater Runoff
Where spillage of hazardous materials inside or outside a building
cannot be avoided or where fugitive dust from manufacturing, storage,
or handling operations accumulates and is subsequently washed away
by rainwater, impoundment into cisterns has been proposed. Curbs
around paved areas, sloping floors with sumps and drainage troughs,
and systems of gutters and downspouts to temporary storage cisterns,
although costly, are technically feasible and probably necessary for
new facilities that will handle toxic materials.
Brief Selected Annotated Bibliography
To be truly useful, a technical bibliography should be of limited length
and contain only carefully selected references which are thought to
contain material specific to the general subject and likely to be helpful
to knowledgeable practitioners who already possess a functional core
of technical expertise. In contrast, a historical bibliography should
contain virtually every pertinent reference known to the author to
facilitate future efforts in research, development, or writing. The
writer feels that brief comments can save readers a great deal of
unnecessary time, effort, and expense in searching out disappointing
material by steering them succinctly to essential, important supporting
and amplifying information.
 Page 465
1. J. L. Alden, Design of Industrial Exhaust Systems, 3rd ed.,
Industrial Press, New York, 1959. Dated, but a good primer with
much useful information on industrial air handling and exhaust
systems.
2. Committee on Industrial Ventilation, Industrial Ventilation,
numerous editions, American Conference of Governmental Industrial
Hygienists, Lansing, Mich. A basic tutorial with broad general
coverage and detail.
3. J. A. Danielson, ed., Air Pollution Engineering Manual, 2nd ed.,
U.S. Environmental Protection Agency, Research Triangle Park, N.C.,
1973. Dated; much material specific to certain industries (not
including the fertilizer industry) and to types of control equipment.
4. J. M. Dallavalle, Micromeritics: The Technology of Fine Particles,
2nd ed., Pitman, New York, 1948. Dated; much useful information.
5. R. M. Harrison and R. Perry, eds., Handbook of Air Pollution
Analysis, 2nd ed., Chapman & Hall, London, 1986. Complete
treatment of monitoring, sampling, and analysis.
6. J. M. Marchello and J. J. Kelly, Gas Cleaning for Air Quality
Control, Marcel Dekker, New York, 1975. Covers major types of
control equipment and control practice. Very useful.
7. G. Nonhebel, ed., Gas Purification Processes, George Newnes,
London, 1964. Dated; thorough, covers some chemical as well as
mechanical methods of purification. Very useful.
8. W. Strauss, Industrial Gas Cleaning, Pergamon Press, London,
1970. Dated; much useful information.
9. ASHRAE Guide and Data Book, various editions, American Society
of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta,
Ga. Very broad coverage. Industrial section; some industry-specific
information (not including fertilizer).
 Page 467

35
Ammonium Phosphate Solutions and Suspensions
Mark T. Holt
Tennessee Valley Authority
Muscle Shoals, Alabama
I
Introduction
Since the early 1970s the fluid fertilizer industry has incorporated use
of the Tennessee Valley Authority (TVA) pipe reactor process for
production of ammonium polyphosphate (APP) solutions by the
ammoniation of superphosphoric acid (SPA) [1]. The process is a
method of producing fertilizer composed of nitrogen and phosphorus.
Nitrogen is supplied to the process in the form of anhydrous ammonia.
Phosphorus is supplied as wet-process superphosphoric acid. The only
other raw material is water, which is used to dilute the plant nutrients
to stable, standard grades of commercial fertilizer. The standard grade
manufactured by this process is 10340, which equates to 10%
nitrogen, 34% phosphorus pentoxide, and 0% potassium. The 10340
product made through this process of ammoniating superphosphoric
acid has a high polyphosphate content (usually over 70% of the P2O5
present as polyphosphate) and because of the simplicity of the
process, was readily accepted throughout the fluid fertilizer industry
in the United States. Currently, about 1.8 million metric tons of 10340
are produced annually in the United States [2].
II
Description of Major Pieces of Equipment and Raw Material
The production procedure is a three-step process: (1) ammonia and
acid react, (2) reaction products are brought to correct proportions,
and (3) the product is cooled. Figure 1 shows a flowsheet of the 10340
process as it is operated today at more than 150 plants across the
United States. Phosphoric acid is fed directly from a railcar by means
of a variable-speed, positive-displacement metering pump. About 70
to 80% of the ammonia is vaporized to a temperature of 48.9°C in a
shell-and-tube heat exchanger using 60.0°C 10340 product on the tube
side. The vaporized ammonia is fed directly to the pipe reactor, where
it reacts with the phosphoric acid. The acid and ammonia enter the
 Page 468

Figure 1
10340 flowsheet.
water-jacketed 15.24-cm-diameter pipe reactor, where a reaction
temperature of about 343.3°C is attained. The melt from the pipe is
then discharged below the liquid level in the hot well where it is
quenched. The remaining 20 to 30% of the needed ammonia is fed as
a vapor and injected into the APP liquid in the sump of the packed-
tower cooler. About 60% of the needed formulation/evaporation water
is fed through the water jacket of the reaction pipe to control pipe skin
temperature before it flows into the process. The balance of needed
water is injected directly into the recirculation system to adjust the
grade of the product. All water metered to the process is used for
evaporation or grade dilution. The liquid in the hot well is pumped to
the top of the packed-tower cooler. Air is pulled through the packing
countercurrent to the flow of product by a fan located above the
demister section. The liquid in the hot well is sprayed over the bed of
packing by spray nozzles. A demister section consisting of woven,
stainless steel wire mesh is located above the spray nozzles.
 Page 469

III
Typical Emissions Discharged from the APP Process
During the production of 10340, gaseous ammonia, fluorine, and
particulate are emitted from the cooling tower into the atmosphere. To
quantify these emissions, TVA conducted emission sampling tests on
five separate commercial 10340 plants during a period of about one
year from August 1990 to September 1991. In all cases, gaseous
ammonia was the primary pollutant emitted from the process [3].
IV
Methods of Measuring the Emissions
At the time the 10340 plants were sampled for emissions, the
Environmental Protection Agency (EPA) did not list an ammonia
sampling procedure in the Code of Federal Regulations (CFR),
although the CFR did list procedures for determining fluorine and
particulate emission rates [4]. TVA worked with the various state
regulatory agencies in which the sampling would be performed to
develop a suitable method for measuring ammonia emissions from the
10340 process.
Figure 2 shows the main components of the stack sampling train as it
was used in all the sampling tests. A Universal Stack Sampler, which
is designed to sample gas streams effluents isokinetically according to
EPA standards, was used [5]. The sampling equipment was calibrated
and pretest preparations were performed prior to each test. The
procedure for sampling ammonia from the exhaust stack of the
cooling tower involves use of a nozzle to extract the stack gas sample,
which is then passed through a filter chamber,
 Figure 2
Sampling train schematic.
 Page 470
which is heated to 121.1°C to collect any particulate matter that is
present. The hot gas stream is then passed through a series of cold
impingers where condensibles are removed, and the gas stream is
cooled before going to the pump, dry gas meter, and flow-sensing
orifice.
Attached to the sample probe is an S-type Pitot tube, which is inserted
through the stack wall. The Pitot tube measures the stack gas velocity
by detecting a pressure difference that is proportional to the stack gas
velocity. The cold box that contains the impingers is used to remove
the moisture from the gas stream. The cold box contains three
modified GreenburgSmith impingers and one standard
GreenburgSmith impinger connected in series and surrounded by
crushed ice and water. The first two impingers contain the absorbing
solution that is used to trap the effluent. For ammonia sampling, a
sulfuric acid solution is used as the absorbing solution. The third
impinger is empty and acts as a dry trap for moisture. The fourth
impinger contains silica gel, which is used to dry the gas stream
completely before it exits the cold box. From the fourth impinger, the
dry cool gas flows through an umbilical cord and then to the control
unit.
The control unit is used to adjust and monitor gas sample flow rates to
achieve isokinetic sampling conditions. Within the control unit, a
rotary-vane vacuum pump provides gas suction, and two control
valves adjust the gas flow. Thermocouples are used to measure the gas
temperature at the probe, filter, impinger outlet, and inlet and outlet to
the gas meter. The gas flows from the dry gas meter through a
calibrated orifice and is vented to the atmosphere. A dual inclined
manometer measures both the pressure differential across the orifice
and the velocity pressure from the Pitot tube.
V
Emission Sampling Results
Using the sampling procedure outline above, five test sites were
sampled. Afterward, the results of the emission losses from each of
the sites were determined. Table 1 shows these results. At site 3,
particulate sampling was not performed. At site 4, only emissions of
ammonia were determined.
