Tungsten

Properties, Chemistry, Technology of the

Element, Alloys, and Chemical Compounds

Tungsten
Properties, Chemistry, Technology of the
Element, Alloys, and Chemical Compounds

Erik Lassner and

Wolf-Dieter Schubert
Vienna University of Technology
Vienna, Austria

Springer Science+Business Media, L L C

L i b r a r y of Congress C a t a l o g l n g - l n - P u b l I c a t I o n Data

Lassner, Erik.
Tungsten : p r o p e r t i e s , c h e m i s t r y , t e c h n o l o g y of the element,
a l l o y s , and c h e m i c a l compounds / E r i k L a s s n e r and W o l f - D i e t e r
Schubert.
p.
cm.
I n c l u d e s b i b l i o g r a p h i c a l r e f e r e n c e s and i n d e x .
ISBN 978-1-4613-7225-7
ISBN 978-1-4615-4907-9 (eBook)
DOI 10.1007/978-1-4615-4907-9
I . Schubert, Wolf-Dieter.
II. Title.
1. T u n g s t e n .
QD181.W1L37 1998
620. T 8 9 3 4 d c 2 1
98-45787
CIP

ISBN 978-1-4613-7225-7
1999 Springer Science+Business Media New York
Originally published by Kluwer Academic / Plenum Publishers, New York in 1999
Softcover reprint of the hardcover 1st edition 1999

All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher

To a group of pioneers in the field of Power Metallurgy and Tungsten

Science: Friedrich Benesovsky, Gustav F. Huttig, Richard Kieffer,
Tivadar Millner, Hans Novotny, Paul Schwarzkopf, and
Karl Sedlatschek. They educated a generation of scientists and
technicians who spread their knowledge throughout the world.
And to our friend, the late Bernhard F. Kieffer.

Preface
Why does someone write a book about Tungsten? There are several reasons and precedents
for this, the most important of which is that the last book on tungsten was written more
than 20 years ago, in 1977, by St. W H. Yih and Ch T. Wang. During the intervening
period there have been many new scientific and technological developments and innovations, so it was not only our opinion but the view of many other members of the "tungsten
family" that it was time to start writing a new book about tungsten. Preparations of the new
book began in 1994.
further impetus to the project was provided by the realization that in spite of this new
knowledge having been presented at seminars or published in the technical press, a general
acknowledgement of it by the majority of technicians and scientists is still far from being
realized. It is our hope that this book will significantly contribute to a broader acceptance
of recent scientific and technological innovations.
An important prerequisite for such a project is the availability of a recently retired,
experienced person willing to devote his time and talents to the tedious part of the exercise.
Erik, who retired in 1993, was both highly motivated and eager to start; and it was a
relatively easy task for Erik to persuade Wolf-Dieter to participate in the project. The fact
that both of us have for many years enjoyed a close cooperation in tungsten-related
research projects leading to the publication of several joint papers only facilitated the
decision. Moreover, as authors of "Tungsten, Tungsten Alloys, and Tungsten Compounds"
in the latest edition of Ullmann s Encyclopedia of Industrial Chemistry, we have also had
some common experience in writing.
Another important prerequisite was that we both had extensive experience in
tungsten-related research and technology during our professional work. Wolf-Dieter's
scientific career and extensive research activities combined with Erik's long-term industrial
and development knowledge contributed to a fruitful cooperation in preparing the manuscript for this book.
Nevertheless, in the writing of this book we also learned a lot more about tungsten.
Richard Kieffer (the father of Bernhard Kieffer and for many years one of the leading
experts on tungsten) always said: "If you want to inform yourself only roughly about
something, present a paper about the subject; but, if you want to gain an in-depth
knowledge about it - you have to write a book." So it seems that during the writing of this
book (it is now February 1998) we grew into experts.
vii

viii

PREFACE

Encouragement, a very important aid in a project of this kind, was given to us first of
all by Bernhard F. Kieffer, who unfortunately passed away prematurely. Bernhard not only
encouraged us to write this book, but also encouraged Plenum Publishing Corporation to
publish it.
Sometimes it was not easy to decide what should be included and what should be
omitted; or the extent of detail to which each item should be discussed. We decided in
general to emphasize the technologically related matters and to refrain from discussing
purely scientific principles. Naturally, these decisions are always subjective and chapters
dealing with those subjects with which we are more familiar are undoubtedly treated more
precisely than others. This is natural, because otherwise the individual chapters should
each have been written by an expert on that particular topic. As our original manuscript
already exceeded the maximum allowed by the publisher, such an expert treatise would
have exceeded the publisher's maximum several times over. Nevertheless, we hope that our
book gets a good reception from the reader.
Returning once again to the original question: "Why write a book?" somebody once
said it is to establish a "monument" to the author(s). If this should be the truth, we want
many people to come and glance at it. However, besides constructing a monument, it is our
intention to summarize the current status of the science and technology of tungsten for
those interested in extending their knowledge of this subject. A book always provides for
easier and more efficient learning than searching for a large number of original publications, which must first of all be identified.
The preparation of this manuscript has sometimes exacted its toll on our families, we
want to especially thank our wives for their patience and understanding.
Vienna and Graz, February 1998

Acknowledgments
Without the support and help of numerous colleagues and friends as well as companies, it
would not have been possible to write the book in its present form.
For valuable information, literature, photographs, data sheets, etc., we are kindly
indebted to many people in many countries:
Austria
K. Voigt and R. Cantz, Bohlerit, Kapfenberg
M. Schwarzkopf, G. Leichtfried, A. Singer, and R. Piircher, Plansee, Reutte
H. W6hrle, Plansee Tizit, Reutte
P. Putz, Sandvik Austria, Vienna
F. Sattler, Tamrock Voest-Alpine Bergtechnik, Zeltweg
M. Spross, B. Zeiler, O. Grau, and E. Moos,
Wolfram Bergbau-und Huttenges., St. Peter i.S
I. Begsteiger, Porzellanfabrik Frauenthal GmbH
Th. Nagl, Austrian Energy and Environment
F. Koch, Bohler Edelstahl GmbH, Kapfenberg
G. Groboth, Austrian Research Center Seibersdorf
E. Hengge, Technical University Graz
B. Lux, P. Ettmayer, H. Danninger, and R. Haubner, Vienna University of Technology
Belgium
D. Lison, University of Louvain, Brussels
China
Zhao Quinsheng and Zou Zhiqiang, Central South University of Technology, Changsha
Czech Republic
V. Dufek, Prague
D. Rafaja, Charles University Prague
Germany
K. Dreyer, Widia, Essen
D. Ermel, ALD Vacuum Technologies, Erlensee
H. Frick, Kemmer Priizision, Schwabisch Gmund
ix

:I

ACKNOWLEDGMENTS

1. Holz, Industrievertretungen GmbH., Wachtersbach

H. Kolaska, Fachverband Pulvennetallurgie, Hagen
E. Kiibel, WI Hartmetall, Miinsingen
G. Marsen, Osram, Schwabmiinchen
H. Palme, University of Cologne
D. Pratschke, Gebr. Leitz, Oberkochen
G. Gille, Starck, Goslar
H. Westennann, United Hardmetal, Horb;
Elino Industrieofenbau, DUren

Hungary
L. Bartha, O. Horacsek, and 1. Neugebauer, Hungarian Academy of Science, Budapest
Israel
R. Gem, Ashot Ashkelon, Ashkelon
R. Porat and M. Leidennan, ISCAR; Tefen
Japan
T. Tanase, Mitsubishi Materials Corporation, Omiya
T. Nomura, Sumitomo Electric Industries, Itami
Y. Yamamoto, Tokyo Tungsten, Toyamaken.
K. Kitamura, Toshiba Tungaloy, Iwaki City
Luxemburg
1. P. Lanners, Cerametal, Mamer
Principality of Liechtenstein
Th. Kraus, Triesen
Sweden
L. Rohlin, AB Sandvik Coromant, Stockholm
U. Fischer, AB Sandvik Rock Tools, Sandviken
B. Uhrenius, Sandvik Hard Materials, Stockholm
C. G. Granquist, University of Uppsala
The Netherlands
F. Mertens and G.vd. Kerkhof, Philips Lighting, Maarheeze
UK

B. Williams, EPMA, Shrewsbury

S. Jones, Royal Ordnance Speciality Metals, Wolverhampton
M. Maby, Secretary General, International Tungsten Industry Association, London

USA
B. North, M. Greenfield, and G. J. Wolfe, Kennametal, Latrobe
B. F. Kieffer, M. Ostennann, J. Oakes, and K. Horten,
Teledyne Advanced Materials, Huntsville
E. Rudy, SINTEX, Oregon
C. L. Conner, The Dow Chemical Company, Midland

ACKNOWLEDGMENTS

xi

E. A. Arney, USBM

1. Stone, Denver
Ch. W Miller, Jr., Harper International, Lancaster
We gratefully acknowledge the critical reading of various parts of the manuscript by
the following people:
M. Spross and B. Zeiler, Wolfram Bergbau- und Hiittenges., St. Peter i.S., Austria
G. Leichtfried, Plansee, Reutte, Austria

K. Mereiter, E. Zobetz, H. Mayer, P. Mohn, and J. Redinger, Vienna University of

Technology, Austria
W Kiesl, R. Stickler, and R. Podloucky, University of Vienna, Austria
M. Shale, William Rowland, Sheffield, UK
E. Bennett, NPL, Teddington, UK

H. C. Starck, Goslar, Germany

O. Foglar, Schladming, Austria
A particular vote of thanks is due to A. Bartl and W. Prohaska (Institute for Chemical
Technology of Inorganic Materials, Vienna, Austria) for the preparation of numerous
drawings, tables, copies, and photographs.
We owe special thanks to Mrs. Ann Stanly of Teledyne Metalworking Products for
correcting the manuscript in regard to the English, which is not our mother tongue.
We gratefully acknowledge the financial support of Te$edyne Advanced Materials,
Huntsville, AL, USA

Contents
CHAPTER 1. The Element Tungsten: Its Properties
1.1. Analogous to Atom Related Physical Properties . . . . . . . . . . . . . . . . . .
1.1.1. Nucleus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.2. Electron Configuration. . . . . . . . . . . . . . . . . . . . . . . . . . ... .
1.1.3. Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.104. Thermodynamic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2. Bulk Tungsten Metal Related Physical Properties . . . . . . . . . . . . . . . . .
1.2.1. Electronic Structure and Bonding . . . . . . . . . . . . . . . . . . . . . . .
1.2.2. Structural Properties . . . . . . . . . . . . . . . . . . . . . ......... .
1.2.3. Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.204. Thermal Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.5. Electromagnetic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.6. Optical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.7. Electron Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.8. Acoustic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3. Chemical Properties of Tungsten Metal . . . . . . . . . . . . . . . . . . . . . . .
1.3.1. General Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
1.3.2. Reactions with Nonmetals . . . . . . . . . . . . . . . . . . . . . . . . . ..
1.3.3. Reactions with Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
1.3.4. Reactions with Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.5. Reactions with Aqueous Solutions . . . . . . . . . . . . . . . . . . . . . .
1.3.6. Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.7. Reactions with Organic Compounds . . . . . . . . . . . . . . . . . . . . .
References for Chapter 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1
2

2
3
6
7
7
11
16
30
34
36

40
42
42
42
46
46
50
53
55
56
56

CHAPTER 2. Tungsten History: From Genesis to the 20th Century Products

2.1.
2.2
2.3.
2.4.

The Formation of Tungsten Atoms . . . . . . . . . . . . . . . . . . . . . . . .

How Tungsten Atoms Came on Earth . . . . . . . . . . . . . . . . . . . . . .
Average Abundance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geology: Formation of Ore Deposits. . . . . . . . . . . . . . . . . . . . . . .
xiii

..
..
"
..

61
63
64
65

xiv

CONTENTS

2.5. Minerals. . . . . . . . . . . . . . . . . . . . .
2.6. Ore Deposits and Reserves. . . . . . . . .
2.7. Early Discoveries of Ores, Compounds,
2.8. Technically Important Discoveries. . . .
2.9. Industrial Evolution . . . . . . . . . . . . .
References for Chapter 2 . . . . . . . . . . . . .

...
...
and
...
...
...

.. .........
.. .........
of the Element
...........
...........
.. .........

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

..
..
..
..
..
..

69
70
77
79
80
83

3.1. Oxidation of Tungsten Metal by Air or Oxygen. . . . . . . . . . . . . . . . ..

3.2. Reaction of Tungsten with Water . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.3. Reduction of Tungsten Oxides by Hydrogen. . . . . . . . . . . . . . . . . . ..
3.3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . "
3.3.2. Thermodynamic Considerations. . . . . . . . . . . . . . . . . . . . . . ..
3.3.3. Formation of a Volatile Tungsten Oxide Hydrate
[W0 2(OHh]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.3.4. Kinetic Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.3.5. Influence of Foreign Elements. . . . . . . . . . . . . . . . . . . . . . . ..
3.4. Reduction of Tungsten Oxides by Carbon or Carbon-Containing
Compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.5. Reduction of Tungsten Halides. . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.5.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.5.2. Reduction of WF6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.3. Reduction of WCl 6 by Hydrogen. . . . . . . . . . . . . . . . . . . . . ..
3.5.4. Reduction of WCl 6 by Carbon-Containing Reagents. . . . . . . . ..
3.6. Reaction of Tungsten with Carbon or Carbon-Containing Compounds
(Carburization). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.6.1. Carburization of Tungsten Powder by Solid Carbon ........ "
3.6.2. Formation of Tungsten Carbides in Melts. . . . . . . . . . . . . . . ..
3.6.3. Preparation of Coarse Tungsten Carbide Powder in an Auxiliary
Melt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.4. Preparation of Tungsten Carbide in Salt Melts. . . . . . . . . . . . ..
3.7. Chemistry of Aqueous Tungsten Solutions . . . . . . . . . . . . . . . . . . . . .
3.8. Electrochemistry of Tungsten. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.8.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.8.2. Cathode Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . "
3.8.3. Anodic Processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.8.4. Molten Salts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
3.8.5. Practical Applications of Electrochemical Processes . . . . . . . . ..
3.8.6. Electrochromism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
References for Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

85
86
88
88
89

CHAPTER 3. Important Aspects of Thngsten Chemistry

91
91
103
107
110
110
III
113
113
114
115
118
118
119
119
123
123
125
125
126
126
126
129

CHAPTER 4. Tungsten Compounds and Their Application

4.1. Tungsten and Metals: Intermetallic Compounds and Phases ......... , 133
4.2. Tungsten and Nonmetals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 133

CONTENTS

4.2.1. Tungsten and Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

4.2.2. Tungsten and Carbon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.2.3. Tungsten and Silicon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.2.4. Tungsten and Nitrogen. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.2.5. Tungsten and Phosphorus ............................
4.2.6. Tungsten and Arsenic ..............................
4.2.7. Tungsten and Oxygen. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.2.8. Tungsten and Sulfur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.2.9. Tungsten and Selenium .............................
4.2. 10. Tungsten and Tellurium .............................
4.2.11. Tungsten and Fluorine . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.2.12.Tungsten and Chlorine ..............................
4.2.13.Tungsten and Bromine ..............................
4.2.14.Tungsten and Iodine ...............................
4.3. Mixed Ligand and Coordination Compounds of Tungsten. . . . . . . . . ..
4.3.1 Mixed Ligand Compounds. . . . . . . . . . . . . . . . . . . . . . . . . ..
4.3.2. Coordination Compounds. . . . . . . . . . . . . . . . . . . . . . . . . . ..
4.4. Organometallic Tungsten Compounds . . . . . . . . . . . . . . . . . . . . . . ..
References for Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

xv

135
139
143
143
144
145
145
165
167
167
167
169
171
172
173
173
174
176
176

CHAPTER 5. Industrial Production

5.1. Mining and Ore Beneficiation. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

5.1.1. Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.1.2. Ore Beneficiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.2. Hydrometallurgy ......................................
5.2.1. Introduction including Ecological and Economical
Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.2.2. Raw Materials and Their Treatment .....................
5.2.3. Digestion-Dissolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.2.4. Purification Steps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.2.5. Conversion of Sodium to Ammonium Isopolytungstate Solution ..
5.2.6. Modem Methods in Chinese Plants .....................
5.2.7. Ammonium Paratungstate Crystallization .................
5.3. The Highly Pure Intermediates .............................
5.3.1. Ammonium Paratungstate (APT) .......................
5.3.2. Tungsten Trioxide .................................
5.3.3. Tungsten Blue Oxide (TBO) ..........................
5.3.4. Tungstic Acid. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.3.5. Ammonium Metatungstate (AMT) ......................
5.4. Tungsten Metal Powder Production ..........................
5.4.1. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.2. Push-Type Furnace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.3. Rotary Furnace ...................................
5.4.4. Tungsten Powder ..................................

179
179
179
184
184
187
191
194
197
201
205
208
208
212
212
214
214
215
215
221
223
225

CONTENTS

xvi

5.4.5. Reduction of Doped Tungsten Oxides. . . . . . . . . . . . . . . . . . . .

5.5. Powder Metallurgy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.5.1. Compacting .....................................
5.5.2. Sintering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.5.3. Fabrication of Tungsten .............................
5.6. Alternative Processes ...................................
5.6.1. Electron-Beam Zone Melting . . . . . . . . . . . . . . . . . . . . . . . . .
5.6.2. Plasma Spraying. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
5.6.3. Chemical Vapor Deposition ..........................
5.6.4. Physical Vapor Deposition (Sputtering). . . . . . . . . . . . . . . . . ..
5.7. Special Tungsten Forms and Qualities ........................
5.7.1. Globular Tungsten Powder ...........................
5.7.2. Coarse Tungsten Powder ............................
5.7.3. Single Crystals ...................................
5.7.4. Tungsten Thin Films ...............................
5.7.5. Porous Tungsten ..................................
5.7.6. High-Purity Tungsten ...............................
References for Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

229
231
.231
233
239
244
244
244
244
245
245
245
245
246
246
246
247
250

CHAPTER 6. Tungsten Alloys

6.1. Substitutional Alloys (Solid-Solution Alloys) ....................

6.1.1. Tungsten-Rhenium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . ..
6.1.2. Tungsten-Tantalum Alloys . . . . . . . . . . . . . . . . . . . . . . . . . ..
6.1.3. Tungsten-Titanium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . ..
6.2. Dispersion-Strengthened and Precipitation-Hardening Alloys. . . . . . . ..
6.2.1. Non-Sag Tungsten. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
6.2.2. Alloys with Oxide Dispersoids . . . . . . . . . . . . . . . . . . . . . . ..
6.2.3. Alloys with Carbide Dispersoids . . . . . . . . . . . . . . . . . . . . . ..
6.3. Tungsten Composites ...................................
6.3.1. Tungsten Heavy Metal Alloys .........................
6.3.2. Tungsten Copper and Tungsten Silver ....................
6.3.3. Tungsten-Fiber-Reinforced Composites ...................
References for Chapter 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

256
256
258
259
259
259
266
268
270
270
276
278
281

CHAPTER 7. Tungsten and Tungsten Alloy Products

7.1.
7.2.
7.3.
7.4.
7.5.

Lighting Application. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Electrical Engineering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
High-Temperature Technique ..............................
Welding, Cutting, Plasma Spraying, Spark Erosion, and Vapor
Deposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
7.6. X-Rays, Radiation, Medical Engineering ......................
7.7. Mechanical and Engine Engineering .........................

283
286
289
291
293
297
298

nil

CONTENTS

7.8. Leisure Time and Sports Equipment ........................

7.9. Chemical Industry and Metallurgy ..........................
7.10. Space Aviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.11. Military Applications ................................. "
7.12. Aviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.13. Laser Technique ......................................
References for Chapter 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

300
301
302
302
304
304
305

CHAPTER 8. Tungsten in Melting Metallurgy

8.1. Tungsten in Steel ............................... , . . . . ..
8.1.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
8.1.2. Master Alloys. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
8.1.3. Tungsten Alloyed Steels .............................
8.2. Superalloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
8.3. Stellite Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References for Chapter 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

307
307
309
312
317
318
320

CHAPTER 9. Tungsten in Hardmetals

9.1. Introduction ........................................ "
9.2. Raw Materials ........................................
9.2.1. Tungsten Carbide Powder ............................
9.2.2. Other Carbide Powders .............................
9.2.3. Binder Metals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
9.2.4. Other Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3. Hardmetal Production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
9.3.1. Preparation of Powder Grades. . . . . . . . . . . . . . . . . . . . . . . ..
9.3.2. Powder Consolidation ..............................
9.3.3. Sintering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
9.3.4. Postsinter Treatments ...............................
9.4. Hardmetal Qualities and Applications . . . . . . . . . . .. . . . . . . . . . . ..
References for Chapter 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

321
324
324
343
344
344
345
345
348
348
351
353
362

CHAPTER 10. Tungsten in Catalysis

10.1. Metallic Tungsten. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.1.1. Survey .....................................
10.1.2. Hydrodesulfurlzation and Hydrodenitrogenation. . . . . . . . . .
10.2. Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.2.1. Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.2.2. DENOX SCR Catalysts. . . . . . . . . . . . . . . . . . . . . . . . . .
10.3. Halides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4. Carbon Compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..
"
..
..
..
..
..
..

365
365
365
366
366
366
368
368

CONTENTS

xviii

10.4.1. Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
10.4.2. Preparation of Carbide Catalysts. . . . . . . . . . . . . . . . . . . . ..
10.4.3. WC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
10.4.4. W2C.........................................
10.4.5. Tungsten Carbide Catalysts in Fischer-Tropsch Synthesis .....
10.5. Organotungsten Compounds: Catalysts in ROMP and ADMET ......
10.6. Tungstates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
10.7. Chalkogenides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References for Chapter 10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

368
369
369
370
371
371
374
374
374

CHAPTER 11. Tungsten Scrap Recycling

11.1. Introduction and General Considerations. . . . . . . . . . . . . . . . . . . . ..
11.2. Tungsten Recycling Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
11.2.1. Hydrometallurgy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2.2. Melting Metallurgy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
11.2.3. Direct Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2.4. Semidirect Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
References for Chapter 11 ....................................

377
380
380
381
381
384
385

CHAPTER 12. Ecology

12.1.
12.2.
12.3.
12.4.
12.5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mining and Ore Dressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Chemical Conversion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Powder Metallurgy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Environmental Considerations about the Substitution of High Speed
Steel by Hardmetals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References for Chapter 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

387
388
392
394
394
394

CHAPTER 13. Economy

13.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.2. World Mine Production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
13.3. Price of Tungsten. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
13.4. Supply and Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.5. Consumption and Use. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . ..
13.6. The "International Tungsten Industry Association" (ITIA) ..........
References for Chapter 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

395
396
399
401
403
406
407

CHAPTER 14. Tungsten and Living Organisms

14.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
14.2. Tungsten in Bacteria (Tungsten Enzymes) . . . . . . . . . . . . . . . . . . . . . 410

CONTENTS

14.3. Tungsten and Animals ..................................

14.3.l. Rats .........................................
14.3.2. Mice ........................................
14.3.3. Guinea Pigs ....................................
14.3.4. Rabbits .......................................
14.4. Tungsten and Humans ..................................
14.4.1. Important Data ..................................
14.4.2. Hardmeta1 Disease ...............................
14.4.3. Beneficial Influences ..............................
References for Chapter 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xix

412
412
412
413
413
413
413
413
414
416

Index . ................................................... 417

Tungsten
Properties, Chemistry, Technology of the
Element, Alloys, and Chemical Compounds

1
The Element Tungsten
Its Properties

Tungsten is a metallic transition element. Its position in the Periodic Table is characterized
by:
Period
Group
Atomic Number
Average Relative Atomic Mass

6
6* (6A)t
74
183.85 0.03 (not regarding geological exceptional
samples) (12C = 12.0000)

1.1. ANALOGOUS TO ATOM RELATED PHYSICAL PROPERTIES [1.1,1.2]

Important figures related to the atom are:
Atomic Radius
Ionization Potential
Electron Affinity (M -+ M-)
Ionic Radii

Metallic
Covalent

w-

w+

W2 +
W3 +
W4 +

W5 +
W6 +

Atomic Volume (W /D)

137.0 pm (coordination number 8)

125.0pm (single bond; valence 6)
121.0pm (double bond; valence 6)
7.98eV
0.816 0.008 eV
226.5 pm
136 pm
117pm
10lpm
90 pm, 66 pm (CN 6) [1.3]
80 pm, 62 pm (CN 6) [1.3]
74 pm (crystal, CN 6) [1.3]
69 pm (ionic, CN6) [1.3]
9.2

New IUPAC proposal 1985, which is used henceforth exclusively.

tOld IUPAC recommendation.

CHAPTER I

Electronegativity

1. 7 (Pauling)
1.40 (Allred)
4.40 eV (Pearson; absolute)
4.35 (Slater)
9.85 (Clementi)
14.22 (Froese-Fischer)

Effective Nuclear Charge

1.1.1. Nucleus
By definition, tungsten contains 74 protons in its nucleus besides 84 to 116 neutrons.
Thirty-five isotopes (including isomers) are known. Five of them are naturally occurring;
the rest can be formed artificially and are unstable. Their half-life varies between
milliseconds and more than 200 days. The characteristics of the natural and two of the
more important artificial isotopes are listed in Table 1.1. More detailed information about
the tungsten radioisotopes can be found elsewhere [1.4].
The cross-section for absorption of thermal neutrons is 18.4 barn (v = 2200 m . S-I ).
The Nuclear Magnetic Resonance characteristics for 183Ware:
(Relative Sensitivity eH = 1.00)
Recepticity (13C = 1.00)
Gyromagnetic Ratio/rad T- I . s-I
Frequency eH = 100 MHz; 2.3488T)/MHz
Reference: WF6

7.20 X 10- 4
0.0589
1.1145 x 107
4.161

1.1.2. Electron Configuration

The electron configuration of the unexcited tungsten atom is defined by [Xe]
The term symbol is 5Do, which means that K, L, M, and N shells are
completed while, 0 and P shells are incomplete. The neutral atom contains 74 electrons.
Their energy states and quantum numbers are given in Table 1.2.
Ionization energy in eV [1.5].

4F 14 5d4 6s 2 .

wo ~ W+ ~ W2+ ~ W3+ ~ W4+ ~ W5+ ~ W6+

7.89

17.7

(24)

(35)

(48)

(61)

TABLE 1.1. Important Tungsten Isotopes [1.1,1.2]

Number of Number of
Symbol protons
neutrons
180W
182W
183W
184W
185W
186W
187W

74
74
74
74
74
74
74

106
108
109
110
111
112
113

Atomic
mass

Natural
abundance

TI/2

179.946701
181.948202
182.950220
183.950928
184.953416
185.954357
186.957153

0.13
26.3
14.3
30.67
0
28.6
0

Stable
Stable
Stable
Stable
75.1 d
Stable
23.9h

Magnetic
Decay type Nucl. nuclear
energy (keV) spin momentum Application

~-(0.433);
~-(I.312);

0+
0+
1/20+
Y 3/20+
Y 3/2-

0.11778

NMR
Tracer

0.688

Tracer

THE ELEMENT TUNGSTEN

TABLE 1.2. Energy States of Electrons in the Neutral Tungsten Atom

Energy states

Electron shell
K

L
M

Spectral name

Quantum
number n

Quantum
number kj

Number
electrons

Is
2s
2p
3s
3p
3d
4s
4p
4d
4f
5s
5p
5d
6s

2
2
3
3
3
4
4
4
4
5
5
5
6

II
II
21> 22
II
2122
32,3 3
II
21> 22
32,3 3
43,44
II
21> 22
32,3 3
II

2
2
2,4
2
2,4
4, 6
2
2,4
4, 6
6,8
2
2,4
4
2

The electron configuration, especially the 5d4 niveaus, is responsible for the typical
tungsten-related physical and chemical properties.

1.1.3. Spectra
1.1.3.1. X-Ray Emission [1.6-1.8]. A typical X-ray emission spectrum from a W
target X-ray tube at 50 kV is shown in Fig. 1.1. It consists of a continuous X-ray spectrum,
or white radiation ("Bremsstrahlung"), on which are superposed a few characteristic lines
(L lines), resulting from the direct ionization by the impinging electrons. Table 1.3 shows
the strongest characteristic lines of the extended X-ray spectrum. For more details and the
o spectra, the reader is referred elsewhere [1.1].
The X-ray spectrum is used for the analytical determination of the element, preferably
in X-ray fluorescence analysis and electron-beam microanalysis (L and M lines). These
methods have gained much importance, primarily because there is no necessity for any
prior chemical separations. Moreover, the matrix influences are much less compared to
other methods.
Due to its high atomic number, tungsten gives an excellent X-ray output, making it a
preferred material for stationary and rotating anodes of X-ray tubes, which are used for
medical diagnosis (see also Section 7.6).
1.1.3.2. X-Ray Absorption [1.l,1.8]. The attenuation of X-radiation, which passes
through a material of thickness t (cm), is given by 1 = 10 exp( -Ilt~ with 10 being the
intensity of the incident beam, 1 the intensity of the transmitted beam, and Il the linear
absorption (attenuation) coefficient. Most tabulations refer to the mass absorption
coefficient (Ill p), with p being the density of the absorber and having units cm2I g.
The mass absorption coefficient for X-rays in tungsten is given in Fig. 1.2 as a
function of X-ray wavelengths. The coefficient varies with wavelength, exhibiting sharp

CHAPTER I

characteristic lines
continuous
spectrum

>-

w
c

CD

.E

0.5

1.0

Wavelength A

<A)

2.0

1.5

FIGURE 1.1. Typical emission spectrum from a W target X-ray tube at 50 kV [1.8].

discontinuities in the otherwise smooth curve (absorption edges). These correspond to

photon energies, which are determined by the energies of the respective K, L, and M
shells. The absorption edges are at 0.17837 A (69.509keV) for the K series, at 1.2155 A
(10.200 keY) for the L series, and at 6.83A (1.814 keY) for the M series. The complete set
of absorption wavelengths (energies) for tungsten is shown in Table 1.4.
Due to its high density, tungsten is a very effective X-ray absorber. Therefore,
tungsten and some of its alloys are important materials for X-ray radiation shielding.

TABLE 1.3. Wavelengths oflmportant K-, L_a, and M-Emissionb Series

Transition

A (A)

K~l
K~2

K-L2
K-L3
K-M3
K-N3, K-N2

0.213813
0.208992
0.184363
0.17950

LIl2
Lill

Lr~

LrMs

1.48742
1.47635
1.28176
1.24458

8.3356
8.3981
9.6730
9.9620

6.978
6.969
6.743

1.7768
1.7791
1.8387

Spectra1line name
KIl2
Kill

L~l
L~2

Lr~

MIl2
Mill

Ms-N6
Ms-N7

M~

~-N6

L3-N S

International Tables for Crystallography [1.6].

b R.

Jenkins (1973) [1.7].

Energy (keV)
57.99
59.32
67.25
69.07

TIlE ELEMENT TUNGSTEN

10'

~----r-----~--r-~'------r-----r---r~-'

c::

.~ 103

;:
CD

o
o

c::

~Q. 102

...o

CO

.c

III
CO

l3

10

0.1

0.2

0.4 0.60.8 1

2 1.5

810

Wavelength A (A)
100

60

40 30

20 15 10
6
Energy E (keV)

FIGURE 1.2. Mass absorption coefficient of tungsten as a function of wavelength and radiation energy [1.8].

Values of (~/ p) for tungsten are listed below for several important X-radiations [1.6].

Mo Ka. (A. = 0.7107 A):

94 cm2 . g-1

Cu Ka. (A. = 1.5418 A):

168 cm2 g-1

Co Ka. (A. = 1.7905 A): 246 cm2 g-1

Fe Ka. (A. = 1.9373 A): 301 cm2 g-1

TABLE 1.4. X-Ray Absorption Wavelengths for Tungsten [1.1]
Series

Name

Wavelength (A)

Energy (keY)

0.17837

69.508
12.100
11.538
10.200

L.
L.
L3

1.02467
1.07450
1.21550

M.
M2
M3

4.407
4.815
5.435
6.590
6.830

Ms

2.813
2.575
2.281
1.880
1.814

1.1.3.3. Auger Transitions [1.9]. The electron energies (eV) of the principal Auger
peaks for tungsten are shown in Fig. 1.3 for an operating voltage of 5 keV.

CHAPTER I

Tungsten, W : : 74

,,5.1

"II

663

1310

1511

1523

1796

38

'89 179
1738

200

400

600

800

1000

1200

ELECTRON ENERGY,e'l/

'400

'600

'600

2000

FIGURE 1.3. Electron energies of the principal Auger peaks of tungsten [1.9].

1.1.3.4. Atom Spectrum [1.1]. The atom spectrum of tungsten is very complex. It
consists of approximately 6800 lines between 2000 and 10,000 A. The strongest lines are
tabulated below.
Wavelength (run)
429.461
407.436
400.875
255.135
207.911
202.998

Species

I (used in AAS)
II
II

These lines are used for the spectrographic determination of the element (AAS, DC arc).
Due to the large number of lines, spectral analysis of trace impurities in a tungsten
matrix is only possible in a reducing atmosphere after conversion into tungsten carbide and
when using a carrier distillation method.
1.1.4. Thermodynamic Functions [1.10]

Values for heat capacity, Gibbs energy function, entropy, and enthalpy increment have
been calculated for an ideal gas of monoatomic tungsten. They are listed in Table 1.5.

THE ELEMENT TUNGSTEN

TABLE 1.5. Thermodynamic Functions (Tungsten Gas) [1.1,1.10]

Temperature

Heat capacity

Cp

Enthalpy
increment

Gibbs energy
function

H -H298

Entropy
S
(J. K- I . mol-I)
173.95 I
180.464
191.097
200.903
209.790
226.137
236.227
243.205
248.693
253.422
257.735
261.784
265.629
269.290
272.774

T
(K)

(J . K- I . mol-I)

(kJ mol-I)

-(G - H298 )/T

(J. K- I . mol-I)

298.15
400.00
600.00
800.00
1000.00
1500.00
2000.00
2500.00
3000.00
3500.00
4000.00
4500.00
5000.00
5500.00
6000.00

21.304
23.169
30.366
37.738
41.226
37.684
32.614
30.299
30.150
31.392
33.326
35.474
37.511
39.285
40.775

0.000
2.262
7.556
14.409
22.387
42.480
59.955
75.569
90.618
105.966
122.127
139.327
157.582
176.793
196.818

173.951
174.809
178.504
182.892
187.403
197.817
206.249
212.977
218.487
223.146
227.203
230.823
234.113
237.146
239.971

Values for 298.15 K are named standard Enthalpy and standard entropy, the latter valid at
a pressure of 1 bar; to convert J to cal, divide by 4.184.

1.2. BULK TUNGSTEN METAL RELATED PHYSICAL PROPERTIES

1.2.1. Electronic Structure and Bonding [1.1, 1.11-1.13]
Tungsten has been of keen theoretical interest for electron band-structure calculations
[1.14-1.25], not only because of its important technical use but also because it exhibits
many interesting properties. Density functional theory [1.11], based on the ab initio
(nonempirical) principle, was used to detennine the electronic part of the total energy of
the metal and its cohesive energy on a strict quantitative level. It provides information on
structural and elastic properties of the metal, such as the lattice parameter, the equilibrium
volume, the bulk modulus, and the elastic constants. Investigations have been perfonned
for both the stable (bcc) as well as hypothetical lattice configurations (fcc, hcp, tetragonal
distortion).

1.2.1.1. Total Energy. The bulk equilibrium total energy (including all 74 electrons)
was calculated to be -439,457 eV [1.17,1.20], and the total energy of the isolated atom
-439,447 eV [1.17,1.20].
1.2.1.2. Cohesive Energy.

Cohesive energies can be obtained from ab initio

calculations as the difference between crystalline bulk and the total energy of the free
atom (d4 s2 configuration). Tungsten has the largest cohesive energy of all elements (Fig.
1.4) [1.12], including diamond (carbon). The cohesive energy can be interpreted as the
bonding energy of the metal in the bcc crystal lattice at absolute zero (T = 0 K). Values
obtained from different calculations vary between 7.9 eV /atom [1.14] and 10.09 eV/atom

CHAPTER I
10r---------------------~

.e

,!!!

>.

e>
CD

Ar

CD

~ 10

'iii

CD

.s::.
0

Nb

Xe

An

FIGURE 1.4. Cohesive energy across the short periods (upper panel) and long periods (lower panel) [1.12].

[1.20], depending on the theoretical approach [1.14-1.16, 1.19-1.21]. The "experimental"

value of the cohesive energy, which is derived from extrapolations of the sublimation enthalpy (Mis) to absolute zero (T = 0 K), is quoted as 8.geV/atom (~859kJ/mol) [1.1].

1.2.1.3. Structural Energy (Lattice Stability). The total energy ofW was calculated
for the bcc, fcc and hcp structures at various atomic volumes close to the experimental
volume. The result of such a calculation is shown in Fig. 1.5 [1.22]. From this, it follows
that the bcc modification has the lowest energy and thus is the stable crystal structure ofW.
The cohesive energy decreases in the order of bcc ~ fcc ~ hcp. The difference
I1E = Etot(fcc) - Etot(bcc), or I1E = Etot(hcp) - Etot(bcc) is named structural energy or
lattice stability. Values for I1E are in the range of 0.45-0.57 eV/atom [1.17, 1.19, 1.22] for
fcc-bcc, which corresponds to an enthalpy difference Hfcc - H bcc between 44 and
55kJ/mol. The value for hcp-bcc (0.6eV/atom; 57.9kJ/mol) is only slightly higher
[1.22]. The similarity between the fcc and hcp results is due to the close relation between
these two lattices.
Lattice stabilities can also be obtained from computer calculations of phase diagrams
(Pt-W, Ir-W; so-called CALPHAD assessments) [1.25-1.27]. This approach is semi-

THE ELEMENT TUNGSTEN

-273.7

-273.9

-~
E
0

-274.1

Q)

>.

-274.3

OJ
~

Q)

c:

UJ

-274.5

-274.7

-274.9
85

95
105
Volume (a.uJ

115

FIGURE 1.5. Total energy of W as a function of atomic volumes for bec, fcc, and hcp (with ideal cia ratio)
structures [1.22]. The energy is given in the pseudopotential formalism by restricting the calculations to the
valence electrons only. Note that bcc W has the smallest volume per atom (i.e., the strongest bonding). 1 au =
0.529A.

empirical and has its roots in chemical thermodynamics. Early calculations have given
misleading low results for the enthalpy difference (!pee/bee = 10 kJfmol), but recent work
resulted in more realistic estimates, and a generally better agreement with ab initio results
(Hfee/bce = 30 kJfmol) [1.26]. Differing from the ab initio calculations, which refer to
absolute zero, CALPHAD calculations are based on high-temperature data.

1.2.1.4. The Density of States. When the free atoms bind together to form a crystal,
their energy levels are broadened and form the density of states, which can be seen as the
solid-state analogy to the free-atom energy spectrum. The densities of states (DOS), as
obtained from electronic structure calculations, are shown in Fig. 1.6 [1.19] for both the
stable bcc and hypothetical fcc W The DOS is not uniform throughout the band but
displays considerable structure that is characteristic of the given crystal lattice [1.12]. This
can be analyzed into low-lying bonding and anti bonding states at higher energy. In bcc
tungsten, all the bonding states are filled with electrons and all the antibonding states are
empty. The Fermi energy Ef , which is the energy of the highest occupied state, lies exactly
in between the bonding and antibonding states. This bimodal behavior is characteristic for
the strong unsaturated covalent bonds between the valence 5d orbitals [1.12]. It also
reflects the strong stability of the bcc lattice with respect to the fcc structure, where this
clear decomposition between bonding and antibonding states is not present.
1.2.1.5. The Nature of Metallic Bonding. The bonding in tungsten arises from the
strong unsaturated, covalent bonds between the valence 5d orbitals (localized "t2g" -like

CHAPTER 1

10

CD

...CD

60

(a) Ibee

wi

t: 50
CIS
.....

.t:.

~ 40

CIS

iii 30
o

>- 20

'iQ

lii

10
O~~~~~~~~~~~~~

-0.4

CD
CD

...

60

-0.3

(b)

-0.2 -0.1
0.0
Energy (hartree)

0.1

0.2

Ifee wi

t: 50
CIS

.t:.

.....

~ 40

CIS

iii 30
o

'?: 20

'iQ

lii

10

o~~~~~~~~~~~
-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

Energy (hartreel
. . . . , S ; - - - , p ; - - - , d; and---, f.

FIGURE 1.6. The density of states for (a) bee and (b) fee tungsten [1.19]; 1 hartree =27.211 eY.

states); while, sand p bonds have lower contributions, forming a diffuse (delocalized,
nearly free electron-like) background. According to the local spin density theory the
ground-state configuration in the bcc tungsten crystal is d5s' rather than d4 s2 in the free
atom [1.1,1.14], demonstrating the half-filled d band in the solid.

i.2.i.6. Ab Initio Calculations of Structural and Elastic Properties. Equilibrium

lattice constants, equilibrium volumes, as well as bulk and shear moduli can be assessed
based on ab initio electron-structure calculations. They are obtained from the calculated
total energies as a function of volume in the bcc or fcc crystal structure and from respective
volume-conserving distortions of the lattice. In most cases, they agree well with experiments (Table 1.6).
Hypothetical properties can be calculated for the fcc modification (Table 1.6)
[1.17,1.21,1.27]. It should be noted that both C' and C44 for the fcc structure are
negative, which means that the fcc lattice is unstable with respect to the respective
distortions.

11

THE ELEMENT TUNGSTEN

TABLE 1.6. Comparison of Equilibrium Lattice Constants and Bulk/Shear Moduli of bee and fcc
Tungsten
Lattice parameter
(A)

Reference

Bulk modulus
(GPa)

Shear modulus
(GPa)

bee tungsten
305-310
297
345
318
340
333

3.16524
3.163
3.15
3.164
3.16
3.13

Experimental [1.6,1.30,1.31]
Bylander et al. [1.l6]
Jansen et al. [1.l7]
Wei et al. [1.20]
Mattheiss et al. [1.21]
Chan et al. [1.22]
Guillermet et al. [1.25]

C'

= 163.8; C44 = 163.1

C' = 178; C44 = 155

fcc tungsten
274
309

3.99
4.00

Jansen et al. [1.17]

Mattheiss et al. [1.21]
Guillermet et al. [1.25]

C'

= -232; C44 = -256

Theoretical predictions can be made of structural phase transitions under very high
pressure (compressions down to 30 % of the ambient volume), and bcc tungsten was
demonstrated to be the only stable modification over the whole pressure range [1.28].
Recently, electronic structure calculations were also applied to assess the linear coefficient
of thermal expansion [1.29].

1.2.1.7. Experimental Determination of Binding Energies. High-accuracy determinations of core level binding energies were carried out on clean tungsten surfaces by ESCA
measurements [1.32]. The obtained binding energies (taken as the peak maxima) are
summarized in Table 1. 7. Experimental valence-band spectra were obtained by highresolution angle-resolved X-ray photoemission spectroscopy [1.33]. At room temperature,
the following peaks were observed below the Fermi level: -4.8 eV (s-like character), -3.2
and -2.3 eV (mainly d-like character), and 0.6 eV Good agreement was obtained between
experimental values and theoretical predictions [1.33].
TABLE 1.7. Binding Energies for Tungsten in eV'
4s
594.3(8)

4pl/2
490.8(5)

4p3/2
423.7(3)

4d3/2
256.0(3)

4d5/ 2
243.5(3)

5s
75.5(8)

5PI/2
47(1)

5p3/2
37.2(4)

5f5/ 2
33.48(10)

5f7/ 2
31.32(10)

"Calibrated against the Au 4f7/2 level (84.00eV) [1.32].

1.2.2. Structural Properties

1.2.2.1. Crystal/ographic Properties [1.1,1.6]. Besides amorphous tungsten, three
modifications (cx-, p-, and y-tungsten) are known. Table 1.8 provides crystallographic
parameters and related data, such as density, molar volume, and specific volume.

CHAPTER I

12

TABLE 1.8. Polymorphism of Tungsten: Crystallographic and Related Data [1.1]

a.-W

f3-W

y-W

A2
Z = 2 atoms in 000 and
position

AI5 (Cr3Si)
Z = 8 atoms, 2 in 2a sites: 000 and
position; 6 atoms in 6c
sites: 0,
0,
0,
0,
0;
0
O~ -Pm3n
25.25 and 2.826 A

Al

5.05 A
(5.037-5.09) depending on
preparation method
18.9 g/cm3
19.1 g/cm3

4.13 A

Modification
Model type

! ' ! '!
!,!,!; i, ! '! ' !;
!, i; !,!, H,

! '! ' !

Space group
Shortest interatomic
distance
Lattice parametef
(25C)

o~

2.741

3.16524 A
O.OOO04

Density calc.
pycnometric
Specific volume at
melting point
Molar volume
a International

- 1m3m (No. 229)

19.246 g/em3
19.250 g/cm3
0.057---{).062 em3

15.8 g/cm3

9.53 em3

Tablesfor Crystallography (\995) [1.6].

a-Tungsten is the only stable modification. It has a body-centered cubic lattice of

space group O~ - Im3m (No. 229). A diffraction pattern is shown in Fig. 1.7, together
with a crystal structure model.

...~
hkl
110
200
211
220
310
222
321
400
"Cu K.,

ic:

..

Relatiw intens~ [%]

dIAl
100
2.2382
1.5826
15
1.2922
23
1.1191
8
1.0009
11
0.9137
4
0.8459
18
0.7913
2
(l. =1.5405981 AI.