Table 1 10340 Production Emission
Sampling Results
Ammonia Fluorine Particulate
Plant emissions emissions emissions
location (kg/h) (kg/h) (kg/h)
Site 1 54.4 0.599 0.762
89.8 1.234 1.500
Site 2 20.4 0.069 1.18
21.3 0.140 1.32
Site 3 55.3 1.057
61.2
Site 4 21.7
26.3
19.1
Site 5 15.4 0.014 4.390
12.2 0.011 0.689
13.1 0.024 2.280
 Page 471

VI
Process Variables That Affect Emission Rates
Table 1 shows that the plants at sites 1 and 3 had much higher losses
of ammonia than those at plants at the other three sites. There are
several factors that would affect the emission rates from the 10340
plant sites [6]. The discharge end of the pipe reactor should be
submerged below the liquid level in the hot well at least 30.5 cm to
minimize the loss of any unreacted ammonia that discharges from the
pipe reactor. The secondary ammonia sparger should be located in the
sump of the packed tower instead of in the hot well, and it should be
located away from the suction of the recycle pump. The amount of
ammonia to the secondary ammonia sparger should be limited to no
more than 20% of the total amount required for the process. The
remainder of the ammonia should be added directly to the pipe
reactor. The recycle liquid sprayed on top of the bed of packing in the
cooling tower should be distributed so that it wets the entire bed of
packing; however, the liquid spray should not be allowed to hit the
walls of the tower. If possible, the depth of packing material in the
cooling tower should be increased without creating an excessive
increase in the pressure drop through the tower. Add the water of
formulation directly to the top of the packing in the tower to allow it
to flow down through the bed of packing. The water will help to
absorb gaseous ammonia as it rises through the packing. Finally,
check the airflow through the tower so that there is an equal
distribution of air at the inlet of the tower. This will help to ensure that
the inlet air is not being distributed unequally across the tower, which
would reduce the efficiency of the tower. These proposed
modifications to the 10340 APP process should help to reduce
emissions and allow the process to operate in a more environmentally
acceptable manner.
References
1. F. P. Achorn and H. L. Kimbrough, Latest developments in
commercial use of the pipe reactor process, Solutions, JulyAugust
1974.
2. 10340 Production Estimate for 1991, Tennessee Valley Valley
Authority, Muscle Shoals, Ala. 1991.
3. Results of Stack Gas Sampling for Mears Fertilizer Company,
Incorporated, of El Dorado, Kansas, September 910, 1991, Tennessee
Valley Authority, Muscle Shoals, Ala., 1991.
4. Code of Federal Regulations, No. 40, Pt. 60, Appl A, Methods 5
and 13B, Environmental Protection Agency, Washington, D.C., 1990.
5. Universal Stack Sampler Operation Manual, Anderson
Instruments, Inc., Atlanta, Ga., January 1990.
6. C. E. Breed, and M. T. Holt, Environmental update for production
of 10340, paper presented at the 204th National Meeting of the
American Chemical Society, Washington, D.C., August 2328, 1992.
 Page 473

36
Hot and Cold Blend Liquid Fertilizer Plants
Derek A. Palgrave
Technical Consultant
Bury St. Edmunds
Suffolk, England
I
Introduction
Although the processing of fluid fertilizers is not normally associated
with major pollution problems, the many operations involved [1] can
give rise to significant levels of contamination that require control
measures. There is no doubt that the inappropriate handling of solid
raw materials may generate a great deal of dust and charge the air with
unacceptable levels of particulate matter. Liquid and gaseous raw
materials tend to be transferred by pipe, so they are readily contained.
Nevertheless, the integrity of pipework may not be complete, so
leakage into the environment is a possibility.
Perhaps the greatest risk of a pollution incident is during the mixing of
raw materials, when high temperatures resulting from the heat of
neutralization of, say, ammonia and phosphoric acid may cause
boiling and the development of high vapor pressure. This could lead
to surges with possible overflows and emissions of fumes and gases at
toxic levels [2]. Subsequent handling of hot fluids through strainers
and other ancillary equipment presents less of a pollution risk, but the
possible influence of residual fumes should not be disregarded. The
development of unwanted contaminants on strainers, over a period of
time, may need to be dealt with as a distinct solid waste disposal
operation.
In general, excessive noise is not a feature of fluid fertilizer
manufacture, but it is necessary to use quite a lot of mechanical
equipment, including pumps, impellers, fans, and vehicles, all of
which generate some noise. Regular maintenance will ensure that
noise levels remain within reasonable limits.
II
Dust Control
The solid raw materials handled vary from granular products, such as
MAP, through a variety of crystalline materials, such as potassium
chloride, to some very finely divided
 Page 474
clays, such as attapulgite or bentonite. The method of handling can
affect the dust potential markedly. The use of front-end loaders, which
drive into a pile of raw material to fill the bucket prior to tipping it
either into an elevator hopper or directly into a mixing vessel, may be
expected to create a very dusty atmosphere. Provided that the latter is
contained within a building equipped with extractor fans and dust
filters, this may be acceptable. Personnel working inside the building
need to be located in a suitably designed enclosure, with a clean,
controlled atmosphere. This includes the cab of the front-end loader. It
is not practicable or acceptable to rely on the use of personal dust
masks.
Where possible, totally enclosed systems should be used, especially
for raw materials with the finest particles. The pneumatic handling of
clay is widely practiced, and in many instances it is possible to pre-gel
clay with water by feeding it directly from storage into the gelling
unit.
III
Fume Control
The primary emissions liable to occur are ammonia, water vapor, and
certain fluorine compounds. Almost all arise during the neutralization
stage, so it is important to operate under strictly controlled conditions
using fully trained personnel. As ammonia in one form or another is
often a raw material, its handling prior to its use in manufacture is
particularly important.
A
Ammonia
Whenever deliveries of ammonia are made into storage, the making
and breaking of connections between delivery vehicles and permanent
storage vessels, however well conducted, tends to be associated with
an almost inevitable but finite gaseous release. All personnel involved
wear full protective equipment and are fully trained so that a product
transfer operation does not become a pollution incident. What is
perhaps more important is the integrity of the ammonia installation,
which has to be regularly monitored and tested. It should be noted that
the transfer of aqueous ammonia into a tank is normally accompanied
by the displacement of a corresponding volume of air containing
ammonia fumes. This must be vented through a scrubber operating on
water or dilute acid. The resultant liquors may be recycled into the
neutralization operations.
The vapor pressure of ammonia over solutions of ammonium salts in
water is a function of concentration, temperature, and pH [3]. During
neutralization it is important to control the feed rates of the reactants
so as to keep down both temperature and pH. Provided that the ratio
of acid to ammonia is not allowed to drift, the pH can be kept neutral
or slightly acidic. The temperature may be reduced by either physical
or thermochemical cooling. Physical cooling may be effected by
recycling the contents of the mixing vessel through an external cooler
or by placing cooling coils in the reaction zone. Thermochemical
cooling may be employed when supplementary plant nutrients such as
potassium chloride or urea are introduced, thereby extracting their
heats of solution. Similar considerations apply when nitrogen
solutions are being manufactured by the hot-mix technique, from
high-temperature ammonium nitrate liquor. The latter, which is often
transported in insulated tankers, normally contains free ammonia, and
when discharged, can give rise to significant ammonia emissions.
These can be minimized by discharging into a ureawater slurry.
The control of gaseous ammonia loss may be influenced by the
geometry of the
 Page 475
mixing vessel. By bringing ammonia and phosphoric acid together in
a pipe reactor [1] to make either ammonium orthophosphate or
polyphosphate, the scope for loss is considerably less than it would be
in an open-topped vessel. There may some conflict between the
requirements of tipping large quantities of solid raw materials into an
open-topped vessel and the need to minimize ammonia emissions.
This may well be resolved by operating with both in series.
Routine continuous measurement of temperature and pH during the
neutralization stage is essential [4]. Ideally, preset alarm points should
be included either to provide a visual/audible warning or to activate a
control circuit that cuts off or reduces the raw material feeds. Some
form of continuous monitoring of the atmosphere in the vicinity of the
neutralizers is also desirable.
B
Water Vapor
The emission of water vapor would not itself be regarded as a
significant pollutant. If it originates in the processing operation, it may
be in admixture with ammonia and possibly, some fluorine
compounds. Adequate steps to deal with them is likely to remove the
water vapor also. However, water vapor emerging from a secondary
cooling system, such as a cooling tower, can be something of a
nuisance to neighbors in the immediate vicinity. More careful siting of
such a facility away from sensitive boundaries would seem to be the
best option.