29 [deg]"
40.262

58.251
73.184
86.996
100.832
114.923
131.171
153.535
radiation

3.18524 A

.!.

fc:

...:;;

..
;;;

20

30

40

50

60

70

110

...~

100

90

DIffrIIctIon angle 29

110

120

130

140

FIGURE 1.7. X-ray diffraction diagram of a.-tungsten; 29 values (Cu Ka.\); according to International Tables for
Crystallography [1.6].

13

THE ELEMENT TUNGSTEN

p- Tungsten is a metastable phase and converts to (X-tungsten when heated above 600
to 700 DC. It has a cubic A 15 lattice (space group O~ - Pm3n; structure type Cr3Si) and
was first characterized structurally as a suboxide (W30) [1.34]. As such, it is still listed in
the International Powder Diffraction File (41-1230). A more recent diffraction diagram is
shown in Fig. 1.8 together with the proposed crystal structure [1.35, 1.36]. p-Tungsten
forms at least partially during low-temperature hydrogen reduction of tungsten oxides,
tungsten bronzes, or ammonium paratungstate besides (X-tungsten. Several foreign
elements, such as P, As, AI, K, etc., promote its formation and also have a stabilizing
effect at higher temperatures. Moreover, p-tungsten forms during electrolytic reduction of
W0 3 in phosphate melts and in thin films, produced by evaporation, sputtering, or
chemical vapor deposition.
y-Tungsen (face-centered cubic structure of Al type) was only found in thin sputtered
layers and amorphous tungsten at the very beginning of sputtering. y-tungsten too will
form the (X-modification when heated ~ 700 DC.
1.2.2.2. Density [1.35]. The density of tungsten, equal to that of gold, is among the
highest of all metals. Measured densities of sintered materials may vary significantly, since
full density is only achieved after heavy working. Exact measurements were therefore
carried out on foils, thin wires, single crystals, etc. A summary of the respective values,
obtained by different methods, is given below.

CI

-...

~
c

.."

I!
f

.!.
~
c

...

hkl
002
012
112
222
023
123
004
024
124
233
0251234
125
044

......
;!;

f;1

...

CI

'"~

CI

30

40

50

60

70

80

90

...'"

...~

CI

20

5.0512 A

...!:!

CI
CI

.5

relative intens!!): [%J

33.9
100.0
76.1
8.7
22.5
38.9
15.8
11.3
21.9
10.1
27.3
17.6
18.7
13.6
12.4
35.8

d[AJ
29 [d!!llJ"
35.514
2.5257
39.873
2.2591
43.866
2.0623
63.773
1.4582
66.707
1.4010
1.3501
69.579
75.172
1.2629
85.992
1.1296
88.660
1.1023
91.325
1.0770
110.405
0.9380
113.276
0.9223
119.221
0.8930
0061244
132.392
0.8419
136.113
016
0.8305
1161235
140.104
0.8195
" Cu K", radiation (A = 1.5405981 A).

100

110

CI

5;
CI

-~

CI

120

130

140

Diffraction angle 29 [deg]

FIGURE 1.8. X-ray diffraction diagram of ~-tungsten;

2e values (eu Kcx\);

according to A. Bartl [1.36].

CHAPTER 1

14

Density ofr::J.-W in g/cm 3

Calculated from X-ray lattice parameter measurements:
19.246 0.003 (25C),
19.254 0.0046 (25C)
19.256 0.02 (25C)
19.316 (77K)
Extrapolated from 40-180 K to 298 K:
19.250
Measured by hydrostatic weighting:
19.250 0.004 (20C) on zone refined single crystals,
19.253 0.11 (25C) on the same material,
19.3615 (25 0c) on less pure samples.
Measured by X-ray reflection and X-ray refraction:
19.262 (room temperature)
At the melting temperature:
16.740.6.

Density of P- W
19.1 (Pycnometrically),
18.9 (calculated from X-ray parameter).

Density ofy-W film

15.8 (grazing X-ray reflection).
The apparent density is used: (a) in case of deagglomerated powders for a rough and
simple grain-size estimation; (b) in case of compacts (green density) or sintered specimens
to determine the residual porosity.

1.2.2.3. Lattice Defects [1.1,1.37]. Generally, lattice imperfections can be divided

into four groups:

point defects (vacancies, self interstitial atoms, impurity atoms),

line defects (dislocations),
plane defects (stacking faults, grain boundaries, twins), and
volume defects (clusters, voids, and bubbles; segregations, microcracks, second
phases and domains, etc.).
Even very pure and long-time annealed tungsten always contains lattice defects, such
as vacancies and dislocations. The concentration of defects depends on temperature and
deformation of the material, but also on the purity.
Vacancies. The formation energy of a monovacancy in tungsten Ef is between 2.801
and 4.47eV/atom (270-430kJ.mol- 1). The vacancy concentration is given by
c = 760 exp(-72,500/RT) and reaches 3.4% at the melting point. Vacancy migration

THE ELEMENT TUNGSTEN

15

starts at 350-400 C and measured as well as calculated values for the activation energy
are between 1.3 and 2.3 eV /atom (125-222 kJ . mol-I).
Interstitial impurities. In a bcc lattice two interstitial sites can be occupied: the
octahedral and the tetrahedral. Oxygen, nitrogen, and carbon occupy the octahedral
position and hydrogen the tetrahedral position in the tungsten host lattice. Interstitial
impurities induce lattice distortions by that altering the mechanical properties.
Dislocations. Dislocation densities (dislocation lines/ cm2 ) in electron-bearn-zone
refined tungsten have been measured to be between 5 x 103 and 106 . A diminution from
the center of a rod to its surface was determined. In cold-worked foils, filings, rods, and
thin wires (as worked or annealed) the densities vary between 108 and 10 12
A very complete compilation of the respective literature on all kinds of lattice
imperfections in tungsten is given elsewhere [1.1].
1.2.2.4. Other Structural Properties: Slip, twinning, and cleavage of single crystals.
[1.35]. Single crystal deformation mechanisms are slip (gliding, shear), twinning, and
cleavage with fracture. The dislocations contained are the fundarnental carriers of the
plastic deformation in general.
Slip or gliding. Slip is the basic mechanism in the plastic deformation of tungsten
single crystals. Slip occurs in the most densely packed [111] direction and either in {11O}
or in {122} planes. At elevated temperature also {Ill} occurs in addition.
Twinning. Twinning is a less dominant deformation mechanism and occurs in highpurity tungsten only. Twins are mainly formed on {112} planes in [111] direction. Twins
have been observed in tension below OC and, after rolling, compressing, swaging, and
explosion deformation, up to 1500 0c.
Cleavage. The usual cleavage plane for bcc metals is {l1O}, but cleavage may also
happen along {1l0} planes and twin interfaces.
Surface free energy/surface tension [1.37, 1.38]. Between the surface tension (y) and
the surface free energy (cr) the following relation exists: y = cr +A(dcr/dA), for an areaA.
Cracking, zero creep, and field emission microscopy yield equal values for y and cr.
By inert gas bubble precipitation, grain boundary grooving, and scratch relaxation rate
studies cr can be determined, and y by contact angle and multiphase equilibrium studies.
The calculated as well as experimental values for tungsten scatter largely depending
on method and material. Values of cr measured at single crystal samples are around
4.3Jm- 2 for (110) planes and scatter between 1.7 and 6.3Jm- 2 for (100) planes.
Polycrystalline samples showed cr values between 1.8 and 5.38 J. m -2. The temperature
dependence of cr was determined to be 5.38 J . m -2 at 1200 K, decreasing linearly to
3.88Jm- 2 at l800K. Theoretical calculations resulted in 4.435Jm- 2 for OK and
3.938Jm- 2 for 2100K.
Values of y at the melting temperature were measured between 2.316 and 2.545
J. m- 2 .
Surface diffusion [1.37]. The data for the surface diffusion of tungsten atoms on
their own lattice in a general direction are Do = 4 cm2 / sand Q = 3.14 eV /atom
("-'300kJ.mol- I). The activation energy for migration on {100}, {l1O}, and {2ll}
planes is 2.7geV/atom (,,-,270kJ mol- I).

16

CHAPTER I

1.2.3. Mechanical Properties [1.35, 1.39-1.43}

The mechanical properties of tungsten are only intrinsic as long as highly pure single
crystals are considered. Polycrystalline tungsten, from a technical standpoint, is the most
important form and, by far, the biggest amount of the metal used. These properties are also
strongly influenced by two main factors:
microstructure,
presence, concentration, and combination of impurity elements.
Microstructure. The microstructure of a tungsten sample, which influences the
mechanical properties within a wide range, depends on the type of preparation (powder
metallurgy, arc cast, electron beam melted, zone refined, chemical vapor deposition) and
the subsequent working (deformation, annealing, recrystallization).
In case of powder metallurgical processing (P/M tungsten), the crystal size and shape
can be regulated by the sintering conditions, the type and degree of deformation, and the
intermediate and final annealing processes. In this way it becomes possible to produce
material of desired mechanical properties. Examples for the variety of different microstructures of polycrystaline tungsten are shown in Fig. 1.9.

FIGURE 1.9 Microstructures of tungsten wires: (A) fibrous structure of the as-drawn tungsten wire, (B) eqiliaxed
grain structure of recrystallized pure wire, (C) interlocking grain structure of recrystallized NS-doped wire. By
courtesy of Philips Lighting B.V, Eindhoven, The Netherlands.

THE ELEMENT TUNGSTEN

17

Generally, the following rules exist:

a wrought (as-worked) structure, consisting of deformation direction elongated
crystals, is stronger than an annealed or recrystallized one. In particular, this is
valid for the strength in deformation direction ("texture strengthening"),
the lower the working temperature and the more energy is stored during deformation, the higher is the strength ("work hardening"),
the smaller the grain size, the higher is the strength ("fine-grain strengthening"),
different fabrication processes (extrusion, swaging, rolling, drawing) yield different
properties, and even after final working, there remain origin-related differences in
the properties of arc-melted and P1M materials.
Impurities. Foreign elements can be contained in tungsten in different modes.
Homogeneously or quasi-homogeneously distributed foreign elements can occur either
in solid solution or segregated. In the first case, one has to distinguish between substitution
of tungsten atoms by atoms of the foreign element in the crystal lattice and atoms at
interstitial positions. The solubility for interstitial elements in tungsten, such as hydrogen,
nitrogen, oxygen, silicon, and carbon at room temperature, is very low 0.1 J,1g1 g). Their
solubility at eutectic temperature ranges between 100 and 1000 J,1g1 g. During cooling, the
difference between the two concentrations will be segregated mainly at the grain
boundaries, thereby significantly weakening the grain boundary strength. Generally, in
case of segregations, the impurities can be located either at grain boundaries or within the
crystals (at dislocations or subgrain boundaries).
During heavy working of tungsten, additional low-energy sites (dislocation tangles)
are formed. Segregation of interstitials to these sites increases the grain boundary strength,
thus increasing both ductility and strength [1.41].
The analytical determination of the foreign element concentration and distribution in
tungsten is linked with highly sophisticated equipment (A~ger spectroscopy, secondary ion
mass spectrometry) and with complicated pretreatment of the samples. This may be one
reason why our knowledge about foreign element influences is still quite incomplete.
Heterogeneous impurities (particles of foreign matter in the raw materials) cause
locally foreign phases or voids, often combined with diffusion zones.
In any case, the consequences of the impurities present are various types of structural
defects. Their size can be quite different and can range from atomic distances upward to
some micrometers. The concentration of the impurity plays an important role. In some
cases, traces of an impurity element show a considerable influence on the mechanical
properties. Finally, the combination of impurity elements could change completely the
effect they would cause if present alone. Taking this into account, one may understand that
a comparison of older values of mechanical properties with those obtained today may
sometimes show bigger differences. The reason for that is that, generally, the impurity
content of today's tungsten is quite lower than 40 or 50 years ago.
Generally speaking, the indication of absolute values for the mechanical properties of
polycrystalline tungsten is not appropriate as long as the related structure, structural
history, type of impurity elements, their concentration, and kind of distribution cannot be
precisely defined. However, for technical samples, this is especially impossible due to the
high cost and the time required for analyses of such type. Therefore, any values given for

CHAPTER 1

18

the mechanical properties should be regarded as lines instead of points and as planes
instead of curves.
The outstanding mechanical properties of polycrystalline tungsten are:
the high strength and yield point at elevated temperature and the high creep
resistance.
1.2.3.1. Elastic Properties [1.30, 1.31, 1.35]. In regard to elasticity, at least below
room temperature, tungsten behaves nearly isotropically; the anisotropy coefficient at
24 QC is A = 1.010 [1.35]. The elastic constants for polycrystalline tungsten at 20 QC are
given below. Their temperature dependence as well as the respective values for singlecrystal elastic constants are shown in Fig. 1.1 0 [lAO], based on ultrasonic measurements
[1.30, 1.31].

Young's modulus, E=39~IOGPa

shear modulus, G = 156-177 GPa
bulk modulus, K=305-310GPa
Poisson's ratio, v = 0.280--0.30
The compressional modulus L of polycrystalline tungsten can be derived from [1.30]

L = (5.2415

10 12 )

(3.7399 x 10 8 )T - (4.598 x 104 )T2

(dyn . cm - 2; for conversion to GPa, multiply by 100; T in QC).

1.2.3.2. Hardness. Hardness measurement is easy to perform; this is why the
determination is used worldwide in production control as a measure of strength. As
long as the measures are taken in one plant, any comparison might be suitable. However,
comparison between different testing facilities and on different samples, respectively,
measured at different load, might be sometimes doubtful and misleading, especially if
methods are not standardized. Furthermore, different methods are in use to determine the
hardness (Vickers, Brinell, Rockwell, Knoop, and microhardness), and the results cannot
easily be compared with each other.
Today, Vickers hardness is the most common method of determination. Table 1.9
compares the results of Vickers microhardness measurements obtained from two differently produced types of single crystals and several polycrystalline tungsten samples.
The following conclusions can be drawn:

there exists a load dependence (the higher the load, the lower the hardness),
measures on single crystals exhibit a slight dependence on crystal orientation; the
average of the three crystal orientation measures compares very well with those of
polycrystalline samples (annealed or recrystallized),
the hardness of swaged rods "as-worked" is significantly higher than that of the
subsequently annealed material (463 kg mm- 2 {} 357 kg . mm- 2)
Macro Vickers hardness for polycrystalline tungsten is given as (HV 30) [1.39]:
OC

recrystallized
worked/deformed
400C
800C

450
300
up to 650
240
190

145

150

330

270

265
800
1200 1600 2000
Temperature,OC

125

120

320

340

275

130

350

280

400

.~ 360
~

135

.x:

285 ~

c:

370

,
,,

400

------J

.... --- , ,

'-,

--....1-

"

1\

\,

---....J-

.-;

4~

",'
.

,,

0.28

0.29

0.30

Q.

(/)

(/)

-c:

(/)

-l

0.31

....
~

'"-l

m
m

-l

Eii

-'

...'
,, ..'
,, ..'
, .'
'

- - - - Vc
Vp

- - - Ep

800
1200 1600 2000
Temperature,OC

----

'~

".,. ~'

\\

\~

~\\

- - Ec

FIGURE I.IO. Shear modulus G, bulk modulus K, Young's modulus E, and Poisson's ratio v of tungsten vs. temperature, as calculated from single-crystal elastic constants
(Gc' Kc. E c' and vc) [1.31], and from measurements on polycrystalline tungsten (Gc Kc. E c' and vc). [1.30] Taken from Metals Handbook [1.40].

C/)

"8

"8

.,;

CJ 380

.,;
~
:5

8!.

CJ

290 :5

295

(lJ

Q.

390

300

155

~ 140

a;

~0

.,;

CJ

Q.

(lJ

400

305

160

410

310

165

420

315

170

20

CHAPTER I

TABLE 1.9. Microhardness (kg mm -2) of Tungsten of Different Manufacturing History [1.39]
Load
Material

Plane of
indentation

IOOg

200g

300g

400g

362
389
410
379
349
390
382
377

352
385
394
363
350
383
393
379

350
380
380
357
340
373
387
367

386
474
379
393

374
475
363
384

363
463
357
372

Single-crystal tungsten
Grown on a Pintch wire

Zone melted

(100)
(110)
(III)
Average
(100)
(110)
(Ill)

Average

360
395
408
392
361
405
396
387

Polycrystalline tungsten
Electron beam melted
Swaged as-worked
Swaged and annealed
Rolled and recrystallized

397
498
392
401

The hardness of polycrystalline tungsten increases appreciably as the grain size is

reduced. It was shown that hardness and average grain size follow a Hall-Petch relationship:

with Ho = 350 kg . mrn- 2 , KH = 10 kg . mrn- 3 / 2 , and I being the average grain diameter.
Hardness values of up to 1200 kg . mrn - 2 (2000 g load) were obtained on submicrongrained materials [1.44].
The temperature dependence for work-hardened and annealed tungsten is shown in
Fig. 1.11 [l.40, 1.45]. At about 1600 e, both curves coincide, i.e., at this temperature, the
tungsten completely recrystallizes as it is heated to the testing temperature.
Doped tungsten (AKS-, Thoria-doped) retains deformation hardness up to higher
temperatures. Thoria-doped tungsten may even have 450kg.mrn- 2 at 1900 0 e [1.37].

1.2.3.3. Low-Temperature Brittleness. At low temperatures, tungsten behaves in a

brittle manner during loading, if subjected to tensile stresses. This behavior changes above
a certain transition temperature, which is called the ductile-to-brittle transition temperature
(DBTT). This temperature is not a sharp, reproducible value, but a certain range (for
example 200-300 0C), which depends on mechanical, structural, and chemical conditions.
The most important influencing parameters are summarized in Table 1.10. The DBTT is an
important material characteristic for the "low temperature" use and fabricability (shearing,
punching, bending, etc. operations) of the metal.
The brittleness of polycrystalline tungsten, at low temperatures, is attributed to the
weakness of the grain boundaries, which leads to initiation of cracking in both wrought
and recrystallized tungsten. Pure, single crystalline tungsten (free of grain boundaries)

THE ELEMENT TUNGSTEN

21

450 ~-'-----r---r-I-"---I""""'I----'
400 H - - - i - - - + -

Work hardened _
Annealed

350.r-~-~-~--~-4--~

300~-~-~-~--~-4--~

>

250H-~-r--~----~---4----+---~

200rr~)r-\r----r----+----+----+---~

I \

\0 ~

\\

150r-~~J--+--r---r--4--4

100

~~
50r--+--,~_~~n~-+_~~-4
~

O~--~--~----~--~----~~~

500

1000

1500

2000

2500 3000

lemperature,OC
FIGURE l.ll. Temperature dependence of the hardness of tungsten [1.40]; source [1.45].

remains ductile down to at least 20 K. In this form, its fracture behavior is very sensitive to
notches or other surface defects.
Originally, it was assumed that cracking only occurs along grain boundaries (i.e.,
intergranular), but later, a high percentage of cleavage (transgranular fracture) was also
observed. Today, we know that, depending upon deformation history, purity, and stress
state, the fracture will be more or less on one side or the other.
TABLE 1.10. Parameters which Modify the Ductile-to-Brittle Transition Temperature [1.43]
Mechanical

Microstructural

Stress state
(tensile/compressive/shear stresses)

Grain size/grain area

Cold work
Texture; grain shape
Twins; subboundaries
Dislocations
Surface roughness
(noches, surface cracks)

Strain rate

Chemical
Distribution of interstitial/substititial
impurities (C, N, 0, P, S, Re)

22

CHAPTER 1

The main structural influences can be described as follows. Any plastic deformation
decreases the transition temperature by fining the structure. The higher the degree of
deformation, the lower is the DBTT. An "as- rolled" tungsten sheet shows a DBTT of
about 100--200 QC. Fine wires or foils are even ductile at room temperature. Any annealing
increases the transition temperature, and recrystallized samples show a maximum, mainly
due to the coarsening of the structure (360 QC).
Grain coarsening has a very important consequence. Simultaneously with grain
growth, the internal grain boundary area is reduced, and hence the concentration of the
grain boundary impurities is raised. Since the grain boundaries are stronger when the
segregations are distributed on a larger area, the ductility decreases.
Today it is assumed that the brittle to ductile transition occurs in two phases:
grain interior:
grain boundaries:

30-40C
28Q-330C

In general, the transformation temperature is increased by the presence of small

amounts of interstitially soluble elements, such as 0, C, and N, which lead to intergranular
precipitations, thereby further weakening the grain boundary strength. However, in regard
to the respective impurities and their influence on the DBTT, a lot of contradicting
statements exist. In the beginning, oxygen segregations at grain boundaries were thought
to be most detrimental. Hydrogen and nitrogen came in addition, and later, in particular
phosphorus and carbon were made responsible for embrittlement. Some metallic elements
were said to favor the ductility. Unfortunately, most of these investigations disregarded
that, in principle, not just one but several elements are present at grain boundaries. Only
their concentrations vary within a wide range. Now it seems well established that the
element combination and their mutual concentrations are of importance. This means,
for example, that if phosphorus is present with no additional metallic element, the
grain boundary strength is low. However, in presence of metallic elements and of
phosphorus, samples show a high percentage of cleavage (a consequence of high boundary
strength).
Additions of rhenium, iridium, and ruthenium can cause the transformation temperature to fall below room temperature, even for slightly deformed products. The ductilizing
effect of Re is explained by alloy softening for low Re alloys and through scavenging
effects (oxygen removal from grain boundaries and dislocations) for high Re alloys
[1.31,1.35].
Drawn wires and rolled sheet exhibit elongated fibrous grains, which can alleviate the
embrittlement caused by interstitials. However, "delamination" can occur in heavily
worked materials during bending along the fiber boundaries. The splitting mode can be
related to contaminated {IOO} cleavage planes [1.41,1.43].
Noncoherent dispersions, such as thoria or zirconia, were shown to have a slightly
ductilizing effect at low temperature, due to their grain-refining effect. Dispersion
strengthened NS-W wires (see Section 6.2.1) behave in a ductile manner during bending,
even in the recrystallized condition as a result of their interlocking, elongated grain
structure [1.42].

1.2.3.4. Plasticity, Strength, and Creep [1.35,1.46,1.47]. Strength, plastic, and

creep properties of tungsten are of paramount interest for its widespread use as structural
material. They have therefore been the subject of a large number of investigations. Most of

23

THE ELEMENT TUNGSTEN

the fundamental work on the deformation of pure tungsten single crystals and polycrystalline tungsten was carried out in the late sixties and early seventies, but recent work based
on dislocation theory of plasticity and work hardening (considering its structural defects)
has significantly contributed to a better understanding of the deformation behavior
[1.35, 1.46-1.49]. It is well beyond the scope of this book to discuss the manifold,
complex metal physical principles which govern the deformation of bcc tungsten; therefore, only the general principles are discussed below. For more information, the reader is
referred to other excellent compilations [1.35, 1.46, 1.47].
Since the early days of tungsten metal manufacture, the metal was considered to be
brittle and difficult to work. Inherent brittleness was at first attributed to the low degree of
purity which could be obtained by P1M techniques and the strong influence of interstitials
on the fracture strength of the metal. In contrast, high-purity single crystals, obtained by
electron beam zone refining, behaved in a ductile manner during deformation, even at the
temperature of liquid helium.
Although these findings are correct in principle, they are only part of the "truth" in
terms of the characteristic deformation behavior of bcc tungsten. Strictly speaking,
tungsten cannot be regarded as "ideally" ductile, in the same way as the fcc metals
[1.48]. This is well reflected through Fig. 1.12, which shows the temperature dependence
of the flow stress of a pure tungsten single crystal (orientation (11O)) [1.48]. The strong
increase in the flow stress with temperature above a certain critical temperature Tc' as
indicated in the figure, is characteristic for all bcc metals [1.47-1.50]. Even for pure
tungsten, Tc is well above room temperature (about 600-800 K ~ 330-530C)
[1.35,1.46-1.49], which explains the difficulty in metal forming at ambient temperatures.
It must be regarded as intrinsic to the bcc crystal lattice and its structural defects.
Above the critical temperature, tungsten behaves during deformation similarly to fcc
metals, showing a characteristic three-stage or four-stage work hardening ("high-temperature work hardening") [1.49]. Below this temperature, parabolic stress-strain curves

400

ta

a.

:::E 300
~

uS

rJl

...<D

200

en

100
0

200

400

600

800

Temperature (K)
FIGURE 1.12. Critical flow stress (0) vs. temperature of a high-purity tungsten single crystal ((110)) [1.48];
tensile deformation rate, 1.6 x 10 - 5 s- I.

24

CHAPTER I

TABLE 1.11. Effect of Orientation and Temperature on the Critical Resolved Shear Stresses for Slip
of Tungsten Single Crystals in Tension [1.50]

(K)

Orientation of
tensile axis

Initial yield
stress
(kg.mm- 2 )

Critical resolved
shear stress on
observed slip
system (kg. mm - 2)

Observed slip
systems

448
293
199
77

[001]
[001]
[001]
[001]

14.3
23.2
38.0
60.0

6.1
9.5
15.5
24.6

(101)[1 II]
and equivalent
systems. Also
(112)[II I]

(112)[111] and
three other
equivalent systems

423
293
199
77

[Oil]
[Oil]
[Oil]
[Oil]

25.4
46.5
84.5
124.0

6.6
11.0
20.0
29.3

(112)[111]
and three other
equivalent systems

(112)[111] and
one other equivalent
system

293
199
77

[111]
[III]
[III]

34.5
50.7
81.0

9.4
16.0
25.6

(110)[1 II]
and equivalent
systems

(211)[111] and
two other
equivalent systems

293

[112]

23.1

10.8

(112)[111]

(011)[111] and
one other equivalent
system

293

qc"

22.9

10.8

(110)[1 II]

(101)[111]

Temperature

Systems with the

highest resolved
shear stress

("low-temperature work hardening") are obtained, as shown in Fig. 1.13 for W single
crystals with various orientations [1.50]. It is striking, that the yield stress and critical shear
stress vary for different orientations, which is a further characteristic difference between
bcc and fcc metals [1.49). Values for both parameters are summarized in Table 1.11 for
different temperatures, together with the operating slip systems [1.50] ..
Micropiasticity and dislocation concepts. Considerable dislocation motion can occur
at stresses far below the conventional macroscopic yield stress, which can be followed by
high-sensitivity strain transducers. Based on the hysteresis loops in load-unload cycling
experiments, the true elastic limit for pure, polycrystalline (arc-cast) tungsten was
measured as cr E3.43 0.8MPa [1.51). It is widely agreed today that the movement of
edge and non-screw dislocations is responsible for this microdeformation (microplasticity). Above the yield point, (Ill) screw dislocations contribute to the macroscopic
deformation (work hardening; double-kink mechanism). The increased activation energy
for the gliding of (Ill) screw dislocations, as compared to edge dislocations, is an
important characteristics of bcc tungsten. Its activation threshold value is responsible for
the strong increase in flow strength with decreasing temperature [1.48). Thus, at low
deformation temperatures, the screw dislocations control the macroscopic flow stress. The
rate-determining step is believed to be the formation of kink pairs in screw dislocations
[1.47-1.49).
Deformation of polycrystalline tungsten. Stress/strain curves of high-purity single
and polycrystalline tungsten (swaged and subsequently annealed for 1 h at 2000 Qq,
obtained at room temperature and 400 QC are shoWn in Fig. 1.14 [1.52]. Whereas the
polycrystalline material exhibited pronounced ductility at 400 QC (elongation 53 %), it

2S

THE ELEMENT TUNGSTEN

800

a;

a.. 600
~

II)
II)

CD

L-

I I)

"'iii
E
La

400

200

0.04

0.08

0.12

0.16

Normal strain
FIGURE 1.I3. Stress--strain curves of tungsten single crystals with various orientations, tested at room
temperature [1.50].

1000

a;

800

a..

:t
II)
II)

CD
I I)

c:
.;:

CD
CD

400

;:;;"0"".'
'\
25C

.5
Cl
c:
w 200

,,
I

V
o

Single crystal
25C

'/

600

L-

Cl

/ '"'\

,
,,

"

...... -

--

--\-- -----

20

Polycrystal
__ /400o C
,
Single crystal _
/4000C
\,

--

40
Elongation

60

80

(%)

FIGURE 1.I4. Stress--strain curves of polycrystal and single-crystal tungsten at 25C and 400C [1.52]; strain
rate 2%min- 1

26

CHAPTER I

behaved in a completely brittle manner (elongation 0.0-0.3 %) during RT testing. In

contrast, the single crystals behaved in a ductile manner under both test conditions
(elongation 27 %, resp. 80 %). These results indicate that the presence of grain boundaries
significantly lowers the DBTT and remarkably deteriorates the low-temperature plastic
properties of the metal.
Deformation curves of heavily worked tungsten ribbons are presented in Fig. 1.15
[1.53] and show a quite different picture. In this case, plastic flow occurred even at 200 K,
indicating a pronounced decrease in the DBTT as a result of cold working ("ductile
tungsten"). This behavior is due to changes in microstructure (texture formation), but also
to the formation of a large amount of edge dislocations, which move rather easily during
deformation and thus provide better ductility [1.48]. Again, an increase in yield stress with
decreasing temperature is obvious, as is characteristic for bcc metals.
The grain size and grain structure have a great influence on both strength and ductility
(elongation). These can be controlled by sintering conditions, type of deformation, degree
of deformation, and annealing processes both during and after machining (intermediate
annealing, stress-relieve annealing, soft annealing, recrystallization annealing).
For maximum deformability, tungsten has to be partly recrystallized (primary
recrystallization), forming a columnar structure with good retention of strength and
ductility. Higher temperatures must be avoided since secondary recrystallization occurs
with significant grain coarsening, leading to a pronounced embrittlement [1.54]. The
temperature for onset of this coarsening is very sensitive to the nature of the starting
material and its deformation history [1.55].
The hot strength of tungsten, like its elongation, depends on its deformation and
annealed condition, and can vary widely at a given temperature (Fig. 1.16) [1.56]. The

8?

300 K

b
uS
(/)

...
en

Q)

400 K

600 K
800 K

Strain rate
, 2.23.10. 5 (sec")

10

12

Strain. E (%)
FIGURE 1.15. Stress-strain curves of as-received tungsten ribbons at various defonnation temperatures [1.53);
strain rate 2.23 x 10 - 5 S - I.

THE ELEMENT TUNGSTEN

27

1200

.....
as
a..

~
......

800

.s::.

CD

c:
Q)

rIJ

400

enc:

{E-

400

800
Temperature

1200

1600

(oC)

FIGURE 1.16. Hot tensile strength of tungsten sheet [1.56]; thickness 1 nun, strain rate 15 % . min -I.

upper limit corresponds to a stress-relieved sheet, and the lower limit to a recrystallized
sheet. At '" 1500 DC, both limiting curves coincide, i.e., at this temperature the tungsten
completely recrystallizes as it is heated to the tensile testing temperature.
The hot strength of pure tungsten is exceeded only by that of rhenium. It can be
significantly further increased by dispersion or precipitation hardening. Alloys based on
W-Re-HfC are the strongest human-made metallic materials at temperatures up to
2700K.
Strength of tungsten wires, whiskers, and theoretical strength. The room-temperature
ultimate tensile strength (UTS) of polycrystalline rods is commonly in the range of 5801470 MPa [1.37]. Cold deformation may increase it significantly. The strength of wires
(exhibiting a characteristic fiber structure in the as-drawn condition) varies with diameter.
The following strengths values are quoted [1.37]: 1765 MPa (600 J.1m); 2450 MPa
(300 J.1m); 2940 MPa (100 J.1m); 3920 MPa (20 J.1m).
The tensile strength of tungsten whiskers, grown by vapor-phase transport, was
reported as up to 27.5 GPa [1.57].
For the theoretical strength, the stress required to cause decohesion in a perfect single
crystal can be assessed by

where E is Youngs's modulus, 'Ys the specific surface energy, and ao the interatomic
spacing [1.46]. Calculations of the ideal strength of tungsten (tension along (100
revealed 90.8 GPa at 0 K and 61 GPa at room temperature. The ideal shear strength of
tungsten (shear plane {11 O} and the direction ( 111 )) was calculated as 16.5-19 GPa
[1.58,1.59]. Both values well reflect the high bonding energy of W in the bcc lattice.
Influence of impurities/alloying. Impurities, in particular interstitials (C, N, 0), are
partly responsible for the high DBTT of tungsten. On the other hand, certain amounts of

28

CHAPTER I

interstitials or solutes (Re, Jr) can improve the low-temperature ductiliy. Both these effects
are explained by the interaction of the foreign atoms with the dislocation network, which
can shift the critical temperature Tc to lower or higher values. It is assumed that the
propagation of dislocations by kink nucleation and lateral motion along the screw
dislocation line is either retarded by the distortion field of the foreign atoms (solidsolution hardening), or enhanced, by generating new kink pairs for their movement (solidsolution softening) [1.47,1.49].
The most effective way to improve the low-temperature plastic properties of tungsten
is by alloying with rhenium [1.60]. For more information see section 6.1.1.

Creep. Creep properties of polycrystalline tungsten, in particular of thin wires, are of

utmost importance for their high-temperature application.
For polycrystalline tungsten, the creep rate can be expressed by [1.61-1.64]
E

= Ban exp ( -

~)

where E is the steady-state creep rate, a the applied stress, !lll the activation enthalpy,
T the temperature (K), and B, n, and k are constants.
Values for B, n, and !lll vary significantly in the literature, depending on the
respective material and testing condition [1.62-1.64]. They are characteristic for the
respective dominating creep mechanism.
For large samples of pure tungsten at temperatures >2200 C (0.65 Tm), the stress
exponent n = 5 and !lll = 585 kJ . mol-I. This activation enthalpy is similar to that of
self-diffusion in tungsten. It was therefore assumed that under these conditions the
deformation of the grains by dislocation climb or glide processes is the rate-controlling
reaction [1.63].
For thin wires, having a grain size comparable to the creep specimen diameter
("bamboo"-grain structure), a stress exponent of n = 2, and an activation enthalpy of
276 kJ . mol-I were obtained, as characteristic for grain boundary diffusion. In this case,
the creep deformation occurred almost entirely by grain boundary sliding (offsetting)
without extensive grain deformation [1.62,1.63].
In general, different independent deformation mechanisms will simultaneously
contribute to the macroscopic deformation, such as dislocation glide (at lower temperatures), dislocation creep (at > 0.05 Tm ), and diffusional creep (Coble creep, or NabarroHerring creep), depending on temperature, magnitude of applied stress, grain size, grain
morphology, purity, and specimen diameter.
Competition between the various mechanisms can be described by so-called deformation-mechanism maps, as shown in Fig. 1.17 for pure W with a grain size of 10 J.lm
[1.35, 1.66]. In industrial practice these maps, however, are of limited use, because the
predominance areas for the respective mechanisms alter significantly with changes in
microstructure (grain size, subgrain size, grain aspect ratio), which may even occur during
deformation [1.64].

High temperature creep resistant tungsten alloys. The microstructure is a crucial

parameter in determining the magnitude of creep. It can be remarkably influenced by the
presence of second phases (Th0 2, Y203, La203, HfC, potassium bubbles), as demonstrated in Fig. 1.18 [1.62] for both non-sag doped and thoriated tungsten. These additives

29

THE ELEMENT TUNGSTEN

Temperature (DC)

100

cD

en

3000
105

--

~
l:::5

2000

1000

--

-- - -

theoretical strength
I

dislocation glide

10- 2

103

....

creep

en
J!1

iii
c:

CD

10- 4

10'

"C
CD

en
en

....CD
iii
J!1

iii

Coble creep

(ij

E
....

to

a..

CD

c:

10- 6

10- 1

10- 8

~--~--~----~--~--~----~--~--~--~~~

.2

.8

.6

.4

1.0

Homologous temperature TIT M

FIGURE 1.\7. Defonnation-mechanism map of pure tungsten with grain size of IOIJlIl (p = 4
nonnalized tensile stress refers to the shear modulus (Il); TM is the melting point [1.65].

1010 cm- 2);

150r-------------------------~

uo-w

(0" A

'

21 MPa)

......

100

:;:;

ttl
Ol

..Q

Th- W

(0" A

'

11.6 MPa)

CD

0 50

AKS- W

(0" A

'

10
20
Test time (min)

21 MPa)

30

FIGURE 1.\8. Creep curves for pure tungsten, thoriated tungsten, and nonsag tungsten wires at 2500C,
showing the outstanding creep resistance of nonsag tungsten [J .62].

30

CHAPTER I

affect the recrystallization grain morphology and the retained dislocation substructure. In
particular, non-sag tungsten is significantly more creep-resistant than pure tungsten,
mainly as a result of the interlocking grain structure, which forms on recrystallization
and which prevents both grain boundary sliding and diffusional creep. In addition, the fine
dispersion of potassium bubbles contributes to the outstanding creep resistance through
pinning of dislocations [1.62, 1.66].
Non-sag tungsten wires are the most creep-resistant wires, with the exception of
monocrystalline tungsten. They are therefore used for sag-free lamp filaments at service
temperatures of up to 3000C (0.88 Tm ), and shear stresses in the range of 0.5 to 10 MPa
[1.66].

1.2.4. Thermal Properties

Melting point:

3422 15C (3695 K) [1.67],

3390 40C (3663 K) [1.68],
3423

30C (3696) [1.69].

The high melting point (highest of all metals) is the most prominent and important
property in regard to all applications as refractory metal. It is a consequence of the electron
density of states. Small amounts of impurities, such as carbon, lower the melting point.
The molar volume increases by 8 % on melting. This is the largest expansion
observed for bcc metals [1.70].
The melting curve ofW has been determined to 5GPa [1.71].

Enthalpy offusion: 46 4 kJ . mol-I [1.68].

Entropy offusion: 14 J . mol- I . K - I [1.70].
Enthalpy of sublimination: 858.9 4.6 kJ . mol- I [1.35].
Vapor Pressure. Tungsten has the lowest vapor pressure of all metals. Within the
temperature range from 2600 to 3100 K, it obeys the following equation [1.76]:
log p[Pa] = -45395T- 1 + 12.8767
At 2000C, the vapor pressure is 8.15 x 1O- 8 Pa; at 3000C it is 10- 1 Pa. Experimental
data for p over liquid W are not available.
The rate of evaporation in vacuum is about 6.2 x 10 - II g. cm - 2 S- I at 2000C,
about 7 x 10- 8 g. cm -2. S-I at 2500C, and about 2.5 x 10- 5 g. cm- 2 . S-I at 3000C
[1.72]. It is markedly reduced by an inert gas atmosphere (Ar, Kr). Therefore, modem
incandescent lamps contain inert gas fillings to avoid enhanced wall-blackening (the rate of
evaporation in vacuum is about 500 times larger as compared to an Ar atmosphere of 1.2
bar) [1.73].

Boiling Point: calculated from rates of evaporation of solid tungsten,

5663 C (5936 K) [1.74],
5700 200C [1.64].

THE ELEMENT TUNGSTEN

31

Critical temperature: 13400 1400 K [1.75].

Critical pressure: (3.37 0.85) x 108 Pa [1.75].
Critical density: 4.31 g. cm -3 [1.35].
Thermal expansion. At room temperature, values between 4.32 and 4.68
x 10 - 6 K - 1 were obtained for the linear coefficient of expansion (1., depending on the
material (P1M sheet, arc-cast sheet, etc.) and the type of measurement. Values for low and
high temperatures are listed in Table 1.12. The linear coefficient of expansion can also be
calculated according to the following equations [1.76]:
temperature range: 293-1395 K
(1.

= 4.266 x 1O-6(T - 293) + 8.479 x lO-lO(T - 293)2 - 1.974 x lO-13(T - 293)3

temperature range 1395-2495 K

(1.

= 0.00548 + 5.416 x 1O-6(T - 1395) + 1.952

x lO-lO(T - 1395)2

+ 4.422

x 1O-13(T - 1395)3

temperature range 2495-3600 K

(1.

= 0.01226 + 7.451 x 1O-6(T - 2495) + 1.654 x 1O-9(T - 2495)2

+ 7.568

x 1O-14(T - 2495)3
The very low thermal expansion of tungsten makes it compatible with glass and
ceramics in high temperature applications.

Thermodynamic jUnctions [1.35]. Thermodynamic functions for solid tungsten are

listed in Table 1.13. For more details and data for liquid tungsten, see elsewhere [1.1 0].
TABLE 1.12. Thermal Expansion Coefficient for
Low and High Temperatures [1.35]
T (K)

106 . CX(K-l)

T (K)

106 . cx(K- 1)

10
15
20
25
30
40
50
60
70
80
100
130

0.006
0.019
0.048
0.102
0.20
0.53
0.90
1.43
1.88
2.30
2.82
3.42

160
190
220
260
300
600
1000
1400
1800
2200
2600
3000
3400

3.82
4.06
4.20
4.32
4.49
4.75
5.02
5.46
6.11
6.89
7.76
9.05
11.60

32

CHAPTER I

TABLE 1.13. Thennodynamic Functions [1.35]

T

Hr -H298 .S

(K)

(kJ'mol- l )

298.15
500
1000
1500
2000
2500
3000
3500
3600

0
5.022-5.024
18.23-18.25
32.33-32.62
47.59-48.11
64.52-64.78
83.1 0--83.88
105.19
II 0.5-111.8

Hr - Ho
S
S - So
Cp
Cv
(kJmol- l ) (kJ.mol-I.K- I) (kJ.mol-I.K- I) (kJmol- 1 K- I) (kJ.mol-I.K- I)

4.970
10.007
23.142
37.249
52.599
69.476
88.310
109.845
114.574

32.640
45.516
63.658
75.068
83.881
91.399
98.254
104.880
106.212

32.618
45.457
63.686
75.099
83.585
91.397
98.442
106.89

24.1 0--24.42
24.33-25.44
27.19-27.60
29.23-29.86
31.37-32.13
34.67-36.00
39.25-41.80
46.49-50.85
48.11-54.68

23.96
25.16
26.70
28.12
29.80
35.91

Analyses of thennodynamic properties of tungsten at high temperatures are available

[1.77,1.78].
The heat capacity at low temperatures is:
25
30
Cp (J. mol-i. K- I) 0.73
1.35
TinK
120
140
Cp (J . mol-I. K- i ) 18.28 19.87

TinK

35
2.22
160
21.01

40
3.30
180
21.86

50
5.82
200
22.54

60
8.39
220
23.04

70
10.74
260
23.81

80
100
12.81 16.04
300
24.35

The heat capacity of liquid tungsten is 35.564 J mol-I. K -I [1.1 0]

Self-diffusion [1.35]. At a certain temperature, the diffusion process is characterized

by the diffusion coefficient D. Its temperature dependency is given by the Arrhenius
equation D = Do exp( -Q/R . T) with Do a constant in cm2 g-I, T the absolute
temperature, R the molar gas constant in J . K- 1 mol-I, and Q the activation enthalpy
in kJ mol-I.
Activation enthalpies for the lattice (volume) diffusion were derived between 586 and
628 kJ . mol- I for single crystals and between 502 and 586 kJ . mol- I for polycrystalline
tungsten. Accordingly, the following equations were set up [1.79]:
Self-diffusion in tungsten single crystals: D = 42.8 exp( - 640/RI).
Self diffusion in polycrystalline tungsten: D = 54 exp( - 504/RI).
Over the range of 1900 to 2800 o e, a linear relationship between log D and 1/ Twas
obtained for polycrystalline tungsten.
The self-diffusion parameters Do and Q are influenced by the impurity content of the
diffusion zone. Higher values for both were obtained for impure tungsten. This effect is
more pronounced at low temperatures and vanishes above 2043 K. (It is assumed that
impurities attract vacancies and the higher vacancy concentration disturbs the diffusion
process).
The fact that lower Q values were obtained for polycrystalline tungsten than for single
crystals, especially for T < 0.7Tm, is due to a more or less significant contribution of grain
boundary diffusion to the total bulk diffusion. Observed rates of volume diffusion,

33

THE ELEMENT TUNGSTEN

obtained on polycrystalline tungsten, will therefore always depend on the microstructure of

the material under investigation.
Grain boundary diffosion is the dominant mechanism in polycrystalline tungsten
below 2100C. For grain boundary diffusion, activation enthalpies between 377 and
460kJ mol-I were measured.
Thermal conductivity and diffusivity: [1.35,1.80,1.81]. At room temperature, the
thermal conductivity coefficient A is about 1. 75 W . cm- I . K-I. Its temperature dependence is shown in Fig. 1.19. Some typical data are given in Table 1.14. In the lowtemperature range, it is strongly dependent on the RRR (residual resistivity ratio Pol Pr),
which is also an indication of the purity. Table 1.15 shows the differences between samples
of varying purity.
Between 1200 and 2800 K, A obeys the following equation [1.80]:

At the melting point, A drops from 0.895 (solid) to 0.705 (liquid) [1.35].
Sintered and arc-cast tungsten specimens differ in A values at elevated temperatures.
The thermal conductivity of single crystals and polycrystalline tungsten, both having the
same degree of purity, coincide at high temperature.
There exist certified NBS standards of arc-cast as well as sintered tungsten as
reference material for the temperature range of 4 to 3000 K [1.83].
The high thermal conductivity of tungsten (very suitable for use as heat sinks) in
combination with the low specific heat results in high cooling rates during hot-working of
the metal, which makes the handling during working more difficult.
The thermal diffusivity of tungsten at 300 K is 0.662 cm2 S-I, and decreases to
0.246 cm2 . S-I at the melting point. At low temperatures 200 K) the values are strongly

100
80

U/

~20

""

Ii

"
1/

~4

f-

1l-

....

....

T.

[~ f-

0.4
0.31

2 3 4 56 810

2 3 4 56 8 JOZ

T(K)

2 3 4 56 810 3

-+

2 3 45

FIGURE 1.19. Temperature dependence of the thermal conductivity A of well-annealed high-purity tungsten
[1.35]; source [1.81].

34

CHAPTER I

TABLE 1.14. Thennal Conductivity

Coefficient at Different Temperatures
[1.35,1.81]
Temperature
(K)

A.
(W. em-I. K- 1)

0
10
50
100
300
500
1000
2000
3400

0
97.1
4.28
2.08
1.74
1.46
1.18
1.00
0.90

dependent on the impurity level and the density oflattice defects. For more information see
elsewhere [1.82].
TABLE 1.15. Thennal Conductivity Coefficient in the
Low-Temperature Range and Dependence on the
Residual Resistivity Ratio [1.35, 1.80]

Temperature (K)
I

5
10
20
50
100
150
200
300

RRR 100

RRR 300

0.50
2.49
4.88
7.99
3.57
2.17
1.97
1.89
1.74

1.51
7.49
14.04
17.34
3.98
2.24
2.01
1.91
1.76

1.2.5. Electromagnetic Properties [1.35]

Electrical Resistivity (J.t!l. cm). In the low-temperature range <0.07 K the electrical
resistivity Po is independent of the temperature and will be only determined by the
impurity level.
In order to calculate p within different temperature ranges, equations have been set
up:
up to 40K

35

THE ELEMENT TUNGSTEN

40-90K
p = 0.14407 - 1.16651 x 1O-2 T + 2.41437 x 1O-4 T2 - 3.66335 x 1O-9

90-750K
p = l.06871

+ 2.06884 x

- 5.14195 x lO- I2

1O- 2 T + l.27971 x 1O- 6 T2 + 8.53101

1O- 9 T 3

300-1240K
p = 4.33471

1O- 12 T2 + 2.19691 x 1O- 8 T - l.64011 x 10-6

1O- 12 T2 + 4.67093 x 1O- 8 T - l.94101 x 10-5

1240-2570K
p = 4.06012

Moreover, the resistivity is influenced by all kinds of lattice defects (vacancies,

dislocations, grain boundaries, etc.) and impurities.
Some selected values near and above room temperature are listed in Table 1.16. The
resistivity of liquid tungsten close to the melting point and at 5000 K was given as
131 ~. cm and 160 ~. cm, respectively [l.84].
The resistivity of thin films depends on microstructure, impurity content, and surface
roughness. These properties are the consequence of deposition conditions, substrate
temperature, and annealing. Therefore, a wide scatter of values was observed, for example,
between 6 and 20 ~ . cm at 30 K [1.35]. Under certain deposition conditions, metastable
P-W forms, which has a significantly higher bulk resistivity (approximately 10 times that
of the thermodynamically stable r:J..-W phase).

Superconductivity [1.35]. Tungsten is a Type I superconductor with a transition

temperature of 0.0154 0.0005 K. The critical magnetic field strength Hc(T ~ 0) is
1.15 0.03 Oe. (9l.5 A m -I). Impurities only show a minor influence on the transition
TABLE 1.16. Electrical Resistivity of Tungsten at
Near-Room and at Elevated Temperature [1.35]
T(K)

p(!1il. em)

T(K)

p(!1il. em)

273
293
298
300

4.82
5.28
5.40
5.54

400
500
600
800
1000
1200
1400

8.05
10.70
13.35
18.85
24.75
30.90
37.20

36

CHAPTER 1

temperature:
RRR 57000
RRR 7500

0.0160 K
0.0154 K

In thin films, due to structural influences or the presence of ~- or y-tungsten, much

higher critical temperatures (up to 6 K) were found.
Thermoelectric effects. A temperature gradient generates in tungsten a small potential difference (the Thomson effect). This difference, called the thermoelectric power
S (IlV IK), drops with increasing temperature. Its temperature dependence is tabulated in
Table 1.17.
Thermocouples. For temperature measurements up to 2300C, thermocouples
consisting of tungsten and tungsten-rhenium alloys are employed widely. The most
common combinations are W versus W26Re, W5Re versus W26Re, and W3Re versus
W25Re. Electromotive forces built up by each combination for different temperatures are
shown in Table 1.18.
W-Re thermocouples can only be used in neutral or reducing atmospheres. Other
combinations, such as W IPt, W ITa, or W IMo, have gained no commercial importance.
Magnetoelectric effects [1.35]. The Hall coefficient at room temperature varies
between 10 and 12 x 10- 11 m3 . A-I. S-I.
The Magnetic susceptibility is given either as Xmol (10- 6 cm3 . mol-I) or as Xg
6
(10- cm3 . g-I). It increases with temperature. Values for different temperatures are given
in Table 1.19.
1.2.6. Optical Properties

The optical properties of tungsten have been studied in more detail than those of any
other metal or material, because tungsten is not only used as an incandescent lamp
filament, but also as a comparison temperature standard in specially constructed strip
lamps.
TABLE 1.17. Temperature Dependence of
Thermoelectric Power of Tungsten [1.35]
T
(K)

10
20
50
80
100
150
200
250
273
300

S
(IlV . K)

+ 0.05
-

0.28
2.78
3.70
4.04
2.45
1.41
0.10
+ 0.56
+ 1.44

T
(K)

(IlV' K)

400
600
800
1000
1200
1400
1500
1600
1800

4.62
10.75
15.51
18.46
20.06
20.63
20.70
20.61
19.15

37

THE ELEMENT TUNGSTEN

TABLE 1.18. Tungsten and Tungsten-Rhenium Thermocouple Electromotive

Forces at Various Temperatures for a Reference Junction at OC [1.39]
Electromotive force (mV)
Temperature eC)

Wvs. W26Re

W5Re vs. W26Re

W3Re vs. W25Re

0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300

0.000
0.344
1.006
1.985
3.282
4.793
6.488
8.331
10.300
12.319
14.393
16.497
18.648
20.767
22.814
24.841
26.849
28.842
30.814
32.589
34.246
35.851
37.435
38.896

0.000
l.381
2.988
4.768
6.655
8.573
10.508
12.450
14.374
16.266
18.120
19.994
21.724
23.424
25.033
26.583
28.078
29.528
30.922
32.298
33.632
34.915
36.089
36.928

0.000
1.146
2.604
4.289
6.129
8.098
10.092
12.129
14.184
16.226
18.242
20.229
22.192
24.082
25.896
27.686
29.450
31.182
32.874
34.359
35.723
37.037
38.307
39.350

Spectral emissivity A. [1.37,1.85-1.88]. The spectral emissivity of tungsten as a

function of temperature is shown in Fig. 1.20. In the wavelength range of 0.3-0.5 Jlm, the
emissivity has a maximum. The X point of tungsten, where all the emissivity wavelength
isothenns cross, corresponds to Ax = 1.27 f.1m = 0.33).
Total emissivity et The experimental results of et measurements can be well
TABLE 1.19. Temperature Dependence of
the Magnetic Susceptibility of Thngsten
[1.35]
Temperature eC)
-193
-78
Room temperature
Single Crystal
Polycrystalline
500
1600
1850

51.1
53.7
54.0
53.3
58.0
63
68

38

CHAPTER 1

0.5

Cb

:>.

r---------------------~

0.4

.s;
0

e
(/)

Cb

~ 0.3
"0
Cb

1600
1800
2000
2200
2400
2600
2800

Q.

C/)

0.2

0.3

0.2

0.4

K-~~,~

K
K
K
K
K
K

0.5 0.6
0.8 1.0 1.2 1.4
Wavelength A (/Lm)

1.8

2.2 2.6

FIGURE 1.20. Spectral emissivity of tungsten according to de Vos [1.37].

represented by the following empirical equation (valid from 400 to 3600K) [1.88]:

el

= (-2.6875 x 10-2) + (l.819696 x 1O-4 )T -

(2.l946163 x 1O-8 )T2

A comparison between measured and calculated results is ~ven in Table 1.20. Values
in the temperature range from 1200 to 3200 K are compru:ed for single crystals and
polycrystaUine tungsten in Table 1.21.

Pyrometry. It was shown -that the emissivity of tungsten depends on wavelength. It is

TABLE 1.20. Total Emissivity and Radiated Heat of Tungsten at
Various Temperatures [1.39]
Emissivity

True temperature

eC)

(K)

Radiated heat
P(W cm-2)

3327
2927
2527
2127
1727
1327
927
527
127

3600
3200
2800
2400
2000
1600
1200
800
400

327.4
197.0
108.2
53.3
22.6
7.72
1.87
0.238
0.0042

Calculated

Measured

0.344
0.331
0.311
0.283
0.249
0.208
0.160
0.105
0.0424

0.354
0.341
0.323
0.296
0.260
0.207
0.143
0.088
0.042

THE ELEMENT TUNGSTEN

39

TABLE 1.21. Temperature Dependence of Total Tungsten

Emissivity for Single Crystals and Polycrystalline Material
[1.35]
Temperature (K)
1200
1400
1600
1800
2000
2200
2400
2600
2800
3000
3200

Single crystal

(e t )

0.126
0.163
1.196
0.226
0.251
0.273
0.292
0.307
0.319
0.329
0.337

Polycrystal

(et )

0.139
0.173
0.204
0.233
0.257
0.278
0.298
0.312
0.324
0.333

higher in the blue and less in the red, in comparison to a uniform gray body. Therefore, the
brightDess temperature of tungsten at O.651lm (where an optical pyrometer reads
temperature) is higher than the true temperature of a gray surface in equilibrium with
tungsten. Related corrections must be applied. Table 1.22 shows a comparison between
brightness temperature and true temperature as well as color temperature. The last
represents the true temperature of a black body radiating with the same intensity
distribution of wavelength as tungsten (it would exhibit the same color). The more intense
radiation (emissivity) in the blue is the reason why the color temperature is higher than the
true temperature.

TABLE 1.22. Temperature Scale for Tungsten [1.39]

Brightness temperature (K)

True temperature (K)

Color temperature (K)

1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2320
2400
2500
2600
2700
2800

1577
1674
1771
1866
1961
2055
2149
2242
2335
2427
2520
2612
2703
2793
2886
2977

2080
2200
2320
2450
2580
2700
2840
2960
3100
3240

40

CHAPTER I

TABLE 1.23. Work Function for Different Metals and Tungsten/Adsorbate Combinations
[1.90] and Electronegativity of the Metals According to Pauling
Metal

Work function
(eV)

Metal/adsorbate

Electronegativity

Work function
(eV)

0.79
1.00
1.33
1.50
1.70
2.28

1.81
3.2
3.31
4.19
4.52
5.32

Cs on W
Baon W
Thon W
La on W
Ceon W
Zr on W b

1.36
1.56
2.6
2.70
2.70
3.13

Cs
Ca
Tha
Ta
W

Pt
a
b

From Ref. 1.93.

From Re( 1.89.

1.2.7. Electron Emission

The energy required by an electron to escape from the surface of a crystalline solid is
called the work function (<p) of the material. It is a characteristic parameter for its electron
emission behavior. The work function of metals is in the range of ~ 2 to 6 eV and correlates
mainly with the electronegativity of the element (Table 1.23).
For polycrystalline tungsten, it is commonly quoted as 4.50-4.56 eV
[1.38,1.89,1.90]. In case of single crystals, it varies for different crystallographic
orientations and ranges between 4.21eV for the (310) plane and 5.7geV for the (110)
plane [1.90]. The higher the atom density of the emitting plane, the higher is its work
function (<P110 ><P100 > <PIlI) Calculated single-plane work function data are presented
elsewhere [1.38].
The temperature coefficient of the work function of polycrystalline tungsten (d<p/dT)
is in the range of 60-110 ~eV . K- I [1.38].
Thermionic emission. The number of electrons which escape from the metal surface
increases rapidly with temperature (thermionic emission). In general, the higher the
temperature and the lower the work function, the higher is the electron emissivity. The
current density can be calculated by the Richardson-Dushman equation (in the absence of
an external electrical field), according to i = AT2 exp( -<P / kT), where A is the Richardson
constant (A cm- 2 . K- 2), Tis the temperature (K), and <P is the work function (eV). For
pure tungsten A = 60.2 (A cm- 2 . K- 2 ) [1.91]. The thermionic current (A cm- 2 ) can
then be calculated as: i = 60.2T2 exp( -52230/T) [1.37].
The work function of tungsten is considerably changed by adsorption of foreign
elements on the crystal surface. It can be both increased, as in case of oxygen
contamination, or considerably decreased, as in case of Th, Ba, and Cs additions (Table
1.23). In general, the work function is increased if an electron transfer occurs from
tungsten to the adsorbate (i.e., if the adsorbate is more electronegative than tungsten), or,
vice versa, it is decreased if the adsorbate is less electronegative and electrons are thus
transferred from the adsorbate to tungsten. In the latter case, a minimum in the work
function is obtained for a monolayer of the adsorbate, covering the tungsten surface [1.92J.
For example, the work function of thoriated tungsten in the activated form (where a
monolayer of thorium metal forms on the tungsten surface as a result of a high-temperature
treatment) is significantly lower (2.6 eV) than that of bulk thorium metal (3.38 eV) [1.93].

41

THE ELEMENT TUNGSTEN

106
,,/

E
......

./
./

10

>-

thoriated
tungsten

iii
r:::

Q)

"C

c:
......

10 2

Q)

::::l

r:::
0

iii

en

10

E
UJ

10. 2
800

1200

,/

/
1600

//
/

//
/

v,ungsten

2000

2400

2800

Temperature (Kl
FIGURE 1.21. Variation of emission current density with tungsten for pure and thoriated tungsten electrodes
[1.93); according to Ref. 1.92.

Due to the lower work function of W-Th, W-Cs, or W-8a, much higher thermionic
current densities can be obtained as compared to pure tungsten (Fig. 1.21). This is of high
practical importance for the widespread use of tungsten as an electron emitter material. On
the one hand, significantly higher current densities can be achieved at a certain service
temperature. On the other hand, significantly lower service temperatures can be applied for
an equal current density, which results in considerably longer lifetimes of the respective
emitter parts.
Enhancement of the thermionic emission of tungsten by addition of thoria was first
observed by 1. Langmuir in 1914 [1.94]. Already, it was assumed that metallic thorium is
responsible for this improvement. How this metallic layer forms on the W surface has
never been fully clarified, although several explanations have been given in the past (see
also Section 6.2.2). In common with all these theories is the fact that the material must be
activated prior to use. This is done by heating it to above 2000C and then holding it at a
somewhat lower temperature. During this treatment, metallic thorium is formed and
thorium atoms diffuse from the bulk to the surface, forming a monolayer. The emission
efficiency can be enhanced by heating the material in a hydrocarbon vapor. In this case, it
is assumed that W2C is acting as a reducing agent for the formation of metallic thorium.
Tungsten is the most important metal for thermoemission applications, not only
because of its high emissivity but because of its high thermal and chemical stability
(extremely low vapor pressure at service temperature; high hot strength and rigidity;
excellent corrosion resistance against metal and oxide vapors).
Tungsten-base materials are used for cathodes in power grit tubes, radio valves, X-ray
tubes, as electrode material for gas-discharge lamps, electric arc welding, electron guns,
electron microscopes, and plasma generators. The electron emission characteristic is also
an important property for their use as electrical contact materials. Finally, tungsten is a

42

CHAPTER 1

potential candidate for crucial parts in thermionic energy conversion systems and cesiumion engines for extraterrestrial applications [1.95].
In most of these applications, either thoriated tungsten is used (service temperature
1700-1800 DC), or impregnated cathodes, based on BaOI Al20 3 /CaO additions (so-called
dispenser cathodes; service temperature 900-1250 DC). However, because of its radioactive nature, thoriated tungsten is currently being more and more replaced by tungsten
with rare earth (RE) oxide additions or w-zr0 2 (see also Section 6.2.2).
Field emission. For only a few specific applications (for example, in field emission
electron microscopy) is tungsten used as a cold emitter. In this case, the electrons are
extracted from the tungsten surface by applying an external electrical field in the order of
5 x 107 V lern. This field is obtained by using tungsten single crystal wires with tip radii of
about 100 nm (produced by electrolytic etching) and applying extraction tensions of up to
6kY. Field emission cathodes can only be used in ultrahigh vacuum (l0-8Pa).
Photoelectric emission. Electron emission can also be caused by incident photons
(photoelectric emission). The energy required to remove an electron by light energy in
terms of wavelength is called the threshold wavelength. For thoriated tungsten, it has been
quoted as 319 and 438 nm, respectively, depending on the measurement device [1.96].
For more information on electron emission, including interactions of tungsten with
charged particles and atoms (secondary emission), the reader is referred to the compilation
in a recently published handbook [1.38].
1.2.8. Acoustic Properties [1.97]
The sonic velocities in tungsten are 5180m S-I for longitudinal (compression)
waves, 2870m S-I for transverse (shear) waves, and 2650m S-I for surface waves. The
acoustic impedance for longitudinal waves is 9.98 g . cm- 2 . S-I.

1.3. CHEMICAL PROPERTIES OF TUNGSTEN METAL

1.3.1. General Remarks
Although the main directing properties in tungsten's applications are of a pure
physical nature (high melting point, high density, low vapor pressure, etc.), the chemical
properties too are of special importance because they determine and limit the application
fields of the metal in regard to diverse environments. It must be stated in advance that these
properties are quite opposite.
On the one hand, tungsten can be considered as a rather inert metal, which is resistant
to many elements and compounds. It is compatible with most ceramics and glasses up to
high temperatures and shows good resistance to many molten metals. Tungsten is stable to
mineral acids in the cold and is only slightly attacked at higher temperatures.
On the other hand, it reacts with numerous agents as well as elements and chemical
compounds. For example, at room temperature, it is strongly attacked by fluorine. Below
100C, it dissolves in hydrofluoric-nitric acid mixtures, aqua regia, and alkali solutions
containing oxidizing agents. Hydrogen peroxide is a good solvent for tungsten powder. At
elevated temperatures, the number of reacting agents increases. At 250 DC, it reacts with
chlorine, phosphoric acid, potassium hydroxide, and sodium nitrate or nitrite. At 500 DC,
the attack by oxygen or hydrogen chloride becomes vigorous. At 800 DC it reacts with

THE ELEMENT TUNGSTEN

43

ammonia and at 900C with carbon monoxide, bromine, iodine, and carbon disulfide.
With carbon or carbon-containing compounds, reactions occur above 1000 DC. Reactions
of this type are important in the preparation of tungsten carbide (in regard to quantity, the
most important tungsten compound) and are treated in more detail in Section 3.6.
Tungsten metal is stable in dry and humid air only at moderate temperature. It starts to
oxidize at about 400C. The oxide layer is not dense and does not offer any protection
against further oxidation. Above 700 C the oxidation rate increases rapidly, and above
900C, sublimation of the oxide takes place, resulting in catastrophic oxidation of the
metal. Any moisture content of the air enhances the volatility of the oxide.
Although tungsten is the metal with the highest melting point, its sensitivity toward
oxidation is a big disadvantage. Therefore, all high-temperature applications are limited to
a protective atmosphere or vacuum.
Bulk tungsten does not react with water but will be oxidized by water vapor at
elevated temperature, e.g., at 600 DC.
Due to the importance of the reactions with oxygen, air, and water, Sections 3.1 and
3.2 have been specially devoted to this subject.
This chapter deals primarily with the reactions of bulk tungsten and tungsten powder.
A vast literature exists about their reactions. However, it is not so much the reactions of the
bulk metal which attracted the interest of scientists and technicians as its reactions at the
surface. Scientific as well as practical reasons boosted that research owing to the
widespread application of tungsten as a catalyst and as an electron or ion emitting
source. It would be far beyond the scope of this book to deal with surface reactions in
detail. In this regard, reference is made to the very complete compilation elsewhere [1.991.101].
Figure 1.22 presents a Periodic Table of the elements with information about the
reactivity of tungsten toward elements, formation of compounds, and solubility of tungsten
in metal melts. As can be seen, tungsten reacts with the majority of nonmetallic elements,
with the exception of hydrogen, inert gases, and nitrogen. Molecular nitrogen, due to its
high bonding strength, does not react with tungsten directly, but tungsten nitrides can be
prepared via reaction with dry ammonia. Reactions with nonmetals are described in the
following subsection.
The oxidation state of tungsten in compounds may vary between - 2 and + 6, the
latter being the most common. Lower oxidation state compounds exhibit basic properties
while higher ones are acidic. The maximum coordination number is 8, but it may attain 13
in coordination compounds with cyclic organic ligands. Chapter 4 treats tungsten
compounds in detail.
However, in regard to the reactions of tungsten with metallic elements, the situation is
quite different. A large number of metals exist which fail to react with tungsten, like the
alkali metals, the alkaline earth metals with the exception of beryllium, the rare earth
metals with the exception of cerium, and especially the elements scandium, yttrium,
lanthanum, copper, silver, gold, zinc, cadmium, mercury, indium, thallium, tin, lead,
antimony, bismuth, and polonium.
"Not reacting with tungsten" does not imply surface reactions, such as surface
bonding of metals, which occurs with all metals. Some of these layers may change
considerably specific surface-related properties of tungsten, like the work function, as in
the case of thorium, lanthanum, barium, etc. This is of great technical interest for highenergy electron-emitting cathodes or welding electrodes.

Mg

00

Be

Sc

Ti

neglectible
slight
moderate

Cr

Molten Metal

Mn

W Solubility in the \ .

...-:'

Ni

Co

00

Fe

00

Cu

unlimited
extensive
neglectable or none

Solid Solution

oomp,"o'(,) " ph".(,)

formation of intermetallic

no reaction with W

00

+ lom,!;oo oloompo"'(')

Zn

Ga

00

As

Ge

Si

AI

Se

S
+

0
+

- ...

-------

Ac

.. -

...

Th

..
00

Ce

Nd

00

00

Pm Sm

00

00

Eu

00

Gd

Tb

Dy

00

Ho

00

Er

Tm

Yb

Rn

At

Lu

Xe

Kr

Ar

Ne

Br

CI

'"He
-

Pa

...

.
~--

Np

. .

Pu

--~---

-----

Am Cm

--------

Bk

----

Cf

------

Es

L-~.

Fm

Md

No

Lr

. .. - .. - .. - .. - .. - .. - .. - .. - ...- ... - ...- ...- ... -

Pr

00

00

La

00

00

FIGURE 1.22. Periodic Table; reactions of W with the elements.

00

- . - .. .. + +
..
..
.
.
.
..
.
...

Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te
- - ..
-... - . - . - . - . - . +
..
.
..
.
..
..
..
..
Cs Ba
Hf Ta W Re Os Ir
Pt Au Hg TI Pb Bi Po
- .. ...
.. .. .. .. - . - . - . - . - . Fr Ra
-

K Ca

Na

Li

::c

:>
"0

"'"'""

45

THE ELEMENT TUNGSTEN

Group

~
Clr ~

oo
~

::l

~Q)

c.

4000Df
(
2000 ..
\

50 100

10

IDDD

~~DD

D~~D~
Atomic percent tungsten

FIGURE 1.23. Solid solubilities in the transition metal-tungsten a\losy [1.101].

Metallic elements reacting with tungsten besides beryllium and cerium are concentrated in groups 4 to 7 of the Periodic Table. Here, we find binary intermetallic compounds
with close homogeneity ranges and phases with very broad ranges, besides solid solubility
ranges toward the element-rich sides of the respective binary system. In some cases,
especially within groups 4 to 6, only a continuous series of solid solubility exists. A survey
on solid solubilities is presented in Fig. 1.23.
Table 1.24 presents data on compound types in the various binary systems. Reactions
with metals are dealt with in more detail in Section 1.3.3.

TABLE 1.24. Types of Compounds and Intermetallic Phases in Binary Tungsten Systemsa
Formula

at% W

Number of systems

M 22 W
M l2 W
M6C
MsW

4.3
7.7
14
17
18-20
25
28-31
33.3
40
42-46
50
62
62.5
66.7
75

I
2
3
4
8
8

~W

M3W
MSW2
M 2W
M3W2
M 7W 6 (/l)
MW
MI_xW
M3WS
MW2
MW3
cr
E
a Excluding

nitrogen and oxygen.

13
1
2
10
2
5
2
5
3

Element
Be
AI, Be
Br, CI, F
AI, CI, Br, F
AI, B, CI, Br, F, I, Ni, Rh
Br, I, Co, Os, Pd, Re, Rh
B
As, Be, Br, CI, Fe, Ge, I, P, Pt, S. Se, Si, Te
As
Co, Fe
B, C, Fe, Ir, Ni, P, Pt, Re, Ru, Te
C
Ge, Si
B, C, Hf, Ni, Zr
Cr, P
Ir, Os, Re, Ru, Te
Ir, Pt, Rh

46

CHAPTER I

TABLE 1.25. Reaction of Tungsten Metal with

Acids and Alkalis
Temperature
Reagent

20C

10001lOC

HF
HN0 3
H2 SO4
HCl
H3P04
H2 0 2
NH4 0H
KOH
NaOH
HCI+ HN03
HF+ HN03
KOH +H2 0 2

None
Slight attack
None
None
None
None
None
None
None
Oxidation
Dissolution
Slight attack

None
Oxidation
Slight attack
Slight attack
Slight attack
Dissolution
None
None
None
Dissolution
Dissolution
Dissolution

Finally, Sections 1.3 .4-1.3.6 are concerned with reactions of tungsten with
compounds, aqueous solutions, and miscellaneous substances, which are of importance
for diverse applications. The reactivity of tungsten in aqueous solutions is interesting in
relation to analytical procedures and chemical engineering. It will be shown in Table 1.25
that aqueous solutions of nonoxidizing acids and alkalis, as well as ammonia, do not attack
tungsten. On the other hand, the presence of oxidizing agents in combination with acids or
alkalis (nitric acid, sodium nitrate, sodium nitrite, hydrogen peroxide, potassium peroxide)
results in rapid dissolution.
Tungsten resists the attack of molten alkalis. Here, too, the addition of oxidizing
compounds leads to vigorous dissolution.
Due to its complexity, the chemistry of aqueous solutions is treated in Section 3.7.
1.3.2. Reactions with Nonmetals [1.39,1.99,1.102]
Table 1.26 presents a review of the reacting elements, the related reaction conditions,
and the products formed.
Reference is made to Chapter 4 for preparation possibilities concerning the
compounds described here and further possible tungsten compounds, their properties,
and their technical importance.
1.3.3. Reactions with Metals [1.39,1.99-1.101]
This section not only includes reactions where chemical compounds of close
homogeneity ranges are formed (intermetallic binary compounds), but also cases in
which phases with broad homogeneity ranges exist or only mutual solid solubility was
observed.
Group 1: Alkali Metals (Li, Na, K, Rb, Cs). These metals do not react with tungsten
and do not form any binary compounds. Due to the very low solubility of tungsten in
molten alkali metals, only a slight attack of tungsten metal specimens was observed in

THE ELEMENT TUNGSTEN

47

TABLE 1.26. Reactions of Tungsten Metal with Nonmetals

Element

Reaction conditions

Remarks

Products

Boron

W + B (amorphous) powder
mixture compacts
500C in H2/l hour
800-l200C in Ar/2 hours

Individual boride formation

depends on W/B ratio

W2B, WB, W2Bs, WB4

Carbon

Reaction between 800-1900C

800C
1050C
1550C
1900C

Depending on reactant form and W2C, WC

atmosphere
CVD carbon layer on W foil
Fine powder mixture
Polycryst. Wire in C powder
Monocryst. W wire in C
powder

Silicon

WlSi powder mixtures at

elevated temperature and
protective atmosphere

Properties of starting materials

and experimental conditions
determine reaction products

WSi2, W3Si2, W2Si3, WsSi3

Phorphorus W/P (red) mixtures at

700-950C

Under vacuum

WP2

Arsenic

W/As mixture at 620C

Protective atmosphere

WAs2

Oxygen

Reaction starts at

Above 500C oxide layer forms W03

cracks, at 800 C sublimation of
W03

Sulfur

Molten S or vapor attacks slowly,

powder mixtures start reacting at
400C

Selenium

W/Se mixtures start reacting at

480-520C

Exothermic

WSe2

Tellurium

Mixture react slowly at 600700C

In vacuum

WTe2

Fluorine

Reacts already at room

temperature

Reaction products volatile, no

fonnation of protective layer
In presence of oxygen

Main product WF6

Chlorine

Attack starts at 250-300 C

~400 C

WS2

above 2000K
Bromine

Iodine

Reaction starts at 450-500 C

Iodine vapor starts reacting

between 550-700C

In presence of 0 or H2O
Reaction product depends on
reaction conditions
In presence of 0 or H2O

o and H20

enhance the attack

WOF4
Main product at low and
medium temp.: WCIt;
WCI2 is dominant
WOC4, W02Cl2
WBr6, WBrs, WBr4
WOBr4, W02Br2
W6Il2

48

CHAPTER I

long-time corrosion tests. The corrosion is temperature-dependent and is further influenced

by the presence of oxygen traces, which lead to an enhanced dissolution.
Tungsten exhibits an excellent chemical stability against cesium vapor at elevated
temperature up to 1900C. This is of importance for the application of tungsten in ionic
propulsion systems in spacecrafts.
Group 11 (Cu, Ag, Au). These metals do not react with tungsten, and no binary
compounds are formed. The solubility of tungsten in the molten metals is negligible.
Pseudo-alloys of Cu-W and Ag-W, combining the beneficial properties of both
metals, play an important role as electrical contact materials and electrode materials for
spark erosion machines.
Group 2: Alkaline Earth Metals (Be, Mg, Ca, Sr, Ba). Between 750 and 1000C,
tungsten reacts with Beryllium forming WBe12 and WBe22' Above 1000C, only WBe22
is stable. Reaction of solid tungsten with Be vapor leads to diffusion zones, with WBe2
formed in the inner zone and both WBe12 and WBe22 in the outer zone.
Magnesium, calcium, strontium, and barium do not react with tungsten, and no binary
compounds are known. The solubility of tungsten in the respective molten metal is very
low.
BaO- resp. BaC03-doped porous tungsten play an important role as cathodes in highenergy electron tubes. The emitting element is metallic Ba, which forms upon heating
above 1000C.
Group 12 (Zn, Cd, Hg). These elements do not react with tungsten, and no binary
compounds are known.
Group 3 (Sc, Y, La) and Rare Earth Metals. These metals do not form intermediates
with tungsten. Only cerium is said to form an intermetallic compound of the composition
W2Ce with tungsten, but this is still doubtful.
At higher temperature, tungsten exhibits a certain solubility in the respective metal
melt. The solubility in RE metals at 2200 K decreases with increasing atomic radius from
Sc 4.5 at% to La<O.l at%.
Group 13 (AI, Ga, In, T/). Tungsten dissolves in molten aluminum. Three binary
compounds WAL 12 , WAL5, and WAL4 are known and form by peritectic reactions.
No intermediates are formed with gallium under atmospheric pressure. At 400800C and increased pressure (77kbar), W2Ga5 and W2Ga3 form. Tungsten exhibits a
slight solubility in molten Ga at 900-1100 0c.
Indium and thallium do not react with tungsten, and no compounds were found.
Solubility in the respective metal melt is negligible.
Group 4 (Ti, Zr, Hf) and Th. No intermetallic phases form with titanium. Tungsten is
only slightly soluble in cx-Ti, but "-'25 at % Ti dissolve in W at 740 C. ~-Ti forms a
continuous solid solution with W at temperatures above 1250C. Below 1250C, a
miscibility gap occurs.
Tungsten reacts with zirconium to form ZrW2. The solid solubility of Zr in W is
3.5-10.0 at% (~2000C), depending on the respective investigation. The maximum
solubility of W in cx-Zr is 0.5 at % and in ~-Zr 4 at %.
Also, hafnium forms an intermetallic compound HfW 2. Within the system, the
following solubilities were determined: 0.9 at% Win cx-Hf(l560C), 13.5 at% W in
~-Hf(1950C), 4 at% Hf in W (800C), and 9 at% Hf in W (2512 0c)

THE ELEMENT TUNGSTEN

49

No compounds are known with thorium. The solid solubilities ofW in Th and ofTh
in W are negligible. The solubility of W in liquid Th corresponds to 1.2 at % at 1695 C
and to 8.0 at % at 2509C.
Group 14 (Ge, Sn, Pb). Germanium reacts with tungsten at elevated temperature only
under increased pressure (77kbar, 1500-2500C). Two binary compounds are known:
W5Ge3 and WGe2' No solid solubility exists between the two metals and the solubility of
W in molten Ge is negligible.
Tin and lead do not react with tungsten and do not form compounds, while the
solubility of tungsten in the respective metal is negligible.
Group 5 (V, Nb, Ta). There exists unrestricted solid solubility in all three cases, and
no binary compounds are formed.
Group 15 (Sb, Bi). Antimony reacts sluggishly, and no compounds are formed. With
bismuth there is no corrosion of tungsten at 1000C, and no intermetallic compound is
formed.
Group 6 (Cr, Mo) and U. In the case of chromium and molybdenum unrestricted
solid solubility exists, while it is very restricted (:s 1 at %) for uranium. Below 1677 C, a
miscibility gap occurs in the system Cr-W No compounds are known with one of the three
metals.
Group 7 (Mn, Te, Re). No binary compounds are formed with manganese, and no
homogeneous product could be prepared. Technetium reacts with W, forming a solid
solution and a cr-phase (WTc3)' Tungsten-technetium alloys are of practical interest as
superconductors.
The solid solubility of W in rhenium is 11 at % at 1600C and 20 at % at 2825 C.
The solid solubility ofRe in W is 27.5 at% at 1600 C and 37 at% at 3000C. Besides the
solid solutions, two intermetallic phases exist: a tetragonal cr-phase with a broad
homogeneity range and a cubic phase with a narrow homogeneity range.
Solid solution alloys ofW and Re in the range of2-25 wt % Re play an important role
as rotating anodes in X-ray tubes for diagnostic purposes and as wires for thermocouples.
These materials are produced by powder metallurgical techniques. Inappropriate mixing of
the two powder components may result in a partial cr-phase formation, which is
undesirable due to embrittlement.
Groups 8-10 (Fe, Co, Ni, Ru, Rh, Pd, Os, Ir, Pt)
Iron group metals. Iron and tungsten form two stable intermediate phases: Fe? W6 (/l)
and FeW (b), as well as a metastable transitional phase, Fe2 W (A). The solubility of W
in fcc Fe is <1.5 at% but up to 14.3 at% in bcc Fe at the peritectic temperature of
1548 C. The maximum solubility of Fe in bcc W is ""2.6 at%. Above 2 wt% W, no y-Fe
(fcc) exists.
The element-rich solid solubility alloys are of importance in the so-called "heavy
metal" alloys. The tungsten phase formed during liquid-phase sintering is saturated with
iron (and nickel) and, vice versa, the nickel-iron binder phase with tungsten.
Cobalt and W form two intermetallic compounds: C0 3W and C07W6. The maximum
solubility ofW in fcc Co is 17.5 at% (at 1471 0c) and <0.5 at% in hcp Co. The maximum
solubility of Co in W is about 0.9 at %.
Three intermediate phases are known in the nickel-tungsten system: NiW2, NiW, and
N4W Up to 17.5 at% W dissolves in fcc Ni (at 1495C). The solubility ofW in Ni is
about 0.3 at% Wat 1495C.

50

CHAPTER 1

Platinum metals. Tungsten reacts with ruthenium, rhodium, palladium, osmium,

iridium, and platinum, forming extended areas of solid solutions. The maximum solubility
of tungsten in the respective metal decreases from 55 at% (Os) to 48 at% (Ru), 25 at%
(Pt), 22 at % (Pd) to 19 at % (Rh, Ir). The solubility of the metals in tungsten decreases
from Ru (23 at%) to Os (18.5 at%), Ir (10 at%), Rh (6 at%), Pd (5 at%) to Pt (<3.5 at%).
The following intermediate phases exist: W3Ru2 (0'), Rh3 W (E), OsW3 (0). Furthermore, two phases are known to form with platinum [a tetragonal y-phase (33-37 at% W)
and a hexagonal E-phase (45-52 at% W)] as well as with iridium [an E-phase (homogeneity range between 22 and 47 at% W) besides a o-phase (IrW3)].
Tungsten/palladium alloys between 26 and 44 at % W are superconducting.
1.3.4. Reactions with Compounds [1.4]
1.3.4.1. Reactions with Nonmetal Compounds
Water (H20). Water reacts with tungsten from room temperature up to 2000C. At
higher temperatures it dissociates, and oxygen is the reactant.
In principle, tungsten is oxidized by water. The reaction rate is determined by the
temperature and the pH20/pH2 ratio. As a reaction product, the presence ofH2 always has
to be considered. The reaction rate as well as the 0: W ratio of the oxide formed increases
with temperature and pH20/pH2 ratio. Water increases the volatility of the tungsten oxides
by the formation of the volatile oxide hydrate W02(OHh. For more details, see Sections
3.2 and 3.3.
Ammonia (NH3)' At room temperature, ammonia does not react with tungsten.
Ammonia containing moist air shows enhanced corrosion in comparison to ammonia
free. A quantity of 1-21lm tungsten powder in contact with a fast stream of ammonia
shows an uptake of up to 3.5 wt% N at 600 to 700C, increasing to 3.7 wt% N at 800C.
The reaction products are W2N and y-W3N4' both fcc.
At lower temperatures the reaction is restricted to the surface, and at higher
temperatures the rather unstable nitrides decompose. There is no attack by ammonia
from there up to the melting point. Due to the very low solubility of nitrogen in the
tungsten crystal lattice, volume reactions proceed mainly via grain boundary diffusion.
Tungsten is an efficient catalyst for both synthesis and decomposition of ammonia. At
elevated temperature, partial or complete dissociation ofNH3 occurs, and hydrogen, and at
high temperature also nitrogen, are released into the gas phase.
Nitrogen oxides (N20, NO, N02Y. They behave as oxidants and at elevated temperature form tungsten oxides, whereby nitrogen is released.
Hydrogen fluoride (HF). Tungsten powder reacts only slightly with HF at 300C,
and the reaction product is undefined. Up to 600 C, there is only little corrosion. The
oxidic passivation layer on tungsten converts into low-valent, nonvolatile fluoride which
protects the surface. At higher temperature, they disproportionate forming volatile
fluorides, and the corrosion rate is markedly increased.
Hydrogen chloride (HCI). Tungsten metal is stable against dry HCI up to 700C and
even higher.
Hydrogen bromide (HEr). The reaction ofHBr with tungsten plays an important role
in modem incandescent lamps (see Chapter 7).
Hydrogen sulfide (H2S), Tungsten powder reacts at 350-500C according to
W + 2H2 S ~ WS 2 + 2H2. At low temperature, the hexagonal modification forms.

THE ELEMENT TUNGSTEN

51

Sulfur dioxide (SOJJ. S02 behaves as an oxidant. At red heat tungsten is oxidized to
lower oxides, whereby elemental sulfur is formed: W + 2S02 -+ 1In (W0 3) +! 02 + S2'
Sulfur hexafluoride (SF6). At 500-700 C the ternary compound WSF4 is formed,
besides tungsten tetrafluoride: 3W + 2SF6 -+ 2WSF 4 + WF4 .
Disulphur dichloride (S2ClJJ. In presence of air, WOC14 and, under exclusion, WC1 6
are formed.
Hydrogen selenide (H2Se). At 300-400C, WSe is formed and, between 500 and
800C, the diselenide. In both cases hydrogen is released. Above 850C, decomposition
of the diselenide starts.
Boron nitride (BN). BN reacts with tungsten powder at 1500C and with dense
tungsten at 1600C to form W2B and WB.
Boron trichloride (BCI3). It boronizes tungsten at 900 to 1000C.
Carbon monoxide (CO). Between 80 and 200C, tungsten hexacarbonyl forms.
From 1000C upward, bulk tungsten is carburized: W + 2CO -+ WC + CO2 resp.
2W + 2CO -+ W2C + CO 2, Nanocrystalline powders start to carburize already at about
550C.
Carbon dioxide (COJJ. At elevated temperature, CO 2 acts as an oxidizing agent and,
depending on temperature and partial pressure, lower or higher tungsten oxides will be
formed: xW + yC02 -+ WxOy + ycO.
Hydrogen cyanide (HCN). From 730 to 2200C tungsten reacts as follows:
W + HCN -+ WC + ! H2 + ! N2. Above 2200C, W 2C and, at 2600C, in addition
graphite are formed.
Carbon tetrafluoride (CF4), At elevated temperature, WF6 is formed.
Carbon tetrachloride (CCI4). At 600C, tungsten powder reacts to WC14 and WCk
Carbon disulfide (DSJJ. Tungsten reacts at elevated temperature to WS 2.
Boron carbide (B4C). In the contact zone between the two components,
tungsten boride layers can be detected after heat treatment between 1100 and 1600C.
Depending on the reaction conditions, (X-WB or W2B is formed. Additions to NaF as
activator.
Silicon dioxide (SiOJJ. If the protecting atmosphere is free of oxidizing species,
tungsten is very stable against liquid Si02 up to 2500C. Above 2000C, slight oxidation
starts: 2W + 3Si02 -+ 2W03 + 3Si.
Trisilicon tetranitride (Si3N4)' In vacuum at 1300C, tungsten silicides are formed.
Silicon tetrachloride (SiCl4). Tungsten reacts at 1000-1200C in hydrogen atmosphere according to: W + 2SiCl4 + 4H2 -+ WSi2 + 8HC!. The reaction product forms a
protective layer on tungsten.
Silicon tetrabromide (SiBr 4). Reacts analogously.
Silicon carbide (SiC). Between 1100 and 1900C, WC and tungsten silicides are
formed.
Hydrogen phosphide (PH3). Tungsten reacts at 850C to WP.
Phosphoric acid concentrated (H3P04)' Tungsten shows a slight attack depending on
temperature and time.

1.3.4.2. Reactions with Metal Compounds.

Alkali hydroxide melts. These compounds do not attack as long as any oxidizing
material (including air) is excluded. Any oxidic scale on the tungsten metal surface will be

52

CHAPTER 1

dissolved. The presence of oxidizing agents like nitrates, nitrites, chlorates, etc. leads to a
vigorous attack.
For example, for technical purposes, the dissolution of tungsten scrap material
mixtures of NaOH with sodium nitrate or sodium nitrite have been widely used. The
dissolution occurs due to the following reactions:
W + 6NaN02 ~ Na2 W04 + 2Na20 + 6NO
W + 6NaN03 ~ Na2 W04 + 2Na20 + 6N02
Because NO as well as N02 react with the excessive NaOH to form nitrite, less nitrogen
oxides are evolved than expected.
Sodium oxide (Na 20). Above 500C, it reacts with tungsten according to:
W + 4Na20 ~ Na2 W04 + 6Na.
Sodium peroxide (Na 20j. Gives a violent raction.
Sodium nitrate (NaN0 3). Gives a violent raction.
Sodium nitrite (NaNoj. Gives a violent reaction.
Sodium chloride (NaC/). Gives a slight attack.
Sodium sulfide (Na 2S). There is no reaction up to 1500C.
Sodium sulfate (Na2S04). Reacts from 320C upwards: 3W + 2Na2S04 ~
2Na2 W04 + WS 2
Sodium borides. Form borides.
Sodium carbonate (Na2C03). From 950C onward tungstate is slowly formed:
W + 3Na2C03 ~ Na2 W04 + 2Na20 + 3 CO.
Potassium nitrate (KN03). Reacts like NaN0 3 .
Potassium perchlorate (KCl03). Gives a violent reaction.
Beryllium oxide (BeO). Oxidizes tungsten above 1800C.
Magnesium oxide (MgO). Reacts with tungsten from 2000C onward. The reaction
products are Mg, MgW0 3 , MgW04 , WO b and W0 3 .
Calcium oxide (CaO). Reacts with tungsten at ::: 1900C under formation ofCaW03
and CaW04 .
Stronium oxide (SrO). Reacts with tungsten at ::: 850C under formation of Sr3W0 6
and SrW04 . Metallic Sr forms according to the equation: 2 Sr3 W0 6 + W ~
3SrW04 + 3Sr.
Barium oxide (BaO). At ::: 800C Ba3 W0 6 is formed: 6BaO + W ~
Ba3 W0 6 + 3Ba. Above 1000C, the reaction ceases, because the bronze compound
dissociates thermally into W0 3 and Ba, which evaporates. Above 1400C, the tungsten
surface is clean.
Barium carbonate (BaC03). Reacts with tungsten in the temperature range between
600 and 800C according to: 3BaC03 + W ~ Ba3 W0 6 + 3CO.
Aluminum oxide (A 120 3). Compatibility is given up to 1900C. No or very slow
reaction occurs with molten alumina according to: 3W + Al20 3 ~ 3W02 + 4AI
resp.W + Al20 3 ~ W0 3 + 2Al. The reaction products are in both cases volatile. Seamless tungsten crucibles are used to grow large sapphire single crystals.
Zirconium oxide (ZrOj. Is compatible with tungsten up to 1600C in vacuum and
up to 1400-1500C in protective atmospheres. Starting temperature indications in the
literature are quite different due to differences in zirconia stabilization (2000-2390C).

THE ELEMENT TUNGSTEN

53

Thorium oxide (ThOJJ. Is compatible up to 2200 C in vacuum and up to 2000C in

protective atmosphere. Indications about the stability at higher temperatures are quite
different. The reaction: 2W + a-Th0 2 --* 2WOx + a-Th takes place at a temperature
of >2000K [1.37].
Alkali chromates. Binary powder mixtures of tungsten and potassium dichromate are
easy to ignite (650-660C) and bum over a wide range of compositioin (20-90 % W).
There is only little change In weight (gasless pyrotechnical system):
W + K2Cr20 7 --* K2 W04 + Cr203'
Uranium dioxide (UOJJ. No reaction with tungsten occurs up to the melting point of
uranium dioxide (approx. 2760C). The molten oxide penetrates tungsten along grain
boundaries.