C
Fluorine Compounds
The neutralization of wet-process phosphoric acid may be
accompanied by the evolution of fumes containing hydrofluoric acid,
silicon tetrafluoride, and fluorosilicic acid, or possibly some
ammonium derivatives. The more concentrated the phosphoric acid,
the less fluorine is likely to be present since the concentration process
tends to be associated with defluorination. Where appropriate, wet
scrubbing should be installed and the liquors blended with the finished
products.
IV
Liquid Control
Plant design is normally such that all liquids are readily contained so
that the only liquids released stem from occasional leaks from valves,
flanges, pump and impeller seals, and so on. The possibility of vessel
overflow has been mentioned, but if operating procedures are properly
observed by qualified staff, this is an unlikely scenario. Clearly,
regular inspection and maintenance are essential, and this should
include checking for corrosion and erosion hot spots, which might
lead to sources of leakage [5].
In most countries, the enforcement authorities view containment of
potential leaks as a very high priority. The focus is normally on
storage vessels, but all process vessels come under scrutiny. Bunding
of individual tanks is widely recommended, but in some instances the
concept of site bunding has been accepted. This requires the
construction of a very large lagoon at a convenient low area of the
site, into which a series of gullies connect with the remainder of the
site. All areas where leaks are liable to occur are provided with
impermeable surfaces [6] and are curbed so as to deflect any spillage
into the gully system. The regular accumulation of rainwater in the
lagoon is utilized as feedwater for further fluid fertilizer manufacture.
Furthermore, any minor spillages or
 Page 476
leaks find their way to the lagoon and are automatically recycled. The
purge from cooling systems and boilers can be accommodated in the
same way.
V
Solid Waste
Fluid fertilizer manufacture gives rise to very little solid waste. In
plants where only clear liquids or solutions are produced, insoluble
components may become a problem. However, such insolubles
usually contain nutrients, so they may be incorporated either in solid
or suspension grades. Not all insoluble components are amenable to
this treatment. For example, items such as large stones or metallic
objects which have found their way into bulk solid raw materials are
genuine solid waste and need to be disposed of by conventional
means.
The material developed through cleaning operations should not be
overlooked. Tanks that have been on wet-process phosphoric acid
duty often contain significant quantities of gypsum which need to be
disposed of from time to time. It may not be possible to ensure that it
is suitable for utilization as a soil conditioner. Tanks that have been
used for storage over a period of years tend to accumulate large
quantities of corrosion products. Disposal arrangements for these need
to be established.
VI
Noise
Although it is not a chemical pollutant, noise is often an unwanted
part of our environment. Fluid fertilizer plants have a limited capacity
to generate noise by virtue of the equipment they employ. Pumps,
agitators, impellers, and cyclones all produce sound during normal
operation, but the levels rarely approach legal limits. Badly
maintained units can be the exception, such as the situation
encountered with a cavitating pump. Vehicle noise in a confined
space, such as a front-end loader in a raw material warehouse, can be
severe and it may be necessary to sound-insulate the cab to make it
safe for the operator to use. It is not sufficient to supply ear protection
for continuous use.
Occasionally, the installation of a fan to extract dust from a warehouse
introduces a noise problem and it may be necessary to modify the
installation or replace the unit with a different model to overcome this
difficulty. Noise monitoring should be undertaken and the results
recorded on a regular basis.
References
1. F. Achom, Manufacture of mixed fluids, Chapter 3 in Fluid
Fertilizer Science and Technology, Marcel Dekker, New York, 1991.
2. R. G. Wells and K. T. Johnson, Health, safety and environmental
aspects of fluid fertilizers, Chapter 21 in Fluid Fertilizer Science and
Technology, D. A. Palgrave, ed., Marcel Dekker, New York, 1991.
3. D. A. Palgrave, Miscellaneous physical properties of fluid
fertilizers, Chapter 7 in Fluid Fertilizer Science and Technology, D.
A. Palgrave, ed., Marcel Dekker, New York, 1991.
4. F. J. Johnson and E. J. Huber, Jr., Quality of fluid fertilizers,
Chapter 20 in Fluid Fertilizer Science and Technology, D. A.
Palgrave, ed., Marcel Dekker, New York, 1991.
5. H. R. Horsman and P. S. McRight, Corrosion mitigation in fluid
fertilizers, Chapter 9 in Fluid Fertilizer Science and Technology, D.
A. Palgrave, ed., Marcel Dekker, New York, 1991.
6. A. Peitscher, Exploring containment materials, Solutions, Vol. 35,
No. 5, 1991, pp. 1618.
 Page 477

37
Containment Systems for Liquid Fertilizer Plants
Horace C. Mann, Jr.*
Tennessee Valley Authority
Muscle Shoals, Alabama
I
Introduction
Containment systems for liquid fertilizer facilities should be designed
and operated to keep raw materials and products within the facility
boundaries and out of the air and out of surface water and
groundwater. Theoretically, only the product should leave the site. In
actuality, a properly constructed and operated facility will have
product and some waste leaving the site in separate streams, but the
waste stream should be small in comparison and should be disposed
of in an environmentally correct manner. Disposal of waste is
expensive, and therefore it is economically advantageous to keep the
volume of waste generated as small as possible and to reuse it on site
whenever possible. The following sections contain suggestions for
design of the various parts of an integrated containment system that
should prevent loss of materials into the environment. In most
instances a number of designs will work and a variety of materials of
construction will be suitable. Specific details regarding the design and
construction of containment systems are given in a number of
handbooks [1,2] or are available from reputable environmental
engineering firms familiar with agrichemical facilities [3].
Emphasis should be placed on designing systems that are both safe
and convenient for the employees and that allow efficient flow of
people, vehicles, and materials in, around, and out of the facility. An
example of a fully integrated dealership that handles fluid and solid
fertilizers and pesticides is shown in Fig. 1.
II
Loading and Unloading Systems
All loading and unloading systems should be large enough to contain
the entire contents of the tanker in case a massive rupture were to
occur. Usually, the systems are built large
*Retired.
 Page 478

Figure 1
Dealership for handling solid and fluid fertilizers and pesticides.
 Page 479
enough to accommodate at least two tankers loading and/or unloading
simultaneously. The pads should be strong enough to support the fully
loaded vehicles and sloped to allow any spillage to drain to a central
sump. The sump should be small and emptied daily or more
frequently, when necessary.
Figure 2 illustrates loading pads that are used at one of TVA's model
demonstration site facilities. The load-out section, mix plant, and area
1 are inside a building, while area 2 is outside. The area 1 load-out
section can be used for filling the bulk pesticide storage tanks, loading
tankers to take product to the field, or washing applicators or tankers.
A picture of this portion of the facility is shown in Fig. 3.
Alternatively, to reduce the loading pad containment volume some
pads are constructed adjacent to and connected to a tank farm
containment area. The pads should be built so that small spills will
flow into the pad sump, but a massive rupture would flow into the
tank farm containment area.
Ramps on and off the pads are necessary to keep the pad lips from
cracking and to help keep dirt, gravel, and so on, off the pad. Curbs
are usually built on the outer perimeter of the pad for containment
purposes. If possible, roofs should be built over the pads to minimize
the amount of stormwater that must be captured and used. Piping
generally should be either overhead or in trenches and exposed to
show any leakage, yet easily accessible for repair or replacement. The
electrical system and controls should be shielded to prevent corrosion
and in a location to provide good visibility of the equipment being
loaded or unloaded.
III
Reaction and Mixing Systems
Most batch reaction and mixing systems are usually either under roof
or in a building, while the continuous reaction systems which, for
instance, produce ammonium polyphosphate fluids are usually located
outside. Both systems should have sufficient containment to retain
100% of the volume of fluid in the largest tank or reaction vessel in
the area, plus the submerged volume occupied by the other tanks or
equipment in the contained area. An extra 10 to 25% should be added
for stormwater when the system is outside. Containment in an
adjoining tank farm area can be used as part of the volume, when
applicable.
When properly sealed, the building walls can be considered a dike
wall. The door sills could be elevated to provide a continuous wall
around the perimeter. Ramps to the door sills would provide safe
access to and from the building. The floors should be sealed to
prevent leakage. In addition, all piping should be exposed to show any
leakage, but supported to prevent sagging or breakage. Electrical
systems should conform to good safety and engineering practices.
The production equipment that is located in the open should have
containment that will not only capture any spills or tank rupture
within the area, but should also be able to retain storm water. The
storm water can be released if uncontaminated, but should be used or
disposed of in an environmentally safe manner if it contains
contaminants. All piping should be exposed to show any leakage that
may occur.
Off-gases from the reaction units should be sampled and analyzed to
determine if what is released is within allowable environmental limits.