1.3.5. Reactions with Aqueous Solutions

Many aqueous solutions of chemical compounds only attack the surface of tungsten
metal slightly. This may be compared with the attack of pure water, especially if diluted
solutions are considered. Typical examples of this group of compounds are nitric acid,
hydrochlotic acid, sulfuric acid, sodium and potassium hydroxide, sodium chloride or
ammonium chloride (see Table 1.27), and many others. A completely different behavior is
shown with compounds containing a complexing anion (forming very stable coordination
compounds with tungsten). A prominent member of this group is hydrofluoric acid (see
Table 1.27). The attack is about one magnitude higher than in the case of hydrochloric
acid. Solutions of compounds which combine complexing and oxidizing ability are very
useful dissolution reagents, as, for example, hydrogen peroxide or mixtures of hydrofluoric
and nitric acid.

TABLE 1.27. Corrosion of Tungsten by Diverse Aqueous Solutions

HN03
10%

HF
3%

Boiling reflux
nonaerated
highly agitated

0.20

75C reflux
slightly aerated
slightly agitated

a,b

HCI
10%

H2SO4
10%

KOH
10%

2.80

0.13

0.15

0.41

0.38

0.03

0.25

2.00

0.08

0.65

5.20

0.80

0.72

20C static
immersed
nonnally aerated

0.002

0.032

0.005

0.0055

0.55

0.0041

0.0066

20C dipping
highly aerated
highly agitated

0.02

0.34

0.008

2.14

0.020

0.022

a
b

NaCI
3%

N~CI

Gmelin Handbook of Inorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. A7, Springer-Verlag, Heidelberg (1987).
Corrosion rate in g . m- 2 . d- I .

CHAPTER 1

54

Hydrogen peroxide (H20J. Tungsten dissolves without inhibition. The dissolution

rate is proportional to the peroxide concentration. The dissolution rate for bulk tungsten
corresponds to 1.6 mg W /1 min at 20C and 14 g H2 0 2 /1. Under the same conditions,
tungsten powder dissolves at a rate of 300 mg/l min. Increased temperature enhances the
dissolution rate up to 60C. Higher temperatures slow down the reaction because of
peroxide decomposition. Addition of nitric acid, hydrofluoric acid, sulfuric acid, phosphoric acid, acetic acid, or sodium hydroxide does not influence the reaction. The solution
shows a bright yellow color. The reaction is said to proceed as follows:
W + 2H20 2 --+ W0 2 + 2H20
W0 2 + H20 2 --+ H2 W04

3H2W04 + 2H20 2 --+ H2 W30

12

+ 4H20

or
2W02 + 6H20 2 --+ H2 W20))

+ 5H20

3H2W20 11 --+ 2H2 W3 0 12

+ H20

NaOH solutions containing 5 vol % of hydrogen peroxide are proposed as etchants.

1.3.5.1. Acids
Nitric acid (HN03). At room temperature, tungsten is rather resistant against all
concentrations and also against hot concentrated acid. Hot and diluted nitric acid attacks
slowly (see Table 1.27).
Hydrofluoric acid (HF). The dissolution rate is about one magnitude higher than that
for nitric acid and the attack is faster in hot solution. Any addition of oxidizing agents
increases the dissolution rate considerably (see Table 1.27).
Hydrochloric acid (HCI). Tungsten shows rather good resistivity against HCl at all
concentrations and temperatures. The corrosion rate is somewhat smaller than for nitric
acid (see Table 1.27).
Hydrobromic acid (HEr). Shows no significant effect.
Hydroiodic acid (HJ). Tungsten is dissolved slowly in concentrated solutions.
Periodic acid (HJ04) Does not react with tungsten.
Sulfuric acid (H2S04). The resistivity of tungsten is comparable to that for hydrochloric acid over a wide range of concentration and temperature. Increased attack occurs
by fuming sulfuric acid (see Table 1.27).
Phosphoric acid (H3P04). Tungsten is highly resistant to diluted phosphoric acid as
long as air is excluded. It dissolves easily in concentrated acid.
1.3.5.2. Important Acid Mixtures to Dissolve Tungsten
40 vol% conc. HN03 and 60 vol% conc. HF. Already at room temperature, a
vigorous attack of tungsten takes place. The metal surface is oxidized by N02 and the
tungsten oxide dissolves in HF, whereby tungsten fluoride or oxofluoride ions are formed
which show a high solubility.
HNO"HCI Mixture (aqua regia). Tungsten is attacked only slowly in cold, but
rapidly in hot, solutions. Due to the fact that there is no complexing agent like fluoride as
in the former acid mixture, an oxide layer is formed which slows down the reaction.
HF-H2SO"HN03 mixtures. Tungsten is dissolved quickly at all concentrations.

55

THE ELEMENT TUNGSTEN

1.3.5.3. Aklaline Solutions. In general, tungsten shows less resistivity to alkaline

than to acidic solutions, especially in the presence of air. Corrosion rate increases with
increasing OH- concentration. Oxidizing agents including oxygen or air raise the
dissolution rate.
Sodium hydroxide (NaOH). In the absence of oxygen, dilute solutions do not attack
tungsten at temperatures up to 10 0c. Concentrated solutions show a slight attack.
Potassium hydroxide (KOH). Reacts analogously to NaOH (see Table 1.27).
Ammonium hydroxide (NH40H). Is more corrosive than KOH and NaOH.
1.3.5.4. Salts. Sodium chloride (NaCl). Slightly corrodes (see Table 1.27).
Sodium hypochlorite (NaCIO). Dissolves tungsten slowly according to: W + H2 0 +
3CIO- --+ WO~- + 3Cl- + 2H+.
Ammonium chloride (NH4Cl). Slightly corrodes (see Table 1.27).
Iron trichloride (FeCI3). Is strongly corrosive for tungsten.
Potassium hexacyanoforrate(III) (K3{Fe(CN)J)' Alkaline solutions are in use as
etchants in metallography and for meal finishing (Murakami etch):
W + 6K3[Fe(CN)6] + 8KOH --+ K2 W04 + 6K4 [Fe(CN)6] + 4H2 0
The K3[Fe(CN)6]:KOH molar ratio for polishing action is ?: 2.5, and for etching (revealing
details in microstructure) <2.5.
Cu2 + salts. Show an etching effect in ammoniacal solution preferentially on W(100)
faces (Millner-Sass solution).

1.3.6. Miscellaneous
Ceramics [1.56]. Compatibility is given for the following ceramic materials up to the
indicated temperature (valid in vacuum, in protective atmosphere 100-200C lower):
Alumina (Ah03)
Beryllia (BeO)
Magnesia (MgO)
Thoria (Th0 2)
Zirconia (zr0 2)
Sillimanit
Magnesit (brick)
Fire Brick
3BeO2CaO
3BeO2MgO
Ah03 . 4BeO 4MgO
Ah03 . 4BeO . MgO

"-'1900C
"-'2000C
"-'1500C
"-'2200C
"-'1600C
"-'1700C
"-'1600C
"-'1200C
1400-1500C
1800-1900C
1600-1700C
"-'1900C

Glass melts [1.56,1.101]. Tungsten is very stable against glass melts up to 1400C.
Above 1400C, a slight attack occurs. The wetting behavior of glass melts besides glass
quality and temperature depends on the atmosphere. Oxidizing conditions cause yellow
(W0 3) or yellow-green colored glasses or opal glasses due to crystallization.
Due to its excellent corrosion resistance, tungsten is used as electrode material for
melting kaolin.
Rock and magmatic melts {l.lOl]. Tungsten is severely oxidized and dissolved by

56

CHAPTER I

those melts: The solubility at ISOO K in Bandelier tuff melts is 0.4-1.6 wt % and is similar
for Janez basalt at 1900 K and 3.S wt % at 1900 K for other basalts. The solubility in
basalts increases with increasing Fe3+ /Fe 2 + ratio. Tholeitic basalt lava from 1971 Kilauea
eruption oxidizes tungsten strongly in an H 2 0, CO 2 , CO, S03 atmosphere.

1.3.7. Reactions with Organic Compounds

The numerous reactions which yield in the formation of organometallic compounds
are not considered here (see Chapter 4).
In general, organic compounds like hydrocarbons react with tungsten at elevated
temperature via dissociation of the organic molecule and subsequent carbide formation.
Methane is the most prominent compound in this field and should act as a
representative for all the others here. It also plays a very important role in the technical
carburization process of tungsten as an intermediate which enables the gaseous transport of
carbon.
Methane (CH4). Bulk tungsten reacts with methane at temperatures above 1000 K
forming W 2 C and WC according to: 2W + CH 4 ~ W2 C + 2H2 or W + CH4 ~
WC+2H2
The thermochemical data for the reactions in the W-CH4 -H 2 system are:
2W +CH4 ~ W2C +2H2
I1Go = 13470 - 2S.7T
log P(CH4 = 2 . log P(H2) - 89.S + 2490/T
CH4 ~ C (in W) + 2H2
I1Go = 34270 - 2S.8T
log p(CH 4) = 2 . log P(H2) + log c(C) - IO.S2 + 9460/T
the units of I1GO are cal .. mol- 1 K - 1, P is in Torr, and c in at %.
The equilibrium pressure of the methane dissociation is only slightly higher than that
of the carbide formation. In order to suppress excessive free-carbon formation, the CH4
pressure has to be maintained above the minimum pressure for the carburization, but below
that resulting in free-carbon formation.

REFERENCES FOR CHAPTER 1

1.1.
1.2.

1.3.
1.4.
1.5.
1.6.
1.7.
1.8.

Gmelin Handbook ofInorganic Chemistry, 8 ed., Syst. No. 54, Tungsten, Suppl. Vol. A2, Springer-Verlag,
Heidelberg (1987).
1. Emsley, Die Elemente, Walter de Gruyter, Berlin, New York (1994).
R. D. Shannon, Acta Crystallogr. A32 (1976), 751-767.
1. C. Bailar Jr, H. 1. Emeleus, R. Nyholm, and A. F. Trotman-Dickenson, Eds. Comprehensive Inorganic
Chemistry, Vol. 3, p. 746, Pergamon Press, Oxford (1973).
G. V Samsonov, Handbook of the Physiochemical Properties of the Elements, Plenum Press, New York
(1968).
A. 1. C. Wilson, ed., International Tables for Crystallography, Vol. C, Kluwer Academic Publishers,
Dordrecht/Boston/London (1995).
R. Jenkins, X-Ray Spectrom. 2 (1973), 207.
R. Tertian and F. Claisse, Principles of Quantitative X-ray Fluorescence Anlaysis, Heyden, London
(1982).

THE ELEMENT TUNGSTEN

1.9.
1.10.
1.11.
1.12.
1.13.
1.14.
1.15.
1.16.
1.17.
1.18.
1.19.
1.20.
1.21.
1.22.
1.23.
1.24.
1.25.
1.26.
1.27.
1.28.
1.29.
1.30.
1.31.
1.32.
1.33.
1.34.
1.35.
1.36.
1.37.
1.38.
1.39.
1.40.
1.41.
1.42.
1.43.
1.44.
1.45.
1.46.
1.47.
1.48.
1.49.
1.50.
1.51.
1.52.
1.53.
1.54.
1.55.

57

L. E. Davis, N. C. MacDonald, P. W Palmberg, G. E. Riach, and R. E. Weber, Handbook of Auger

Electron Spectroscopy, Physical Electronics Industries, Inc., Eden Prairie, Minnesota 55343 (1976).
I. Barin, Thermochemical Data of Pure Substances, Vol. II, VCH Veriagges.m.b.H., Weinheim (1995).
P. Hohenberg and W Kohn, Phys. Rev. B 136 (1964), 864.
D. Pettifor, Bonding and Structure of Molecules and Solids, Clarendon Press, Oxford (1995).
W A. Harrison, Electronic Structure and the Properties of Solids, WHo Freeman and Company, San
Francisco (1980).
A. Zunger and M. Cohen, Phys. Rev. B 19 (1979), 568-583.
D. M. Bylander and L. Kleinman, Phys. Rev. B 27 (1982), 3152-3159.
D. M. Bylander and L. Kleinman, Phys. Rev. B 29 (1984), 1534-1539.
H. 1. F. Jansen and A. 1. Freeman, Phys. Rev. B 30 (1984), 561-569.
H. L. Skriver, Phys. Rev. B 31 (1985), 1909-1923.
1. W Davenport, M. Weinert, and R. E. Watson, Phys. Rev. B 32 (1985),4876-4882.
S. H. Wei, H. Krakauer, and M. Weinert, Phys. Rev. B 32 (1985), 7792-7797.
L. F. Mattheiss and D. R. Hamann, Phys. Rev. B 33 (1986), 823-840.
C.T. Chan, D. Vanderbilt, and S. G. Louie, Phys. Rev. B 33 (1986), 7941-7946.
M. Korling and 1. Haglund, Phys. Rev. B 45 (1992), 13293-13297.
V. Ozolins and M. Koriing, Phys. Rev. B 48 (1993), 18304-18307.
A. F. Guillermet, V. Ozolins, G. Grimvall, and M. Korling, Phys. Rev. B 51 (1995), 10364-10373.
G. Grimvall, M. Tiessen, and A. F. Guillermet, Phys. Rev. B 36 (1987), 7817-7826.
A. F. Guillermet and G. Grimvall, Phys. Rev. B 44 (1991), 4332-4340.
P. SOderiind, R. Ahuja, O. Eriksson, 8., Johansson, and 1. M. Wills, Phys. Rev. B 49 (1994), 9365-9371.
P. Soderlind and B. Johansson, Thermochim. Acta 218 (1993), 145-153.
R. Lowrie and A. M. Gonas, J Appl. Phys. 36 (1965), 4505-4509.
R. Lowrie and A. M. Gonas, J. App/. Phys. 38 (1967), 2189-2192.
R. Nyholm, A. Berndtsson, and N. Mfutensson, J. Phys. C. 13 (1980), Ll091-Ll094.
Z. Hussain, C. S. Fadley, S. Kono, and L. F. Wagner, Phys. Rev. B 22 (1980), 3750-3765.
G. Hagg and N. Schonberg, Acta Crystallogr. 7 (1954), 351.
Gmelin Handbook ofInorganic Chemistry, 8 ed., Syst. No. 54, Tungsten, Suppl. Vol. A3, Springer-Verlag,
Heidelberg (1989).
A. Bartl, Doctoral Thesis, TV Vienna, Austria (1997).
G. D. Rieck, Tungsten and Its Compounds, Pergamon Press, Oxford (1967).
Gmelin Handbook of Inorganic Chemistry, 8th ed., No. 54, Tungsten, Suppl. Vol. A4, Springer-Verlag,
Heidelberg (1993).
S. W H. Yih and C. T. Wang, Tungsten, Plenum Press, New York, London (1979).
Metals Handbook, lOth ed., Vol. 2, pp. 1169-1172, ASM International, Metals Park, Ohio (1995).
1. P. Wittenauer, T. G. Nieh, and 1. Wadsworth, Advanced Materials & Processes 9 (1992), 28-37.
E. Pink and R. Eck, "Refractory Metals and their Alloys," in: Materials Science and Technology, Vol. 9
(R. W Cahn, P. Haasen, and E. 1. Kramer, eds.), VCH, Weinheim (1992).
1. P. Morniroli, "Low Temperature Embrittlement of Undoped and Doped Tungsten," in: The Metallurgy
of Doped/Non-sag Tungsten (E. Pink and 1. Barha, eds.), Elsevier, London (1989).
U. K. Vashi, R. W Armstrong, and G. E. Zima, Metall. Trans. 1 (1970), 1769-1771.
G. S. Pisarenko, V. A. Borisenko, and Y. A. Kashtalyan,Sov. Powder Metall. Met. Ceram. 5(1962),371-374.
H. Mughrabi, "Microstructure and Mechanical Properties," in: Materials Science and Technology (R. W
Cahn, P. Hassen, and E. 1. Kramer, eds.), pp. 1-17, VCH, Weinheim (1993).
H. Neuhauser and C. Schwink, "Solid Solution Strengthening," in: Materials Science and Technology (R.
W Cahn, P. Haasen, and E. 1. Kramer, eds.), pp. 191-250, VCH, Weinheim (1993).
H. Schultz, Z. Metallkd. 78 (1987), 469-475.
B. Sestak and A. Seeger, Z. Metallkd. 69 (1978), 195-202,355-363,425-432.
A. S. Argon and S. R. Maloof, Acta Metall. 14 (1966), 1449-1462.
1. D. Meakin, Can. J. Phys. 45 (1967), 1121-1134.
B. C. Allen, D. 1. Maykuth, and R. I. Jaffee, J. Inst. Met. 90 (1961), 120-128.
O. Boser, J. Less-Common Met. 23 (1971), 427-435.
E. Lugscheider, R. Eck, and P. Ettmayer, Radex Rundsch. 1/2 (1983), 52-84.
G. L. Davis, Metallurgia 58 (1958), 177-184.

58
1.56.
1.57.
1.58.
1.59.
1.60.
1.61.
1.62.
1.63.
1.64.

1.65.
1.66.

1.67.
1.68.
1.69.
1.70.
1.71.
I. 72.
1.73.
1.74.
1.75.
1.76.
1.77.
1.78.
1.79.
1.80.
1.81.
1.82.
1.83.
1.84.
1.85.
1.86.
1.87.
1.88.
1.89.
1.90.
1.91.
1.92.

1.93.
1.94.

1.95.
1.96.
1.97.
1.98.

CHAPTER I
Wolfram, company brochure, Metallwerk Plansee, Reutte, Austria (1980).
A. G. Starliper and H. Kenworthy, Chem. Abstr. 69,21516 (1968).
A. Kelly, W. R. Tyson, and A. H. Cottrell, Phi/os. Mag. [8] 15 (1967), 567-586.
N. H. McMillan, J. Mater. Sci. 7 (1972), 239-254.
P. L. Raffo, J. Less-Common Met. 17 (1969), 133-149.
D. M. Moon and R. Stickler, Philos. Mag. 24 (1971), 1087-1094.
D. M. Moon and R. Stickler, High Temp. High Pressures 3 (1971), 503-518.
S. L. Robinson and O. D. Sherby, Acta Metall. 17 (1969), 109-125.
E. Pink, in: Proceedings of 10th Plansee Seminar (H. M. Ortner, ed.), Vol. I, pp. 221-235, Metallwerk
Plansee, Reutte, Austria (1981).
M. F. Ashby, Acta Metall. 20 (1972), 887-893.
E. Pink and I. Gaal, "Mechanical Properties and Deformation of Non-Sag Tungsten Wires," in: The
Metallurgy ofDoped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 209-233, Elsevier, London,
(1989).
A. Cezairliyan, High Temp. Sci. 4 (1972), 248-252.
G. Wouch, E. L. Gray, R. T. Frost, and A. E. Lord, High Temp. Sci. 10 (1978), 241-259.
E. Rudy, St. Windisch, and 1. R. Hoffman, AFML-TR-65-2, Part I, Vol. VI, Air Force Materials Laboratory,
Res. And Techn. Div., Dayton, Ohio (1966).
P. Gustafson, Int. J. Thermophys. 6 (1985), 395.
D. A. Young, Phase Diagrams of the Elements, University of California Press, Berkeley, p. 174 (1991).
H. A. Pastor, "Microstructure-Independent Physical Properties of Tungsten," in: The Metallurgy of
Doped/Non-Sag Tungsten, (E. Pink and L. Bartha, Eds.), Elsevier, London (1989).
B. Kiihl, "Lichterzeugung," in Ullmanns Encyclopaedie der technischen Chemie, 4th ed., 16 (1978), 227.
D. R. Stull and H. Prophet, JANAF Thermochemical Tables, 2nd ed., NSRDS-NBS37, NBS, Washington
D.C. (June 1971).
W. Fucke and U. Seydel, High Temp. High Pressures 12 (1980), 419-432.
1. A. Mullendore, "Tungsten and Tungsten Alloys," in Kirk-Otrner: Encyclopedia of Chemical Technology, Vol. 23, 3rd ed., Wiley, New York (1983).
G. Grirnvall, M. Thiessen, and A. F. Guillermet, Phys. Rev. B 36 (1987), 7816-7826.
A. F. Guillermet and G. Grirnvall, Phys. Rev. B 44 (1991), 4332-4340.
R. L. Andelin, 1. D., Knight, and M. Kahn, Trans. AIME 233 (1965), 19-24.
C. Y. Ho, R. W. Powell, and P. E. Liley, J. Phys. Chem. Ref Data 3, Supp!. I (1974).
C. Y. Ho, R. W. Powell, and P. E. Liley, J. Phys. Chem. Ref Data 1 (1972),279--421.
Y. S. Touloukian, R. W. Powell, C. Y. Ho, and M. C. Nicoladu, in: Thermophysical Properties of Matter,
Vol. 10, Plenum Press, New York (1973).
1. G. Hust, High Temp. High Pressures 8 (1976), 377-390.
P. D. Desai, T. K. Chu, H. M. James, and C. Y. Ho, J. Phys. Chem. Ref Data 13 (1984), 1069-1096.
L. N. Latyev, V. Ya. Chekhovskoi, and E. N. Shestakov, High Temp. High Pressures 4 (1972), 679-686.
N. Aksyutov and A. K. Pavlyukov, Zhr. Prikl, Spektrosk, 32 (1980), 813-820.
N. Aksyutov, Zhr. Prikl. Spektrosk. 26 (1977), 914-918.
P. Hiernaut, R. Beukers, M. Hoch, and M. Matsui, High Temp. High Pressures 18 (1986), 627-633.
L. W. Swanson and L. C. Crouser, Physl Rev. 163 (1967), 622-641.
G. E. R. Schulze, Metallphysik, Akademieverlag Berlin GmbH (1967).
C. Zwikker, Physica 5 (1925), 249.
R. L. Sproull, Modern Physics: The Quantum Physics of the Atom, Solid and Nucleus, 2nd ed., pp. 435444, Wiley, London (1963).
Gmelin Handbook of Inorganic Chemistry, 8th ed., Syst. No. 44, Thorium, Vo!' A3, pp. 77-84, SpringerVerlag, Berlin (1988).
Gmelinls Handbuch der anorganischen Chemie, Thorium and Isotope, Systemnummer 44, Verlag Chemie
GmbH, Weinheim (1955).
R. N. Wall, D. L. Jacobson, and D. R. Bosch, Metall. Trans. 24A (1993), 951-958.
1. M. Heras, 1. Borrajo, and F. E. Mola, An. Soc. Cient. Argent. 191 (1971), 179-190.
Metals Handbook, 9th ed., Vol. 17, p. 235, ASM International, Metals Park, Ohio.
Gmelin Handbook of Inorganic Chemistry, 8th ed. Syst. No. 54, Tungsten, Suppl. Vol. A5b, SpringerVerlag, Heidelberg (1993).

THE ELEMENT TUNGSTEN

59

Gmelin Handbook of Inorganic Chemistry, 8th ed. Syst. No. 54, Tungsten, Suppl. Vol. A6a, SpringerVerlag, Heidelberg (1991).
1.100. Gmelin Handbook of Inorganic Chemistry, 8th ed. Syst. No. 54, Tungsten, Suppl. Vol. A6b, SpringerVerlag, Heidelberg (1998).
1.101. S. V. Nagender and P. Rama Rao, eds., Phase Diagrams of Binary Tungsten Alloys, Indian Institute of
Metals, Calcutta (1991).
1.1 02. see folio 51
1.99.

2
Tungsten History
From Genesis to the 20th Century Products

2.1. THE FORMATION OF TUNGSTEN ATOMS

[2.1~2.5]

Articles concerned with tungsten history usually start by describing the discovery of
tungsten ores, compounds, and the element. In reality, however, the history of tungsten
began with the formation of the tungsten atoms, a very long time ago.
The explanation-how elements have been formed and are still formed today-will
be among the outstanding achievements in the history of natural science of the 20th
century [2.1]. Therefore, it seems worthwhile to discuss the formation of heavier elements,
like tungsten.
The rough schedule in Table 2.1 should give an idea as to the period of the universal
evolution in which the element tungsten present in the solar system was formed. It was
within the long period of 10 billion years, between the occurrence of the first stars in our
galaxy and the formation of the solar system.
The formation of elements is a long and complicated path of diverse nuclear
reactions, which took place during the evolution of massive stars (>8 times the solar
mass). The "raw material" is always hydrogen, which was formed in addition to the
smaller amounts of deuterium and helium shortly after the Big Bang. During the lifetime
TABLE 2.1. Important Events in the Universe's Evolution (Time Table)
Years after the
Big Bang
(x 109 )
0
0.0003
0.5
10.4

Years ago
(x 109)

Event

15
14.997
14.5
14
4.6

Big Bang
Fonnation of H and He atoms
Fonnation of galaxies
{Fonnation of tungsten
Early stars in Milky Way
atoms within this period
Fonnation of solar systems

61

62

CHAPTER 2

of such a star, its central region passes through subsequent cycles of nuclear fusion
processes. In the first and longest period, called "hydrogen burning," helium is formed by
nuclear fusion of hydrogen. When the "fuel" (hydrogen) is consumed, the core contracts
further, the temperature and pressure consequently rise, and the next fusion reaction
(helium burning) becomes possible. The foregoing burning process moves outward in a
radial direction, forming a shell around the core. The different cycles, following one after
the other, are summarized in Table 2.2.
This cycling can be exhibited schematically as follows:
Nuclear Burning

Core Heating

Fuel Exhaustion

Core Contraction

and causes an onion-skin structure of the interior of the star.

At the end of the silicon burning, the core consists of '" 1.5 solar masses iron and
nickel, while, in the covering shells the foregoing burning processes are still in progress.
The life of the star comes to an end with a super nova explosion. Due to the fact that no
more fuel is available in the core, it starts to collapse. The temperature will rise over
10 1o K, and a gas of free nucleons will be formed. The increased pressure finally balances
the collapse, and a neutron star of about 1 solar mass is born. Consequently, a shock wave
runs through the outer parts of the star combined with high pressure and high temperatures
originating in different nuclear reactions. The chemical composition will be changed by
this reaction to a certain degree. From the original material of the star, only about 1.5 solar
masses remain as a neutron star. The rest is distributed into interstellar space by the blast.
All elements formed within the burning processes by nuclear fusion have even atomic
numbers, because in all cases ex-particles are involved in these reactions. The nuclei of iron
and nickel generated in the last burning process are of optimal composition, which means
that the bonding strength of the nucleons is of maximal value. This is why elements having
higher atomic numbers than nickel, or cobalt-and this includes tungsten-cannot be
formed by nuclear fusion processes, because this would consume energy. All the fusion
TABLE 2.2. Nuclear Fusion Reactions
in the Course of the Evolution of a
Massive Star
Name

Reaction

H burning
He burning
C burning
N burning
o burning
Si burning

H-+He
He-+C,O
C-+Mg, Ne
N-+O, Mg
O-+Si, S
Si-+Fe, Ni

log T
7
8.3
8.7
8.9
9.5
9.9

TUNGSTEN HISTORY

63

reactions described above are exothennic due to a mass defect. Higher elements as well as
elements. having uneven mass numbers can only be fonned by neutron and/or proton
absorption reactions.
Neutron absorption processes occur at different times and places in the course of the
evolution of massive stars. The S-process (slow neutron capture) occurs in the He burning
region (state of a red giant). The R-process (rapid neutron capture) occurs during the super
nova explosion, either within a short distance of the fonning neutron star or in the shell
where He burning took place prior to the blast when the shock wave hits this area.
S-process. Neutrons freed as a result of the nuclear fusion and decay reactions
provide a neutron density of 105 cm - 3. Higher elements are fonned by neutron absorption
and subsequent ~-decay. Hence S-elements at about the solar abundance are generated.
The S-process, depending on the stars mass, lasts between 100 and 100,000 years.
R-process. This occurs during a super nova explosion and lasts only 0.1-10 seconds.
Not far from the neutron star at 5 x 109 K, the neutron density is 1025 cm -3 so enabling
the absorption of neutrons by atoms to a high degree. Again higher elements will fonn by
subsequent multiple ~-decay. If the pressure wave reaches the shell where He burning took
place prior to the blast, a neutron density of 1020 cm - 3 will be generated and the R-process
becomes possible.
Proton absorption (P-process). Besides neutron absorption, also protons can be
captured; these are generated in a super nova, leading to atoms of higher mass and atomic
number. The process is possible in a time interval between 10 and 100 seconds.
It seems likely that during the early life of galaxies, the fonnation of massive stars and
consequently super novae events took place much more often than today. As a result, larger
amounts of their "ash" were distributed to interstellar space. During the above-mentioned
period of 10 billion years, the ratio of hydrogen to heavier elements in the interstellar
medium was continuously decreased, due to star evolutions and explosions.

2.2. HOW TUNGSTEN ATOMS CAME ON EARTH [2.6-2.9]

4.6 billion years ago the solar system was formed by condensation, contraction, and
subsequent collapse of an interstellar cloud. More than 99% of the matter of the solar system
is concentrated in the sun; the rest is contained in planets, meteorites,and comets. This
event stopped the continuous evolutionary change in element concentrations of the system.
Besides hydrogen and helium, the cloud contained the elements fonned in exploded stars, as
described in Section 2.1. One of the heavier elements in this "ash" is tungsten. We do not
know to how many star cycles this material, or part of it, was subjected.
The material remaining outside the sun has undergone one or more of the following
processes: oxidation, accretion, melting, segregation, and fractional crystallization. Planetesimals were fonned out of dust and ice particles; consequently, planets were built up by
collisions of planetesimals. Regarding the earth already at a stage of 85% accretion, a
separation of the iron-nickel core by segregation was tenninated. During the further
evolution of the planet, volcanism leads to chemical reactions and differentiations, which
are still going on today.
Knowledge about the evolution of the solar system is mainly based on meteorite
investigations. With the exception of material brought to earth by spacecrafts from the

64

CHAPTER 2

moon, meteorites represent the only extraterrestrial substances which can be analyzed in
laboratories by man. Most meteorites have been formed 4.6 billion years ago like the earth,
but have not changed since. Due to the fact that the earth is still an active planet, rocks are
continuously changing; older ones are destroyed by melting or corrosion and new ones are
formed by igneous processes. The oldest rocks on earth have an age of 4 billion years
while the majority are only one billion years old.
Many meteorites have never been subjected to processes of planetary differentiation.
These undifferentiated meteorites come from planetesimals that were never molten and
resemble the composition of the solar nebula at the time and place of their formation. They
exhibit approximately solar system composition, are called chondritic meteorites, and
represent one big group of meteorites. The second group consists of the differentiated
meteorites which represent pieces of partially or totally molten parent bodies. Examples of
this group are meteoritic basalts (eucrites) or iron meteorites. The latter are pieces of the
segregated core of the parent body.
Investigations of meteorites in regard to their chemical and mineralogical composition
as well as their microstructure not only offer a view into the material composition and
properties ofthe early solar nebula, but also support an understanding of the formation and
evolution of the early solar system, as well as the formation and internal creation of the
planets and moon and the segregation of the core within planets, etc. (comparison of
normalized abundances tell us about depletions, origination, and dictate further the time
setting for different events). Tiny fragments within inclusions of carbonaceous chondrites
are thought to be of presolar origin and represent the oldest material we have on hand.
Tungsten, in regard to meteorite research, is a very important and interesting element.
As a highly refractory element, it condenses under solar nebula conditions in a common
alloy of refractory metals (tiny nuggets in chondritic meteorite inclusions). However, the
degree of condensation depends on the oxygen concentration and may be incomplete due
to the formation of volatile oxides. In the metallic stage tungsten belongs to the siderophile
elements, but in the oxidized form it is highly lithophile and incompatible. The relative
abundance of tungsten in comparison to refractory elements, which cannot be volatilized
by oxidation, allows calculation of the oxygen concentration during condensation. The low
abundances of siderophile elements like tungsten in the mantle of the earth are the result of
the formation ofan iron-nickel core. During the segregation process, most ofthe metallic elements were removed to the core. The partition coefficient of tungsten between the iron-nickel
melt and the silicate phase is strongly influenced by the oxygen concentration. The more
oxidizing conditions prevail, the more tungsten remains incompatible in the silicate minerals.
The partly oxidizing conditions during core segregation in the earth evolution are
responsible for the fact that we find tungsten deposits in the crust.
2.3. AVERAGE ABUNDANCE [2.2,2.10,2.11]
Spectrographic analyses of stars and galaxies as well as of the sun's photosphere
reveal that the average abundance of the elements within a certain scatter is comparable
throughout the cosmos. As already pointed out, the chondritic meteorites also reflect this
general scheme of abundances, but some differences could be found among them
originating from the various conditions prevalent during their formation. The biggest
anomalies are found in planetesimals, on planets or the moon, where processes like

TUNGSTEN HISTORY

6S

melting, segregation, and differentiated crystallization occurred accompanied by partition,

enrichment, or depletion in diverse areas.
The probability of the fonnation of higher atomic number elements, like tungsten, by
absorption of neutrons and/or protons during a star evolution and subsequent super nova
blast is quite low. Therefore, the abundances of higher atomic number elements are
considerably smaller, as for the elements which were fonned by nuclear fusion reactions in
an evoluting star.
Figure 2.1 shows the dependence of element abundances on atomic number for the
cosmos. In cosmochemistry, such abundances are always presented as a ratio-number of
atoms of the respective element! 106 atoms silicon-in order to eliminate the influence of
the strongly varying concentrations of hydrogen and helium. The remarkable drop in
abundance from atomic number 47 onward indicates the boundary between elements
fonned by nuclear fusion and those built up by subsequent neutron absorption reactions.
Maxima are always related to especially high nuclear stability of the respective element (a
typical example is the iron peak).
Table 2.3 summarizes the tungsten abundances so far known for the cosmos, the solar
system, the sun photosphere, different types of meteorites, and the earth crust and mantle.
It is interesting to note that the tungsten concentration in the sun photosphere is three times
higher than in chondritic meteorites. According to some specialists, it might be a wrong
result. By far, the highest tungsten abundances (x 7000) have been found in the
aforementioned tiny metallic micro inclusions of chondrites.
Accordingly, tungsten can be regarded as a rare element. It is rated 56th of the
elements in tenns of the earth crust (lithosphere) abundance and ranks point 18 among the
metallic elements. Under oxidizing conditions, tungsten fonns the hexavalent ion W6+,
which is either tetrahedrally or octahedrally coordinated by four resp. six oxygen atoms. In
this fonn, it counts among the lithophile (rock-forming) elements; therefore, we can find it
in the rocks of the earth's crust. Under reducing conditions, however, tungsten occurs as a
metal and is siderophile. This is the reason why iron-nickel meteorites contain increased
concentrations of tungsten. These meteorites belong to the segregated core of their parent
bodies. By analogy, and moreover due to the low abundance of tungsten in the earth's
mantle, it can
assumed that the earth's core is enriched in tungsten.

be

2.4. GEOLOGY: FORMATION OF ORE DEPOSITS [2.12-2.15]

The increase in concentration by 3 magnitudes, locally even by 4 magnitudes via
igneous processes, is the most important fact in understanding tungsten's geology. As
already explained, the average abundance of tungsten is approximately 1 ppm (~g/ g)
throughout the solar system, and consequently on earth too. In sharp contrast, one finds
concentrations of one-tenth of a percent and sometimes of 1-2% as average contents of
tungsten ore deposits. All of them are fonned by igneous activities which are best
discussed in tenns of the magmatic hydrothermal model.
This model (see Fig. 2.2) starts with any magma of changing composition and of
different sources. Changes in chemical and physical properties take place. Moving occurs
sometimes to higher levels of the earth crust, and consequently temperature as well as
pressure gradually drop and differential crystallization starts. The first crystallization step

I
I
I

'U

'loB

Be

50
Atomic number. Z

60

Pb

FIGURE 2.1. Cosmic abundance of elements as a function ofatomic number Z [2.4]. Abundances are expressed as numbers of atoms per 106 atoms ofSi and are plotted on a
logarithmic scale.

-3 0

-2

-1

0,

2~UI1\1t".

:94

S!

lll

10

TUNGSTEN HISTORY

67

TABLE 2.3. Abundance of Tungstena

logNw
(log NH = 12)
Cosmos
Solar system
Sun photosphere
Chondrites
Earth mantle
Earth crust
Iron meteorites
Eucrites
Metallic micropartic1es in chondrites
Moon basalt
Moon highland soil
a

Atoms W/106 atoms Si

Ilg Wig

0.2~.3/0.16

0.66
1.11/0.8
0.68

0.127
0.36/0.15
0.13
0.01/0.0164
1.3
0.07-5.0
0.024-0.205
574
0.09
0.30

Additional literature:
Landolt Bomstein, Numerical Data and Functional Relationships in Science and Technology, Group VI, Vol. 2: Astronomy and
Astrophysics, Subvol. a (K. Schaifer.; and H. H. Voigt, eds.) Springer-Verlag, Heidelberg (1981).
N. Grevesse and E. Anders, Solar Element Abundances in Solar Interior and Atmosphere (A. N. Cox, W. C. Livingston, and
M. S. Mathews, eds.), The University of Arizona Press, Tucson (1991).
To convert the Si scale to log N, add 1.554 to the log of the Si scale value.

removes large amounts of Fe, Mg, Ca, and Al from the melt in order to form the so called
basic rocks (basalt and gabbro). By that, the remaining magma is enriched in silica, alkalis,
and other chemical compounds which cannot cocrystallize with the minerals of basic
rocks. In other words, they do not fit into their crystal lattices. When crystallization
proceeds, the silica content of the minerals formed increases in series, like gabbro, diorite,
granodiorite, monzonite, and granite. The last two steps of the model are the pegmatite and
the hydrothermal activity. Following the first two main crystallization phases, a fluid
remains, consisting of any material which could not be built in during the foregoing steps,
such as excessive silica, alkalis, alkaline earth, metal ions, and volatiles. Pegmatite
crystallization includes minerals like quartz, feldspars, and micas and may be regarded
as the last stripping of the aluminum silicate constituents from the residual magmatic fluid.
The remaining "hydrothermal liquid" is enriched in excessive silica, volatiles, and
metals. Metals and volatiles, present in the original magma only in very low concentrations, have now been concentrated as a result of the removal of the main constituents by
the prior crystallization phases. Their concentration in the hydrothermal liquid is high
enough to form ore minerals in the successive crystallization. Where a crystallizing magma
contains tungsten, this metal will be concentrated in the water-rich residual solution
(hydrothermal liquid) because of the notable difference in chemical properties between
anions of hexavalent tungsten and the major ions of the rock-forming minerals. Hexavalent
tungsten has the wrong size and the wrong charge in order to substitute elements like Fe,
Mg, Ca, AI, and Si, which are the major rock-forming components. The same is true for
the tungstate ion, which cannot substitute silicate or aluminate ions.
It is generally believed that tungsten is born in mineralizing fluids chiefly as tungstate
ion, tungstic acid, sodium tungstate, or as heteropoly acid. The relative amount of these
compounds depends on temperature, pH, and silica concentration of the solution. Tungsten

68

CHAPTER 2
Original magma

!
1

I j

Process

Removal of big amounts of Fe, MIl,

Ca, AI snd Inc_ of SIO.
concentration on the residual
mallma

Crystallization product

f-----+

Removal of above eIIIments with

Bltle Rpcka
Buell, Gabbro

Increasing SIO. concentration

Gabbro, Diorite, Granodiorite,

_ _ nita, Granlta

Last strIpplnll of AI..mcatea from

residual magma

Quartz, Feldspar, Micas

Pac!1IIIIItu

'&

Hydrothennal
liquid contaInlnll
ex_1ve ailica
highly ...rIc:hecIln

Dacntylng tempagtym

.5

. . . . . pf

Crystallization at the ~ of origin

of Intrusion and transport ovtr long
dlatsn_ by c:ryataIllzatlon or
reaction with surroundlnll roc:b

f-----+

voIaIIlessnd

different metals

me. IGQIIQmlc om

ISn) CaaaIterite
(WI Tungatllll ores
ICu) CIIaIcopyrIte
(Zn) Sph......
IPb) Galena
IAII) Silver salt
(Au) Gold
ISb) S8bn1te
IHII) Cinnabar
Besides bigger concentratl_
of lIangue mlnerale

FIGURE 2.2. Magmatic hydrothermal model.

will be deposited as wolframite if the calcium concentration of the environment is low, or

as scheelite if it is high (for example, in contact with limestone).
Tungsten ore crystallization can be classified as near-source high-temperature
deposits, because wolframite as well as scheelite range on the high-temperature end of
the series of minerals crystallizing from hydrothermal liquid. Due to the properties of this
liquid, a move to further distant places may also occur.
The physical state of the hydrothermal liquid is a supercriticalliquid (around 600 C
and under extremely high pressure). During intrusion of rocks and combined fracturing,
pressure decrease leads to boiling. Volatiles and liquid may move over extensive distances
from their source. Also, metals may be carried away as volatile compounds. So it is easy to
understand that tungsten ores are not only associated with pegmatites, but also due to. the
above moving processes with other geologic formations. Although the solubility of
tungsten compounds is generally low in pure water, it will be enhanced by other
constituents of the magmatic fluid to permit the transport of the metal during ore-forming
processes and over a wide range of temperatures. A likely form for that transport might be
the heteropoly compound, silico-tungstic acid.
Neutralization of a silico-tungstic acid solution in the presence of Ca2+, Fe2 +, and
2
Mn + ions in laboratory experiments yields successive precipitation of Fe, Mn, and Ca
tungstate. This order parallels a frequently observed mineral paragenesis, where ferberite,
huebnerite, and scheelite are deposited successively.
Tungsten ore deposits have been classified according to their formation, which is
closely related to the properties and composition' of the residual magma and the
hydrothermal liquid as well as to the physical and chemical properties of the surrounding

69

TUNGSTEN HISTORY

rocks and intrusives. A short summary of the different kinds of tungsten ore deposits is
given in Table 2.4.

2.5. MINERALS [2.12,2.13]

Tungsten minerals can be divided into economically important and minor minerals.
The first group, comprising the minerals wolframite, huebnerite, ferberite, and scheelite,
occurs in sufficient abundance to guarantee economic mining and processing. The second
group is of minor importance due to their small amounts, low concentration, and rare
occurrence.
Chemically, nearly all the minerals are tungstates. In the group of the minor minerals,
two exceptions exist: (l) tungstenite-which is a sulfide, and (2) tungstite-which is an
oxide hydrate.
The Wolframite Group. The wolframite group comprises iron and manganese
tungstate minerals of varying composition. The iron tungstate up to a manganese content
TABLE 2.4. Types of Tungsten Deposits
Type 'Of deposit

Tungsten mineral

Formation and typical criteria

Pegmatite

Scheelite
Wolframite

Tungsten minerals crystallize together with Pegmatite

minerals.

Greisen

Wolframite

Tungsten-bearing hydrothermal fluid stays with the

intrusive. Intruded rocks are similar in chemistry to the
intrusive rocks. Granite-like rock made up of quartz,
mica, topaz, tourmaline, fluorite, cassiterite, wolframite.
Typical association of W, Sn, and F.

Scheelite

Hydrothermal liquid invades limestone, causing a violent

reaction of the acidic liquid and the limestone. Close
proximity with the intrusive. Formation of huge amounts
ofCa-silicates (scam or tactite). Majority of the tungsten
deposits.

Contact metasomatic

Vein

Pneumatolytic

Scheelite
Wolframite

Intrusion into fractures already present or formed by

temperature and pressure, followed by crystallization.
Distant from the place of origin.

Scheelite

Transport of volatile tungsten compounds via the gas

phase followed by condensation and reaction with the
surrounding rocks.

Hot spring

Tungsten compounds in aqueous solution reach the

surface. Seldom! Chemism unknown.

Secondary enrichment

Contain tungstite and are very rare. Tungstite is a

weathering product of all four W minerals.

Placer

Scheelite

If tungsten deposits are eroded, placers can be formed

during the transport of the material by water due to the
high gravity and the relative insolubility of the tungsten
minerals.

70

CHAPTER 2

of 20% is called ferberite, the manganese tungstate up to 20% iron is called huebnerite,
and the mixed crystals between these limits are named wolframite, containing 20-80% of
each, iron and manganese tungstate.
At an elevated temperature, iron and manganese tung states form a continuous solid
solution series. Due to the fact that the temperature during mineral crystallization was
sometimes below the limit of miscibility, ferberite and huebnerite can also be found side by
side and not as a mixed crystal, as would be the case at correspondingly high temperatures.
The chemical composition and important physical and mineralogical properties are
summarized in Table 2.5. Well crystallized mineral specimens are shown in Fig. 2.3.
Scheelite. Chemically, scheelite is calcium tungstate (CaW04 ). It is isomorphous to
powellite (CaMo04 ) with which it forms an incomplete series of mixed crystals. Scheelite
contains mostly only minor concentrations of powellite. Its chemical composition as well
as physical and mineralogical properties are given in Table 2.6. A large scheelite single
crystal is shown in Fig. 2.3.
A very important property of scheelite is its bright bluish-white fluorescence under
short-wave UV radiation. The color of the fluorescence radiation is influenced by the
mixed crystal formation with powellite. The color changes with increasing concentration
of molybdenum from blue to cream, pale yellow, and orange. The fluorescence of scheelite
is of great help for prospecting and mining.
Minor Minerals. Chemical as well as physical and mineralogical properties of these
minerals are summarized in Table 2.7.

2.6. ORE DEPOSITS AND RESERVES [2.14-2.23]

The map in Fig. 2.4 reveals that known tungsten deposits are scattered over all the
continents with the exception of Antarctica. The various types of these deposits can be
assigned to three macrotectonic units: Precambrian, Paleozoic orogenic belts, and
Mesozoic-Tertiary orogenic belts.
Deposits of the Mesozoic-Tertiary orogenic belts are the most abundant. Eighty-seven
percent of the reserves, or 93% of the potential reserves, are located in these areas
(orogenic belts of the American Cordillera-Andes, Alpes, Central Asia, Malaysia, and East
Asian Islands). These belts can be regarded as the most important areas for tungsten
mining and also for further prospects.
The second most abundant deposits can be found in Paleozoic orogenic belts. They
represent about 10% of the reserves, or 6% of the potential reserves.
Precambrian deposits are rare. Only 3% of the reserves, or 1% of the potential
reserves, belong to them.
Wolframite. About one-third of the knownworld tungsten reserves are hydrothermal
wolframite-bearing veins and pegmatites. One may distinguish between three different
types of this species:
Plutonic hydrothermal vein deposits, which can be found in China, (Jiangxi;
Guangdong; Guangxi and Hunan provinces), Burma, countries of the former Sovient
Union, Bolivia, Spain, and Portugal.
Subvolcanic vein deposits" located in Bolivia, USA, and Japan.

Formula (pure)
W03 content (%)
Mn content (%)
Fe content (%)
Crystal structure
Lattice parameters
a (A)
b (A)
c (A)
~
Cleavage
Specific gravity (g . cm - 3)
Color
Tenacity
Luster
Fracture
Hardness (mohs)
Magnetism
Streak
Diaphaneity
Common form of occurrence

Property
(Fe, Mn)W04
76.5
3.6-14.5
14.7-3.7
Monoclinic
4.79
5.74
4.99
9026'
Perfect in one direction
7.1-7.5
Dark gray to black
Very brittle
Submetallic to metallic
Uneven
5-5.5
Slightly magnetic
Dark brown
Opaque
Irregular masses radiating groups of bladed crystals

4.71
5.70
5.94
90
Perfect in one direction
7.5
Black
Very brittle
Submetallic to metallic
Uneven
5
Sometimes feebly magnetic
Dark brown
Opaque to translucent in cleavage plates
Well-defined crystals, massive crystalline

Wolframite

FeW04
76.3
0--3.6
18.4--14.7
Monoclinic

Ferberite

TABLE 2.5. Physical and Chemical Properties of Wolframite Group Minerals

Brownish red to greenish yellow

Opaque to translucent
Radiating groups of thin-bladed crystals

4.85
5.77
4.98
9053'
Perfect in one direction
7.2-7.3
Reddish brown to black
Very brittle
Submetallic adamantine
Uneven
5

MnW04
76.6
14.5-18.1
3.7---{)
Monoclinic

Huebnerite

-.J

Vl

::I:

en

72

CHAPTER 2

FIGURE 2.3. Huebnerite, Ferberite, and Scheelite: (a) Huebnerite: ruby red, transparent euhedral crystals; length
about 6 cm; small white tetragonal scheelite crystals grown on top of them; Pasto Bueno Mine, Ancash, Peru. (b)
Ferberite: black, plate-like crystals; dimensions 3 x 5 cm; together with pyrite and arsenophrite; Palca XI Mine,
Puno, Peru.

(Fe,Mn)W(Ta,Nb)20S
(Y,Th,U,Cah(Ti,Fe,W)40 I1

~[(W,Mg,Sb,Fe)031[(OHhIOI(Si04)2]

(AI, Fe)[OHIW041 H2O

W03 2H2O
H(Ca,Pb)Fe3(W04)6IOH20
(Fe,Mnh(Nb,Ta,Ti)2WO IO
CaFeWsV403612H20
PbW04
BhW06
(Zn,Fe,Ca,Mn)W04
(Ca, Y,Sb,Mn)2Ti2(Ta,Nb, W)06(OH)
PbW04
2W203 . H20+ (Th02,Ce203,Zr02)+ H2
WS2
WS 2
W02 H20
(Fe,Ba,Pb)(U02h(OHMW04) 12H2O

~SnW(S,Se)8

AlW20 6(OH)3
Al(W04)(OH) H2O
(Ce,Nd)W20 6(OHh
(Ca,Cu)W04
CuW04
CU2W04(OH)2
2MgO W20 S ' Si02 . nH20
C~F~+Fe~+(W04)7 . 9H2O
H2W04 H2O
P~Fe4(W04)I I(OH)8 6H2O

Fonnula

2.5
2
4
5
5.75

5.5
2.75

2
5.25
2.5
2.5-3.0
3.5

bright yellow
yellow, orange brownish
dark: red brown reddish black
black
black

greenish yellow
yellow brown
yellow
black
dark: brown, brownish yellow
yellowish brown light yellow, gray
pale yellow to green
dark: brown to brownish black
black
reddish brown, yellow, green, red

4.87
4.59
4.02
5.26
6.42
4.5
8.46
7.35
6.70
4.74
8.12

2
I

2.5

pale yellow to brownish yellow

pleochroitic

5.2
4.5

tetr
mono
cub
cub
mono
amorph
ortho
ortho
ortho
mono
tetra
mono
cub
tetra
ortho
hexa
trig
ortho
ortho
trig
mono
ortho

I
3

Hardness
(mohs)

dark lead gray

green, brownish to yellowish

7.06

tetr

7.75
7.73
5.5
4.27
4.47
6.55
4.80

white
white
orange yellow

3.39
4.84
6.27

cub
tric
mono

Density
(g.cm- 3)

Additional literature: A. R. Holzel, Systematics of Minerals, Alexander R. Holzel, Mainz (1989).

Some chemical formulas given are putative.

Alumotungstite
Anthoinite
Cerotungstite
Cuproscheelite
Cupritungstite
Cuprotungstite
Farrallonite
Ferritungstite
Hydrotungstite
Jixianite
Kiddcreekite
Mpororoite
Meymacite
Phyllotungstite
Qitanglinite
Rankachite
Raspite
Russelite
Sammartinite
Scheteligite
Stolzite
Thorotungstite
Tungstenite 2H
Thngstenite 3R
Thngstite
Uranotungstite
Welinite
Wolframoixiolite
Yttrocrrassite

Name

Crystal
structure

TABLE 2.7. Minor Tungsten Mineralsa,b

greenish yellow
yellow

lead gray

reddish brown
pale yellow to grayish

yellowish
dark brown
brown

greenish yellow, greenish gray

white
white

Streak

>-l

(")

...,

TUNGSTEN HISTORY

73

FIGURE 2.3. Huebnerite, Ferberite, and Scheelite: (c) Scheelite: milky, nearly transparent single crystal; height
5.5 cm; Felbertal, Austria. By courtesy of Dr. Niedermayr, Naturhistorisches Museum Wien.

Pegmaticic-pneumato!ytic tin-tungsten deposits, found in Saxonia, Bohemian Erzgebirge, England, Spain, Portugal, Zaire, Australia, and China.
For example, in 1979, half of the world tungsten production came from these three
types of deposits.
Scheelite. Approximately two-thirds of the world tungsten reserves consists of
scheelite deposits. Distinction can be made between:
TABLE 2.6. Physical and Chemical Properties of Scheelite
Formula (pure)
W03 content (%)
Crystal structure
Lattice parameters
a (A)
c (A)

cia
Cleavage
Specific gravity
Color
Tenacity
Luster
Fracture
Hardness (mohs)
Magnetism
Streak
Diaphaneity
Common form of occurrence

CaW04
80.6
Tetragonal
5.257
11.373
2.163
Good in four directions
5.9---{j.1
Pale yellow, brown, commonly white
Very brittle
Vitreous to resinous
Uneven
4.5-5
Nonmagnetic
White
Transparent to translucent
Massive and in small grains, exist sometimes as
pseudomorph after wolframite.

TUNGSTEN HISTORY

IChc'llI,"l'

....

1NOI1"mltl

75

hvlJt c limax]

..

WQI. .....",lt.

m1 nl"'J IlHl1;1l.cb

~l)fecombl l.Jn st,.,lIds

l -.~.j.ncien\ platforms

~p"~ie fold lyslDms

~ II'lIts(,u:oic:'ulno:rolc: lold tV,l t:Jm

FIGURE 2.4. Principal tungsten deposits and mining disricts [2.19].

contact metamorphic deposits, which can be found in USA, Sweden, Brazil, and the
former Soviet Union; and
volcanogenic-sedimentary deposits, located in Austria, South Korea, France, Austria,
Australia, and Canada.
In 1979, nearly half of the world production stemmed from scheelite deposits. Around
1980, the ratio of schee1ite:wolframite deposits in the western world was 73%:27%.
From time to time, summaries about the tungsten ore reserves are published in the
current literature. In principle, it must be taken into account that all those figures are based
on estimates. Moreover, they are influenced by the demand of tungsten, the market
situation, and consequently by the world market price for tungsten.
Generally, figures for the tungsten reserves are always correlated with production.
The higher the demand for tungsten, the bigger the ore reserves are. The reason for this
dependence is that high demand generates high prices, and this again intensifies further
prospecting activities as well as corrected calculations of the potential reserves. For
example, higher prices will bring down the cut-off grade in mining operations. Therefore,
figures concerned with ore reserves are only valid for a certain time and a specific situation
and will change in the future.
A comparison of the global cumulative production of tungsten, the reserves, and the
static lifetime presented in Fig. 2.5 is very informative. From this comparison, it can be
seen that the reserves increased from 700.000 t in 1951 to 2.3 million t in 1978, in spite of
considerable production growth rates. This increases the static lifetime of the reserves from
38 to 49 years.
Table 2.8 informs about the tungsten ore reserves of countries, continents, and the
earth at different times. By far the largest reserves lie in the Peoples Republic of China

76

CHAPTER 2
nllll(:

cumuljhve minE! J:lroduction

since 1946

1000 tonnes W-con1ent

li'e1lme
lungfitien reWNC'S
.."..,';.:OOO;;;.':;o:.;;nn;..,:..;w;.,-c;;.;o:..;nl;;..n.;;.I~~~ (vein)
49

r-.....,..,.,.....

53
45
44

31

32
35
35
35
36
41

41

43
41
43

36

27

38

1.1 .1946

3000

FIGURE 2.5. Cumulative mine production, reserves, and static lifetime of tungsten (1.1. 1946-1. 1.1978) [2.19].
TABLE 2.8. Reserves of Tungsten Ore at Different Times ([10 3 t] tungsten content) [2.23]
1995

1985

1980

Country

Reserves

Reserve base

Reserves

Reserve base

Reserve

Reserve base

Australia
Austria
Bolivia
Brazil
Burma
Canada
China
France
Korea
Portugal
Russia
Thailand
U.S.A.
Others

5
10
53
20
15
260
980
20
58
26
250
30
140
347

129
15
105
20
34
493
1330
20

140
20
70
20
15
670
1230
20
60
40
490
30
290
405

109
18
40
18
32
270
1361
16
82
24
213
18
125
572

259
54
86
41
73
318
2268
2

26
355
30
200
440

130
15
45
20
15
480
1200
20
58
40
280
30
150
317

2200

3300

2800

3500

2598

4158

Total

77

77

27
318
18
327
1290

TUNGSTEN HISTORY

77

followed by Canada. This area in southern China is a special section of the circum pacific
metallogenic belt.
The average tungsten concentration in ore deposits is generally low, which means
mostly between 0.4 and 0.8% W0 3 and only seldom between 1 and 2%. An exception is
the Cantung Mine in the North West Territories of Canada with about 2.5%.
We refrain from any description of mines in regard to geology, mining, lifetime, etc.
This was already done by Yih and Wang [2.12] for some major mines. However, our
steadily changing economy has a marked impact on mining operations and their feasibility.
A typical situation exists since 1988. All the larger mines of the western world had to close
due to extremely low tungsten prices. So, "mothballing" and reopening of tungsten mines
occur relatively often.

2.7. EARLY DISCOVERIES OF ORES, COMPOUNDS, AND OF THE ELEMENT

[2.24-2.31 ]
This period can be regarded as 300 years of scattered laboratory experiments in
European countries (mainly Sweden and Germany), which elucidated the composition of
the ores scheelite, wolframite, the constitution of several tungsten compounds, and a
preparation method for the element itself.
Tungsten minerals were already known during the Middle Ages in the tin mines of the
Saxony-Bohemian area and in Cornwall, long before the discovery of tungsten itself,
because of their negative influence during the tin production. Gregorius Agricola named it
"spuma lupi" (latin), "Wolfsschaum" (german), and "Wolfs Foam" (English). In tin
melting, the presence of wolframite minerals causes a foam which retards a certain amount
oftin, thus decreasing the yield [2.27,2.30].
Because of their high density, tungsten minerals in Cornwall were called "Call" or
"Mock Lead".
Until the end of the 18th century, tungsten minerals were generally named
"Wolfram". The name Wolfram can be deduced from the German words 'IWolf" (English
wolf) and "Rahm" = Geifer (English spittel). The underlying idea is that the wolf eats the
tin. Manifold variations of this name can be found in the old literature, such as Wolform,
Wolfrurnb, Wolfferam, Wolferam, Wolffram, Wolfert, Wolfort, Wolfrig, Wolferan, Wolfish,
Woolfram, Wolframit, and Wolframicum. The word Wolfram has remained and today it is
the official designation of tungsten in the German language.
1564. The German teacher and priest Johann Mathesius was the first to mention the
mineral wolframite in the literature. In his Sarepa [2.29], sixteen prayers for miners are
summarized. Wolfram is mentioned as wolform (1555, third prayer) and later as wolfrumb
(1559, ninth prayer); see Fig. 2.6.
1574. The words wolfferam and wolffram were used a few years later by Lazarus
Ercker, who assumed the mineral to be an arsenic and iron containing tin ore because of its
association with tin.
1757. The mineral Scheelite was first mentioned by A. F. Cronstedt. Also, in regard
to the high density (6g em -3), he named it tungsten, derived from the Swedish tung sten,
which means "heavy stone" (tung stone in English or tungstene in French). At that time, it
was regarded as a calcium-containing iron ore.
1781. C. W Scheele analyzed tungsten (scheelite ore) to be a calcium salt of an

78

CHAPTER 2

eo btbalt nUl' !,tut it jungtn bergfltutlbas mttaJ bUytrlc}' bing

btgztiffet/tin Ulmttlfct,e obtt bttg"tt ,,fJ aUedcy l)lnbfarotnlOit ftill
tuglein metAl inn fid) l)tlt.Wit nrnntn fold)t taubt oOtr It"'t "ten in,.
tlnfcrm frlbtc bttgfWtccf/ glamJ obta: fMrd)Afit/fir Icob41t/fpeyftl
gctufIgilbc/(ttttIJ It11t nfd)uV /bltYfct,",tifl q~4rlJ obtt QU4ttflJl fpatt.
~u j&:cybtrg ljat rna,. fotm "no wtitftn ;in4/3nn btl: l\Aunp
vnbanbcnJ ~ltrn/fd)rollbcu obtt obtrg~l/wtfd)ttl ~inij Sil if!. Jill.
l;untlcrtl bllcl)t tin gtltf/fltllJ g"Utn/fltlU tn(Jrcf .3utf ;it,n bttlJfi\'t't~

bat t6 4U~ fM"dJC~lty 'n~t/"on wolfozrnl mtfpitd'd/f~bfleiu/g)Qf"

foptfI blutffcitt.

FIGURE 2.6. Copied sections of the manuscript of 1. Mathesius, published in 1564 [2.29]. The third and ninth
prayer, mentioning wolform, wolfrumb, wolffshar, wolffsschaum. By courtesy of Dr. V. Dufek, Prague; original
available at the Charles University in Prague.

unknown acid. He digested the ore with potassium carbonate and isolated the acid by
subsequent precipitation with nitric acid. Moreover, he described the digestion of scheelite
by hydrochloric acid, the formation of tungstic acid, and their dissolution in ammonia. In
principle, he applied the same chemical reactions to be used in industrial scales 200 years
later.
T. Bergmann proposed the preparation of tungsten metal by reduction of tungstic acid
with powdered charcoal in 1781.
I 783. Don Juan Jose de Elhuyar and his brother Fausto prepared tungsten metal as
proposed above and named it Wolfram. As already indicated, this name remained only in

TUNGSTEN HISTORY

79

the German and Swedish language while, according to the official IUPAC designation, all
others use Tungsten, although the symbol is W.
1785. The de Elhuyar brothers describe the formation of a grayish-white, hard, and
brittle material, which forms as a reaction between wolfram and pig iron.
1820. A. Breithaupt adopts the word Wolframite for the mineral (Fe,Mn)W04 .
1821. K. C. von Leonhard suggested the name Scheelite (in German and Swedish,
Scheelit) for the mineral CaW04 (formerly tungsten), in this way honoring the famous
Swedish chemist.
1841. R. Oxland was the real founder of systematic tungsten chemistry. He gave
procedures for the preparation of sodium tungstate, tungsten trioxide, and tungsten metal.
He was the first to propose a method for ferrotungsten production forming the basis for
modem high speed steels. Although patents have been granted for those procedures (1847,
1857), an industrial application was not granted mainly due to the enormous price.

2.8. TECHNICALLY IMPORTANT DISCOVERIES [2.24,2.27,2.31-2.33]

1855-57. The use of tungsten for special steels was patented by F. Koller (Austria),
and later by Mushet.
1868. Mushet started to manufacture a self-hardening V-Mn-W steel (containing
5.5% W) in England.
1890. During experiments to produce diamonds in a self-constructed electric arc
furnace, Henry Moissan actually found very hard compounds. None of them were
diamonds, but one was tungsten carbide. This fluke was to have an unbelievable impact
on the further development of tungsten technology on a long-term basis. Tungsten carbide
grew within a period of only 30 years to be the biggest tungsten consumer.
End of the 19th Century. This marked the development of the first high speed steels
(produced by quenching the steel just below the liquidus point) by Taylor and White.
1900. At the Paris World Fair, the Bethlehem Steel Company presented the first high
speed steel (HSS) cutting tools. For several decades after that, steel was the most
prominent tungsten consumer.
1903. According to a patent of A. Just and F. Hanamann (at that time assistant
professors at the Technical University of Vienna), the first tungsten filaments for
incandescent lamps were produced in Hungary. Tungsten filaments were made by forming
a paste from fine tungsten powder after addition of sugar solution and gum. The paste was
then squirted through diamond dies and the wire loops or coils were then sintered by
electric current in hydrogen [2.33]. The elevated carbon content (reaction with carbon of
the binder) was removed during sintering by a moistened hydrogen-nitrogen atmosphere.
1907. Commercial production of incandescent bulbs with these "squirted" tungsten
filaments gradually replaced Edison's carbon filament bulbs due to their much better light
yield and lower energy consumption.
1909. W. D. Coolidge applied for a patent, describing the powder metallurgical
production of ductile tungsten wire. The main features of that patent are still valid for
today's technology. Moreover, the procedure gave birth to large-scale powder metallurgy.
1911. This year saw the commercial production of light bulbs containing the new
ductile tungsten filaments.

80

CHAPTER 2

1922. Investigations of F. Skaupy and co-workers at the Osram Studiengesellschaft

with iron, nickel, or cobalt additions to tungsten carbide were the basis for a patent about
production of cemented carbides.
1923-1925. The outstanding results of the above research enabled the application of
the famous patent of K. SchrOter, DRP 420,689 (WC + 10% Fe, Ni, or Co), followed by
DRP 434,527 (WC up to 20% Co). Diamond, at that time very expensive, was in use as
drawing die for refractory metal wires. Osram substituted it successfully by cemented
carbides. But the inventors did not at all realize the real importance of the new material.
They even missed the deadlines for patent announcements in several countries and the
existing patents were sold to the Friedrich Krupp Aktiengesellschaft.
Although K. SchrOter is named the inventor in the above patents, it was not at all his
sole merit, but the logical consequence of the united efforts of the whole Osram research
group under F. Skaupy [2.32].
It was Krupp, a company experienced in steel technology and machining problems,
who really recognized the potential of cemented carbides (hardmetals).
1927. Krupp offered the first cemented carbide (WC-6%Co) at the Leipzig tool fair.
The brand name was WIDIA, an abbreviation of Wie Diamant (like diamond in German)
and is still known all over the world.
In summary, it can be said that within a short period of less than 40 years at the
change from the 19th to the 20th century, the foundations were laid for an avalanche-like
growth of tungsten technology.

2.9. INDUSTRIAL EVOLUTION [2.34-2.37]

Before and around 1920, tungsten was only important as a steel alloying element and
as a filament in incandescent lamps. At that time, tungsten was added to the steel melt as
tungsten powder derived from tungstic acid by hydrogen reduction (low carbon, 96-98%
W). In principle, this was complicated and expensive. In 1914, ferrotungsten production
was started. The prealloy could be prepared directly from ore concentrates by carbo-,
silico-carbo-, or metallothermic reduction, offering a much cheaper possibility of adding
tungsten to steel. Since 1940, the prealloy "melting base" has also been in use and is
produced by reductive melting of tungsten bearing scrap materials.
In 1935, the first tungsten heavy metal alloys were produced as a material group of
growing importance that would continue for the next four decades.
By 1940, tungsten and its alloys and cOIllJ)ounds already had widespread application
as demonstrated by Wah Chang's Tungsten Tree (Fig. 2.7).
Figure 2.8 illustrates the growth in tungsten demand since 1930. Except for
temporary fluctuations, a steady growth due to increasing industrialization in the world
can be observed. War times were always related to maxima and economical recessions to
minima.
Figure 2.9 demonstrates the shares of the four main tungsten consumers. In slightly
more than 60 years, the No.1 of the past (steel) became No.2, and the cemented carbides
now consume more than half of the total tungsten production.
In the early years of industrial tungsten evolution, around 1930, the largest consumer
of the metal was the steel industry, and only stnall percentages were used for lamp

81

TUNGSTEN HISTORY

GROWTH OF
THE
TUNGSTEN TREE

ORE

FIGURE 2.7. 40 years of the tungsten tree [2.34].

filaments and the just-born cemented carbides. This marked the starting point of a steady
competition between high speed steel and cemented carbides. The share of the latter was
steadily growing; thereby, decreasing the percentage of tungsten applied to high speed
steel production, a process which still exists today.
60,-------------------------------------------------,

50

40

...-

30

20

10

o+----+----+---~----~--~~--~----~--~--~

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

year
FIGURE 2.8. Tungsten world production 1910-1996. By courtesy of ITIA and A. P. Newey, in: Proc. 3rd Int.
Tungsten Symp. Madrid, pp. 19-33 (1985).

82

CHAPTER 2

4% Chemicals

24%
Steel and
Superalloys

FIGURE 2.9. Tungsten consumption by end use (worldwide).

Since 1930, cemented carbides (also called hardmetals) steadily attained a greater
share in tungsten consumption. It is of interest to ask why the demand for cemented
carbide grew so rapidly. Table 2.9 shows a chronological table indicating the most
important events in cemented carbide research and development, a process which is still
under way today. We recognize that what we call cemented carbides or hardmetals are in
reality a very wide palette of materials with different properties. Cemented carbide
properties can be adjusted by several variations and combinations of the components, as
shown in Table 2.10. Hence cemented carbides could be applied widely. Figure 2.10
presents a breakdown of the fields of application of cemented carbides.
The ability to adjust the properties of a tool to the special problem of the material to
be worked marks an important advantage over high speed steel, and explains the
replacement of the latter in many fields.

TABLE 2.9. Evolution of Cemented Carbides (Chronological Table)

1927
1929
1930
1931
1932
1969
1970
1973
1979
1990

WC-6%Co cutting tool (Krupp)

WC-6%Co fine WC Grain (Krupp)
WC-Il%Co (Krupp)
WC-1O%TiC-6%Co (Krupp)
WC-27%TaC-13%Co (Carboloy)
WC-15%Co (Krupp)
WC-TiC-TaC-CO (Firth Sterling)
Coatings (TiC, TiN, Ti(C,N), Ah03)
Improved micrograin qualities
HIPing
Ni binder
Ultrafine qualities/nanograin powder precursors

83

TUNGSTEN HISTORY

TABLE 2.l0. Reason for the Expanding Applicability of Cemented Carbide

Widespread property variation and combination through variation and combination of:
-Carbide/binder ratio (influencing hardness and toughness)
-Binder composition (influencing corrosion resistance)
-Carbide phase(s) grain size (influencing hardness, toughness and strength)
-Second carbide phase composition (influencing wear resistance)

15%
Mining Tools and
Drilling Tools for
Water and Oil

59%
Metal Cutting

Tools

11%
Wood and Plastic
Working Tools

FIGURE 2.10. Field of application of cemented carbides.

Although cemented carbides themselves have been replaced by other materials

(ceramics, cermets, diamonds) in special applications, and although the quality improvements and coatings have increased the lifetime of tools, a steady increase in cemented
carbide demand is evident.

REFERENCES FOR CHAPTER 2

R. Kippenhahn and A. Weigert, Stellar Structure and Evolution, Springer-Verlag, Berlin, Heidelberg
(1990).
2.2. S. Mitton, ed., Cambridge Enzyklopaedie der Astronomie, Orbis Verlag, Munchen (1989).
2.3. W. Hillebrandt and W. Ober, Naturwissenschaften 69 (1982), 205-211.
2.4. N. N. Greenwood and A. Earnshaw, The Chemistry of the Elements, Pergarnon Press, Oxford (1984).
2.5. H. Palme, "Cosmetic Chemistry," in: Reviews in Modern Astronomy (G. Klare, ed.), Vol. I, SprlngerVerlag, Berlin, Heidelberg (1988).
2.6. E. R. D. Scott, Earth Planet, Sci. Lett. 39 (1978), 363-370.
2.7. H. E. Newsom and H. Palme, Earth Planet. Sci. Lett. 69 (1984), 354-364.
2.8. H. Palme and F. Wlotzka, Earth Planet. Sci. Lett. 33 (1976), 45-60.
2.9. H. Palme and W. Rarnmensee, Proc. Lunar Planet. Sci. 128 (1981), 949-964.
2.10. H. Palme, "Chemical Abundances in Meteorites," in: Review of Modern Astronomy (G. Klare, ed.),
Springer-Verlag, Berlin, Heidelberg (1988).
2.11. Landolt Bomstein, Numerical Data and Functional Relationships in Science and Technology, Group VI,

2.1.

84

2.12.
2.13.
2.14.
2.15.
2.16.
2.17.
2.18.
2.19.
2.20.
2.21.
2.22.
2.23.
2.24.
2.25.
2.26.
2.27.
2.28.
2.29.
. 2.30.
2.31.
2.32.
2.33.

2.34.
2.35.
2.36.
2.37.

CHAPTER 2
Vol. 2, Astronomy and Astrophysics, Subvol. a (K. Schaifers and H. M. Voigt, eds.), Springer-Verlag,
Heidelberg (1981).
S. W. H. Yih and C. T. Wang, Tungsten, Sources, Metallurgy, Properties and Application, Plenum Press,
New York (1979).
H. J. Rosier, Lehrbuch der Mineralogie, YEB Deutscher Verlag fiir Grundstoffindustrie, Leipzig (1984).
V. F. Barbanov, Int. Geol. Rev. 13 (1971), 332-344.
K. B. Krauskopf, "Tungsten," in: Handbook of Geochemistry (K. H. Wedepohl, ed.), Springer-Verlag,
Berlin (1970).
H. Grundlach and W. Thonnann, Z. Dtsch. Geol. Ges. 112 (1960) 1-35.
R. F. Horsnail, "The Geology of Tungsten," in: Proc. 1st Int. Tungsten Symp., Stockholm, pp. 18-31,
Mining Journal Books Ltd., London (1979).
A. Smith, in: Proc. 1st Int. Tungsten Symp., Stockholm, pp. 32--63, Mining Journal Books Ltd., London
(1979).
F. Bender, in: Proc. 1st Int. Tungsten Symp., Stockholm, pp. 2-17, Mining Journal Books Ltd., London
(1979).
Wu Weisun, in: Proc. 1st Int. Tungsten Symp., Stockholm, pp. 118-126, Mining Journal Books Ltd.,
London (1979).
R. F. Stevens, Tungsten, Bull. U.S. Bur. of Mines, Minerals, Facts, Problems p. 1161 (1980).
R. F. Stevens, Tungsten, Bull. U.S. Bur. of Mines, Minerals, Facts, Problems, p. 667 (1975).
The Economics of Tungsten, Roskill Infonnation Services, London (1990).
E. Lassner, W. D. Schubert, E. Liidertiz, and H. U. Wolf, "Tungsten, Tungsten Alloys, and Tungsten
Compounds," in: Ullmanns Enciyciopedia of Industrial Chemistry, Vol. A27, pp. 229-266 (1996).
S. Engels and R. Stolz, ABC-Geschichte der Chemie, YEB Deutscher Verlag fiir Grundstoffindustrie,
Leipzig (1989).
I. Kappel, Chem. Ztg. SO (1926), 969-971.
Gmelins Handbuch der Anorganischen Chemie, 8. Auflage, Syst. Nr. 54, Verlag Chemie, Berlin (1933).
V. Dufek, Lazarus Ercker von Schreckenfels, manuscript prepared for the Symposium on Science and
Technology of Rudolfinian Time, National Technical Museum, Prague (1997).
J. Mathesius, Sarepa oder Bergpostill, Niirnberg (1564) .
J. C. Bailar, Jr., H. J. EmeIeus, R. Nyholm, and A. F. Trotman-Dickenson, eds., Comprehensive Inorganic
Chemistry, Pergamon Press, Oxford (\973).
H. Pastor, in: Proc. of the 1996 European Con! on Advances in Hard Materials Production, Stockholm, pp.
3-11, EPMA, Shrewsbury (1996).
H. Kolaska, "The Dawn of the Hard Metal Age," Powder Metall. Int. 24 (1992), 311-314.
J. Gurland and J. D. Knox, "No Child of Chance ... "-From Sintered Tungsten to Cemented Tungsten to
Cemented Tungsten Carbide," in: Tungsten and Refractory Metals 3 (A. Bose and R. J. Dowding, eds.), pp.
219-227, MPIF, Princeton (1996).
Wah Changs, "Unique Plant for Concentrating Tungsten," Eng. Min. J. (Sept. 1944).
A. P. Newey, in: Proc. 3rd Int. Tungsten Symp., Madrid, pp. 19-33, MPR Publishing Services, Shrewsbury
(1985).
IT/A Information; 119-123 Hackford Road, London SW9 OQT, UK.
P. Borchers and W. Miiller, in: Proc. 5th Int. Tungsten Symp., Budapest, pp. 40--55, MPR Publishing
Services, Shrewsbury (1990).

3
Important Aspects of Tungsten
Chemistry
3.1. OXIDATION OF TUNGSTEN METAL BY AIR OR OXYGEN [3.1-3.3]
Tungsten's sensitivity toward oxygen restricts its high-temperature application to protecting atmospheres or vacuum only.
Adsorption. Oxygen is adsorbed on clean tungsten surfaces in a variety of atomic and
molecular states. At low temperature ( <0 0c), oxygen is adsorbed molecularly, but at room
temperature this adsorption is a precursor state to the atomic adsorption. A covered surface
shows an ordered oxygen superstructure. If the temperature is increased, a more extensive
coverage occurs and oxide-like structures are formed. The surface layer can be described
as adsorbed oxide.
The adsorption of oxygen on tungsten surfaces has been investigated most extensively. Relevant data can be found elsewhere [3.1].
Diffusion. Oxygen diffuses from the surface into the crystal lattice forming a bcc (ct)
solid solution. The solubility is very low. Calculated diffusion coefficients at 1700C are
given as

D = 10-7 and 3 x 10-8 to 5

10-8 cm2 . S-I for polycrystalline tungsten [3.4, 3.5]

and
D= 7

10-8 to 5

10-8 cm2 . s-I for monocrystalline tungsten [3.4]

Permeation. For the diffusion of oxygen through tungsten in the temperature range
1950-2300C, the following relationship was found:
P _ {510 exp(-40,000/RT) [3.5], resp. 380 exp (-38, 100/RT) [3.6] for P/M tungsten
850 exp( -44,000/ RT) [3.5] for arc-cast tungsten
where P=permeation coefficient in (mm3 O2 , cm- 2 . min-I. atm- I/ 2 and R=gas
constant in cal mol-I. K- 1 with Tin K.
Oxidation [3.3). The oxidation of tungsten depends strongly on temperature and on
the oxygen partial pressure at elevated temperatures. A clean-looking tungsten surface
between room temperature and 370C contains oxide.
85

86

CHAPTER 3

It was shown by ESCA investigations that the oxidation of tungsten by oxygen or air
starts at room temperature. The oxide formed in oxygen as well as in dry or humid air is
always tungsten trioxide. The thickness of the oxide layer increases slowly with increasing
temperature up to 200 DC, but increases rapidly above 200 DC. Increase in humidity also
accelerates the oxidation rate.
From 327--400 DC, a firmly adherent blue and thin oxide film will be formed. It acts as
a protective layer. Its formation rate is determined by diffusion and proceeds according to a
parabolic rate law. The film composition is given as W02.75, but this is in contradiction to
above ESCA findings. The color is not real but is caused by interference.
Above 500 DC the oxide layer cracks, and above 600 DC, W0 3 forms on the blue layer.
The W0 3 is permeable to oxygen. The W0 3 formation rate is determined by the tungsten
ion transport to the phase boundary W02.75jW0 3 As long as the inner layer has not yet
attained its maximum thickness, the growth follows a mixed parabolic linear rate law.
When the maximum thickness is reached, it is strictly linear.
Sublimation of tungsten trioxide starts at 750 DC and becomes substantial above
900 DC. Above 1300 DC, the W0 3 sublimation rate corresponds to the oxidation rate, and
the surface remains free of oxide.
At elevated temperature, the oxygen pressure strongly influences the reaction rate.
The tungsten consumed by oxidation in the temperature range between 700 DC and
1300 DC and oxygen pressures from 0.0013 to 20.8 atm can be calculated as follows [3.7]:

_ d(mjA) = (5.89 x 106 ) exp (12, 170 P

dt
T

_ dx = (3.05
dt

1/2)

102 ) exp (- 12, 170

resp.

PI/2)

where m j A is the weight change per unit area, dx j dt the depth of oxidation per unit time, P
the pressure in atmospheres, and Tthe absolute temperature. Units in mg cm- 2 . h- I resp.
cmh- I
At high temperature (>2100 K) and low oxygen partial pressure, gaseous W02 is
formed more rapidly than W0 3 . Thermal dissociation slows down the oxidation and, above
2500 K, volatilization of the metal occurs.
Oxidation o/Tungsten Metal Powder. Tungsten powder, of average grain size> 111m,
reacts like bulk tungsten. Finer powder qualities, depending on grain size and preparation
method, can be pyrophoric.
Tungsten powder that is reduced by hydrogen at low temperature from tungstic acid
or tungstic-acid-derived W0 3 contains large concentrations of ~-W and burns immediately
when in contact with air. If such a powder is saturated by an inert gas containing only a low
concentration of oxygen, it is no longer pyrophoric. It is assumed that the high density of
lattice defects is responsible for the pyrophoricity.
3.2. REACTION OF TUNGSTEN WITH WATER [3.3,3.8]
This section is closely related to the following one (3.3.) The basic equation for the
reactions involved is

87

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

This section deals with reactions from left to right, while Section 3.3 deals with
reactions in the opposite direction. The hydrogen partial pressure in Section 3.2 is
normally low, because hydrogen is a reaction product. In Section 3.3 hydrogen is the
reducing agent, which is usually applied in excess, and accordingly the partial pressure is
higher.
Water reacts with tungsten below the boiling point. In principle, water acts as an
oxidizing agent. According to ESCA investigations W02 , W0 3, and oxide hydrates are
formed [3.3]:
W + 2H2 0

--+

WO z + 2Hz

W + 3H2 0

--+

W0 3 + 3Hz

--+

W0 3 . xHzO

W0 3 + xHzO

(x=I-2)

The reaction with distilled water is very slow at low temperature: the corrosion rate is
3.8J.Lg W . m- z . h- I at 38C. The reaction is accompanied by the formation of a
nonprotective oxide film, which is in equilibrium with 1O-3moljl WO~- in solution. At
elevated temperature and pressure (150-360C and 70-80 atm), the corrosion rate
increases to 1.6 g W . m- z . h- I .
.
Reaction with water vapor or humid air (60-95% relative humidity) in the temperature range 20-500 C results in the formation of a W0 3 layer. No lower oxides or hydrates
could be found by ESCA. The thickness of the oxide layer increases with increasing
humidity. The reaction rate depends on temperature and [HzO]/[H z] partial pressure ratio
and is more rapid than with liquid water. Water molecules are adsorbed at the tungsten
surface and dissociate. The oxygen atoms diffuse into the tungsten metal, forming at first a
solid solution and then the oxide compound, while, hydrogen escapes as element.
Depending on temperature and the PH20/PH2 ratio, all known oxides can be formed.
As long as both parameters are low, the reaction product is W0 2 An increase in both
parameters yields in the formation of the higher oxides W02.7Z, WO Z.9 and W0 3. Loss by
evaporation shows a linear increase with PH2 0 between 1050 and 1275C. Changes of
volatility occur at the equilibrium pressures of the instability points of the oxides formed.
For example, the point of coexistence of W /W 30 jWOz corresponds to 1150 K and a
log PH 20/PH2 ratio of 0.13.
W03 evaporation at 1480-1950K occurs at [3.9]:
PH 2 0 = 6.3-20

10-4 Torr

An argon stream saturated with water vapor (21C) flowing over tungsten at 1470 K
produces red-violet W 18049 at the cooler parts of the furnace. At higher humidity, blue
WO Z.9, and yellow W03 will be formed.
At temperatures >600C, gaseous W0 3 HzO = WO Z(OH)2 is the most volatile
compound formed in the system W-O-H. It is responsible for all vapor-phase transport
processes at temperatures where volatilization of oxides can be neglected:

The logarithm (log) of the equilibrium constant

88

CHAPTER 3

plotted versus I/T(K) is linear and given by

10gK

=T -

584~ (370) + 1216 (0.23)

The standard free energy change for the reaction between 1200 and 1500C is given by:
AF = 26,700 - 5.56 . T( 500 cal) and the standard entropy of W03 . H20 at 1600 c as
SO = 136.7 1.05 eu [3.10].
Tungsten sheet or powder reacts with water vapor between 700 and 1200C to form
W0 2. Above 900C, volatilization via the oxide hydrate occurs. In a moist Nr H2
atmosphere, after one hour at 1700C, the loss in W is:
and
During the reaction of a tungsten filament with water vapor, the following species can
be found at temperatures between 1580 and 3200K: WO, W02, W0 3, (W03)2, (W03h,
(W03)4 [3.11]. From 2250 c upward, oxygen replaces the water molecule because of its
dissociation. The trimeric and quadruple species form between 2200 and 3200C:
3W + 9H20 <=> 9H2 + (W03)3
4W + 12H20 <=> 12H2 + (W03)4
Knowledge of these reactions is important in many respects. Tungsten is used for hightemperature engineering components, as heater or cathode in high vacuum techniques, and
as filaments in incandescent lamps. Water vapor may be responsible for:

increased burn-off of cathodes,

corrosion of contacts,
water cycle in incandescent lamps,
destruction of tungsten in contact with alumina (water vapor oxidizes W to W0 2,
which reacts with Ah03 to form AlW04 ).

3.3. REDUCTION OF TUNGSTEN OXIDES BY HYDROGEN [3.12-3.15]

3.3.1. Introduction
Today's only technically important method for tungsten powder production is the
hydrogen reduction of tungsten oxides according to the overall equation

at temperatures between 600-11 00 C in a streaming hydrogen atmosphere. As starting

materials besides the trioxide, tungsten blue oxide (W0 3_x ) or tungstic acid (H2 W04 ) are in
use.

89

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

The reduction process is not only a chemical conversion from oxide to metal. It is
combined with a chemical vapor transport of tungsten which is responsible for the final
powder characteristics, and which is a peculiarity of the W-O-H system. By changing the
reduction parameters the powder properties like average grain size, grain size distribution,
grain shape, agglomeration, etc. can be regulated within a considerable wide range and this
makes the process unique in a certain sense.
Besides chemical conversion and chemical vapor transport, the reduction process is a
purification step, too. Trace impurities, always present in the oxide, may evaporate. On the
other hand, foreign phases can be incorporated during the CVT growth of tungsten, finally
leading to inclusions in the tungsten powder particles. This is of special interest in the
production of "non-sag" tungsten wire used for incandescent lamps.
The reduction of W0 3 to W as described by the simple equation above is in reality a
rather complex process. Most of the scientific background and understanding is relatively
new, although the process has been in use on a technical scale for more than 80 years.
In the past an attempt was made to describe the reduction process in terms of
temperature, time, diffusion-path lengths, and oxide bulk density, based on theoretical
diffusion equations [3.16], but such models are not used in practice due to the complexity
of parameter interactions. Industrial production today is still based on extensive empirical
knowledge, supported by in-depth understanding of the basic aspects governing the
reduction process, accompanied by a feel for the matter.
The following subsections will present the major aspects of the reduction process in
order to understand the chemical and physical processes that arise during the reduction
sequence and determine the quality of the final metal powder. The industrial procedure
including technical details will be presented in Section 5.4.

3.3.2. Thermodynamic Considerations

The W-O-H System. Six condensed phases appear in the W-O-H system: W0 3 ,
W02.9 (W200S8), W02.72 (WlS049), W02, \J..-W, and ~-w. The latter is metastable and was
earlier described as W3 0. In regard to their properties we refer to Section 4.2.7. Phase
relations were studied by several authors [3.17-3.24], and thermodynamic data were
evaluated for both the solid [3.18,3.22] and gaseous oxides (W40 1b W3 0 9, W 30 8, W20 6 ,
W0 2(OH)z) [3.10,3.19,3.25].
Standard Gibbs functions of various reactions of the W-O system are given below for
the temperature range of 700 to 1000C [3.23]:
W(s) + 02(g)
2
0.72 . W02(s)

+*

W0 2(s)

= -574,547 + 170.0 . T

I'1GO = -499,067

+ 125.6 T

+ Oz(g) +* 0.18' W02.9(S)

I'1GO = -567,852

+ 202.6 T

+ 02(g) +* 20 . W0 3(s)

I'1GO = -558,689

+ 223.9 T

+ 2(g) +* 0.72 . W02.n (s)

0.18' W02.n (s)

I'1GO

20 . W0 2.9(s)

90

CHAPTER 3

Based on the Gibbs functions, the temperatures for the invariant three-phase equilibria can
be calculated:
620C [3.23]/560 C [3.26] for the equilibrium:
280C [3.23]/260 C [3.26] for the equilibrium:
10 . W0 2.9(s)
1480 C [3.26] for the equilibrium:

W02 (s)

*+

*+

9 . W0 3(s) + W0 2 (s)

2
2.72 . W2.7is)

0.72

+ 2.72 . W(s)

Vapor pressures above W0 3 and W0 2 are listed as a function of temperature elsewhere

[3.19,3.25].
The stability regions of the oxides are determined by the temperature and the
PH 20/PH2 ratio, as demonstrated in Fig. 3.1 [3.17]. Any of the equilibrium oxides can
be produced by annealing metallic tungsten or W0 3 under the respective equilibriumtemperature and humidity conditions.
In one of the older equilibrium studies [3.20] an additional oxide was described
(W3 0) to occur at low-temperature/humidity conditions. However, this compound was not
confirmed in any of the later investigations [3.22-3.24], and it is well accepted today that it
(W30/~-W) is a metastable transient and not a stable equilibrium phase.

1.0

0.5

-0.5

,.~.........-.-.....
,,
,,
,,

,,

C/...-W

WOz

,,
,,

,,

-1.0

0.8

1.0

,,

"

10 J r'(K)

1.2

1.4

FIGURE 3.1. Equilibrium diagram for the W-O-H system [3.17].

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

91

3.3.3. Formation of a Volatile Tungsten Oxide Hydrate [W02(OHhl

Tungsten oxides and tungsten metal exhibit an apparent increase in vapor pressure at
elevated temperatures in the presence of water vapor [3.27-3.31]. This increase is due to a
reaction between the respective oxide or the metal and the water vapor, to form a volatile
tungsten oxide hydrate. The reactions can be described by the following equations:
W0 3 (s) + H20(g) ++ W02(OHMg)

~. W 20 0 58 (S) + IIH20(g) ++ 10 . W02(OHMg) + H2(g)

W 18 0 49 (S) + 23H20(g) ++ 18 W02(OHMg) + 5H20(g)
W02(s) + 2H20(g) ++ W0 2(OHMg) + H2(g)
W(s) + 4H20(g) ++ W0 2(OHMg) + 3H2(g)
The presence of the monomere, volatile oxide hydrate [W03 H20 resp. W02(OHh]
was proven by mass spectroscopy [3.28], and thermodynamic data are available for all of
the above phase equilibria [3.10,3.29]. Based on these free energy data as well as on those
of the solid oxides [3.23], the equilibrium partial pressure of the volatile compound can be
calculated as a function of humidity. The result of such a calculation is shown in Fig. 3.2
for a temperature of 1000 QC [3.32]. In addition, the equilibrium pressures of the other
volatile tungsten compounds are also presented. From these calculations it is evident, that
the oxide hydrate is by far the most volatile tungsten compound in the W-O-H system.
The oxide hydrate is responsible for the chemical vapor transport of tungsten, which
occurs throughout the reduction sequence, and which decisively codetermines the physical
and chemical properties of the metal powder. Its actual partial pressure during reduction
will depend on both the temperature and the prevailing humidity. It will be lower toward
the end of reduction, due to the steady decrease in the oxygen partial pressure (humidity)
of the system as the reduction proceeds.
Chemical vapor transport of tungsten also occurs during equilibrium annealing of the
oxides with hydrogen-water mixtures (i.e., overall not-reducing conditions), as demonstrated in Fig. 3.3 for the annealing of W02 at 800 QC for I hour, respectively 6 hours in
humid hydrogen (dew point: 72 QC). In such a case a significant particle coarsening occurs
via the vapor phase, leading to the evaporation of smaller grains and growth of the larger
grains (phenomenologically similar to an Ostwald ripening process).