In some cases it may be necessary to install scrubbers to catch the
gases if they are above the regulated value.
 Page 480

Figure 2
Fluid fertilizer production facility.
 Page 481

Figure 3
Load-out section for product or for washing applicators and tankers.
 Page 482

IV
Raw Material and Product Storage
All storage tanks should be located in a secondary containment area
that will hold the entire contents of the largest tank plus the
submerged volume of the other tanks. When not under roof the
equivalent of about a 6-in. rainfall should be included in the
containment volume. Some mechanical method of preventing tanks
from floating should be incorporated in the design, as well as leak
detectors under the pads.
All of the tanks containing fertilizer should be grouped in one area
and those containing pesticides in another area (as shown in Fig. 2).
Ideally, pesticide storage either in small containers, minibulks, or
tanks should be inside a secure building. The bulk pesticides should
be in a separate diked area, as shown in Fig. 2. There can be a
common wall between the fertilizer and pesticide containment areas if
the pesticide cannot flow into the fertilizer area. It is usually
acceptable to use the pesticide containment volume as part of the
volume needed for fertilizer containment.
Tanks with capacities of 100,000 gallons or more offer unique
challenges for containment. These tanks are too large to be moved
easily to install an impervious layer and leak detectors under the
tanks. Some companies are reputed to be able to float and move these
tanks either on water or a cushion of air, or in some cases they can be
jacked-up to allow installation of the secondary containment.
Alternative procedures are either to install a second steel bottom
separated from the original bottom by a layer of inert porous material
or to install a bladder inside the tank on top of the porous layer. The
original tank bottom than becomes the secondary containment under
the tank and the porous layer becomes a leak detector. Openings are
cut into the tank wall below the new bottom or bladder to show when
a leak has developed. An impervious surface, either concrete,
synthetic sheets, or claysoil mixtures, is then installed between the
tank outer wall and the dike wall. More details about installing
containment around and under large storage tanks are given in Refs. 1,
2, 4, and 5.
Most fluid fertilizer facilities use some solid raw materials, such as
potash, urea, micronutrients, and secondary nutrients. These solids
should be stored on an impervious surface either under-roof or if
stored outside, covered with a rain-resistant cover. When stored
outside, provisions should be made to prevent any runoff from leaving
the area until it can be tested for contamination.
V
Washing and Parking Areas
A number of fluid fertilizer dealerships use their loading and
unloading areas for washing areas also (as shown in Fig. 2). Some
install separate wash pads. In either instance, the pads should be
impervious to the materials being washed onto them, have curbs or
gutters to prevent runoff from leaving the pad, and have a sump to
collect the wash water for storage in an adjacent tank or tanks.
If at all possible, it is better to take enough water to the field to wash
the equipment in the field. The wash water should be applied to the
field at a rate not to exceed the recommended labeled rate for the
particular fertilizer and agrichemical used. Washing or application
should not occur near wells, sink holes, or other areas that might allow
the agrichemical or fertilizer to enter directly into the surface or
groundwater.
If the applicator or other equipment is washed at the facility, the wash
water should be stored and reused either as (1) makeup water in other
fluid mixes, (2) for dust
 Page 483
suppression on solid fertilizer, or (3) as a carrier for other pesticide
formulations. If there is an excess of wash water, it can usually be
applied to nearby land.
Parking areas should be evaluated and constructed from the standpoint
of whether the equipment would contain agrichemicals or whether the
equipment was empty and clean. The main concern in any parking
area is that a valve or sight gauge on a loaded vehicle could be
damaged either accidentally or by vandalism and the tank contents
emptied onto the parking area. Any loaded vehicle parked overnight
should have discharge and drain valves locked shut.
Areas where loaded or partially loaded equipment will be parked
should be paved with a material impervious to what may be in the
vehicle and constructed to contain any loss of fluid that may occur.
The parking area should be sloped so that any spillage, rainwater, or
snow melt would flow to a sump where it can be captured and reused
if contaminated. Curbs and gutters would be useful to aid in
containment. Areas where cleaned and empty vehicles are parked do
not have to have containment.
VI
Operational Areas
These areas would typically include (1) the office and weigh scales,
(2) shop, and (3) any miscellaneous areas not described above.
Wherever there is the possibility of spillage that can enter the surface
or groundwater, containment should be installed. Oils, greases,
antifreeze, and other chemicals used around shops should be
contained as well as agrichemicals. Waste oil and used vehicle
batteries can generally be disposed of through recyclers. Used tires
can normally be handled by tire dealerships.
References
1. Environmental Handbook for Fertilizer and Agrichemical Dealers,
Bulletin Y-223, TVA/NFERC, Muscle Shoals, Ala.
2. Pesticides and Fertilizer Containment Facilities Handbook,
Bulletin MWPS-37, Mid West Plan Service, Iowa State University,
Ames, Iowa.
3. World Environmental Directory (ISBN-0-916742-05-9).
4. Secondary Containment of Large Fertilizer Storage Tanks, Circular
Z-307, TVA/NFERC, Muscle Shoals, Ala.
5. Containment of Fertilizers and Pesticides at Retail Operations,
Circular Z-291, TVA/NFERC, Muscle Shoals, Ala.
 Page 485

38
Remediation Measures for Existing Plant Facilities and
Decommissioned Facilities
Susanna Tomlinson*
Tennessee Valley Authority
Muscle Shoals, Alabama
I
Introduction
The process of cleaning and restoring contaminated soil and water,
with the possibility of associated air contamination, is just beginning
to be understood as a complex and costly undertaking. The regulatory
steps themselves are lengthy and many months may be consumed
before remedial action can actually begin. Gathering sufficient data to
characterize the extent of the contamination fully and finding accurate
cost figures on which to base the feasibility studies and decisions have
been difficult. However, as more sites are addressed, more remedial
techniques are commercialized, and more case studies and other
documentation are published, the task of remediation is becoming
more manageable and systematic.
II
Site Identification
Sites are identified for remediation in a number of ways. Obviously,
any release of a hazardous substance during manufacture, storage, or
transport will trigger an investigation into the source and extent of the
contamination. State environmental agencies and individual citizens
have reported existing dump sites. Past disposal practices of industrial
operations may be scrutinized, especially if claims of a threat of
hazardous substance release are advanced. Current disposal practices
may also be investigated. Request for a RCRA Part B permit may
disclose waste disposal activities that will be questioned. Mining
activities and techniques are monitored closely. Even property
transfers are cause for investigation.
III
Superfund Activities
The Comprehensive Environmental Response, Compensation and
Liability Act (CERCLA or Superfund) responds to and cleans up past
hazardous waste activities and responds
*Retired.
 Page 486
to current releases. There is no formal system of site identification and
the CERCLA database depends on states and individuals to list
potential sites for consideration. Once a site has been selected for
action, the regulatory process can include the following sequence of
events [1].
1. Preliminary assessment (PA). The PA sets priorities for site
inspection and consists of a review of existing information, including
a description of hazardous release or the threat of release, the source
and nature of release, and possible routes of exposure.
2. Site inspection (SI). The SI is conducted to determine the potential
need for removal action and to collect or develop additional data,
including developing field sampling and analysis and quality
assurance plans and conducting field sampling. A report is prepared
following the conclusion of the site inspection that summarizes the
nature of the contamination, describes the threats to humans and the
environment, and recommends further action if warranted. At this
point the site may be proposed for the National Priorities List (NPL),
a list of priority releases for long-term remedial evaluation and
response.
3. Remedial investigation/feasibility study (RI/FS). The purpose of the
RI/FS is to select appropriate remedies ''that are protective of human
health and the environment, that maintain protection over time, and
that minimize untreated waste" [1]. The RI/FS process can take
several years and cost millions of dollars before cleanup begins. The
process should result in a more fully developed characterization of the
site and a definition of the extent of contamination that is present.
Some of the factors that are addressed in the field investigations are
the physical characteristics of the site, soils, geology, hydrogeology,
and ecology; characteristics of air, surface water, and groundwater;
waste characteristics including quantities, toxicity, persistence, and
mobility; actual and potential exposures routes; and actual or potential
receptorshuman, biota, and environment.
Andrachek and Sullivan state in Pollution Engineering [2] that the bill
for remediation services in 1991 was $2.7 billion. It is during the
feasibility study that the methods of remediation must be defined and
the costs of the options calculated and controlled. Citing 12 areas
where careful examination of environmental remediation can reduce
expenditures, Andracheck and Sullivan make a strong case for proper
management of the remediation process in order to protect the
financial resources of the responding company.
Identifying available treatment and containment technologies can be
accomplished with the help of an experienced contractor, but
information is more often obtained from a thorough literature search.