3.3.4. Kinetic Considerations

Oxide Reduction in a Powder Bed. Industrial reduction of tungsten oxides is carried
out either in a static (push-type kilns) or a dynamic (rotary furnace) powder bed. During
reduction (accompanied by water formation) this powder bed (by its porosity) exerts
considerable diffusion resistance against the removal of water from the layer. The higher
the layer, the greater the diffusion resistance and the more slowly the reaction water will be
removed. Even increasing the hydrogen flow over the powder layer to ve.ry high rates will
mainly affect the rate of removal from the uppermost part of the layer and not as much
from the interior. The reactions inside the layer remain diffusion-controlled.

92

CHAPTER 3

-17

-1

-15

-13

-11

~I~I il i I~

:e
al

-5

c.
Ol

.Q

-10

FIGURE 3.2. Equilibrium partial pressure at 100C of various tungsten oxide and hydroxide species [3.32)

The influence of the powder layer height and its porosity on the retention of the
reaction water is of fundamental importance for the technical process. The diffusion
resistance determines a dynamic humidity level within the layer, and hence the nucleation
and growth conditions for all phases which form or might form (from a thermodynamic
point of view) during reduction. It also determines the reduction path, which can be quite
different, depending on the prevailing humidity conditions. Finally, the layer humidity is
an important contribution to all tungsten vapor transport reactions which occur during the
reduction sequence and which are decisive for the final properties of the tungsten powder,
since the total tungsten concentration of the vapor phase (i.e., the sum of volatile tungsten
species), and thus the vapor transport capacity, is markedly enhanced by the water vapor
(see Fig. 3.2).
The Reduction Sequence. Under industrial conditions (layer heights of several
millimeters up to centimeters; temperatures between 700 and 1100 0c) the transport of
the water vapor out of the layer is the overall rate-determining step. Any increase in the
diffusion resistance, for example by an increase in the layer height, will lead to a
significant prolongation in reduction time ("transport controlled reaction"). Under such
conditions the humidity in the layer is always high enough to prevent short-cuts in the
reduction sequence (for example, W0 3 -+ W). The reduction process proceeds in distinct
stages, through the formation of all stable oxides:
W0 3 (yellow) -+ W02.9 (blue) -+ W0 2.nCviolet) -+ W0 2 (brown) -+ W(gray)

An example for this reduction sequence is shown in Fig. 3.4, where the water
formation during reduction of pure W0 3 is followed by continuous thermal conductivity
measurements of the reaction gas. The reduction water curve can be separated into four
segments, according to the stepwise character of the reduction process (W03 -+ W02.9 ,
W0 2.9 -+ W02.72 , W0 2.72 -+ W0 2 , W0 2 -+ W). Furthermore, it is demonstrated that an

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

93

FIGURE 3.3. W0 2 fonned at 800C and a hydrogen dew point of 72 C [3.15]: annealing time 1 h (upper
image); 6 h (lower image).

increase in the oxide layer height leads to a significant prolongation of the reduction time,
as characteristic for transport or mixed chemical and transport-controlled processes
[3.14].
Chemical reaction control is observed only at low reduction temperatures, as well as
in the case of thin powder layers ( < 1 mm) and high hydrogen flow (i.e., optimal material
exchange). Under such dry reduction conditions, direct formation of the metal (rJ,- W or ~
W) from the higher oxides becomes possible (W0 2.9 -+ W, W0 2.72 -+ W).
At low reduction temperatures (500-750C) different transitions can occur simultaneously in the powder layer, indicating a strong influence oflocal differences in the oxygen

94

CHAPTER 3
160

-;:- 40

11

"-

..

100

I!

a.

80

ii

1:
1\1

60

W~.72

W02.72

.... W02

\~

a. 40

0N

::I:

....

20

GI

fH W02.9

20

W02 .... W

1\

II

W~

Reduction time (min)

.... W02.9

160

i.a

140

..I!.

100

ii

60

powder layer height variations: 4 mm .... 8 mm

.5.120

I!
~

a.

i!
1\1

80

a. 40

0N

::I:

Bmm

20
0
0

30

60

90

120

150

180

Reduction time (min)

FIGURE 3.4. Reduction water curves of pure W03 at SOOC under near-industrial conditions; influence of
powder layer'height variations (4mm-Smm). Under these "humid" conditions the reduction occurs stepwise
(W03 ~ W029 ~ W02.72 ~ W02 ~ W) within the. powder layer. The dynamic humidity is always high
enough to prevent shortcuts in the reduction sequence (e.g., W02.72 ~ W).

partial pressure (local humidity) within the layer, or even within a reacting particle, on the
nucleation and growth rates of different phases. An example for this reduction sequence is
given in Fig. 3.5 (reduction water curve of W0 3 at 650 QC).
An overview of the broad variety of experimentally proven reduction sequences is
given in Fig. 3.6. The main reaction paths, as observed under near-industrial reduction
conditions, are marked therein by thicker connecting lines. For more information on the
individual reactions the reader is referred elsewhere [3.15].
Metastable ~-W is always formed out of W02 .9 and W02.72 (and from tungsten
bronzes), but never out of W02 , which might indicate that structural relationships between
the individual phases are of importance for phase formation, in particular at low reduction
temperatures.
Reduction Mechanisms. Based on morphological investigations, two fundamentally
different reduction mechanisms can be postulated for the reduction process.

9S

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

..

90

wo, -

'1:'80

-e
~

;,

Powder layer height: 4 mm

WOU

70

80

ea. 5040

ii

~72
I .

1:30

W02-

",..w

120

:z:

10
0
0

10

20

15

25

30

Reduction time (min)

FIGURE 3.5. Reduction water curve of pure WOJ at 650C. Different reactions occur simultaneously within the
powder layer or even within an oxide particle, depending on the local oxygen partial pressure.

>750C

<750C

FIGURE 3.6. Reduction paths associated with the reduction of pure WOJ ; main transitions are marked with
thicker lines [3.15].

(1) Diffusion in the solid state (oxygen transport)

The reduction occurs by direct oxygen removal from the solid oxides (solid-state
diffusion). The basic underlying mechanism is not known (diffusion of 0, OH, H20) and is
likely to vary for different for different phase transitions. On the final reduction the metal
remains pseudomorphous to the starting oxide, forming a polycrystalline metal sponge.
Solid-state reactions are characteristic for low reduction temperatures 750C) and the
early W0 3 ~ W02.9 transition ("crystallographic shear" transition).
An example for a solid-state transition is the formation of metastable IJ-W out of
W02.9 (Fig. 3.7):

96

CHAPTER 3

FIGURE 3.7. SEM image of~-tungsten fonned at 550 C. The metal is pseudomorphous to the starting oxide
(W03 , APT origin); the pseudomorphs are built-up of nanocrystalline grains.

(2) Chemical vapor transport (tungsten transport)

The reduction is combined with a chemical vapor transport of tungsten via the volatile
W0 2(OH)2' The morphology of the solid phases changes significantly. The reaction can be
separated into three sequential stages:
Stage 1: Formation of the volatile oxide hydrate by a surface reaction of the oxide
with water vapor:

W0 2(s) + 2H 20(g) -+ W0 2(OHh

+ 2H2

(oxidation)

Stage 2: Transport of the WOiOHh from the higher to the lower oxide (resp. to W
metal).
Stage 3: Reduction of the volatile W0 2(OHh at the surface of the growing oxide
resp. tungsten metal:

An important example for this mechanism is the formation of small, individual grains
of r:t-W from the W0 2 matrix.
Which of the two mechanisms prevails during the reduction sequence is determined
by the interaction among temperature, humidity, and the kinetics of phase formation
[3.15]. It is very likely, that solid-state reactions govern all the phase nucleation processes,
i.e., also in the case where subsequent growth occurs via chemical vapor transport (CVT).
CVT reactions occur under industrial reduction conditions throughout the reduction
sequence, thereby significantly altering the morphology of the intermediate and final
reduction products (Fig. 3.8). SEM images of the most important phase transitions are
shown in Fig. 3.9. Note that the gap formation between the reacting particles as well as the

97

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

formation of specific crystal habits for the individual phases (W0 2.72 needles, W0 2
rosettes, W cubes, icositetrahedrons, etc.) are characteristic for CVT.
During the reduction, the formation of W0 2(OHh takes place by the reaction of the
oxide with the water which forms during reduction. Above a substance which is stable at
the. conditions applied (e.g., r:J..-W; see Fig. 3.10), a lower W0 2(OHh partial pressure is
built up than above a reacting unstable substance (e.g., W02). This difference in vapor
pressure provides the driving force for a chemical vapor transport of tungsten, which
finally leads to a dissolution (oxidation) of the unstable substance through the formation of
W0 2 (OHh out ofW0 2 and the growth of the stable substance through the deposition ofW
by decomposition (reduction) of W0 2(OH)z. This CVT growth is shown schematically in
Fig. 3.11 for the W02 -+ W transition. It is, however, also valid for the other transitions.
A schematic presentation of the reduction sequence under near-equilibrium conditions is presented in Fig. 3.12, together with the equilibrium vapor pressures of the volatile
W0 2(OHh. Under such conditions the tungsten is transported three times via the vapor
phase, as indicated in Fig. 3.13, before it is reduced to the metal. This behavior (which is
also valid for humid industrial conditions) is a fascinating peculiarity of the W-O-H
reduction system. Only during the early W0 3 -+ W02 .9 transition was no CVT observed.
Nucleation and Crystal Growth. When a new phase forms during the reduction
process, the first step is phase nucleation. If the structure of the new phase does not match
the host matrix, phase nucleation problems can occur (incoherent nucleation). The

BLUE OXIDE

W0 2 ,

W0 272

W0 2

d
10 I , m

FIGURE 3.8. Morphological changes associated with the hydrogen reduction of tungsten blue oxide under nearindustrial conditions (1000 0C); by courtesy of Wolfram Bergbau and Hiittenges.m.b.H.

FIGURE 3.9. Transition W02 .72 -+ W02 (left) and W02 -+ W (right). Note the gap which fonns between the reacting particles. It is characteristic for the CVT growth of the
phases. In case of tungsten, small single crystals are fonned.

99

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

0.0020

T = 1000C
-"-" - "- WO,
--- --- -WO U
W:t.12
- . - . - . - .W0
- 2
---- W

0.0015
~

o
N
o
;:

0.0010
35 vol % H20

0.0005

,...-

.'" ................ . ..... .

0;.....--:- ........ - .. _ ... "

0.0000

,:,,;,-,'-":---::I;;.-+----+------"I-------4

~....

20

40

80

60

100

Percentage of H2 0 (vol %)

FIGURE 3.10. Calculated W0 2(OHh partial pressure of different tungsten oxides at 1000 DC vs. humidity; at a
certain humidity (e.g., 35 vol% H20) the phase with the lowest W02(OHh partial pressure is the stable one (e.g.,
W) [3.38].

W transport via vapor phase

W deposition

W0 2 dissolution

tJ.p causing transport

by diffusion

FIGURE 3.11. Chemical vapor of transport of tungsten via the volatile (W02(OHh; schematic presentation.

CHAPTER 3

100

1::

10-'

WOz+W

W02(OH)2
vapor
pressure
(bar]

time of reduction

FIGURE 3.12. Changes in the [H20l/[H21ratio and W02(OHh-vapor pressure during reduction; 3-step chemical
vapor transport via W02(OHh; IOOOC.

-2

w
Reduction steps
FIGURE 3.13. Changes in O(W ratio during the CVT reduction (schematic).

formation ofW from W02 is a good example for this aspect. When W0 2 is reduced by wet
hydrogen, the reaction comes to a standstill already at comparatively low humidities,
although from a thermodynamic point of view the reaction should proceed even at much
higher humidities (see Fig. 3.14). Obviously, under these wet conditions no W nucleation
takes place. However, tungsten growth by CVT is possible, which can be easily
demonstrated by adding nucleation aids, such as metallic tungsten. In this case, a slow
further overall reduction occurs, but the W is deposited mainly on already present W
particles. Similar reactions take place on the walls of the reduction boats, where W is
readily deposited on the wall surface.
The strong influence of the humidity on the nucleation rate of the metal has an
important consequence for the technical process. Besides the reduction temperature, it is

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

W nucleation problems

-.::

co
.0
E

300
250

200

....

150

0
I

(J)
~

I/J
I/J

101

calculated

(J)
100
....Q.

~....

co

a.

50
0

650

750
Reduction temperature

800
(DC)

FIGURE 3.14. Humidity limits for the reaction W0 2 ---+ W, calculated from phase equlibria and experimentally
observed [3.38].

the dynamic humidity within the layer which determines the nucleation and growth
conditions. The lower the humidity during the W0 2 ---+ W transition (i.e., the lower the
powder layer and the closer to the top of the layer), the higher the nucleation rate and hence
the lower the growth of the individual grains, resulting in smaller average grain sizes. The
reverse is true for high humidities, where only few nuclei form, leading to a coarse powder.
Due to the varying growth conditions for the individual particles the powder always
exhibits a certain grain size distribution. Grain growth of the metal occurs, as described
above, by CVT through the formation of W02 (OHh.

Mutual Interactions of Oxide Properties and Reduction Conditions. Figure 3.15

[3.13] shows schematically the influence of different reduction parameters on tungsten
nucleation and growth. It is divided into three parts. In the first part, the raw material
aspects and the more industrial-technical aspects concerning the furnace parameters
are considered. The space dimensions range from meters down to centimeters. The
second part considers the powder layer itself, where the dimensions range from centimeters downward. Gas permeability of the powder layer and macrogas convection
dominate and determine the local events. The third part considers the birth of the
individual tungsten grains which happens in the dimensions near 1-100 11m. The microporosity within the oxide agglomerates as well as the local conditions for mass transport
(diffusion, microconvection), crystal growth, and nucleation determine the ongoing
chemical reactions.
Alternative processes. In the case of a transport-controlled process, the overall
reduction rate can be significantly increased if a continuous renewal of the gas in contact
with the solid oxide is readily achieved, i.e., when the oxide particles are not in close
contact with each other, as, for example, in a fluidized bed, or a laminar flow reactor, where
the reaction is virtually instantaneous. Under these conditions, however, the change in the
water removal rate will have a strong impact on the powder properties, in particular on the
average grain size. It is very unlikely that under such conditions the production of coarse
powder (i.e., the formation of small W single crystals of size ~ 10 11m) can be successful.
Thus the inherent ability of the powder bed to retard the water vapor within the layer and to
build up locally humid conditions is an important aspect in industrial powder manufacture,

102

CHAPTER 3

"

QUALITY OF II POWDER
-grain size and. shape
-grain s1ze distribution
-apparent dens! ty
-degree and size of agglomeration

=-

!:

....
!;

"

i ~
5" c

."

FIGURE 3.15. Influence of different reduction parameters on tungsten nucleation and growth [3.35]

requiring a variety of different powder sizes (ranging from submicron to 60 J.UD) and
specific size distributions. This has to be considered when discussing alternative methods
for W powder production.
The importance of the powder layer in the hydrogen reduction of tungsten oxides,
although long known, is still today sometimes not altogether understood and therefore its
influence is here, once again, briefly summarized.
The capability of retaining water vapor within the layer is to a certain extent the
reason why it can be used as a steering parameter for the final tungsten metal powder
properties. The water vapor retention capacity of the layer depends on the layer height and
its porosity. The layer height can be chosen within limits in connection with the
temperature. The porosity depends on the oxide powder properties. The macroporosity
is a consequence of the agglomerate size of the oxide powder (which is related to the size
of the crystallized APT particles prior to decomposition) and the microporosity is due to

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

103

the porosity of a single agglomerate (APT pseudomorph). Both parameters combine to

yield the overall porosity of the layer. It is noteworthy that due to morphological changes
during reduction (formation of different intermediate oxides by CVT) the porosity may
also change. The following rules are, however, valid:
The powder layer height is directly proportional to the water vapor concentration
inside the layer, and thus to the average grain size of the W powder (length of
diffusion path).
The layer porosity is inversely proportional to the water vapor concentration and
average grain size (diffusion rate).
As regards the influence on other powder properties, the reader is referred to Section 5.4
(industrial W powder manufacture).
In addition, it must be stressed here that this reduction process is also unique in so far
as it presents the possibility to produce single crystalline powder of a highly refractory
metal at temperatures between 800 and 1l00C (less than one-third of the melting
temperature) by CVT.

3.3.5. Influence of Foreign Elements [3.33-3.38]

Traces elements have considerable influence on the quality of the W powder. They
originate either from raw materials or are introduced during manufacture. They can,
however, also be added intentionally prior to reduction, if specific powder properties are
desired.
During reduction the foreign elements are either evaporated, or remain partially or
completely in the as-reduced powder. In general, the higher the reduction temperature and
the lower the humidity, the stronger is their evaporation. There are, however, exceptions to
this rule in case the oxides are volatile (for example, phosphorous). The prese!1ce of two or
more foreign elements at the same time may change the influence of a single element
drastically by forming more stable compounds. Foreign elements can be contained as
oxides, tungstates, silicates, phosphates, metals, or intermetallic compounds, either as
separate particles or incorporated into the tungsten grain. During reduction they can be
solid or liquid. In case of low concentrations, foreign element atoms can also be present in
solid solution.
Foreign elements may also influence the reduction process, either by altering the
reduction sequence and/or by changing the nucleation and growth conditions for the
individual phases. In particular, the latter aspect is of great technical importance. Most
interactions known today refer to the W02 ~ W transition. Certain trace elements
enhance the reduction rate during this stage (Li, Na, K, Mn, Fe, Co, Ni) while others
retard it (B, AI, Cr). Crystal growth can be significantly promoted (by alkali metals), but
also reduced (Co, AI, V, Cr), or even suppressed (B).
Not only is the grain size affected by foreign element traces, but also size distribution,
grain morphology, agglomeration, apparent density, and compactability of the W powder.
Several foreign elements can stabilize the formation 'of metastable ~-W, shifting the
transition temperature from about 630C to above 800 C [3.39]. Since all the stabilizers
have a very high affinity to oxygen (Be, AI, Th, Sr, Zr, B, P), it was argued, that these
elements form two-dimensional oxide compounds on the surface of the (3-w, and that the

104

CHAPTER 3

~- w -+ r:t.- W transformation may occur only after these compounds are decomposed by
hydrogen [3.40].
Elements or element combinations of industrial importance are discussed below.
Alkali Metals. Alkali metal compounds significantly enhance crystal growth during
reduction at both low (700 0c) and high (1000 0c) temperatures, and increase the reduction
rate. The higher the temperature, the more pronounced are these effects. Alkali additions
are used on an industrial scale to produce extraordinarily large grained W powder (see
Section 5.4).
The metals (Li, Na, K) as well as the tungstates and hydroxides (which form during
reduction) are liquid at the usual temperatures. No evaporation occurs below 800C but,
above 800 DC, a partial or total evaporation takes place, depending on the respective
element and the nature of the compound (tungstate, hydroxide, silicate, borate, etc.).
Lithium as Li 3 B03 , for example, or sodium as silicate are less volatile than LiOH or
NaOH. In contrast to Na and K, Li cannot be volatilized due to its high boiling point.
The effect of the addition of 0.1 mol of alkali metal to tungsten blue oxide is shown in
Table 3.1. It can be seen that the influence in grain growth is considerable. Moreover, there
are differences between the different metals which are mainly based on the properties of
the metal. At the end of the reduction process the respective alkali tungstate is reduced and
at 1000 C the metal is evaporated. The vapor pressure of Li is the lowest of the three
metals and therefore the amount remaining is the highest.
The acceleration of the reduction process is attributed to a catalytic action of the
liquid alkali compounds on the reaction of W0 2 with water vapor forming the volatile
oxide hydrate. This step (evaporation) is said to be the rate-limiting stage in the chemical
vapor transport reaction [3.35]. The overall reaction (3) can be described as the sum of
reactions (I) and (2), according to the following:

(1) K2 W04 (I) + 2H 20(g) -+ W0 2(OH)2(g) + 2KOH (I)

(2) W02 (s)

+ 2KOH (I) -+

K2 W0 4 (1).+ H 2(g)

(3) W0 2(s) + 2H20(g) -+ W0 2(OHh(g) + H2(g)

Equimolar concentrations ofLi, Na, and K have about the same effect on grain coarsening,
which supports this theory. Nevertheless, the presence of a volatile alkali-tungsten-oxygen
compound has also been made responsible for grain coarsening [3.41].
TABLE 3.1. Effect of Alkali Compounds on Tungsten Crystal Growth (addition of 0.01 mol alkali
hydroxide)
Element
Temperature (0C)
Deagglomerated W
Powder particle
Size (11m)
Concentration of
doping elements in
% of addition

Li

Undoped

750
reduction
incomplete

1000
II

750
5.4

84

Na

1000
46

68

750
3.5

31

1000
25

750
2.3

80

1000
24

37

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

105

Aluminum. Aluminum additions severely disturb the W crystal growth, leading to the
build-up of AIOx phases in the tungsten lattice. The reaction rate is significantly decreased,
and consequently the nucleation is favored, resulting in a decreased average grain size.
About 50 J.lg/g are built-in. At higher concentrations, AIOx phases are deposited onto the
crystal surfaces. The growing W crystals are rough-faced during the incorporation period,
but then they become smooth and finally have a completely regular shape in the as-reduced
condition (Fig. 3.16).

FIGURE 3.16. Influence of aluminum on the morphology of tungsten particles; at first, cubic crystals are fonned
with rough surfaces (upper figure) which become smooth during further crystal growth (lower figure); 200 ppm Al
added to the oxide; 1000 C.

106

CHAPTER 3

The influence of Al can be completely suppressed by the addition of sodium

compounds. In this case all Al remains outside of the tungsten crystals as an Na-AI phase.
Iron, Nickel, and Cobalt. Traces of iron, nickel, or cobalt may originate from boat
and furnace materials. They are also present in large amounts in recycled oxides of heavy
metal scrap. High concentrations lead to an activated sintering of the powder and thus
significantly alter the powder properties, in particular the agglomeration. During the early
part of reduction iron and cobalt form tungstates, which are then reduced to intermetallic
phases (Fe7 W6, C07W 6). The latter phase can be incorporated into the tungsten grain.
Nickel is present in the metallic state. Small amounts of Ni (200 /lg/ g) lead to a
characteristic whisker growth of W metal (Fig. 3.17). If such whiskers are present in a
regular powder charge, they are the first witnesses of an enhanced corrosion of the
reduction boat (made of Inconel).
Phosphorus. Phosphorus is a detrimental impurity in most sintered tungsten materials since it commonly occurs in all technical powder grades in the range of 10 to 20 /lg/ g.
It is a strong p- W stabilizer and, as such, responsible for enhancing the agglomeration of
fine-grained W powders under dry reduction conditions. Phosphorous additions to the
tungsten blue oxide lead to the formation of phosphate "bronzes" of the general formula
P4 0 g(W03hm, with m between 2 and 16.
AKS or NS Doping. This type of oxide doping is used for the production of non-sag
filaments in incandescent lamps. The doping compounds are potassium silicate and
aluminum chloride. The term AKS refers to the initials for the German names of the
three elements Aluminium, Kalium (potassium), and Silizium (for details see Chapter 5.4).
During hydrogen reduction of AKS-doped tungsten oxides, incorporation of the dopants
occurs by CVT overgrowth during the W0 2 ~ W transition in the form of aluminosilicates. Less than 10% of the dopants added remain in the W powder after an HCl/HF-acid
leaching. On sectioning the W powder particles the dopant phases appear as holes in the W

FIGURE 3.17. Tungsten whisker. Whiskers fonn if small amounts of nickel (200-1000 ppm) are present; higher
amouhts lead to an activated sintering of the W particles.

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

107

matrix (Fig. 3.18). The mechanism behind this incorporation is still under discussion
[3.42] but seems to be linked to the intermediate formation and transformation of
potassium tungsten bronzes.

3.4. REDUCTION OF TUNGSTEN OXIDES BY CARBON OR

CARBON-CONTAINING COMPOUNDS. [3.2,3.43,3.44]
The reduction of tungsten oxides by carbon or carbon-containing compounds can be
easily performed. Statements about the starting temperature for the reaction between W0 3
and solid carbon (carbon blacks, graphite) vary in the current literature between 655C
and 783 DC. Differences in W0 3 and C properties (particle size of the powders, preparation
history, crystallinity, etc.) as well as in atmospheres may be responsible for that. The
temperature range coincides with the beginning of self-conductivity and sublimation of
W0 3. Carbon monoxide starts to react with W0 3 at 535C (reduction pressure 1 bar,
PcoJpco equilibrium ratio 8.52) [3.45].
The reduction/ carburization sequence can be assessed from so-called predominance
area diagrams (Kellogg diagrams), which can be derived from free energy of formation
data ofthe compounds occurring in the W-C-O system (W03, W0 2.9 , W02.72, W0 2, W,
W2C, WC, CO, CO 2) [3.45-3.47]. Such a diagram is shown in Fig. 3.19 for llOOC
[3.46]. It predicts that at llOOC the reduction/carburization proceeds via the oxide
phases (W0 2.9 ~ W0 2.72 ~ W0 2 ~ W ~ W(2)C)' with the possibility of a direct
carburization of W0 2.72 and W0 2.
The reduction kinetics are based on an adsorption-autocatalytic reaction. Reduction
enthalpy and temperature are linked by a linear relationship. The calculated activation
energy is about 75 kJ . mol- '. Carbide formation occurs only when all oxygen is removed.

FIGURE 3.18. Polished section of a large tungsten particle showing several holes in the matrix that formed on
polishing; these holes were originally filled with potassium aluminosilicates.

CHAPTER 3

108

In PCOI PC~
4.7

2.7

0.7

-1.3

-3.3

-5.3

-38

-34

-30

-26

-22

-18

-1
-2

In a(C)

-3

-4
-5

In

P02

FIGURE 3.19. Kellogg diagram for the W--C-O system at 1100C [3.46].

The apparent activation energy at the beginning of the reaction is 121 kJ/mol W0 3
and increases during further progress to 205 kJ . mol- I W0 3 , due to the fact that the
reaction becomes diffusion-controlled (diffusion of oxygen to the particle surface).
The assumed reaction mechanism is as follows: direct reaction of W0 3 and C at
contact zones and, for the main portion, sublimation of W0 3 to the carbon particles,
reduction at the surface of the carbon particles, and desorption of CO and CO 2 .
Although the reducing agent-carbon-is very cheap, and carbon reduction was the
early basis of tungsten powder production, so far none of the numerous carbothermic
procedures has been established in the production of pure tungsten.
The reasons for that are:
The carbon balance is difficult to control; furthermore, the metal forms carbides.
Although the theoretical carbon consumption to reduce W0 3 to W is 15.5 wt%,
only 12-14 wt% are normally sufficient at 1200 0 C to 1400 0 C because the reaction
does not only proceed via W0 3 + 3C --+ W + 3CO but also to a certain extent via
2W03 + 3C --+ W + 3C02 . Therefore, in most cases the reduction is either
incomplete or the metal powder is contaminated by carbides.
Carbon is always a source of increased tungsten contamination because it contains
impurities like Ca, Si, Fe, S, and P.
Another disadvantage of the carbothermic reduction is the limited possibility of
steering the grain size by varying the reduction conditions as compared to the
hydrogen reduction process.
Carbothermic reduction of W0 3 or ore concentrates is of technical importance in
melting metallurgy-preparation of ferrotungsten, melting base, and cast carbide (see
Chapter 8).
The reduction of W0 3 by carbon in the solid state, however, has gained technical
interest and importance in the case when the desired final product is tungsten carbide. This

109

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

process is named "Direct Carburization" and the basic reaction is described by the
following equation:

+ 3CO

W0 3 +4C -+ WC

In contrast to the conventional procedure for WC production-W03 reduction by

hydrogen followed by W + C mixing and carburization-this method allows the formation of WC in only one step, while, carbon acts as carburizing and as reducing agent.
The above equation is only of theoretical value because, besides the main reaction, the
following reactions also occur:
W0 3 + ~C -+ WC

+ ~C02

C + CO 2 -+ 2CO

Consequently, in this procedure, an uncertainty in the carbon demand is given. This means
that, depending on the operating conditions, the carbon content of the final product may
vary significantly.
Under industrial conditions, the process can be performed in two steps, one following
the other [3.44,3.48]. The first step-reduction-affords the complete exclusion of
hydrogen and is performed in nitrogen atmosphere. In the presence of hydrogen, water
vapor would form which generates tungsten crystal growth and consumption of carbon
(C + H20 -+ CO + H2). This carbon consumption is not only restricted to the carbon
black or graphite used as reagent (disturbance of the carbon balance), but also to the
furnace carbon tube (shortening its lifetime). The second step-carburization-needs a
hydrogen atmosphere, which supports the carbon transport via methane to the tungsten
particle surface (see also Section 3.6).
Accordingly, the direct carburization process starts with a pelletized mixture of W0 3
and graphite passing step-by-step two rotary furnaces, the first operating at 950-1250 C
under nitrogen (reduction) and the second at 1400C under hydrogen (carburization).
Direct carburization is advantageous mainly in the production of submicron tungsten
carbide powders. The idea behind the process is to circumvent the expensive hydrogen
reduction of fine tungsten powder by excluding water vapor during the reduction step,
which is responsible for grain growth reactions and low capacity. Roughly speaking, the
grain size of the final carbides is related to the size of the intermediate oxides [3.48].
Only recently was a new carbothermal reduction process developed in which the WC
is synthesized by a rapid carbothermal reduction of tungsten oxides in a vertical graphite
transport reactor (RCR entrainment process) [3.49]. Rapid heating of the W0 3 /C mixture
driven by thermal radiation allows conversion of the mixture into a carbide precursor
(WC I _x ) within very short reaction times (a matter of seconds). In a second step, additional
carbon is added to the carbide precursor to form a mixture, which then undergoes a second
heat treatment to convert the precursor into substantially pure we.
Carbothermal reduction of tungsten oxides with carbon monoxide [3.47], or gas
mixtures of CO/C02 , CO/H 2 , CH4 /H 2 [3.50], C2H4 /H 2 , and C2 H4 /H 2 [3.51], as well as
by reaction between metal oxide vapor and solid carbon [3.52] have recently attracted
attention for producing high surface area tungsten carbides (up to 100 m2 / g), for use as
catalyst (see Section 10.4), and for nanophase WC/Co composite powders (see also
Section 9.2.1.4) [3.53].

CHAPTER 3

110

3.5. REDUCTION OF TUNGSTEN HALIDES

3.5.1. Introduction [3.2,3.54-3.57]

The direct chlorination of tungsten raw materials and subsequent reduction of the
chloride to tungsten metal would offer a much simpler and cheaper production path in
comparison to the rather complicated one via hydrometallurgy to APT and a two-stage
reduction to tungsten metal. Although several promising attempts have been made, it is
interesting to note that chlorination chemistry of tungsten has not found its way into largescale production, unlike other refractory metals such as zirconium, hafuium, titanium,
niobium, and tantalum. None of the processes mentioned below survived for long or are in
use today.
Chlorination of scheelite or wolframite ore concentrates according to the equations
2CaW04
2FeW04

+ 3C + 6Clz --+ 2CaCIz + 2WOCl4 + 3COz

+ 3C + 7Clz --+ 2FeCl3 + 2WOCl4 + 3COz

and subsequent treatment of the chlorides offers a selective separation of most other
chlorides formed with the exception of molybdenum but bears certain disadvantages:
It mainly yields in the formation of WOCl 4 and not in the desired WCk
Therefore, an additional conversion step from WOCl4 to WCl 6 is necessary; the
separation from other chlorides can be done only in the WCl6 stage by fractional
distillation.
Calcium as well as iron or manganese form molten chlorides at the reaction
temperature, causing difficulties in reactor design.
Finally, hazards associated with HCI, Ch, COCh, and chlorides as well as
environmental considerations are further limiting factors.
In order to circumvent the formation ofWOCI4 , the following path was described and
was even more complicated and more expensive [3.2]. A low melting tungsten silicide was
prepared from scrap or ore concentrates in advance by melting with aluminum or silicon in
an arc furnace at 1800 to 2000 0c, The alloy settles from the slag and can be separated.
During the subsequent chlorination of the alloy WCl6 and SiCl4 form, which can be easily
separated due to their different condensation temperatures.
Based on a patent of Jonsson and Svanstrom [3.56], chlorination of tungsten scrap
and subsequent fractional distillation to purify the crude WCl 6 was used by Sandvik Co.
The most advanced installation was made in the USA in the early 1970s. L. Ramquist
of the Axel Johnson Institue of Industrial Research in Sweden described an on-line,
computer-controlled industrial-scale process for tungsten powder production from ore
concentrates and scrap. The process included chlorination, treatment of the chlorides
formed to prepare pure WCI 6 , and reduction by hydrogen to tungsten powder. Based on
this concept, P. R. Mallory and Co. Inc. in the USA installed a new plant, which was,
however, closed shortly after start-up.
The main complication in producing larger amounts of tungsten powder by hydrogen
reduction of WCl6 is the fact that the powder surface is covered by chlorine (and HCI),
which is gradually desorbed during the further handling. Due to the fine powder grain size,

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

111

the specific surface is quite high and the adsorbed amount of chlorine is accordingly high
too. The consequences were hazard to workers as well as corrosion problems. The chlorine
layer at the tungsten particle surface is one of the reasons why these fine powders are not
pyrophoric besides their perfect crystal structure.
The two important halide compounds which are technically in use today are WF6 and
WCI6. They can be easily prepared by direct combination of the elements. High purity is
achieved by using pure starting materials. The total exclusion of moisture is very
important.
WF6 is the predominant precursor for the production of tungsten thin layers in
microelectronics via CVD. Because of its electromigration resistance and low resistivity of
5.6 J.1!l. cm, tungsten is and will be the most promising material for integrated circuit
metallization.
WCl6 is the starting material for the preparation of extremely fine grained tungsten
and tungsten carbide powders (small scale only), for the deposition of tungsten on
substrates to prepare intricate forms by CVD, and for the preparation of high-purity
tungsten whiskers of high strength. Furthermore, it is also in use, like WF 6, as a precursor
for thin layers in microelectronic devices.
CVD tungsten exhibits several advantageous properties, such as high purity, resistivity to grain growth (low energy grain boundaries and absence of stored strain energy),
near-theoretical density, and high ductility. Depending on deposition conditions and
substrate properties, different types of layers can be produced; such as single crystal
layers, layers of a determined crystal orientation, columnar crystallites rectangular to the
substrate surface, and so on. The crystal orientation can be influenced by the WF6/H2 ratio
and the temperature: low temperature and low ratio favor (100) orientation while high
values favor (111) orientation.
3.5.2. Reduction of WF6 [3.58]

Today, high-purity WF6 is available at low cost. This makes it the most important
tungsten source for thin layers in microelectronic devices prepared by CVD. Different
reduction agents can be used. Hydrogen, silicon, and silane are the most common reducing
agents for WF6 in integrated-circuit metallization processes. Moreover, reduction can be
achieved too by ShH6, B2~' PH3 , SiH2Ch, and GeH4 at much lower deposition
temperatures. Reduction with hydrogen or silane allows the deposition of tungsten in
the temperature range 300-500C. Reduction by silicon is important to deposit tungsten
selectively on Si surfaces.
The WFoSi System. Depending on temperature, the following two reactions are
possible:
below 400C:

2WF6 + 3Si ~ 2W + 3SiF4

above 400 C:

WF6

+ 3Si ~ W + 3SiF2

Accordingly, twice as much Si is consumed at higher temperature. Reduction ofWF6

by a silicon substrate allows for selective tungsten deposition in the contact region of
highly doped silicon. No deposition occurs on insulator surfaces such as Si02 or ShN4 . It
is a replacement reaction which is highly sensitive to the substrate pretreatment. The

112

CHAPTER 3

reaction is very fast and stops as soon as a critical tungsten thickness is reached (1015 ~). The thickness depends on the reactor setup, the amount of oxygen on the silicon
surface, and pretreatment. It is assumed that a continuous tungsten layer of a certain
thickness will inhibit silicon diffusion to the tungsten surface, consequently the penetration
of WF6 to the WlSi interface is stopped.
The WFoH2 System. The reaction WF6 + 3H2 ---+ W + 6HF is possible from 300 to
800C and is also used in thin-film production. There are several disadvantages in
comparison to the other reduction methods. The HF formed during the reaction may cause
defects, like encroachment or wormholes. The layers show poor adhesion on native Si02,
which is always present on Si. Therefore, tungsten is not directly deposited on Si but on a
bilayer. One layer provides an ohmic contact with Si, and the other acts as an adhesion
promotor for W.
The WFo SiH4 System. The advantages over the WF6-H2 system are: the deposition
rate is higher, the film is smoother, good adhesion prevents encroachment and wormhole
formation and suppresses silicon substrate consumption. A disadvantage is the dependence
of the deposition on the Si~/WF6 ratio and on the temperature. Depending on these
parameters, either Wor W silicide can be deposited and blanket or selective deposition on
a phase can occur. A ratio up to 1.6 yields in selective W deposition and higher in blanket
deposition. As long as the ratio is smaller than 1, tungsten is deposited as r::J.-W and,
between 1 and 3, as ~-w. The amount of silicon incorporation increases with temperature.
The dependence of film properties on deposition conditions and on the ratio is given in
Table 3.2.
The reaction rate of the WF6 with silane is very fast. The hazard of an explosion may
be overcome by low total pressure, low silane partial pressure, and fast flow rate. Silane
reduction is only used for a short duration to initiate W nucleation and is followed by
WF 6/H2 reduction.
The WFo<JeH4 System. Deposition at 300-400C yields in layers of higher
resistivity (approximately 200 J.1!l. cm) which consists of ~-W stabilized by a Ge
incorporation of 10-15 at"lo. Above 400C, the ~-W and Ge concentrations decrease
and pure r::J.-W is deposited with less than 1 at"10 Ge. The layers show good adhesion to Si
and Si02 with no encroachment and wormholes.

TABLE 3.2. Deposition Conditions and Tungsten Film Properties using WF6 Precursor and
Different Reducing Agents [3.58]
Temperature
("C)

Carrier and
reducing agent

Pressure
(torr)

420--800
260-400
288-403
300
250-550
500-570
350-550
<400
>500

H2/Ar, H2He
H2/Ar
WF6:H2= 1:15
Sif4
Sif4
H2/Ar
H2/Ar
Gef4
Gef4

1-10,760
0.1-1
0.2-10,760
UHV
Vacuum

3.5
0.5-4.5

1.5
15

Substrate
Si, Si02, Sapphire
Si, Si02, TaSh
Si (100)
Si (100)
Si (100)
GaAs

InP
Si, Si02
Si, Si02

Film resistivity
(J.l. em) and purity

6-15; 100% r:t-W

10-15
6
30
7.5-15; low in Si
50; 2% C, 0
55
200; ~-w, 10-15% Ge
10; r:t-W, 1% Ge

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

113

3.5.3. Reduction of WCl6 by Hydrogen [3.58-3.64]

The reduction of WCl6 by hydrogen is the basis for preparing tungsten powder by
CVD. The reaction may be performed in low-pressure reaction chambers, in chlorine/hydrogen flames, plasma jets, or fluidized bed reactors. These methods offer the
possibility to prepare very fine grained tungsten powders with close grain size distribution
and smooth particle surfaces. Grain sizes are in the range of 10-100 nm and can be
regulated by the reaction conditions. With the exception of plasma jet originated samples,
these powders are not pyrophoric in contrast to those reduced from oxides by hydrogen
and having comparable grain size. This can be explained, on the one hand, by the more
perfect crystal constitution and smooth surfaces of powders produced by CVD. On the
other hand, the adsorbed chlorine at the powder particle surface prevents a rapid reaction
with oxygen or water vapor. The fact that nanosized powders can be produced which are
not pyrophoric is a big advantage, but, on the other hand, the adsorbed chlorine has until
now prevented the application of these production methods on a bigger scale.
A typical procedure using a low-pressure reaction chamber is as follows: WCl 6 is
heated to evaporate and is carried to the reactor by a stream of argon. The argon flow rate
regulates the WCl 6 input. By varying the WCl 6 feed rate, the hydrogen flow rate, and the
temperature (600-920C), different average particle sizes can be obtained (16-46nm).
The principle in preparing such small grains is that, by controlling the dilution, only a
minimum number of atoms and molecules involved in the reaction may form agglomerates.
The preparation of 30-60 nm tungsten powder by feeding WCl 6 into a hydrogen/
chlorine flame burning in a CVD reactor yields also spherical particle shape and close
grain size distribution. Variations in flame temperature, flow rate of reactants, and
residence time of particles allow grain-size regulation.
Fine CVD-derived tungsten powders have been carburized successfully and further
processed to cemented carbides (Axel Johnson Process).
Single crystals can be prepared by reductive dissociation ofWCl6 at a tungsten wire.
The same method can also be used to produce bigger tungsten rods (200 mm length and
6 mm diameter).
An advantage of WC1 6 for thin-layer production in microelectronic application is that
Clz and HCl cause much weaker encroachment and corrosive effects on Si and
semiconductors in contrast to F2 and HE A disadvantage is the lower vapor pressure of
WCl 6 compared to WF 6 .
The WCI 6-H2 system works at 600-900C. Lower temperature, high total pressure,
and high WCl6 concentration favor the process and the selectivity. The chloride process is
more affected by the oxide layers on substrates than is the fluoride process (slower etching
of Si02). This is why thickness reproducibility is worse. Also, film composition shows
poor reproducibility (cx- and ~-tungsten).

3.5.4. Reduction of WCl 6 by Carbon-Containing Reagents [3.65-3.67]

The basic idea in these reactions was always to produce fine and stable WC powder
(0.01-0.1 11m) in one step starting from WCI 6 , and to circumvent the reduction to W by
hydrogen and the subsequent carburization. Chemical vapor reactions are ideal for

114

CHAPTER 3

producing fine powders consisting of discrete particles. Particle size distribution can be
kept quite close and agglomeration is low.
For the system WCI6-CHc H2, the following reactions can occur:
(1) WCI6 + CH4 + H2 -+ WC + 6HCI
(2) WCI6 + !CH4 + 2H2 -+ ! W2C + 6HCI
(3) WCI6 + 3H2 -+ W + 6HCI
(4) W +CH4 -+ WC +2H2
(5) W+!CH4 -+!W2C+H2
(6) WCI6 + !H2 -+ WCIs + HCI
(7) WCI 6 + H2 -+ WCI4 + 2HCI
(8) WCI 6 + 2H2 -+ WCI 2 + 4HCI
(9) WCI 6 + ~ H2 -+ WCI + 5HCI
(10) CH4 -+ C + 2H2
The log Kp values (equilibrium constants) for the direct carburization to W2C as well
as to WC for equations (1) and (2) are quite high (>20). Also high is the log Kp value for
the hydrogen reduction of WC16 to W (17-20) (equation 3); while the log Kp for the
carburization ofW by C~ (equations 4 and 5) to W2C or WC, respectively, are low (2-4).
All lower tungsten chlorides can be easily reduced by hydrogen. The high equilibrium
constants for reactions 1, 2, and 3 are why direct carbide formation and hydrogen reduction
take place at the same time. W formed by hydrogen reduction is subsequently carburized
by CH4 . This step can be the rate-determining stage in the formation of the carbide,
particularly at comparatively low temperatures (::; 1400 QC). Moreover, the high equilibrium constants are a prerequisite for the powder formation. At low temperature mainly
W2C is formed and above 1400 QC, WC is the only product. Powder particle size can be
influenced by the reaction conditions. Powders have been produced in high-temperature
reactors, tube reactors, and hydrogen arc plasmas. Typical grain sizes are between 5 and
100 nm. The main and important drawback of the direct formation of WC via the vapor
phase is the difficulty in producing monophase WC without the presence of either carbon
black (equation 10), or W2C, or both of them. This is only possible by proper control of the
vapor phase, the temperature, and the exposure time, and so far has not been successfully
mastered in large-scale production.
Besides methane, other hydrocarbons were used as reducing and carburizing agents,
such as C3HS and C2H4 . They do not offer any advantage.

3.6. REACTION OF TUNGSTEN WITH CARBON OR CARBON-CONTAINING

COMPOUNDS (CARBURIZATION)
This section deals primarily with the carburization of tungsten powder, the most
widely used process for producing WC on a technical scale. Also, the carburization
performed in melts is discussed briefly owing to historical reasons and the production of
hard facing alloys. Finally, the formation of coarse tungsten carbide crystals in auxiliary
melts ("Menstruum WC)" is discussed.

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

115

3.6.1. The Carburization of Tungsten Powder by Solid Carbon

Tungsten reacts with carbon (carbon black, graphite, CVD-produced carbon layers,
etc.), as well as with a variety of carbon-containing compounds to form carbides (W2C,
WC, "(-WCI_x) at elevated temperatures. Indications of a lower temperature limit for these
reactions are extremely widespread in the current literature. For example, carbides were
detectable after long-time high-energy ball milling of a W + C powder mixture (mechanosynthesis) [3.68]. The overall temperature in that procedure should have been somewhere
between room temperature and 100C. However, no knowledge is available about the local
temperature when a ball hits a W + C increment. Also, at room temperature, W2C
formation was observed in sputter-deposited W/C multilayers [3.69]. Thin layers ofW and
C form W2C and WC at 800C. Technically, the usual carburization temperatures for
W + C powder mixtures range between 1050 C and 2100 DC, depending on the powder
particle size. Bulk tungsten, like sheet or wire, embedded in carbon powder starts
carburizing at 1550C, while, single-crystal wires react only at 1900C [3.70].
This wide range in temperatures can be understood because the carburization of
tungsten is influenced by the properties of the starting materials (W and C) as well as
several other parameters.
Properties of Tungsten. In the case of bulk tungsten, one has to distinguish between
polycrystalline metal, which carburizes more easily than monocrystalline metal. This is
because the diffusivity of carbon in tungsten is strongly increased by lattice defects.
For tungsten powder, the grain size is determining. Smaller particles carburize more
readily than larger particles, due to their higher surface area. The rate of carburization is
inversely proportional to the square of the nominal particle size [3.71]. Since the powders
always consist of a distribution of particles, the larger particles determine the time for
complete carburization. The coarser the tungsten particles, the higher is the temperature
required for complete carburization and the longer is the time required [3.72].
Furthermore, the purity of tungsten has an influence on the carburization. Certain
impurities, such as cobalt, accelerate the carburization, and others, like the alkali metals,
retard it [3.73].
Properties of Carbon Black/Graphite. The fineness, chemical reactivity, and purity
influence the carburization process. The finer the powder, the better the contact with the
tungsten particles and the higher the reactivity. Regarding the impurity level, artificially
produced graphite will be advantageous owing to its higher purity. But this is only
important in regard to harmful impurities like alkali metals, sulfur, calcium, magnesium,
etc. In regard to hydrogen, carbon black is much more impure than graphite. Hydrogen in
carbon black is contained chemically bound to carbon. These bondings are relicts of the
raw material (hydrocarbons), C-H bonds which have not been cracked during the burning
process. The hydrogen promotes the carburization, as described below. Moreover, this
hydrogen content of carbon black makes it possible to carburize W + C mixtures in big,
closed graphite containers by induction heating in an open furnace without any specially
provided protective atmosphere.
The Atmosphere. Industrial carburization of tungsten powder is commonly carried
out in hydrogen or hydrogen-containing gas mixtures. Hydrogen reacts with the solid
carbon in the W + C mixture to form methane (at temperatures ~1600C) or acetylene
(> 1600 0C), which transport the carbon in tungsten-carbon black mixes over larger

CHAPTER 3

116

distances. At the surfaces of the reacting particles (w, W 2C, WC), the hydrocarbons are
dissociated (cracked) and the freed carbon atoms diffuse into the particles. The diffusion of
the carbon through the carbide phases is the rate-determining step of carburization.
Reactions also occur with the carbon parts of the furnace and the graphite boats, but
these are of minor importance. Excessive carbon losses occur in flowing hydrogen if the
carburization is carried out in open graphite boats. These losses can be prevented by using
graphite covers, thus forming a quasi-stationary atmosphere above the W + C powder
mixture.
Carburization in vacuum, argon, or helium is always very slow and incomplete. The
carbon transport in this case is only possible via surface diffusion, and this is a very slow
process to overcome larger distances [3.71]. In this case, the supply of carbon to the
surface of the reacting particles is becoming rate-determining. Carbon monoxide may also
act as a carbon carrier.
Availability of Carbon at the Surface of Tungsten Particles. If we consider a tungsten
surface covered with carbon particles, it will depend on the particle size of the carbon
source, whether the surface is densely or loosely covered. Finer powder will result in a
more dense layer. On' the other hand, finer powders show a higher tendency to form
agglomerates. Therefore, proper mixing of the two components, which guarantees the
break-up of these agglomerates, is important (long-time mixing in ball mills). In addition,
to ensure a dense and close contact of carbon particles to the tungsten surface, compacting
prior to the carburization can be applied.
Upon heating, only the carbon atoms in contact or in the near-neighborhood react
with tungsten. All others must to be transported to the tungsten surface, either by surface
diffusion (slow) or by chemical vapor transport (CVT), i.e., formation and subsequent
dissociation of methane (fast). Thus, during the initial stage of the carburization, the
availability of carbon atoms is quite high and becomes less later.
Figure 3.20 shows the progress of carburization under different atmospheres with
time for a lO-~m W/carbon black powder mixture (containing 6.2 wt%C) on heating to
lSOOC, and then holding for l.5 hours [3.74]. Under hydrogen or carbon monoxide

heating up: RT _

1800 C

Isothermal hold

stoichiometric WC
.... 6

i5

.to .. _________ "- __ _

,. 'V

.
C

.84
1\1

U3

":ci
C

81
H

0
0

0.5

1.5

2.5

lie

H... hydrogen

C...carbon monoxide

.to

V..vacuum

Carburlzation time (h)

3.5

4.5

FIGURE 3.20. Caburization progress with time under different atmospheres [3.74].

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

117

(50 mbar), the carburization was completed on heating the mixture to the isothermal hold
(combined carbon 6.13 wfllo), while, under vacuum condition 0.2 mbar), even after 1
hour at 1800C, about 1.3 wt% of the carbon remained unreacted. In the latter case there
was evidently no further carbon supply to the particle surfaces.
The Carburization Process. Unfortunately, only scarce data about diffusion of carbon
in W, W2C, and WC are available in the current literature. The diffusion coefficient of
carbon in metallic tungsten ranges from 9 x 10-8 cm2 . S-I at 1500C to
4.5 x 10-6 cm2 . S-I at 2400C [3.75]. Activation energies are stated between 37.8 and
53.5 kcal mol-I [3.69,3.75-3.78], depending on temperature, tungsten quality, and
method of measurement. Diffusion coefficients for carbon in W2C were determined as
8.1 x 10- 10 cm2 . S-I at 1526C, and 5 x 10-6 cm2 . S-I at 2400C, based on layergrowth measurements [3.79]. The respective activation energy is cited between 100111kcalmol- 1 [3.79-3.81]. No diffusion coefficients were found for the diffusion of
carbon through WC, but the activation energy is given as 58 kcal mol- 1 [3.71] which,
however, seems comparatively low.
The diffusion of carbon in tungsten and its carbides proceeds via a vacancy
mechanism rather than via the grain boundaries. This is because the carbon atoms are
located in the interstices between the metal atoms and diffuse by squeezing between the
host atoms to move from one interstitial site to another. In the hexagonal carbide phases of
densest packing, the octahedral interstices are either half filled (W2C) or completely filled
(WC). The anomalously low WC layer-growth rate, as compared to W2C or other
transition metal monocarbides, might therefore originate in the fact that WC exhibits a
narrow homogeneity range, and all sites on the carbon sublattice are occupied [3.79].
The Carburization Sequence. As a consequence of the high carbon supply in the
initial stage of carburization, a thin layer of WC is always formed at first [3.72]. The
carbon transport from the outer to the inner surface of that layer is then only possible by
diffusion through it. The layer is polycrystalline, even when the tungsten part or particle
was a single crystal. The smooth interface indicates that preferentially fast grain-boundary
diffusion does not occur. The further carbon supply to the interior is insufficient for WC
formation, and consequently W2C forms. No significant growth of the WC layer occurs,
because all the incoming carbon is consumed for W2C formation, building up a second
layer inside the particle. This layer grows in the radial direction and comprises large
crystals. The carbon diffusion in W2C is faster than in WC. Consequently, during this
stage, only the W2C layer is growing while the WC layer thickness remains about constant.
After all the tungsten metal in the interior is transformed to W2C, the WC shell starts
growing further in the radial direction until the whole particle consists ofWC. Diffusion of
carbon through the growing WC shell is the rate-determining stage of the carburization
process.
Figure 3.21 shows the change of the product composition with carburization time in
hydrogen for a tungsten--carbon black mixture at 1119 DC. After only 10 minutes at this
temperature most of the metal was transformed to W2C, which was then only slowly
converted to WC. This result is characteristic for the temperature range of 1050 to 1850 DC
and W particle sizes of 1.3 to 20 Ilm [3.71]. It is remarkable that W2C forms during
carburization even at 900C, which is well below its eutectoid decomposition temperature
of 1250 dc.
Simultaneous with the carburization reaction, a diffusion of impurity element traces
takes place. It can be assumed that these atoms migrate preferably along grain boundaries

118

CHAPTER 3

100
80
II

8'

c
II

...

60

II

Co

40

~C

01

'ij
~

20

20

.,,---

60
Time in min

40

80

100

FIGURE 3.21. Reaction products from nominal 1.30 ~m tungsten in W-C powder mixtures heated in H2 at
1119C for different periods of time [3.69,3.71].

to the surface, where they evaporate. The driving force for that process is the concentration
gradient between the clean surface and the impure interior. Hence purification of the grain
boundaries takes place, enabling their migration. The WC crystals grow and, depending on
temperature and time, finally single-crystal WC particles or particles consisting of some
very large WC crystals form by coalescence.
The average grain size of tungsten powder to be carburized for cemented carbide
production varies between submicron and 50-100 11m. Time and temperature, according to
the diffusion processes discussed above, must be adjusted in relation to the grain size.
Finer powders up to 211m can be carburized quantitatively at 1400 C and heating periods
of 1-2 hours. In order to keep the heating time at about the same length also for coarser
powders up to 10 11m, temperatures between 1600 C and 1700 C must be applied, and for
very coarse tungsten powders (>20l1m) 1800-2000C is necessary.

3.6.2. Formation of Tungsten Carbides in Melts

This classical production method, first used by H. Moissan in 1896, is a fusion
process in electric furnaces. Either W2C or WC-W2C mixtures can be produced. The latter
containing 3.5-4.0% C are technically important as castings and hard facing alloys and are
produced in large-scale operations (see Chapter 9). Mixtures of tungsten powder and
carbon are heated in graphite tube furnaces up to 3000-3200C. The melt is cast either
directly into wear parts or onto cold plates. In the latter case, the material is then ground to
powder. The usual composition is 70% W2C and 30% We.

3.6.3. Preparation of Coarse Tungsten Carbide Powder in an Auxiliary Melt

The solubility of WC in iron metal melts (Fe, Ni, Co) is relatively low. Supersaturation of a respective melt by W and C results in crystallization of WC. On slow

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

119

cooling, well faceted, big crystals (100 11m up to several mm) are formed. In comparison to
the above-described WC particles formed by solid-solid or solid-gas reaction, these
crystals are by far more perfect. In regard to the technical importance of the process, see
Section 9.2.1.4.

3.6.4. Preparation of Tungsten Carbide in Salt Melts [3.82]

Tungsten dissolved in a NaClfNaF halide melt, which is used to extract tungsten from
ore concentrates, can be crystallized as WC by sparkling the melt with carbon monoxide.
The advantage in comparison to the above-described method is that no acid is necessary to
dissolve the auxiliary "bath" material. Only water is needed.

3.7. CHEMISTRY OF AQUEOUS TUNGSTEN SOLUTIONS [3.83-3.88]

A characteristic feature of the tungstate ion which only occurs monomerically in
alkaline or neutral solutions is its tendency to form condensed, complexed isopolytungstate
ions in slightly acidic medium. The ability of polycondensation is common for group 5 and
6 elements and is more pronounced for elements of higher atomic mass within the same
group. The dimeric dichromate ion is a typical example.
Different techniques have been applied to elucidate the tungstate species forming in
neutral and slightly acidic solutions, such as polarography, paper chromatography, light
absorption, turbidimetry, molecular weight, ultracentrifugation, Raman, infrared and
absorption spectroscopy, radiochemistry, precipitation, and solvent extraction.
Nevertheless, the present knowledge of tungstate species in aqueous solutions is still
rather limited, because what can be determined is the charge of the ion present at a certain
pH and its condensation grade. Proposed structures in the literature are based exclusively
on investigations of solid crystalline compounds. Deducting the structure of a dissolved
ion species from that of the corresponding water soluble, solid, crystallized compound is
not possible, because of a lack in knowledge about the hydration of tungstate ions in
aqueous solutions. The situation is described typically by the following two examples.
Paratungstate A ion is written in the literature sometimes as [HW602o(OHh]5- , or as
[HW602d5- . H20 and [HW602d5-.
At pH :::6.2, the monomeric tungstate ion exists, commonly written in the' form
WO~-. The assumption derived from the solid tung states is that the tungsten ion is
coordinated tetrahedrally by the four oxygen ions. Already in the crystallized stage, the
ions are hydrated as in the case of Na2 W04 . 2H 20. It can be assumed that in aqueous
solutions with their large excess of water molecules the degree of hydration might be
higher. The hydration by two water molecules in case of monotungstate leads to the
proposed octahedral structure [W0 2(OH)4]2-, but more likely seems to be the tetrahydrate
with eight equivalent hydroxo ligands [W(OH)gf-.
As already explained, the monomeric tungstate ion is only stable above pH 6.2 as far
as solutions are considered which are free of complexing agents. At lower pH values the
tungstate ions start to condense to polytungstate species. The condensation grade increases
with decreasing pH until finally, at pH approximately 1, insoluble, highly polymeric,
amorphous tungstic acid is precipitated.

CHAPTER 3

120

Normally, polycondensation is defined by the reaction of anionic species and protons,

such as
2WO~-

+ 2H+

---+ WzO~-

+ H20

However, it seems unlikely that two negatively charged particles can react with each other.
Therefore, polycondensation was also explained as follows: in analogy to molybdenum it is assumed that also cationic tungsten entities are in equilibrium with the anionic
tungstates, although their concentration is low. On acidification, their concentration will
increase because hydroxo ligands of a portion of the anionic tungstate undergo protonation, in which case cationic entities result. These species reacted instantaneously with
some of the remaining anionic hydroxo ions to form polycondensed tungstates. Accordingly, the reaction of positively and negatively charged tungsten ions is the basic
mechanism of the polycondensation, as will be shown in the following simplified
equations (each in a hydrated as well as dehydrated version):
Protonation of the hydroxo complex (tungstate anion):
[W(OH)sf[W04f-

+ 4H+

+ 4H+

---+ [W(OHMH 20)4f+

---+ [wozf+

+ 2HzO

Condensation: aquo ligands of the cationic species are substituted by commonly shared
hydroxo groups of the anionic entity (aquo ligands are more loosely bound as hydroxo
ligands):

"OZOH5-0"
2[ COH)6W, /.w.. . . . /W(OH)6]
o (OHh a

2[W04f- + [W02f+ ---+ [03W-O-WOZ-O-W03fNot only can polycondensation be explained by the existence of positively
charged tungsten entities, but also the reaction of tungstate containing solutions with
anions like fluoride or anions of chelating agents like polycarboxylic acid, etc., always
keeping in mind that a reaction of two negatively charged particles is not possible. Some
chelate complexes are formed with dimeric tungsten species, and this is the proof for a
dimeric cationic tungsten entity which is also analogous with molybdenum.
In principle, the polycondensation may also be regarded as a dehydration process
caused by the increase in proton concentration. In the starting species of the polycondensation process, the fully hydrated monotungstate, the W:H 20 ratio is 1:4 (PH ~ 6.2). In the
various polytungstates prevailing between pH 6.2 and 1, the ratio is somewhere between
1:4 and 1:2, and attains exactly 1:2 in the precipitated tungstic acid at pH 1.
Considering polytungstates as being formed by addition of protons to tungstate ions,
a general, overall equation for their formation can be written in the form

The respective formulas and corresponding values of A, B, and C are given for different
polytungstate species in Table 3.3.

121

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

TABLE 3.3. Isopolytungstate Formulas

and Corresponding Coefficients [3.84]
Coefficient
A

4
8
14

4
I
2

16
18
24

2
12

Formula

6
12
4
16
20
24
26
24
24

[WI2046fo[W 30 1l ]4-

t-

[H4W30 13
[HIO W 12 0 46 ]1O[HW60 21 ]5[W 12 0 40 ]8[H2WI2040]6[H3 W60 2d3[W3 2H 20 ]

It can be seen that preferential condensation grades exist, the most important of them
being W6 and W 12 . In a neighborhood close to the starting pH for the condensation, a
trimeric ion also exists and a dimeric is postulated.
The reason why investigations in this field are difficult and in most cases not very
precise is the existence of equilibria between the different species, and, as was described
above, they do not differ in condensation grade alone but can also be positively or
negatively charged. Another complicating fact is that equilibration rates are fast in some
cases and slow or even very slow in others. This means that although the equilibria are pHdependent, there will always be more than one species present in solution at the same time
due to the equilibrium constants. In addition, the equilibria are influenced by the
temperature, the concentration, and the presence of a third ion or compounds. Higher
concentration shifts the equilibrium to higher condensation grades.
Figure 3.22 presents a scheme of the species in equilibrium between pH 6 and 1. The
number of tungstates is surely confusing, but certainly still incomplete. Once more, it must
be stated that the indices of R, OR, and 0 are always based on assumptions and are
therefore of less importance.
The hexameric para(A)- and the pseudo-meta-tungstate as well as tungstic acid are
formed instantaneously if the pH is lowered to 6 or 4 or I, respectively. The reactions to the
dodecameric ions para(B)- or meta-tungstate are slow or very slow, and the equilibration
may last for days up to one month. In the literature this is often called the "aging" of
tungsten solutions-meaning the change in properties with time.
Here is an example: A sodium tungstate solution acidified to pH 2.5 behaves
differently during further processing via solvent extraction in comparison to the same
solution after some days.
The influence of foreign ions on these equilibria can be different. A variety of ions
form heteropolytungstates. Others shift the polycondensation to the alkaline region like
ammonium. Complexing ions may depress the polycondensation.
Anions like silicate, phosphate, arsenate, etc. form heteropolytungstates. For information about their structure see Section 4.2.7.3.3. The heteropolytungstates are always
higher condensed species in comparison to the isopolytungstates existing in solution under

122

CHAPTER 3

pH~6.2

WO/ / W(OH)sz.

Monomeric tungstate ion

pH between
6.2 and 6.0

HW04 / (H zO)W(OH)7
Hz W04 / (HzO)z W(OH)6
HWO/ / (HzO)3W(OH)s+
WO/+ / (H20)4W(oH)l+

pH-6.0

[WzO(OH)s]Z+
[Wz07(OH)t
[W40 n(OH)4t
[HW40 12(OH)4t
[W60 Z0(OH)z]6-

Monomeric species, short-lived or

thermodynamically unfavorable occurring
in low concentrations

Dimeric, tetrameric and hexameric

tungstates also short-lived and occurring
in low concentrations

Iminutes

pH6

[HW60zo(OH)zt
(Paratungstate A)

days, weeks

[H2W1Z0 4z(OH)Z]I()'
(Paratungstate B)

Preferred hexameric and

dodecameric species

minutes
hours

pH4

[HW60 Z0t
Pseude Metatungstate ~

Polytungstate X

weeks

[HZW120 40
Metatungstate

I
[WIOOnt
I

pH 1

W032HzO (Tungstic acid)

FIGURE 3.22. Scheme of tungsten ions in aqueous solution; updated version of an original scheme published by
Van Put [3.86].

the same conditions. This is often combined with a decrease in solubility or with a change
to colored solutions.
Ammonium ions cause a shift of the condensation pH limit to higher values. This
means that in ammoniacal solutions, polytungstate species are also stable at pH values
where normally only the monotungstate ion exists. For this reason no monomeric
ammonium tungstate can be obtained by crystallization from aqueous solutions, only
polytungstates.
The OH ions bound to the central tungsten ion can be easily substituted by ligands
which form bondings of higher strength mainly by forming more stable chelate ring
structures. The intermediate formation of cationic tungsten species is a prerequisite for
reactions of this type. Hence the ability to condense will be suppressed.
The simplest and best known ion in this respect is fluoride. Therefore, tungsten
solutions containing fluoride are stable also in strongly acidic solution, without any
precipitation of tungstic acid.

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

123

Another well-known but bidentate ion is the peroxo group. Both fluoride and peroxo
ions can occupy up to 8 coordination centers at the tungsten ion. Hydrofluoric acid as
complexing agent in combination with an oxidizing acid or hydrogen peroxide (which
represents both properties) are very important solvents for analytical samples of tungsten
metal and several of its compounds.
Tungsten species in aqueous solutions combine with a variety of chelating agents
offering oxygen, sulfur, selenium, and nitrogen as donor atoms. With tungsten as central
ion in a chelate ring, the following systems have been reported:
-O-W-O-O-W-S-O-W-N-S-W-S-S-W-NImportant examples are chelates formed with polycarboxylic acids or polyhydroxypolycarboxylic acids like oxalic, tartaric, or citric acid, which can be used to stabilize
acidic tungsten solutions (to prevent polycondensation), or chromogenic agents which
react to colored chelates, like f.e. toluene-3,4-dithiol. Several chelates play an important
role in the analytical chemistry of tungsten.

3.8. ELECTROCHEMISTRY OF TUNGSTEN

3.8.1. Introduction [3.89]

The electrochemical behavior of tungsten in aqueous solutions is very closely linked

to two facts: the high affinity of tungsten to oxygen and the complexity of tungsten species
in aqueous solution.
There are an immense number of publications dealing with the electrochemistry of
tungsten, based on practical investigations (a complete compilation is presented elsewhere
[3.90]). The interpretations ofthe results are often contradictory. Their hypothetical nature
is associated with a general lack in knowledge of tungsten chemistry in aqueous solutions
(see Section 3.7). For example, it is never mentioned that for the deposition of tungsten at
the cathode the existence of a positively charged particle is a prerequisite.
In aqueous solution, no oxygen-free tungsten cation exists but only monomeric or
dimeric species like WO~+. The discharge of such a species at the cathode yields an oxidic
deposit.
Tungsten, due to its great affinity to oxygen, may never form a metal/metal ion
electrode system but always a metal/metal oxide/metal ion system, which is quite
irreversible. The oxide coating of the metallic tungsten in its chemical composition,
properties, etc. will be strongly influenced by the ions and equilibria present in solution,
which themselves are closely associated with pH, temperature, concentration, presence of
complexing agents or oxidants, etc. Therefore, not only the chemistry of tungsten in
solution but also the oxide layers at the W electrode are quite complex and not always
understood. For these reasons, no experimental values for standard potentials are available.

124

CHAPTER 3

The steady-state potential measured against a saturated calomel electrode (SCE) is

negative in alkaline and positive in acidic solution, and even higher positive in the
presence of oxidizing agents. Calculated standard potentials from thermodynamic data are
around - O.l V in acidic and around - 0.9 V in alkaline solution against NHE.
The pH dependence (0-10) is linear (about 50 mV /pH). The potential becomes more
negative with increasing temperature due to a partial transformation of a metal oxide to a
metal electrode (the electrode properties shift from semiconductive to metal conductive).
In aqueous solution, the following reactions and corresponding potentials are
assumed:
W02 + 4H+

W2 0 S + 2H+

+ 4e-

--+ W + 2H 2 0 E = -0.119 - 0.0591 . pH

(1)

E = -0.031 - 0.0591 . pH

(2)

E = -0.029 - 0.0591 . pH

(3)

+ 2e- --+ 2W02 + H20

2W03 + 2H+ + 2e- W20 S + H20
WO~- + 2H+ --+ W0 3 + H20
WO~- + 8H+ + 6e- --+ W + 4H2 0

+ 2 pH
(4)
0.0788 pH + 0.098 (log WO~-)

log (WO~-) = -14.05

E = 0.049 -

(5)
WO~-

+ 4H+ + 2e-

--+ W0 2 + 2H 20

E = 0.386 - 0.1182 pH + 0.0295 (log WO~-)

(6)

2WO~-

+ 6H+ + 2e-

--+ W205

+ 2H20

E = 0.801 - 0.1773 pH + 0.0591 (log WO~-)

(7)

Areas of passivation, corrosion, and immunity in aqueous systems can be seen in Fig.
3.23. Accordingly, electrodeposition of tungsten is only possible when the hydrogen
formation is blocked. However, this will not only be influenced in practice by thermodynamic but also by kinetic properties. Therefore, it is only possible to deposit tungsten
2.4
2.0
1.6
1.2

>
W

0.8

passivation

0.4

corrosion

0.0
-0.4
-0.8
-1.2

- 1 .6 L...............L...L....J....-J-o....l..-L....I...............___' - ' -........L...J

-2 0
2
4
6
8 10 12 14 16

pH
FIGURE 3.23. Theoretical regions of passivation, corrosion, and immunity of tungsten at 25C [3.89].

125

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

from aqueous solutions together with alloying elements. Another possibility is deposition
from salt melts where water is completely excluded.
3.8.2. Cathode Reactions [3.89]

An oxide layer is immediately formed during electrolysis of a tungstate solution at a

tungsten cathode. The deposits are high molecular oxides, which vary in degree of
reduction, depending on the conditions of the electrolyte solution. The acidity and
presence of other constituents determine the deposition rate and consequently the layer
thickness.
W0 3 is a n-type semiconductor and is chemically reduced at the cathode. Therefore, a
high potential at the metal-metal oxide interface is initiated, and hence reduction.
W0 3 has a chain structure. The partially reduced ion generated at the metal-metal
oxide interface imparts high conductivity to the whole chain. Consequently, the charge
produced at one end of the chain can be found at its other end (the whole system is an
electrical field). The increase in conductivity of W0 3 is generated by changes in the
electronic structure of the W atoms. Progress in reduction corresponds to decrease in
oxygen concentration, a rising percentage of d5 -configuration atoms, increase in percentage of metallic bonds, and higher conductivity.
The partial W reduction is not only obtained at the metal-metal oxide or at the metal
oxide-electrolyte interface, but the oxide turns blue immediately throughout the whole
layer. This proves the above statement concerning the macromolecular film structure.
When the film thickness has attained a certain thickness, its further reduction is
suppressed, and the generation of hydrogen starts by the discharge of protons. Partially
reduced tungsten oxides may act as catalysts for that reaction. Film thickness and a start of
hydrogen generation depend on the overvoltage of hydrogen, which itself is influenced by
pH, temperature, other ions, or compounds in solution, solvents, etc.
From aqueous solutions, tungsten can only be deposited electrolytically together with
other metals like Co, Fe, Ni, Cu, etc. Hypotheses dealing with the respective deposition
mechanisms are referred to elsewhere [3.89].
3.8.3. Anodic Processes [3.90]

Metallic tungsten as an anode in basic solutions is dissolved electrolytically (oxidized

to the hexavalent state). The six-electron stoichiometry of the overall reaction
W + 801r -+ WO~-

+ 6e- + 4HzO

masks a very complex mechanism, which can be described by the following equations
(s stands for solid and aqu for dissolved):
W(s) + 20H- -+ WO+(s) + 3e- + HzO
WO+(s) + 2011- -+ WOz(s) + e- + HzO
WOz(s) + OH- -+ W0 3H(s) + e-- (rate det. step)
W03H(s) + OH- -+ W0 3(s) + e- + HzO (fast)
W0 3 (s) + OH- -+ HW0 4-(aqu)
HW0 4 -(aqu) -+ WO~-(aqu)

+ HzO

126

CHAPTER 3

This corresponds to a stepwise oxidation from the metallic (elemental) state, via
W3+, W4+, W5+ to W6+.
3.8.4. Molten Salts [3.89]

As in aqueous solution also, in molten salts, too, tungsten behaves in an irreversible

manner. The electrochemical activity of tungsten in melts depends on its position in the
electromotive series, but this position changes with solvent and temperature. In NaOH,
NaP0 3 and Na.tP 20 7, it is electronegative, and in Na2B407 and Na3AIF6 less so.
Corresponding examples are given in Table 3.4.
In contrast to aqueous solutions, tungsten can be deposited from melts as metal at the
cathode.
TABLE 3.4. Position of Tungsten in the Electromotive Series of Metals in Different Molten Salts
Solvent

Temp ("C)

NaOH
NaP03
Na2B407
N34P2 0 7
Na3AIF6

340
720
840
1000
1000

Series
W, Mo, Cd, Pb, Bi, Cu, Ag, Au, Pt, Fe, Ni, Ta, Nb, Zr
Na, W, Zn, Cd, Fe, Pb, Co, Ni, Cu, Mo, Bi, Sb
~~~~~w,~~m~~mn~~~~

Na, Zn, W, Cd, Co, Ni, Cu, Pb, Fe, Mo, Sb, Nb
AI, Mn, Cr, Nb, W, Fe, Co, Mo, Ni

3.8.5. Practical Applications

0/ Electrochemical Processes

[3.89]

pH-Sensitive Electrodes. The utilization of tungsten or tungsten bronze electrodes in

pH measurement is based on the constant slope of the pH-potential dependency
(50mV/pH for tungsten and 58mV/pH for tungsten bronze).
Electrochemical Treatment o/Tungsten and Tungsten Alloys. Electrolytic etching and
polishing is applied in industry to clean tungsten surfaces, or to precondition surfaces
before coating or metallographic examination. Electrolytes and conditions for tungsten
electropolishing are shown in Table 3.5.
Electr%rming. The electroforming of W-Co or W-Co-Ni parts, which can be used
in space vehicles and aircraft, is rather complicated and comprises the following steps:
mechanical processing of an Al matrix, degreasing, zincate treatment, Ag plating,
electrodeposition of the alloy, chemical removal of the matrix, and heat treatment [3.91].
Electrolytic Dissolution. The electrolytic dissolution of tungsten and tungsten alloy
scrap in alkaline media (NaOH, Na2C03, NH 3, (NH4hC0 3) is used technically to recycle
tungsten into the APT production (see Section 5.2.3.6.).
Production and Refining. The production of tungsten or its refining via electrolysis of
molten salts, although technically possible, has never gained greater importance.
3.8.6. Electrochromism [3.92]

Electrochromism is the property of a material or a system to change its optical

properties (color) reversibly if an external potential is applied. It is associated with ion
insertion/extraction processes, which can be presented schematically by:
MeOn +x/+ +xe- --+ /xMeO n

127

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

TABLE 3.5. Electrolytes and Conditions for Tungsten Electropolishing [3.89]

Anodic current
density (A dm -2)

Voltage
(V)

66-300
3-6

25
20

10-15

50

400

21

40-60

KOH (50-100 g .1- 1)

50-1000

0.5-20

25

Na2C03 (ISO g .1- 1)

(50-ISO g .1- 1)

9
39
50-100

6-10
6-12
0.5-20

44-45
40-46
25

Na3P04 (50-100 g . 1-1)

(50-150 g .1- 1)

10-200
30-1000

0.5-20
0.5-20

25
50

50

35-45

50-1000

0.5-20

25

Electrolyte
NaOH (100 g .1-1)
NaOH (40 g .1- 1)
Na2W04 (40 g .1- 1)
K2Cr207 (0.5 g. 1-1)
NaOH(5 g .1- 1)
Na2W04 (30 g .1- 1)

NaCN (70 g .1- 1)

Ethanol (300 g .1- 1)
Na2S (50 g .1- 1)
Methanol (93.75 vol%)
Hydrofluoric acid (1.25 vol%)
Sulfuric acid (5.00 vol%)

15-20
62

Methanol (590ml)
Sulfuric acid (25 ml)

Temperature

eC)

Time
(s)

10
5-7

15-20
15-20

Methanol (590ml)
2-Butoxyethanol (350 ml)
Perchloric acid (60ml)

15-20

15-20

where Me = metal atom, I = single charged small ion, e = electron, and n depends on the
oxide type.
Electrochromism is in principle a device property, although the optical function can
sometimes be caused by a single layer. The basic design of an electrochromic device,
presented in Fig. 3.24, consists of several layers. The substrate (mostly glass) is covered by
a transparent, conducting film in contact with a film of the electrochromic substance.
These films are followed by a layer of a fast ion conductor (electrolyte), an ion storage film,
and another transparent conductor. The electrochromic and ion storage layers are
conductors for ions and electrons while, the ion conductor has zero conductance for
electrons.
W0 3 was the first discovered compound with electrochromism [3.93], and today it
remains the most viable option for the respective devices [3.94-3.97]. The basic reaction
can be presented by the following equation:
W03

+ xI+ + xecolorless

--+ Ix W03

where I stands for H + or alkali and x is about 0.3.

blue

CHAPTER 3

128

r----

(---I
I
I
I
I

I
I
L __ _

+
FIGURE 3.24. Basic design of an electrochromic device, indicating transport of positive ions under the action of
an electric field [3.94].

In comparison to bulk W0 3, thin films always exhibit a high amount of microstructural disorder. This can be further influenced by the preparation mode.
Disordered W0 3 films transform from an optically transparent to an absorbing state
under ion insertion (intercalation) (cathodic coloration). The octahedral W0 6 units are
favorable for both ion as well as electron transport. The pertinent crystal structure allows
long-range diffusion through tunnels or between layers (about W0 3 structures, see Section
4.2.7). Thin films show a cluster-type microstructure and a column-type macrostructure.
This type of coordination leads to electronic bands, responsible for the electrochromic
properties.
The electron insertion/extraction occurring jointly with the ionic movement is
responsible for the actual electrochromism. In highly disordered W0 3 , the optical effects
are associated with the amount of W6 + transformed to W 5 + .
In these thin films, the electrochromic processes are entirely reversible.
W0 3 films can be prepared by evaporation, sputtering, CVD, plasma-enhanced CVD,
anodization, and the sol-gel technique. As already mentioned, the film properties depend
strongly on the preparation mode and parameters.
In devices, diverse types of electrolytes have been used, such as liquid, solid inorganic
bulk type, or thin film and solid organic polymer electrolytes.
Electrochromic devices are potential candidates for numerous applications:
(i) Modulation of the diffuse reflectance can be used in nonemissive display
devices.

129

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

100~------

glass
ITO
Woxide
Ta 2 0 5 :H 2 0
Ni oxide
AI

50

O~--------------------------

0.4

0.8

0.6

Wavelength,

1.0

(/Lm)

FIGURE 3.25. Spectral reflectance in colored and bleached states for an asymmetric electrochromic device with
hydrated dielectric thin-film ion conductors; the design is sketched in the inset [3.94].

(ii) Modulation of the specular reflectance can be applied in antidazzling rear-view

mirrors in automobiles.
(iii) Modulation of the luminous transmittance could be used in future buildings for
superior daylight (optimal working conditions and saving of electrical lighting, as
well as in sunglasses and windows of cars, trucks, and ships).
(iv) Modulation of solar energy transmittance for energy-efficient architecture and
diminished need for airconditioning (smart windows).
An "all solid state" electrochromic device as used in antidazzling mirrors is shown
schematically in Fig. 3.25. It consists of a glass plate coated with a transparent electrode,
covered with the electrochromic W03 film, and attached to a solid electrolyte, a reflector,
and a counterelectrode. An applied voltage drives the charge into the electrochromic layer,
thus causing the change in absorption. Depending on voltage, electrons migrate from the
counterelectrode to the working electrode (absorbing), or vice versa (transparent).
REFERENCES FOR CHAPTER 3
3.1.
3.2.
3.3.
3.4.
3.5.
3.6.
3.7.
3.8.

Gmelin Handbuch der anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. Bl, SpringerVerlag, Berlin (1978).
S. W H. Yih and C. T. Wang, Tungsten, Sources, Metallurgy, Properties and Application, Plenum Press,
New York (1979).
A Warren, A. Nylund, and I. Olefjord, Int. J. Refract. Met. Hard Matter. 14 (1996), 345-353.
C. H. Lee, Nature 203 (1964), 1163.
E. A. Aitken, H. C. Brassfield, P. K. Conn, E. C. Duderstadt, and R. E. Fryxell, Trans. AIME 239 (1967),
1565-1574.
R. E. Fryxell, GEMP 439 (1996), 1-13.
J. N. Ong and W. M. Fassel, Corrosion, 18 (1962) 382T-389T.
Gmelin Handbuch der anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten Suppl. Vol. A7, SpringerVerlag, Heidelberg (1987).

130
3.9.
3.10.
3.11.
3.12.
3.13.
3.14.
3.15.
3.16.
3.17.
3.18.
3.19.
3.20.
3.21.
3.22.
3.23.
3.24.
3.25.
3.26.
3.27.
3.28.
3.29.
3.30.
3.31.
3.32.
3.33.
3.34.
3.35.
3.36.
3.37.
3.38.
3.39.
3.40.
3.41.
3.42.

3.43.
3.44.
3.45.
3.46.
3.47.
3.48.
3.49.
3.50.

CHAPTER 3

T. Hamamura, Bull. Chem. Soc. Jpn. 32 (1959), 1180-1204.

G. R. Belton and R. L. McCarron, J Phys. Chem. 68 (1964), 1852-1856.
Yu. G. Goryachkovskii, V. I. Kostikov, and G. A. Solodkin, Zh. Fiz. Khim. 50 (1976),1959-1962.
R. Haubner, W D. Schubert, E. Lassner, M. Schreiner, and B. Lux, J Refract. Met. Hard Mater. 2 (1983),
108-115.
R. Haubner, W D. Schubert, H., Hellmer, E. Lassner, and B. Lux, J Refract. Met. Hard Mater. 2 1983),
156-162.
H. Hellmer, W D. Schubert, E. Lassner, and B. Lux, in: Proc. 11th Int. Plansee Seminar (H. Bildstein and
H. Ortner, eds.), Vol. 3, pp. 43-86, Metallwerk Plansee GhmbH, Reutle, Austria (1985).
W D. Schubert, J Refract. Met. Hard Mater. 4 (1990), 178-191.
D. Parson, Electrochem. Technol. 3 (1965), 280-283.
C. Choain and F. Marion, Compt. Rend. 252 (1961), 3258-3260.
G. R. St. Pierre, W T. Ebihara, M. J. Pool and R. Speiser, Trans. TMS-AIME 224 (1962), 259-264.
R. J. Ackennann and E. G. Rauh, J Phys. Chem. 67 (1963), 2596-2601.
J. Bousquet and G. Perachon, Camp. Rend. 8 (1964), 934-936.
J. Bousquet and G. Perachon, Camp. Rend. 258 (1964), 3869-3871.
B. Phillips and L. Chang, Trans. TMS-AIME 230 (1964), 1203-1206.
F. E. Rizzo, L. R. Bidwell, and D. F. Frank, Trans. TMS-AIME 39 (1967), 1901-1905.
L. L. Y. Chang and B. Phillips, JAm. Ceram. 52 (1969), 527-533.
P. E. Blackburn, M. Hoch, and H. L. Jolmston, J Phys. Chem. 62 (1958), 769-773.
P. Taskinen, P. Hytoenen, and M. H. Tikkanen, Scand. J Met. 6 (1977), 228-232.
T. Millner and J. Neugebauer, Nature 163 (1949) 601--602.
O. Glemser and H. Ackennann, Z. Anorg. Allg. Chem. 325 (1963), 281-286.
O. Glemser and H. G. Voltz, Natunvissenschaften 43 (1956), 33.
O. Glemser and R. Haeseler, Z. Anorg. Allg. Chem. 316 (1962), 168.
J. Neugebauer, Acta Chim. Acad. Sci. Hung. (1963),247-249.
W Sable and S. Berglund, J Less-Common Met. 79 (1981), 271-280.
H. Mayer, E. Lassner, M. Schreiner, and B. Lux, High Temp. High Pressures 13 (1981), 529-539.
E. Lassner, M. Schreiner, and B. Lux, J Refract. Met. Hard Mater. 1 (1982), 52-97.
R. Haubner, W D. Schubert, E. Lassner, and B. Lux, in: Proc. 11th Plansee Seminar, Vol. 2 (H. Bildstein
and H. Ortner, eds.), pp. 69-97, Verlagsanstalt Tyrolia, Innsbruck, Austria (1985).
R. Haubner, W D. Schubert, E. Lassner, and B. Lux, J Refract. Met. Hard Mater. 6 (1987), 40-45, 111116, 161-167.
R. Haubner, W D. Schubert, E. Lassner, and B. Lux, J Refract. Met. Hard Mater. 7 (1987), 47-56.
W D. Schubert, B. Lux, and B. Zeiler, in: The Chemistry ofNon Sag Tungsten (L. Bartha, E. Lassner, W D.
Schubert, and B. Lux, eds.), Elsevier, Oxford (1995).
A. J. Hegedus, T. Millner, J. Neugebauer, and K. Sasvari, Z. Anorg. Allg. Chem. 281 (1955), 64-82.
W R. Morcom, W L. Worrell, H. G. Sell, and H. I. Kaplan, Metall. Trans. 5 (1974), 155-161.
J. Qvick, Theoretical and Experimental Studies of Trace Elements in Tungsten and Cemented Carbides,
Doctoral Dissertation, Uppsala University, Sweden (1987).
L. Bartha, E. Lassner, W D. Schubert, and B. Lux, eds., The Chemistry of Non Sag Tungsten, Elsevier,
Oxford (1995).
Gmelin Handbuch der anorg. Chemie, 8th ed., Syst. No. 54, Tungsten, Suppl. B2, Springer-Verlag,
Heidelberg (1979).
M. Miyake, A. Hara, T. Sho, and Y. Kawabata, in: Proc. 5th Eur. Symp. Powder Met., Vol. 2, pp. 94-98
(1978).
W L. Worrell, Trans. Met. Soc. AIME233 (1965),1173-1177.
J. K. Fisher and D. R. Moyle, in: Proc. 13th Int. Plansee Seminar (H. Bildstein and R. Eck, eds.), Vol. 2, pp.
425-439, Plansee Metall AG, Reutle, Austria (1993).
A. K. Basu and F. R. Sale, Trans. Met. Soc. AIME 98 (1978), 603--613.
Y. Yamamoto, a. Matsumoto, and Y. Doi, in: Proc. 14th Int. Plansee Seminar (G. Kneringer, P.
Roedhanuner, and P. Wilharitz, eds.), Vol. 2, pp. 596--608, Plansee AG, Reutle, Austria (1997).
S. D. Dunmead, W G. Moore, A. W Weimer, G. A. Eisman, and J. P. Henley, US Patent No. 5,380,688
(1995).
F. H. Ribeiro, R. A. Dalla Betta, M. Boudart, J. Baumgartner, and E. Iglesia, J Catal. 130 (1991),86-105.

IMPORTANT ASPECTS OF TUNGSTEN CHEMISTRY

131

3.51. S. Decker, A. Loefberg, J.-M. Bastin, and A. Frennet, Catal. Lett. 44 (1997), 229-239.
3.52. S. T. Oyama, in: The Chemistry of Transition Metal Carbides and Nitrides (S. T. Oyama, ed.), pp. 1-27,
Chapman & Hall, Glasgow (1996).
3.53. 1. E. McCandlish and P. Seegopaul, in: Proc. Europ. Conf on Advances in Hard Mater. Production,
Stockholm, pp. 93-100, EPMA, Shrewsbury, UK (1996).
3.54. Gmelin Handbuch der anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. AI, SpringerVerlag, Heidelberg (1979).
3.55. Mallory Tungsten Powder Brochure.
3.56. K. A. Jonsson and E. K. A. Svanstrom, Gennan Patent 1,945.154 (1970).
3.57. R. Huenert, G. Winter, W. Kiliani, and D. Greifendorf, J. Refract. Met. Hard Mater. 11 (1992), 331-335.
3.58. A. A. Zinn, "Chemical Vapor Deposition of Tungsten," in: The Chemistry of Metal CVD (T. T. Kodas and
M. J. Hampden-Smith, eds.), VCH, Weinheim (1994).
3.59. Union Carbide Corp., J. B. Culbertson, H. Lamprey, and R. 1. Ripley, U.S. Patent 3.062.638 (1961, 1962).
3.60. J. E. Tress, T. T. Campbell, and F. E. Block, u.s., Bur. Mines, Rep. Invest. No. 6835 (1966).
3.61. 1. Ramqvist, in: Modern Developments in Powder Metallurgy (H. H. Hausner, ed.), Vol. 4, pp. 75-84,
Plenum Press, New York (1971).
3.62. G. Y. Zhao, V. V. Revankar, and V. Hlavacek, J. Less-Common Met. 163 (1990), 269-280.
3.63. V. Revankar, G. Y. Zhao, and V. Hlavacek, Ind. Eng. Res. 30 (1991), 2344--2349.
3.64. H. Ishikawa, S. Watabiki, S. Takajo, and K. Otsuka, Solid State Phenomena 25-26 (1992), 175-178.
3.65. J. Hojo, T. Out, and A. Kato, J. Less-Common Met. S9 (1978), 85-95.
3.66. Soo-Sik Kim and Han-Sam Kim, J. Korean Inst. Met. Mat. 29 (1991), 640-648, 651--658, 659--664.
3.67. E. Neuschwander, J. Less-Common Met. 11 (1966), 365-375.
3.68. P. Matteazzi and G. Le Caer, J. Am. Ceram. Soc. 74 (1991), 1382-1390.
3.69. Gmelin Handbook ofInorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. A5b, pp. 131-153,
Springer-Verlag, Heidelberg (1993).
3.70. W. Geiss and J. A. M. van Liempt, Z. Metallkd. 16 (1924), 317-318.
3.71. 1. V. McCarthy, R. Donelson, and R. F. Hehemann, Metall. Trans. 18A (1987), 969-974.
3.72. Tao Zhengji, J. Refract. Met. Hard Mater. 6 (1987), 221-225.
3.73. E. Lassner, M. Schreiner, and B. Lux, Proc. 10th Plansee Seminar (H. Bildstein and H. Ortner, eds.), Vol. 2,
pp. 761-793, Metallwerk Plansee GmbH, Reutte, Austria (1982).
3.74. A. Tschinkowitz, Diploma-Thesis, Vienna University of Technology (1989).
3.75. I. I. Kovenskii, D!lfUsion in Body-Centered Cubic Metals, (1965) A.S.M., Metals Park, Novelty, Ohio, p. 283.
3.76. A. Shepela, J. Less-Common Met. 26 (1972), 33-43.
3.77. 1. N. Aleksandrov and V. Y. Schelkonogov, Poroshk. Metall. Akad. Nauk Ukr. SSR 22 (1964), 28.
3.78. A. I. Nakonechnikov, 1. V. Paulinov, and V. N. Bykov, Fiz. Met. Metalloved. 22 (1966), 234.
3.79. R. J. Fries, J. E. Cummings, C. G. Hoffman, and S. A. Daily, Proc. 6th Plansee Seminar, pp. 568--607,
Reutte, Austria (1968).
3.80. M. Andrews and S. Dushman, J. Phys. Chem. 29 (1925), 462-471.
3.81. G. S. Kreimer, 1. D. Efros, and E. A. Voronkova, Zh. Tekhn. Fiz. 22 (1952), 858.
3.82. 1. M. Gomes, A. E. Raddatz, and T. G. Carnahan, Proc. 3rd Int. Tungsten Symposium, Madrid, pp. 96-109,
MPR Publishing Services Ltd., Shrewsbury, England (1985).
3.83: N. N. Greenwood and A. Eamshaw, The Chemistry of the Elements, Pergamon Press, Oxford (1984).
3.84. T. K. Kim, R. W. Mooney, and V. Chiola, Sep. Sci. 3 (1968), 467-478.
3.85. V. Cordis, K. H. Tytko, and O. Giemser, Z. Natuiforsch. 306 (1975), 834-841.
3.86. J. W. Van Put, "Crystallization and Processing of Ammonium Paratungstate," in: The Chemistry ofNon Sag
Thngsten (L. Bartha, E. Lassner, W. D. Schubert, and B. Lux eds.), Elsevier Science, Oxford (1995).
3.87. R. Piischel and E. Lassner, in: Chelates in Analytical Chemistry (H. A. Flaschka and A. J. Barnard Jr. eds.),
Marcel Dekker, New York (1967).
3.88. O. Giemser, N. Holmagel, W. Holtze, and E. Schwarzmann, Z. Natuiforsch. 206 (1965), 192-196.
3.89. A. T. Vas'ko, in: Encyclopedia of Electrochemistry of the Elements (A. J. Bard, ed.), pp. 69-126, Marcel
Dekker, New York (1967).
3.90. Gmelin Handbook of Inorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. ASa, SpringerVerlag, Heidelberg (1990).
3.91. M. E. Browning and E. W. Tum, in: Symposium on Electroforming Application; Uses and Properties of
Electroformed Metals, p. 107, Dallas, Texas (1962).

132

CHAPTER 3

3.92. C. G. Granquist, Handbook of Inorganic Electrochromic Materials, Elsevier, Amsterdam (1995).

3.93. T. Kraus, Laboratory Report, Balzers, AG, Liechtenstein, 30.07.1953.
3.94. C. G. Granquist, "Electrochromic Tungsten-Oxide-Based Thin Films: Properties, Chemistry, and Technology, in: Physics of Thin Films (M. Francombe and C. Vossen, eds.), Vol. 17, pp. 301-370, Academic Press,
San Diego (1993).
3.95. C. G. Granquist, Sol. Energy Mater. Sol. Cells 32 (1994), 369-382.
3.96. C. G. Granquist, Solid State Ionics 70/71 (1994),678--685.
3.97. A. Azens, A. Hjelm, D. de Bellac, C. G. Granquist, 1. Barczynska, E. Pentjuss, 1. Gabrusenoks, and 1. M.
Wills, Solid State Ionics 86-88 (1996), 943-948.

4
Tungsten Compounds and Their
Application
The chapter starts with a summary of the tungsten compounds with metallic elements
(intermetallic compounds), and is followed by a more or less detailed description
(depending on the importance of the respective compound) of the tungsten compounds
with nonmetallic elements in sequence from the 14th to the 17 th group of the Periodic
Table. Because of the extremely large number of corresponding compounds known, mixed
ligand compounds and coordination compounds as well as organometallics are treated only
in general in the last two sections.
A short, general discussion at the beginning emphasizes that the chemistry of
tungsten compounds is rather complicated and complex. The following important facts
must be taken into account.
Tungsten forms compounds in the valence states -2, -1,0, +2, +3, +4, +5, and
+6. The most stable compounds belong to the higher valence states +5 and +6.
The stability of lower valence states is less and, in particular, the + 2 state can be
stabilized only by a high degree of metal-to-metal multiple bonds.
Normally the maximal coordination number is 9, but in combination with peroxo
ions or a-bound aromatic ring systems it may attain 13.
The hexavalent oxygen compounds have the tendency to form condensed
compounds with preferential condensation grades of W6 and W 12. The tungsten
atoms in those compounds are connected via oxygen bridges. Higher condensed
compounds form in the presence of heteroions, like, for example, silicate or
phosphate. Several isopolytungstate ions surround the central heteroion. Typical
condensation grades are 6, 9, and 12.
Lower valence compounds ( + 2, +3, and partially +4) are not monomeric but
consist of clusters having a high degree of W-W bondings.
Examples of compounds for different valence states are given in Table 4.1 [4.1]. The
following abbreviations are used throughout the chapter:

133

134

CHAPTER 4

TABLE 4.1. Typical Examples of Tungsten Compounds for the Different Valence States [4.1].
Valence
state

Number of
electrons in 5d

Coordination
number

-2

Trigonal bipyramid

[W(COhf-

-I

Octahedral

[WiCO)101 2-

Octahedral

W(CO)6

+2

5
6
9

Square bipyramid
Octahedral

[W2(CH 3)8t[W2(diarsinehI2l
[W(T)s-CsHs)(CO)3CIW6CI121Cluster

+3

4
6

Tetrahedral
Octahedral

[(R2Nh W=W(NR2)31
[W2CI9l 3-

+4

6
6
8
8
12

Octahedral
Trigonal prismatic
Dodecahedral
Antiprismatic

[WC4fWS2

6
8
13

Octahedral
Dodecahedral

[WF6l[W(CN)81 3[W(T)s -CsHshH3l

4
5
5

Tetrahedral

[W04l 2-

Square pyramid

6
6
7
8

Octahedral
Trigonal prismatic
Pentagonal bipyramid

Tricapped trigonal

[W=(CCMe3)(=CHCMe3)
(CH2CMe3H (PMe2CH2h} 1
[W06l in polytungstates
[W(W2S2H2)31
[WOC4(diarsine)1
[WF8l2[Wl4(Pr2PPhhl

+5

+6

PR
PP

Preparation
Physical properties

D
MH
CO
CR

Density
Microhardness
Color
Crystal structure: Lattice
parameters
Melting point

MP

Fonnula

Structure

[W(CN~14-

[W(Picolinate)4l
[W(T)s-CsHshX2l

[WO~l

CP
HR
STR
I

Boiling point
Enthalpy of formation per mol
(298.15 K)
Chemical properties
Homogeneity range
Stereochemical structure
Importance

Application

BP

1lH298

Thermodynamic parameters (with exception of the enthalpy of formation) are not

considered in this book. They can, however, be easily obtained from the relevant standard
compilations [4.2,4.3]. Data for the following compounds are available: WBr (g), WBrs
(s, I, g), WBr6 (s, g), WC (s), W2C (s), W(CO)6 (s, g), WCI (g), WCh (s, g), WCI4 (s, g),
WCls (s, I, g), WCI 6 (s, I, g), W2Cl IO (g), WF (g), WF6 (g), WO (g), W02 (s, g), W02.72
(s), W02.9 (s), W0 3 (s, I, g), W20 6 (g), W30 S (g), W30 9 (g), W4 0 12 (g), H2W04 (s),

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

135

H2W04 [W02(OHh] (g), WOCl4 (s, I, g), W02Cl2 (s, g), WOF 4 (s, I, g), W02h (g), and
WS 2 (s).

4.1. TUNGSTEN AND METALS: INTERMETALLIC COMPOUNDS AND PHASES

[4.4-4.8]

Tungsten forms intermetallic compounds and phases with the following elements: Be,
~~~~~m~~~~~~~~~~~