A limited number of alternatives are selected as most feasible during
the screening stage of the study. Further examination of these
alternatives must use EPA's nine criteria for evaluation: overall
protection of human health and the environment; compliance with
applicable and relevant or appropriate requirements (ARARs); long-
term effectiveness and permanence; reduction of toxicity, mobility, or
volume through treatment; short-term effectiveness; implementability;
cost; state acceptance; and community acceptance.
When remedial action is decided upon, there is a period for public
comment and perhaps a public hearing. Public comments can affect
the selection of the alternative, which in turn affects the cost of the
project. The final remedy selection is documented in a Record of
Decision (ROD) which the regional EPA Administrator signs.
4. Design and construction of the remedial project. The cleanup may
itself take additional millions of dollars and several years to complete.
If hazardous substances remain after remedial action, EPA must
monitor the site every 5 years and determine if more cleanup steps are
required.
According to the Alabama Department of Environmental
Management, in areas such
 Page 487
as property transfer, some companies have undertaken voluntary
cleanup of contaminated sites. Although voluntary actions are
welcomed in most states, acting without the full knowledge and
cooperation of the state and federal environmental agencies has at
least two associated dangers. No matter how thorough a cleanup may
be, the action needs a letter of No Further Action or other formal
signoff from the regulating agencies to acknowledge the validity of
the project. Especially in cases where there may be some
contamination left at the end of the action, companies may be required
to start the entire remediation process over again. In cases where
third-party lawsuits have been filed, voluntary actions have no
protection without a formal consent agreement, letter of approval, or
other official document.
IV
Remediation Alternatives
Between 1982 and 1991, 70% of all Superfund site Records of
Decision addressing groundwater remediation specified pump and
treat technology for plume containment [3]. Today, second-and third-
generation treatment options such as air sparging reduce both the time
and cost required to close a site. A partial list of treatment methods
has been extracted from literature reviews conducted at TVA's
National Fertilizer and Environmental Research Center [46].
Air sparging or stripping: process of evaporating volatile
contaminants from soil or water by passing a flow of air through the
substrate.
Anaerobic pyrolysis: high-temperature decomposition of waste
materials with little or no oxygen present.
Biodegradation: the use of microorganisms to break down hazardous
chemicals.
Bioremediation: term applied to the controlled use of microorganisms
at hazardous waste sites; can be aerobic or anaerobic degradation
Crystallization: the formation of crystalline substances from solutions
or melts.
Fractional crystallization: separation of a cooling magma into mineral
components as the minerals cool and congeal at different
temperatures.
Fractionation: separation of a mixture in successive stages; see also
Fractional crystallization.
Incineration: high-temperature combustion of wastes or waste
containing materials in the presence of oxygen; the process of burning
a material so that only ashes remain.
In situ vitrification: process that entraps contaminants in a glass
produced by electrically melting waste-containing soils.
Liquid extraction: See Solvent extraction.
Soil washing: process using solventsurfactant combinations to remove
contaminants.
Solidification/stabilization: process that mixes a contaminated
material such as soil with a binder or solidification agent in order to
entrap the contaminants and reduce their mobility or reactivity.
Solvent extraction: dissolving contaminants in a solvent or solvent
mixture and then separating the solvent from the solute; the separation
of materials of different chemical types and solubilities by perferential
extractive action.
Steam stripping: similar to air stripping except that steam is the
extracting medium.
Supercritical fluid extraction: is essentially a solvent extraction
method using materials such as carbon dioxide or light hydrocarbons
at temperatures and pressures above their critical points.
 Page 488
Supercritical water oxidation: a wet-air oxidation process using water
above its critical point.
Thermal desorption: the vaporization of volatile and semivolatile
contaminants from soil at moderate temperatures.
Vapor extraction: process that draws a vacuum through soil to
vaporize and extract volatile contaminents; sometimes referred to as
soil venting.
Nontreatment options, such as removing contaminated soil to a
licensed landfill or capping a hazardous area, were used in the early
days of cleanup operations, but the trend today is toward treatment
that reduces permanently the amount, toxicity, and movement of the
contamination. There is still inconsistency in establishing the goals
and describing the desired results of remediation processes, and many
remediation options that are called treatment only prevent mobility or
reduce volume without minimizing toxicity. Examples are
stabilization, which reduces mobility, or separation, which reduces the
volume of contaminated substance without reducing toxicity
significantly.
Subsurface barriers may be used when it is necessary to isolate and
control contamination before treatment takes place. Some commonly
used barriers are slurry trench walls, grout curtains, vibrating beam
walls, bottom sealing, block displacement, and liners [7].
V
Where to Get Help
Information about remediation is proliferating in today's printed
media. A good technical library will have books such as Luckner and
Schestakow's discussion of methods of modeling migration processes
in soil and groundwater [8] and the ACS Symposium papers edited by
Sabatini and Knox that cover colloidal, interfacial, and surfactant
phenomena [9], books that help characterize and track chemical
releases into the environment. Today's remediation technology can be
researched from sources as diverse as magazines [2,10], to books [11
], to government publications [3,7]. All these sources include lists of
reference material that provide more research possibilities. Much can
be learned from the Superfund Records of Decisions and case studies
[12]. State environmental agencies are more than willing to share past
experiences with those who are just beginning the process. Most of
the technical societies have continuing education courses designed
around the remediation process. The American Institute of Chemical
Engineers offers a course in remediation techniques, design, and cost
engineering [13] and the American Chemical Society's U.S.
Environmental Laws and Regulations course includes discussions of
remediation [14]. There are also many good private-sector education
short courses available.
VI
Waste Reduction
A well-managed environmental program that includes waste reduction
is an important step in pollution control and an integral and ongoing
part of the production process. Waste reduction is cost-effective for
the manufacturing operation itself, can decrease the need for
additional disposal sites, and will help eliminate future contamination
and the need for remediation at disposal locations. In the state of
Alabama, industry can receive help in determining ways to reduce
waste through a confidential noncompliance, nonregulatory
assessment service.
Alabama's Waste Reduction and Technology Transfer Program
(WRATT) is an idea
 Page 489
that grew out of EPA Region IV's Southeastern Regional Hazardous
Waste Round Table meetings. Now established as a foundation,
WRATT's Board of Directors represent the Alabama Department of
Environmental Management, the Tennessee Valley Authority, the
Alabama Chemical Association, and the Business Council of
Alabama. The program uses the skills, knowledge, and experience of
retired engineers and scientists who look for ways to minimize waste
at industrial operations. The stated purpose of the Alabama WRATT
program is to provide confidential and nonregulatory waste reduction
assessments to industries that voluntarily accept the opportunity to
improve the efficiency and environmental impact of their operations.
Services are free, although the foundation does receive donations from
the companies who have requested assessments [15].
A WRATT assessment involves one or more plant visits, including a
detailed plant tour. The WRATT team gathers information on wastes
streams, SARA Title III reporting, OSHA programs, and recycling and
reuse information. A total water balance and material and energy
balances are calculated. The confidential technical report contains the
comments and observations of the team and offers a menu of options
to reduce waste at the facilities. The team will follow up on the report
with a second review in 6 to 12 months. However, implementation of
the suggestions is handled by plant personnel and outside consultants,
contractors, engineering firms, and other for-hire businesses.
This approach to plant efficiency and environmental improvement has
been highly successful. In addition to conducting over 100 waste
reduction assessments in Alabama since the program's inception,
WRATT staff has helped the states of Iowa, Vermont, and New
Hampshire develop and implement similar programs and has provided
direct training assistance to three other states.
References
1. EPA, Code of Federal Regulations, Title 40, Parts 300399, Office
of the Federal Register National Archives and Records
Administration, U.S. Government Printing Office, Washington, D.C.,
1992.
2. R. G. Andrachak and K. M. Sullivan, Cutting the cost of
remediation, Pollution Engineering, Vol. 24, No. 19, November 1,
1992.
3. EPA, In Situ Treatment of Contaminated Ground Water: An
Inventory of Research and Field Demonstrations and Strategies for
Improving Ground-Water Remediation Technologies, Technology
Innovation Office, Office for Solid Waste and Emergency Response,
U.S. Environmental Protection Agency, Washington, D.C., 1992.
4. P. D. Enlow, A Literature Review of Nonbiological Remediation
Technologies Which May Be Applied to Fertilizer/Agrichemical
Dealer Sites, TVA Bulletin Y-213, TVA National Fertilizer and
Environmental Research Center, Muscle Shoals, Ala., 1990.
5. V. M. Norwood, A Literature Review of Waste Treatment
Technologies Which May Be Applicable to Wastes Generated at
Fertilizer/Agrichemical Dealer Sites, TVA Bulletin Y-214, TVA
National Fertilizer and Environmental Research Center, Muscle
Shoals, Ala., 1990.
6. V. M. Norwood and M. E. Randolph, A Literature Review of
Biological Treatment and Bioremediation Technologies Which May
Be Applicable at Fertilizer/Agrichemical Dealer Sites, TVA Bulletin
Y-215, TVA National Fertilizer and Environmental Research Center,
Muscle Shoals, Ala., 1990.
7. EPA, Ground Water, Contamination and Methodology, Technomic,
Lancaster, Pa., 19 .
8. L. Luckner, and W. M. Schestakow, Migration Processes in the Soil
and Groundwater Zone, Lewis Publishers, Chelsea, Mich., 1991.
9. D. A. Sabatini, and R. C. Knox, eds., Transport and Remediation of
Subsurface Contaminants, ACS Symposium Series 491, American
Chemical Society, Washington, D.C., 1992.
 Page 490
10. B. Kamnikar, Bioremediation of contaminated soil, Pollution
Engineering, Vol. 24, No. 19, November 1, 1992.
11. R. Charbeneau, P. B. Bedient, and R. C. Loehr, eds., Groundwater
Remediation, Water Quality Management Library Volume 8,
Technomic, Lancaster, Pa., 1992.
12. U.S. Congress, Office of Technology Assessment, Are We
Cleaning Up? 10 Super-fund Case Studies: Special Report, OTA-
ATE-362, U.S. Government Printing Office, Washington, D.C., 1988.
13. AIChE Continuing Education, Groundwater and Soil
Remediation, American Institute of Chemical Engineers, New York,
1993.
14. ACS Instruction, U.S. Environmental Laws and Regulations,
American Chemical Society, Washington, D.C., 1992.
15. R. O. Nicholson and E. D. Myers, Alabama Waste Reduction and
Technology Transfer Program, Muscle Shoals, Ala., 1993.
 Page 491

39
Remediation Measures for Distribution and Dealer
Facilities
Horace C. Mann, Jr.*
Tennessee Valley Authority
Muscle Shoals, Alabama
I
Spills at the Dealers' Facilities
In this chapter we deal only with facilities that have been in operation
for a number of years and where past practices probably have resulted
in site contamination or in relatively new sites that have experienced a
significant spill. It is assumed that minor spills at relatively new sites
were cleaned up before any contamination left the site. Before
deciding on a remediation procedure, the type and extent of
contamination must be determined. To evaluate the type or types of
contamination present will probably require a thorough site
assessment to determine which raw materials have been used and
products produced at the facility. All potential contaminants need to
be identified. If more than one business had been at the site, this could
require a more extensive investigation. Sampling of the soil and
groundwater can then determine which chemicals are actually present,
their concentration, and the extent to which they have penetrated the
soil and groundwater. Selection of a suitable cleanup procedure can
then be made from a number of options, such as (1) land farming, (2)
pump-and-treat procedures, (3) soil washing, and (4) bioremediation,
to mention a few. The proper cleanup procedure is generally designed
specifically for each site and should be engineered by a
knowledgeable remediation consultant in cooperation with regulatory
authorities who have control over the particular situation. A list of
environmental engineering companies is presented in World
Environmental Directory (ISBN 0-9167-4205-9).
II
Spills While Delivering Product
Most agrichemical spills will occur as the product is being taken to the
field in either nurse tank trucks or in applicators. Most of these
vehicles are equipped with radios that
*Retired.
 Page 492
allow communication with the plant. As soon as it is determined that a
spill has occurred, the appropriate plant manager should be notified
and the area cordoned off to prevent the spillage from being tracked
further away. Any medical emergencies and injuries should be
handled first. The plant manager should notify the appropriate
regulatory authorities and the necessary cleanup contractors. Major
emphasis should be on containing the spill and preventing it from
entering adjoining streams or rivers. Consultation with regulatory
authorities will determine if any remediation is required and the time
frame that will be allowed. As in all other instances, no general plans
can be formulated because each instance is for a specific occurrence
and must be tailored to fit the occasion.
Since it is almost inevitable that most active dealerships will
eventually have a spill or accident, it is suggested that the following
preparations be made.
1. Have available a list of the telephone numbers for:
a. Emergency medical personnel and facilities.
b. State regulatory personnel.
c. State and local emergency response groups.
d. Local contractors that have equipment and personnel available.
e. An environmental consultant and attorney who knows your
facility and the chemicals used.
2. Have available a location where contaminated soil can be "land-
farmed," if it is environmentally agreeable, or know the location and
requirements of a hazardous waste dump site and the requirements to
use it.
3. Stock each truck with appropriate flares and roadside markers to
cordon off the spill area and if possible spill containment equipment
to handle minor emergencies.
4. Know and have available the proper emergency release and
notification telephone numbers to call after the spill is contained. Also
know which reports are necessary to turn in to the appropriate
agencies.
For a more detailed description of responses to use in an emergency,
see the Emergency Response Planning chapter in the Environmental
Handbook for Fertilizer and Agrichemical Dealers [1].
References
1. Environmental Handbook for Fertilizer and Agrichemical Dealers,
Bulletin Y-223, TVA/NFERC, Muscle Shoals, Ala.
 Page 493

Index
A
Absorption:
column, 88
devices, 279
extension, 87
system multisection, 296
Accidental release, 27
Acid rain, 11
Activated carbon, 193, 243, 249, 250, 457, 459
Admissible limits, 122
Adsorption, 85
Aeromixer scrubber, 282
Agrichemical spill, 491
Agriculture:
production, 9
productivity, 9
techniques, 9
Agrolinz, 442
Air Quality Act, 381
Aluminum fluoride, 305, 325331
Amine, 245
Ammonia:
carbon dioxide removal, 65
coal, 36
coke gas, 38
emission, 469, 474
emmissions from ammonium nitrate production, 101
emissions from nitrochalk production, 104
health and safety precautions, 73
heavy oil, 36
loading facilities, 7778
materials of construction for storage, 7576
modern design, 69
naphtha, 3637
natural gas, 3637
neutralization with nitric acid, 100
normal emissions, 53
oxidation, 81
plant, 36, 40
plant safety design, 40
pollution control, 5170
process condensate treatment, 64
process description, 51
production, 3549
purge gas treatment, 58
recovery, 193, 394
refrigeration, 66
scrubbing, 100, 193, 394
 Page 494
[Ammonia]
shift conversion, 63
startup/shutdown emissions, 68
storage, 7377
synthesis, 35, 46, 66
transport facilities, 7780
vapor pressure, 474
Ammoniation granulation plant, 407
Ammonium carbonate, 450, 458
Ammonium cryolite, 325326
Ammonium fluoride, 307, 320, 324325, 328
Ammonium fluorosilicate, 307, 320, 325
Ammonium hydrogen fluoride, 307
Ammonium nitrate:
emissions from processes, 101
evaporation, 103
granulation, 101
haze reduction, 101
nitric phosphate process, 434, 444, 450
prilling, 100
process condensates, 100
 production processes, 100
use, 99
Ammonium phosphate
(see also Monoammonium phosphate, Diammonium phosphate):
solutions, 467471, 473475
suspensions, 467471
Ammonium sulfate, 320, 325326
by FGD, 191
Ammsox process, 183
Apatite, 237241
Aquifer, 345346, 351356, 362
Artisan pressure, 345, 356
Attapulgite clay, 474
B
Backfilling, 362364
Bacteria, 130
growth rate, 131
nitrifying, 130
nitrobacter, 130
nitrosomonas, 130
Bag filter, 395, 407408
Baghouse, 293294
Barium fluorosilicate, 307, 320
BASF, 166, 420
BASF-Odda, 434, 439, 442
Belt conveyors, 387
Beltram Associates, 168
Bentonite clay, 364, 474
Best available technology (BAT), 33
Best available technology techniques not entailing excessive costs
(BATNEEC), 33
Best practical environmental options (BPOE), 33
Biological treatment, 130
Blue baby disease, 11
Bosch, Carl, 35
Brine disposal, 344
Brink-type mist eliminator, 101, 174175, 184
Bubbling scrubber, 280
Bunding, 475
By-products, 1014
C
Cadmium:
in phosphate rock, 148, 254
in phosphoric acid, 256, 264265
pollution in soil, 20, 148
 pollution in water, 11
removal cost, 255, 262
removal from phosphate rock, 255258
removal from phosphoric acid, 258267
Calcination, 158
Calcined phosphate, 449
Calcium ammonium nitrate (see Nitrochalk)
Calcium nitrate, 421, 434, 437, 450
Calcium oxide, 321
Calcium phosphate, 147
CANMET, 351
Capping, 364
Carbon monoxide control, 63
CARIX process, 142
Catalyst disposal, 68
Catalytic reduction, 83, 85
Cathodic protection, 354
Cation exchange, 357
Centrifuge KCl, 350
CERCLA, 485
 Page 495
Chlorides:
in groundwater, 357358
in natural gas, 55
Chutes, 387, 389, 391
Ciba-Geigy process, 182, 183
Clarifier, 239250
Clay:
attapulgite, 474
bentonite, 474
in potash brine, 356
Clean Air Act, 24, 197
Closure, 364
CO shift conversion, 43
CO2, 10, 1314, 17, 43
Colloidal matter, 243, 248
Combustion control, 63
Compliance costs, 2829
Comprehensive Environmental Response, Compensation, and
Recovery Act (CERCLA), 26
Condensate:
contaminants, 121, 124
recycling, 123, 128
 stripping, 124
Condensation, low-temperature, 47
Conductivity profiling, 359
Containment:
liquid fertilizer system, 475, 477479
pesticide storage, 482
pond, 463
raw material and product, 482
storage tanks, 481
wash water, 482
wells, 355356
Contamination:
site, 491
water, 25
Cooling pond, 225
Cross-flow scrubber, 283
Cryogenic distillation, 47
Cryolite, 301331
Crystallizer KCl, 351
Cyclone, 373
D
Deadwood formation, 352
Decisions, EC, 32
Defluorinated phosphates, 159, 302, 313
Demister, 283, 285, 468
Denitrification, 132
Desulfurization:
activated carbon, 52
iron oxide, 52
natural gas, 41
zinc oxide, 54
Diammonium phosphate (DAP), 239242, 248, 393417
Dihydratehemihydrate process, 218
Dihydrate process, 218
Dikes, 344, 347, 350, 356
Directives, EC, 32
Divalent iron, 239249
Dolomite, 343, 349, 352353
Dorr-Oliver process, 166, 170
Dryer emissions (see Emissions)
DSM-Stamicarbon, 434
Duct design, 463
Dust
(see also Emissions):
 ammonium nitrate, 101
characteristics, 463
clay from, 473
collection system, 289294, 405, 461464
control agents, 389
electrostatic precipitator for KCl, 379
filter, 473
from granules, 473
nitrochalk, 105
size, 372
E
EC legislation, 3133
action programs, 33
BAT, 33
BATNEEC, 33
BPOE, 33
enforcement, 32
IPPC, 32
Treaty of Paris, 31
Treaty of Rome, 31
Eco-audit, 33
EFMA, 33
Electron beam process, 96
 Page 496
Electrostatic precipitator, 168, 395
Emergency cleanup, 357
Emergency response planning, 492
Emissions
(see also Dust, Wastewater treatment):
ammonia standards, 24
CO, 3640
CO2, 3642
fluid fertilizer process, 474
granulation plant, 405
guidelines, 25, 401
H2S, 3740
KCl, 343, 365, 370379
liquid fertilizer production, 469470
MAP, DAP, and NPK, 394, 398399, 405417
measurements, 469
mist control, 396, 405
NH3, 6667
NH4Cl, 395
NH4NO3 and nitrochalk, 104105
nitrophosphate, 428430, 433, 437, 440, 444
 noise, 476
NOx (see Nitrogen oxides)
phosphate, 401
sampling, 471
SO2, 1011, 4243, 161184, 189195
water standards, 401
Enclosure, controlled atmosphere, 474
Environmental:
awareness, 1920
laws, Canada, 349, 351
laws, European, 3133
laws, Japan, 401
laws, U.S., 2330
fines, 30
impact, construction, 32
impact, production, 5
protection, 10, 1920
Environmental Protection Agency (EPA), 122, 197, 486, 489
Erosion, 344
Eutrophication, 120
Exhaust fans, 463
F
Fabric collector, 293, 411
Fertilizer production, 2
growth rate, 2
North America, 4
Western Europe, 4
Fertilizer use:
per capita, 2
world, 1, 4
FGD, 189
Fines, environmental, 26, 30
Flocculant, 457, 459
Flooded mine, 375
Flotation, 245, 249
Flue gas desulfurization, 189
Fluid fertilizer, 473475
Fluidized bed furnace, 163, 166167
Fluorine:
absorption efficiency, 276, 283287
compound absorption chemistry, 278
compounds, 13
cryolite, 301
demand, 304305
duct design, 199
emission from SSP, 270305
emission from TSP, 271305
environmental aspects, 299300, 329
fluorspar, 300301
gas, 10
GTSP production, 275277
health effect, 300, 329330
industrial emission, 313
in nitric phosphate, 313
in nitrophosphate, 302, 314
from phosphoric acid, 475
processing, 14, 312
production, 301, 316, 330331
products, 16
purification, 315316
recovery, 274, 278, 283287, 312313, 315317, 394, 401
regulations, 197, 401
scrubber, 197206
in SSP production, 275, 277, 302, 313314
in TSP production, 275, 302, 314
utilization, 299331
 Page 497
[Fluorine]
in wet-process phosphoric acid, 313314
Fluorocarbons, 304, 308309
Fluorosilicic acid, 203207
freezing point, 207
production, 312
quality, 275276
recovery, 203206
storage, 207
use, 306, 314325
Fluorspar, 300304, 315, 324
Foam scrubber, 278, 283, 286, 287
Froth flotation, 247
Fumes (see Emissions)
G
Gastric cancer, 11
Geophysical monitoring, 358
Granular fertilizer
(see also Dust, Containment):
storage, 461464
transport, 461464
Granulation:
ammonium nitrate, 99106
DAP, 239242, 248, 393413
MAP, 239, 393413
mixed fertilizer, 393413
nitric phosphate, 419431
nitrochalk, 99106
powder blending, 393
slurry, 393
SSP, 269295
TSP, 269295
Greenhouse effect, 14
Groundwater
contamination, 344345, 351, 356358, 362
seepage, 227
Gypsum:
cement from, 222
FGD from, 189
phosphogypsum, 209222
pond emergency, 233
radioactivity, 239
soil conditioner, 223
 stack, 230
storage, 211217
sulfuric acid from, 222
transport, 209211
utilization, 10, 13, 15, 218222
H
Haber, Fritz, 35
Haldor Topsoe, 181
Halite, 344
Heavy metals, 11, 13, 148, 253267, 401
Hemihydrate process, 219, 247
Hoboken-type converter, 181
Hoechst-Uhde, 434
Hydrofluoric acid, 306, 322324
Hydrotreating, 55
I
Impoundment of liquids (see also Pond), 463, 475
Integrated pollution prevention and control (IPPC), 32
Ion exchange, 134
acid cationic resins, 135
Duolite resins, 137
leakage, 135
 regeneration, 136
treatment, 134
volume capacity, 136
weak basic resins, 137
K
Kemetic-type cooler, 176
Koppers' Totzek process, 39
K-spar, 344
L
Lagoon (see Pond)
Laws (see Regulations)
Leached zone, 238
Liability, 26
Lime:
handling, 103
use in nitrochalk, 103
Limestone scrubbing, 194
Liner, 346
Liquid membrane, 246
Liquid fertilizer
(see also Solution fertilizer):
containment measures, 475
 hot and cold blend plants, 473
leakage control, 475
Lurgi process, 39, 166, 178, 181184
 Page 498

M
Magnesia scrubbing, 193
Magnesium chloride, 343344
MAP (see Monoammonium phosphate)
Metals
(see also Heavy metals):
in natural gas, 55
Methanation, 46
Methemoglobinemia, 11
Mist control (see Emissions)
Mixer-settler, 243, 250
Modeling, 349
Molecular sieve, 96
Monitoring (see also Emissions), 344, 357358, 381, 475
Monoammonium phosphate (MAP), 239, 393417
Monsanto-Enviro-Chem, 174, 178, 181184
MullerKuhne method, 14
Multicyclones, 293
N
Nitrammonia process, 8889
Nitric acid
 (see also Nitrophosphate):
chemistry, 81
NOx pollution, 8188, 9394, 423426
pollution control methods, 1112, 8391
pollution prevention, 8991
production, 8189
Nitrification, 130
Nitrochalk:
BASF production process, 102
fertilizer use, 99
emissions from production, 104
environmental aspects, 103
external conditioning, 106
internal conditioning, 105
product safety, 105
treatment of waste air, 105
wastewater treatment, 437, 446
Nitrogen fertilizer production (see individual nitrogen products), 6
Nitrogen oxides (NOx):
in ammonia production, 4243, 5663
control, 5663, 93
conversion, 97
 emissions, 9394
fertilizer production, 1012
in nitric acid production in, 8188
removal, 93
selective catalytic reduction (SCR), 93
selective noncatalytic reduction (SNCR), 94
Nitrogen pollution:
control, 119142
regulation, 32, 401
Nitrophosphate, 419460
air pollution, 419431
BASF process, 420
calcium carbonate waste, 445446
dust emission, 428
environmental advantages, 419
environmental aspects, 422431
environmental regulation, 420
fluoride emission, 427
gaseous effluents, 438, 440
liquid effluents, 433, 437, 440
manufacture, 10, 14, 419
materials of construction, 443
 NH3 emission, 426427
NOx pollution, 423426
Odda process, 419
pond, 443
process, 434439
radioactive component removal, 451459
radioactive components, 449460
Romanian process, 433
waste gas treatment, 429430
wastewater, 422
wastewater treatment, 433447
Nitrous acid (HNO2), 82, 88
Noise emission, 476
Noranda-type reactor, 181
Normal superphosphate (see Single superphosphate)
Norsk-Hydro, 434, 441, 450
NOx (see Nitrogen oxides)
O
Odda process, 14, 419420, 434439
Operating license, 357
Organic fluoride, 304, 308309
 Page 499
Organic matter:
in phosphates, 238249
removal from phosphate rock, 243, 249
P
Packed tower, 280, 287
Penalty (see Fines)
Pesticide storage, 482
Phala Boarwa, 14
Phosphate rock:
calcination, 158
chemical analysis, 152
defluorination, 159
deposits, 237238
dissolution, 239240, 433434, 439
drying, 158
dust, 148
environmental and health impacts, 147149
grinding, 157
heavy metals content, 148
installed production capacity, 152
processing, 13, 147, 157
 radioactivity, 237240
reserves, 6
sizing, 157
storage, 153155
transportation, 155157
uranium, 237240
Phosphogypsum (see Gypsum)
Phosphoric acid, 1114, 146, 197267
cadmium removal, 225267
chemical analysis, 240241, 247
color degree, 249
dihydrate-hemihydrate process, 218
dihydrate process, 218, 239240, 248
extraction, 247
heavy metals, 253267
heavy metals removal, 255267
hemihydrate process, 219, 247
phosphogypsum (see Gypsum)
pond water management, 225235
radioactive component removal, 242250
radioactive components, 237250
uranium in, 240250
Pierce Smith converter, 179
Pipe reactor, 475
Pollutants (see also Emissions), 919
Pollution Prevention Act (PPA), 27
Pond:
containment, 463, 475
cooling, 225
gypsum, 225
settling, 433
water balance, 225, 234
water management, 225
Population growth, 2, 5
Potash (see Potassium chloride)
Potash Corporation of Saskatchewan, 350
Potassium chloride, 337392
air emissions, 369385
beneficiation, 340
brine movement, 358
brine pond, 344347, 350355, 362
decommissioning plant, 361
dust, 387, 389
effluents, 338
emissions monitoring, 381
environmental impact, 383
France, 357
Germany, 351, 363
industry, 337342
liquid waste, 351, 355
Manitoba, 352
mine closing, 362
mine ventilation, 380
New Brunswick, 351, 363
producers, 7
refinery ventilation, 380
regulation, 369, 381
runoff, 355
Russia, 344
salinization of water, 338
Saskatchewan, 343365
seepage collection system, 352, 355356, 362
settling pond, 350
size degradation, 387
slimes, 346350, 365
solution mining, 340, 365
storage, 389
sylvite, 337, 343
tails, 343, 347349, 365
 Page 500
[Potassium chloride]
transportation, 391
waste management area, 344, 348351, 355, 361, 365
Potassium fluoride, 307, 326
Potassium fluorosilicate, 328
Potassium sulfate, 338
PSA process, 48
Purge, boilers and cooling systems, 476
Purge gas, 47
Pyrites, 1014, 161162, 166173
R
Radioactivity
(see also Radium, Uranium):
elimination, 242245, 249, 449450
fertilizer in, 241, 248
hemihydrate in, 247
in nitric phosphate, 449452, 450
Radium, 240241, 248, 449451, 460
Rainwater, contaminated, 464
Rare earths, 243, 248, 451452, 457, 460
RCRA, 485
Regulations, 2333, 351
Canada, 349351
EC, 3133
Japan, 401
U.S., 2330
Remediation, 485492
alternatives, 487
dealer facilities, 491
feasibility study, 486
National Priorities List, 486
preliminary assessment, 486
property transfer, 487
Record of Decision, 486
remedial investigation, 486
site identification, 485
site inspection, 486
Resource Conservation and Recovery Act (RCRA), 26, 485
Roasting furnace, pyrites, 166, 167
S
Safe Drinking Water Act (SDWA), 25
Salt pile, 344, 347351, 362
Screw conveyors, 389
Scrubber:
bubbling, 280
cross flow, 283
cyclone, 373
demister, 176, 283, 468
equipment, 278279
flooded bed tower, 415
foam, 278, 283, 286, 287
high-energy, 416
limestone, 194
magnesia, 193
maintenance, 287288
open spray tower, 415
packed tower, 415
SCPB, 282, 294
sodium, 193
urea, 114117
venturi, 416
water impact, 281
wet, 412414
wet centrifugal, 415
Seismic collection system, 352, 355356, 362
Seismic survey, 359
Selective catalytic reduction (SCR), 83, 8687, 93
Selective noncatalytic reduction (SNCR), 94
Septechlorite, 344
Shell groups, 37
Short-cycle system, ARION, 138
Silicon tetrafluoride, 313, 317, 323, 463
Simplot process GTSP, 272, 278, 293
Single superphosphate (SSP), 146, 239242, 269295
Site contamination, 491
Sludge:
activated, 132
recycling, 132
SNOX process, 194
Sodium chloride, 343344, 348352, 365
Sodium fluoride, 304, 315, 325328
Sodium fluorosilicate, 324328
Sodium scrubbing, 193
Sodium tripolyphosphate, 147
Soft-drop system, 389, 391
 Page 501
Soil pollution heavy metals, 253, 255
Solid fertilizer (see Granular fertilizer, specific fertilizer names)
Solution fertilizer, 467483
Solvent extraction:
cadmium, 258267
cost, 458
di(2-ethylhexyl) phosphate, 451, 455, 459
heavy metals, 253267
nitric phosphate, 451453
phenyl phosphate, 457
phosphine oxide (TOPO), 451455, 459
radioactive components, 237250
radium, 451458
tri-n-butyl phosphate, 451455
uranium, 242244, 451458
SOx (see Sulfur oxides)
Spillage, 462
Spill preparations, 492
Spills:
dealer facility at, 492
transportation in, 491
Stack sampling, 471
Steam reforming, 35, 37, 5558
Storage tanks, 481
Strainers, 476
Sulfur elemental by FGD, 192
Sulfur oxides (SOx), 1011, 161184, 189195
Sulfuric acid, 161195
Alon Processing Inc., 181
Ammsox process, 183
BASF, 166
Beltran Associates, 168
Brink mist eliminator, 174, 175, 184
Ciba-Geigy process, 182, 183
containment, 188
CS-I and II (Cost Saver), 174, 176
Dorr-Oliver process, 166, 170
ES (Energy Saver), 174184
by FGD, 192
fluidized-bed furnace, 163, 166167
Haldor Topsoe, 181
Hoboken-type converter, 181
Kemetics cooler, 176
 Koch Engineering, 174
Lurgi, 166, 172, 178, 182
Mitsubishi reactor, 181
Monsanto Enviro-Chem, 174, 178, 181184
Noranda reactor, 181
Pierce Smith converter, 179
pollution, 1014
preconcentration system, 172173
predrying system, 172174
PTFE (polytetraflorethylene), 173174, 177
roasting furnace, 166167
spill control, 188
storage, 187
Sx steel, 176
transport, 188
two-stage electroprecipitators, 168
vanadium catalyst, 168, 178
venturi scrubber, 166, 168, 172, 183
WellmanLord process, 182
Sump, 464
Superfund, 485
Superphosphate:
 ammoniated, 146
enriched, 273
normal or single, 146, 269295, 339342
triple, 146, 239, 242, 247, 270292
Superphosphoric acid, 206
Suspension fertilizer, 467483
Sx steel, 176
Sylvite, 337
Syngas:
production, 36
purification, 38
T
Texaco Development Corporation, 37
process, 39
Toxic metals, 253254
Toxic Substances Control Act (TSCA), 26
Triple superphosphate (TSP), 146, 239242, 247, 269295
U
Uranium:
apatite in, 237238
complex species, 453
cost, 250