In most cases such binary phases are not of importance for the production of alloys,
because they exhibit completely different physical properties in comparison to the
materials of the solid solution regions (higher hardness, brittleness, etc.) which are in
use as alloys. A typical example are a-phase inclusions in powder metallurgically
produced W-I0Re alloy, which lead to a pronounced embrittlement. They are the
consequence of improper mixing of the two components prior to the sintering process.
Re-enriched areas cannot be completely equalized by diffusion during sintering.
In principle, one could imagine that in an alloy a dispersion-strengthening influence
could be achieved by a finely distributed intermetallic compound. Complex intermetallic
phases account for the precipitation hardening in low carbon high-speed tool steels (see
Section 8.1.3) and in superalloys (see Section 8.2). Furthermore, intermetallic phases can
appear in heat-treated WC-Co alloys, increasing the hardness but significantly lowering
the toughness.
Table 4.2 summarizes these compounds and details some properties.
TABLE 4.2. Tungsten Intermetallic Compounds
GroupQ
2

Element

Formula

Be

WBe22

WBe12
WBe2

Properties
4.3 at"110 W
CR: Cubic Zn22Zr-type
a = 11.61-11.64 A
D: 4.25 g.cm- 3 7.7 at% W
CR: Tetragonal D2b Mn12Th-type
a=7.234-7.362 A; c=4.216 A
HR: 28-36 at"110 W
CR: Hexagonal C I4 MgZnrtype
a =4.446-4.578 A; c=7.289-7.429

Zr

W2Zr

66.7 at"110 W
CR: Cubic CIS CU2Mg-type
a = 7.612-7.6195 A

Hf

W2Hf

66.7 at"110 W
CR: Cubic CIS CU2Mg-type
a=7.5825-7.591 A

Tc

WTC3

a-Phase HR: 25-28 at"110 W (1800 0c)

CR: Tetragonal D8b
a =9.479-9.507 A; c=4.952-5.l66 A
(continued)

136

CHAPTER 4
TABLE 4.2. (continued)

GroupQ

Element

Fonnula

Re

K-Phase

Fe

WFe2

W6Fe7

WFe

Ru

as

W3Ru2

w30s
WOs3

Co

WCo3
W6C0 7

Rh

WRh

WRh3

Ir

W3Ir

WIr

WIr3

IO

Ni

WN4

Properties
26.5 at% W
CR: Cubic AI2 (l-Mn-type
a=9.54-9.588 A
cr-Phase HR: 29.0-56.5 at"10 W
CR: Tetragonal D8 b cr-CrFe-type
a=9.55-9.645 A; c=4.98-5.038

A.-Phase metastable 33.3 at% W

CR: Hexagonal C I4 MgZnrtype
a =4.727-4.745 A; c=7.694-7.726 A
~-Phase 42.1 at% W
CR: Hexagonal D8 s Fe7W6-type
a=4.714-4.764 A; c=25.728-25.85 A
o-Phase 50.0 at% W
CR: Orthorhombic MoNi-type
a=7.76 A; b= 12.48 A; c=7.l0 A
p-Phase HR: 59.5-66.5 at"10 W
CR: Tetragonal D8b cr-CrFe-type
a=9.55-9.561 A; c=4.96-4.973

cr-Phase 66.7 at% W

CR: Tetragonal D8b cr-CrFe-type
a =9.6343-9.659 A; c=4.9810-5.001
CR: Hexagonal (existence uncertain)
a=2.75 A; c=4.37 A
x-Phase 25 at% W
CR: Hexagonal DO l9 Ni3Sn-type
a =5.126-5.130 A; c=4.125-4.128
y-Phase 46 at% W
CR: Rhombic D8 5 Fe7W6-type
a = 8.957 A; (l = 30.66
E-Phase HR: 20-52 at"10 W
CR: HCP A3 Mg-type
a=2.708 A; c=4.328 A
E' -Phase HR: 25 at% W
CR: Hexagonal DO I9 Ni 3Sn-type
a=5.453 A; c=4.350 A

a-Phase 75 at% W
CR: Tetragonal D8 b cr-CrFe-type
a=9.672-9.700 A; c=4.99-5.0IO A
E'-Phase
CR: Orthorhombic BI9 ~-AuCd-type
a=4.452 A; b=2.760 A; c=4.811 A
E'-Phase HR: 22-66 at% W
CR: Hexagonal close-packed A3 Mg-type
a=2.74-2.76 A; c=4.389-4.56 A
~-Phase 20.0 at"/o W
CR: Tetragonal Dl. MoN4-type
a=5.730 A; c=3.533 A

(continued)

137

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

TABLE 4.2. (continued)
Group

Element

Properties

Fonnula
WNi

/i-Phase 50.0 a1"l0 W

CR: Orthorhombic MoNi-type
a=7.76 A; b= 12.48 A; c=7.l0
'Y-Phase 66 a1"l0 W
CR: Body-centered tetragonal
a = 10.40 A; c = 10.90 A

'Y-Phase HR: 33-37 at% W

CR: Tetragonally distorted fcc structure

Pt

a=3.895

A; c=3.943 A

E-Phase HR: 45-52 a1"l0 W

CR: Hexagonal close-packed Pt3Mortype

a=2.796 A; c=4.493 A
13

Al

WAl t2

WAls

WAlt

D: 3.88 g. cm-3 measured HR: 7.5-8.0 a1"l0 W

CR: Cubic phase
a =7.580 A
HR: 16.7-18.0 a1"l0 W
D: 5.50 g. cm- 3 measured 5.71 g. cm- 3 X-ray
CR: Hexagonal Ii-phase MoAts-type
a = 4.902 A; c = 8.857-8.860 A
HR: 19.0--21.5 at% W
D: 6.60 gcm-3 measured 6.70 gcm- 3 X-ray
CR: Monoclinic E-phase
a=5.272 A; b= 17.771 A;
c=5.218 A; ~= 100.20

Ga
No intennediates at atmospheric pressure but at 77 Kbar
W2Gas
31 a1"l0 W
(400--800C)
CR: Tetragonal HgsMnrtype
W2Ga3
a=8.948 A; c=2.674 A
(>800 0C)
HR: 39-47 a1"l0 W
CR: Hexagonal

a=3.006
14

A; c=4.738 A

Ge
No intennediates at atmospheric pressure but at ~ 77 Kbar
~-WGe2
66.7 a1"l0 W
CR: Cllb MoSi2-type
a=3.320 A; c=8.l92
Cl-WGe2
66.7 a1"lo W
CR: C23 PbCl2-type
a = 6.399 A; c = 8.544
WSGe3
40 a1"l0 W
CR: D8 m WsSi3-type

a=9.81

a
b

New IUPAC proposal1985.

Lanthanoid series.

A; c=4.92 A

40 at% W
CR: D8 t CrsB3-type
a=6.25 A; c= 11.72
Ce

No details available, existence uncertain

138

CHAPTER 4

4.2. TUNGSTEN AND NONMETALS

4.2.1. Tungsten and Boron [4.4,4.7-4.12]

The following stable intennediates exist: W2B, WB, W2BS-x' and W l-xB3 (earlier
WB 4 , WB l2 , W 2_ x B9 ). For more infonnation on the tungsten-boron system and recent
thennochemical data, see elsewhere [4.10].
PR:

PP:

By solid state reaction between Wand amorphous B powder pressed in compacts at

elevated temperature; pretreatment at 500C in hydrogen for I hour followed by
800-1200 C in argon for 2 hours. The fonnation of the individual boride depends
on the W/B ratio.
By Al reduction of WO r B 2 0 3 mixtures.
By salt melt electrolysis.
By vapor deposition.
They all exhibit high hardness (>9 Mohs), brittleness, and metal-like electrical
conductivity.

Ditungsten Boride: W2B

PP: D: 16.72 g. cm- 3; MP: 2670-2780 C; MH: 2350 kg mm- 2 (100 g);
-Mo 298 =65.3kJ.mol- l ; CR: tetragonal C16, CuAlz-type; a=5.5645.5673 A, c=4.735-4.748 A; HR: narrow; transfonns to WB when heated with
carbon at 1900 C.
Tungsten Boride: WB
Occurs in a low-temperature (cx) and high-temperature modification (~). Values for
the transition temperature scatter in the literature are between 1950 and 2400 0c.

PP:

~-WB

PP:

cx-WB
PP:

D: 16.00 g. cm-3; CO: gray; MP: 2665-2920 C; MH: 2600 kg mm- 2 (50 g);
modulus of elasticity: 608 GPa; electrical resistivity: 25 jln cm;
-Mo 298 = 54.5 kJ mol-I; CR: orthorhombic Bj. CrB-type; a=3.l24-3.19 A,
b=8.4-8.445 A, c=3.06-3.07 A; HR: 48-52 at% B.
D: 16.00 g. cm- 3; CO: gray; -MO 298 = 63 kJ mol-I; CR: tetragonal Bg , cxMoBtype; a=3.101-3.115 A, c= 16.91-16.95 A; HR: 49.0-51.5 at% B.

Ditungsten Pentaboride: W2Bs_x

PP: D: 13.1 g.cm- 3; MP: 2200-2980 C; MH: 2700 kg.mm- 2 (100 g); modulus of
elasticity:
755 GPa;
electrical
~ resistivity:
19 jlQ. cm;
-MO 298 = 192.5 kJ mol-I; CR: hexagonal D8 h , M02BS-type; a = 2.9822.986 A, c= 13.81-13.89 A; HR: 68--69.3 at% Bat 2170 c.
Tungsten Tetraboride: WI _ x B3
D: 8.40 g. cm- 3; MP: 2020-2200 C; -Mo 298 = 82.5 kJ mol-I; CR: tetragonal,
Mo l _xB3-type; a=5.189-5.200 A, c=6.332--6.340 A.

PP:

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

139

4.2.2. Tungsten and Carbon [4.4, 4.7, 4.11, 4.13-4.16]

The binary tungsten-carbon system (Fig. 4.1) [4.13-4.15] is of high technical

importance. It contains three intermediate compounds: W2C(~)' WCi_x(Y), and WC(O),
the latter being the main constituent in most of the commercial cemented carbides
(hardmetals; see Chapter 9). Besides, ternary and even more complex compounds exist
which are of interest in alloyed steels and cemented carbides. Tungsten hexacarbonyl is an
important precursor for organic synthesis.
A thermodynamic evaluation of the W-C system has been presented [4.16].
Weight Percent Carbon
7

L + Graphite

(w)

2600

OU
III

'"'
:l

2200

.-J

<1:1

'"'
a.
III

E 1800

+ Graphite

III

E'-

14-00

10c0~~~~~~~~~~~~~~~~~~~~~~~~~~r+

10

20

30

40

Atomic Percent Carbon

50

60

FIGURE 4.1. Tungsten-carbon phase diagram [4.15].

Ditungsten Carbide: W2 C
PR: It can be produced in pure state by mixing the appropriate amounts of tungsten
powder and carbon black or graphite and subsequent heating in hydrogen atmosphere to at least 1600 C; alternatively, it can be easily produced by melting;
compositions below W2CO.82 cannot be retained at room temperature, even by
quenching in liquid tin [4.17].
PP: D: measured 17.2 g. cm- 3 ; X-ray density: 17.34 g. cm- 3 ; MP: 2785 lO C; MH:
1990 kg. mm- 2 (50 g); modulus of elasticity: 420 GPa; linear coefficient of
expansion: l.2 x 10-6 . K- i (a-axis), 11.4 x lO-6 : K- i (c-axis); electrica.i resistivity: 76-80 Iln cm (room temp.), 125 Iln cm (2000 0C); thermal coefficient of

140

CHAPTER 4

electrical resistivity: 1.95 x 10-3 . K- I; Richardson constant (thermionic emission): 190 A/cm- 2 . K- 2; superconductivity critical temperature: 2.74 K;
-fur 298 = 26.4 kJ . mol-I.
CR: The crystal structure of W2C can be described as a slightly distorted hexagonal
close-packing of tungsten atoms. The carbon atoms occupy half of the octahedral
interstices. They may be distributed in an ordered manner, the type and degree of
ordering depending on temperature. There exist 3 temperature dependent modifications: ~, W, and W'. The transition temperatures are:
2100-2140C
2490C
(earlier yW2C)
hexagonal P6 3 /mmc;, a =2.985-3.000 A, c=4.716-4.730 A; HR: 25.6-34.5 at%
C; statistical occupation of the interstices.
W(earlier ~W2C)
orthorhombic Pbcn, ~-Fe2-N-type; a =4.728 A, b=6.009 A, c=5.193 A; HR:
29.6-32.7 afllo C.
W' (earlier EWlC)
hexagonal; P3ml; E-Fe2N-type; a=5.184 A, c=4.721 A [4.18]; HR: 29.632.5 at% C.
W2C decomposes at 1250 C into Wand WC. It is resistant to many acids, but will
be dissolved by hot, concentrated HN0 3 or a mixture of HN0 3 and HF (1:4). F2
attacks at room temperature and Ch at 200C. It is oxidized by air or oxygen at
500C.
W2C is a component ofthe cast carbide (W2C-WC mixture), which is in use as a
hard facing alloy.
~

CP:

A:

WC,_X
Cubic high-temperature modification; it undergoes a eutectoid decomposition at
2530-2535 C into ~(W2C) + o(WC). At room temperature it can only be obtained by
rapid quenching in liquid tin.
CR: fcc Bl, NaCl-type; a=4.220 A (at 38 at% C).
Tungsten Monocarbide: WC
It is the only stable tungsten carbide at room temperature.
PR:

Reaction of tungsten trioxide with carbon in an inert atmosphere.

Vapor phase deposition.
Electrolysis of fused salts.
Direct combination of the elements (about minimum reaction temperatures, depending on the starting materials; see Section 3.6.1).
Due to the fact that WC melts incongruently, it cannot be produced by melting.
The direct combination of the elements is by far the most widely applied method for
the industrial production ofWC, followed by the reaction ofW03 with carbon (see
Chapter 9).

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

PP:

141

D: measured 15.7 g. cm- 3; X-ray density: 15.77 g. cm- 3; hardness: inherent to its
hexagonal structure, WC single crystals exhibit a pronounced hardness anisotropy;
the room-temperature Knoop hardness of the basal plane (0001) varies from 2200 to
2500, depending on the direction of the long diagonal axis ofthe indenter, and of the
prism planes between only 1000 in the [000 I] direction to 2500 perpendicular to this
direction [4.19]; the hardness anisotropy prevails at least up to 900C [4.20]; for
polycrystalline WC the hardness increases with decreasing grain size (Hall-Petch
relationship); for fine-grained WC it is in the range of 2400-2800 kg mm- 2
(HV30); modulus of elasticity: 670-707 GPa; shear modulus: 262-298 GPa;
compressive modulus: 384 GPa; plasticity [4.21]: WC exhibits a pronounced
plasticity; the formation of slip bands can be observed in the vicinity of hardness
indentations or on compressive testing; the prism planes are the slip planes (1010),
and the slip directions are both (1120) and (0001); the plasticity of WC is an
important aspect for its use as hard component in hardmetals; WC is significantly
tougher than other, covalent hard compounds (fracture toughness: 6-7 MN m- 3/ 2)
[4.22], which contributes to the comparatively high fracture toughness of the WCCo composite; surface energy: 1.7 J . m- 2; -AHo 298 = 40.5 kJ . mol-I; linear
coefficient of thermal expansion: 5.2 x 10-6 . K- 1 (a-axis); 7.3 x 10-6 . K- 1
(c-axis); electrical resistivity: 17-25 Jill. cm; thermal conductivity: 1.2
J. cm- 1 . S-l . K- 1; magnetic susceptibility: +0.005 x 10-6; superconductivity
critical temperature: 10K.

The extremely high modulus of elasticity of WC (only exceeded by diamond and

W2BS), and the high electrical and thermal conductivity are further important criteria for
its use in hardmetals. The latter two properties also reflect the strong metallic component
ofthe mixed covalent (W5d-C2p) metallic bonds in the carbide. Fermi surface properties
of WC and electronic band structure calculations can be found elsewhere [4.23, 4.24].
CR: Hexagonal Bh, WC-type; a = 2.906-2.9066 A, c = 2.8364-2.8374 A. The trigonal
prismatic interstices of the parent metal lattice are completely filled; HR: WC
exhibits a very narrow homogeneity range and is practically stoichiometric. This
means that there are neither tungsten nor carbon vacancies or interstitials other than
those created by thermal activation.
CP: It is highly resistant to many acids. Unlike W2C, it is not attacked by a mixture of
HN03 and HF (1:4) at room temperature, but dissolves on heating. Table 4.3
presents data on the resistance to chemical attack in various aqueous solutions. WC
is resistant against Ch up to 400C, but is attacked by fluorine at room temperature.
Fine WC powders are readily oxidized in air or oxygen at T> 300-500 DC,
depending on the particle size. They are also decarburized in wet hydrogen.
A:
WC is the most important tungsten compound in regard to the percentage of total
tungsten consumption. More than 60% of the tungsten produced (i.e., 20,00024,000 t/year) are used for the cemented carbide production via we. It is also used
as a catalyst. For more details the reader is referred to Chapters 9 and 10.

Ternary Compounds: eta-carbides

Me6W6C, Me2 W4C-Me3 W3C (Me = Co, Fe, Ni).

142

CHAPTER 4

TABLE 4.3. Resistance to Chemical Attack of Tungsten Carbide in Various Media

Insoluble residue (%)

HCI (density 1.19)

HCI (1:1)
H2 S04 (density 1.84)
H2 S04 (1:4)
HN03 (density 1.43)
HN03 (U)
H3P04 (density 1.7)
H3P04 (1:3)
HCI04
HCI04 (1:3)
H2C20 4 (saturated solution)
HCI + HN03 (3:1)
H2S04 + HN03 (1:1)
HN03 + HF (4:1)
H2 S04 + H2C20 4 (1:1)
H2S04 + H3P04 (1:1)
NaOH
20% solution
10"10 solution
20% + bromine water (4:1)
20% + H20 2 (4:1)
20% + K3[Fe(CN)61 (10% solution)

24 h at 20-25 DC

2 h at boiling of the medium

97
96
91
96
63

48
92
I
95

72

10
93
90
40
93
95

91
96
98
98
95
28
92
0
95
96

42
0
70
93

98
97
70
88
68

98
98
60
87
68

Corresponding Co compounds may fonn during liquid-phase sintering of cemented

carbides if the carbon balance is low; analogous Fe compounds (containing also Mo, V,
and Cr) occur in tungsten alloyed steel.
Tungsten Hexacarbonyl: W(CO)6 [4.2S]
PR: By direct reaction of tungsten and carbon monoxide under high pressure.
By Al reduction of WCl6 dissolved in anhydrous ether at a CO pressure of 1 bar at
70 DC with a yield of 90%.
Purification by sublimation in vacuum or by steam distillation.
PP: D: 2.6S g. cm- 3, white crystalline solid with low vapor pressure (13.3 Pa at 20 DC
and 160 Pa at 67 DC); it sublimes in vacuo at 60 to 70 DC and decomposes at
""ISO DC.
CP: Tungsten hexacarbonyl is a zero-valent monomeric compound. The bonding can be
understood as follows: between each carbon atom and the central tungsten a weak (Jbond is fonned by donation of the free pair of electrons of the carbon atom. In
addition, a n-bonding is fonned by back donation of electrons from the Sd level of
the tungsten atom. This explains why the W-C bonding distance (2.06 A) is shorter
than a nonnal single bond. It is slightly soluble in organic solvents like CC4, and
insoluble in water. It is stable in air, water, and acids, but will be decomposed by
strong bases and attacked by halogens.

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

A:

143

It is used as a catalyst in organic synthesis and plays a central role as a precursor to

synthesize a vast variety of coordination as well as organometallic tungsten
compounds. The special importance lays in the possibility to synthesize low-valence
and zero-valent compounds.

4.2.3. Tungsten and Silicon [4.4,4.7, 4.8, 4.11, 4.26}

Two compounds are known in the tungsten-silicon system: WsSh and WSh.
PR:

A:

By reaction of the elements in powder form in a protective atmosphere, like argon or

vacuum at elevated temperature. Starting materials, their size, crystallinity, presence
of defects, oxide films, and experimental conditions determine the reaction product,
mechanism, and kinetics.
By silicothermic reduction of tungsten oxides.
By chemical or physical vapor deposition.
Tungsten silicides form a protective layer at the surface of tungsten and prevent
destructive oxidation at elevated temperatures. The layer fails to protect at lower
temperature (disilicide pest). Tungsten silicide thin films play an important role in
microelectronics (see Section 5.7.5).

Pentatungsten Trisilicide: W5Si3

X-ray
14.56 g.cm- 3 ;
MP:
2370 ec;
PP: D:
measured
12.2g.cm-3 ;
_Me 298 = 125 kJ . mol-I; CR: tetragonal D8m , WsSi3-type; a=9.601 A,

c=4.972

A.

Tungsten Disilicide: WSi2

PP: D: measured 9.30 g. cm-3 ; X-ray 9.75 g. cm-3 ; MP: 2150-2182 ec; MH:
1200 kg mm- 2 (100 g); modulus of elasticity: 430 GPa; electrical resistivity:
12 l. cm; CO: blue gray; _Me 298 = 92 kJ . mol-I; CR: tetragonal CIIb , MoShtype; a=3.211 A, c=7.868 A.
CP: Insoluble in water. Will be attacked by fluorine, chlorine, fused alkalis, and HNOr
HF mixtures. Stable in air up to 900 ec.
A:
Can be used in high-temperature thermocouples in combination with MoSi2 in
oxydizing atmosphere.

4.2.4. Tungsten and Nitrogen [4.11,4.27, 4.28}

There are three nitrides which are considered to be equilibrium phases: W2N, WN,
and WN2 However, no reversible equilibrium transformation has been reported for any of
these phases [4.27J. In addition, twelve other phases have been quoted. Thermodynamic
properties of the W-N system are reported elsewhere [4.28J.
Nitrides usually form by reaction of tungsten with dry ammonia (formation of atomar
nitrogen; "in-situ" nitrogen pressures in the range of 104 _lO s bar). Exceptions are
reactions in a nitrogen jet stream or chemical and physical vapor deposition (plasmaaided formation).

CHAPTER 4

144

The solubility of nitrogen in solid tungsten is very low even at high temperature and
high pressure. Nitrogen is interstitially dissolved in the tungsten lattice.
Ditungsten Nitride: P-W2N (WNO.67-WNl.o)

PR:
PP:

Tungsten powder reacts with ammonia gas at 825-875 dc.

-/lH0 298 = 12.6 kJ mol-I; CR: fcc Bl, NaCl-type; a =4.12--4.14

A.

Tungsten Nitride: o-WN

PR: Tungsten powder reacts at 800C in flowing dry ammonia within a few hours.
PP: -/lH0298 = 24.5 kJ mol-I; CR: hexagonal Bh , WC-type; a = 2.893 A, c =
2.826 A.
Tungsten Dinitride: OR v WN2
PR: Formation during direct heating of tungsten wire in pure nitrogen at 2500 DC.
Evaporating tungsten combines with nitrogen. The nitride deposits on the enclosure
wall.
PP: MP: 400C (decomposition); CO: brownish black; CR: rhombic R3m, OR v_type;
a=2.89 A, c= 16.40 A.

4.2.5. Tungsten and Phosphorus [4.4, 4.8, 4.11]

Three line compounds occur in the W-P system: W3P, WP, and WP 2
Tritungsten Phosphide: W3 P
PR: Observed in arc-melted samples; disappeared on annealing at 800 to 1000 DC;
probably a metastable compound.
PP: CR: tetragonal; a=9.858 A, c=4.800 A.
Tungsten Phosphide: WP
PR: By dissociation of WP 2 or direct synthesis.
PP: D: 12.3 g. cm-3 ; MP: 1450 C (decomposition); CO: gray; CR: orthorhombic B31,
MnP-type: a=5.734 A, b=3.249 A, c=6.222 A.
CP: Insoluble in water, HCl, and alkalis and soluble in a mixture of HF and aqua regia.
Tungsten Diphosphide: WP2
PR: By heating of tungsten powder with red phosphorus in vacuo at 700-950 DC.
By reduction of tungsten trioxide with phosphorus at 550C.
PP: D: 9.17 g. cm- 3 ; CO: black; CR: low-temperature modification Cl-WP2 : monoclinic,
NbAs 2 -type; a = 8.501-8.502 A, b = 3.167-3.1695 A, c = 7.466-7.441 A,
P = 119.33-119.36; high-temperature modification P-WP2: orthorhombic,
MoP 2-type; a=3.166 A, b=I1.161 A, c=4.973 A.
CP: Insoluble in water, acids, alcohol, and ether.

145

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

4.2.6. Tungsten and Arsenic [4.4,4.8,4.11]

Two compounds occur in the W-As system: WAs 2 and W2As 3 .

Ditungsten Triarsenide: W2As3
PR: Heating stoichiometric amounts of the elements in an evacuated, sealed silica tube at
600-1100 C.
PP: CR: monoclinic; a= 13.346 A, b=3.278 A, c=9.599 A; ~= 124.80.
Tungsten Diarsenide: WAs 2
PR: Heating of an element mixture at 620C.
By reaction of arsine with tungsten hexachloride.
Heating the stoichiometric amounts of the elements in an evacuated, sealed silica
tube at 600-1100 C.
PP: D: 6.9 g. cm- 3 ; CO: black; CR: monoclinic, MoAsz-type; a = 9.078-9.085 A,
b=3.3177-3.318 A, c=7.686-7.690 A; ~= 119.42-119.51.
CP: The compound is stable in air, reacts with water above 400C and with chlorine at
500 C. It is insoluble in HF, HCI, and aqueous alkali solutions. It is attacked by
HF + aqua regia mixtures, molten alkalis, concentrated HN0 3, and sulfuric acid.

4.2.7. Tungsten and Oxygen [4.4,4.7,4.29-4.31]

The tungsten-oxygen system is rather complex. Besides the stable stoichiometric

binary oxides (W0 3 , W0 2.9 , W02.72, and W0 2), and the stoichiometric tungstates and
acids, a variety of nonstoichiometric, fully oxidized and reduced compounds exists,
according to the scheme in Fig. 4.2.

Tungsten- Oxygen System

Binary compounds

Nonstolchlomab1c

Ternary compounde

Stolchlomab1c
fullyoxld_

OxIdes

wo.

wo,.,
wo,."

wo.

P-W(W,O)?

Isopolytunga_.
and acids

FIGURE 4.2. Survey of tungsten-Qxygen compounds.

146

CHAPTER 4

(a)
FIGURE 4.3. W06 octahedron; (a) Locations of the centers of the atoms; the black circle is tungsten, the white
circles oxygen. (b) Atoms shown in full size; the tungsten atom is the small circle.

In the fully oxidized state (W6+) the octahedral coordination dominates (W06). The
tungsten atoms are situated in the center of octahedrons and are surrounded by six oxygen
atoms arranged at the corners. As long as all tungsten ions are hexavalent, these
octahedrons are only corner-sharing (Fig. 4.3). However, as the oxygen-to-tungsten ratio
decreases, which occurs on reduction (transition to lower oxides), the octahedra become
more and more intricately joined in combinations of corner-, edge-, and face-sharing.
By the loss of each oxygen atom from the oxide lattice two electrons are added to the
conduction band. Reduced compounds are therefore either semiconducting or conducting.
Besides the tungsten-oxygen octahedra, W04 tetrahedra can also be found in fully
oxidized compounds (tungstates), as well as W07 pentagonal bipyramids in reduced
compounds.
Several hundreds or even thousands of ternary and quaternary tungsten oxides are
known today. They are only partly described in the literature and can be found in the
relevant compilations [4.32-4.35]. They are, however, the top of an iceberg of a much
larger number of compounds which might form, in principle, with a large number of
elements and/or element combinations. In this sense, the crystal chemistry of tungsten
oxides can be regarded as one of the most complex and richest fields in the structural
chemistry of the elements.
4.2.7.1. Stoichiometric Binary Oxides
Tungsten Trioxide: cx-W0 3
By calcination of tungstic acid or APT in air or oxygen:
H2W04 -+ W0 3 + H20 (750-800 0c),
(N~)IO[H2W12042)4H20)-+ 12W03 + IONH3 + IIH 20
By oxidation of tungsten powder (for stoichiometric W0 3 the oxygen partial
pressure has to be higher than in air).
PP: D: 7.21-7.30 g. cm-3 (measured), 7.27 g. cm- 3 (X-ray); MP: 1472 DC; significant
sublimation starts already far below the melting point at >750 DC. The presence of
water vapor enhances the volatility considerably; BP: 1837 c (at this temperature
the sum of the volatile partial pressures corresponds to I bar). The volatile species
are only polymeric molecules (in mol%: 23.7 W20 6, 0.57 W30 g, 58.2 W30 9 , 17.53
W40 11 ); MH: 83-163 kg mm- 2 (50 g); co: yellow (smallest diminution of

PR:

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

147

oxygen changes the color to different types of green); at lower temperature (-27 to
- 50C) it is bluish white and at < - 50 C white. At elevated temperature the color
changes to brownish yellow; -IlHo 298 = 832-853 kJ . mol-I; electrical resistivity: 0.14--0.18 n cm; HR: W02.9873+O.0003'
CR: Several allotropic modifications exist:

> 740C
330 to 740 C
+17 to 330C

- 50 to + 17C

-143 to -50C

tetragonal; P4/nmm- D74h

a = 5.272 A, c = 3.920 A (at 950C)
orthorhombic; Pmnb-D l6 2h
a=7.340 A, b=7.546 A, c=7.728 A.
monoclinic y-W03 ; W0 3 (I); P2dn
a =7.302-7.306 A, b=7.530-7.54l A, c=7.690-7.692 A,
~ = 90.83-90.88
triclinic ~-W03, W0 3 (II); P l-C I i
a=7.30 A, b=7.52 A, c=7.69 A, cx=88.85, ~=90.92,
y = 90.95 (at 10 0C).
monoclinic cx-W03, Pc-C 2 s
a=5.275 A, b=5.155 A, c=7.672 A,
~ = 91.7 (at -70C).

Commercially available W0 3 consists almost entirely of y-W03

The reason for the numerous modifications originates in the structure, which is a
three-dimensional array of comer-sharing metal-oxygen octahedra (Fig. 4.3). The idealized structure (cubic symmetry; Re0 3-type) is drawn in Fig. 4.4, showing the octahedra as
squares in projection. The real structure is distorted. This means that the relatively small
tungsten atoms tend to be displaced from the octahedral center and these displacements are
temperature-dependent. Moreover, the symmetry can be influ~nced by small amounts of
impurities.

FIGURE 4.4. The idealized W03 structure represented as an infinite array of corner-shared octahedra; in the real
structure small distortions of the octahedra reduce the symmetry from cubic to monoclinic at room temperature
[4.30].

CHAPTER 4

148

While still maintaining that every octahedron shares every comer with another, other
arrays than the chess-board-like arrangement are also possible. Low-temperature dehydration of W03 tH20 yields a metastable oxide (hexagonal W0 3, or h-W0 3), with a
structure as shown in Fig. 4.5. Hexagonal W0 3 transforms to monoclinic W03 above
400C. Another metastable modification is the pyrochlore W0 3 , or p-W03 . It can be
prepared from ammonium tungstate by removal of NH4 ions via acid washing. The
flexibility of the W06 comer-linked octahedral framework allows a cubic pyrochlore
structure, too (the pyrochlore structure will be presented under tungsten bronzes). The
material is white and transforms to monoclinic W0 3 above 400C.
CP:

A:

W0 3 is insoluble in acids with the exception of HF, but dissolves easily in aqueous
alkali hydroxide solutions or melts of alkali hydroxides or carbonates. See Chapter 3
for information about reactions of W03 with hydrogen, water, and carbon.
W0 3 is one of the most important, highly pure intermediates for the production of
other tungsten compounds and tungsten metal powder. In the latter application it was
substituted to a large extent by tungsten blue oxide (see Chapter 5.3).
Because of its bright yellow color it is used as a pigment in oil and water colors.
It is employed in a wide variety of catalysts, most recently in catalysts for the control
of air pollution and industrial hygiene (see Chapter 10).
W03 is also used as an addition to V02 in "smart windows." Thermochromism
[4.36] is the ability of a material to be transparent for infrared rays below a certain
transition temperature and to reflect above, due to a crystallographic phase transition. The process is strictly reversible and there is almost no influence on visible
light. V02 has a transition temperature at 68C, which can be lowered by doping
with W03 . Glass coated with a thin film of the above mixed oxide is called a "smart
window" and is a promising candidate for use in automobiles and buildings for heat
protection.

FIGURE 4.5. The hexagonal framework of octahedra in h-W03 [4.30].

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

149

Solid state devices use the photochromism of W0 3 thin films in antidazzling mirrors
(for more details of W03 photochromic properties, see Section 3.8.6).

p-Tungsten Oxide: W02.9 (W2005aJ

Forms intermediately during the hydrogen reduction of W0 3 to W. In this process it
can be obtained in pure state, for example at 500-550 c and a water vapor partial
pressure of 0.6 bar.
By heating corresponding mixtures of Wand W0 3 . The compacts are heated 6 hours
at 800C in argon, then powdered and tempered under the same conditions for 24
hours.
PP: D: 7.15 g. cm- 3 (measured), 7.16 g. cm- 3 (X-ray); CO: blue to deep blue;
electrical resistivity: 5 x 10-3 n cm; -/!JfO 298 = 800-816 kJ . mol-I; CR:
monoclinic P2/m-C I 2h ; a= 12.05 A, b=3.767 A, c=3.59 A, P=94.72; HR:
W02.8975-2.9014 (1000 0C).
W02.9 belongs to the homologous series W n0 3n - 2, which is characterized by a
crystallographic share (CS) structure. The shear planes lie upon {103} planes with
respect to the W0 3 parent structure. An idealized structure ofW02.9 is shown in Fig.
4.7.
It crystallizes mostly in needle shape (the monoclinic b-axis is parallel to the needle
axis).
CP: About its reactions with hydrogen and water, see Chapter 3.
A:
It is partially contained in the technical blue oxide and acts in pure form as an
important catalyst in industrial chemical syntheses (see Chapter 10). Similar to
W0 2.72 , it catalyzes the decomposition of N20 and the selective oxidation of
propene oxygen mixtures to acroleine.
PR:

y-Tungsten Oxide: W02.72 (W/S049)

It also forms during the hydrogen reduction of W0 3 to W and can be obtained in
pure state, for example, in hydrogen atmosphere at 900C and a water vapor partial
pressure of 0.8 bar.
In addition, it can be prepared by heating corresponding powder mixtures of W and
W0 3 in sealed vessels or in argon atmosphere and subsequent tempering.
PP: D: 7.724-7.989 g. cm-3 (measured), 7.78 g. cm- 3 (X-ray); CO: reddish violet, in
whiskers brownish red; electrical resistivity: (2-3) x 10-4 n cm; -/llfO 298 =
743-780 kJ . mol-I; CR: monclinic, P2/m-C I 2h ; a = 18.28 A, b = 3.775 A, c =
13.98 A, p= 115.14.
The crystals are needle-shaped and the monoclinic b-axis is parallel to the needle
axis; HR: W02.7190-2.7224 at 1000 C.
W0 2.72 is not included in the homologous series of CS structures, because besides
edge- and corner-sharing octahedra also pentagonal columns exist (as described
under tungsten bronzes).
CP: About behavior with respect to hydrogen, see Chapter 3. At elevated temperature it
decomposes. It is soluble in alkaline solutions. Propene oxygen mixtures are
oxidized by W0 2.72 to acroleine.
A:
It is partially contained in the technical blue oxide.

PR:

CHAPTER 4

150

Tungsten Dioxide: W02

PR: It forms intermediately during the hydrogen reduction of higher tungsten oxides. In
a pure state it can be obtained, for example, at 900C in hydrogen atmosphere and a
water vapor partial pressure of 0.5 bar. In addition, it can be prepared by reduction
ofW03 with W. The direct oxidation of tungsten by oxygen is not a suitable method
to synthesize W02, since higher oxides are formed, too.
PP: D: 10.82-11.05 g. cm- 3 (measured); 10.82 g. cm- 3 (X-ray); MP: At 1530 c W02
disproportionates to W and W 18049 prior to melting. All indications of melting
points in the older literature are not relevant; CO: brown to violet brown, single
crystals are bronze colored and show metallic luster; electrical resistivity:
2.9 x 10-3 Qcm; -tuf0 298 = 571-595 kJ mol-II; HR: at room temperature
quasi-stoichiometric; W01.97-2.0 at 1357 C; CR: monoclinic, P2i/c MoOrtype;
a =5.550-5.62 A, h=4.89-4.96 A, c=5.571-5.73 A, ~= 118.93-122.1; W0 2
exhibits a characteristic band structure, in which the W06 octahedra are edgesharing. Neighboring bands are corner-sharing. The W atoms within the bands have
alternating shorter (2.53 A) and larger (3.13 A) bonding distances, whiCh indicates a
cation-cation bond interaction and the formation of W-W double bonds between
some of the W atoms. W02 is slightly paramagnetic and is described as both
electrically conducting and semiconducting.
CP: W0 2 dissolves in warm, concentrated mineral acids and in boiling, concentrated
alkali hydroxide solutions. It dissolves rapidly in diluted and conc. H20 2. Nitric acid
oxidizes it to W03. About the behavior with respect to hydrogen, see Chapter 3.
Higher oxides are formed with oxygen at elevated temperature. Reaction with
chlorine leads to the formation of WOCI4 and W02Ch, while with bromine and
iodine only the dioxodihalides are formed.
A:
In a two-step reduction procedure mainly applied in earlier times in NS tungsten
powder production, it was the end product of the first reduction step and the starting
material for the second.
W30 resp. W3W [4.37-4.39]
It was regarded as a suboxide and also as metal modification. Today it is described as
a metastable metal phase, which is stabilized by small amounts of oxygen. A further
stabilization of ~- W occurs by the presence of different foreign elements, such as K, Be, P,
B, As, Ce, Th, and AI. All these elements have in common that they form rather stable
ternary and quaternary tungsten oxides.
~-Tungsten:

PR:

PP:

By electrolysis of fused mixtures of W03 and alkali metal phosphates.

Forms partially during the hydrogen reduction of doped tungsten oxide (NS-W
production).
X-ray pure ~-W forms at 500-600C by the reduction of W200 58 with dry
hydrogen in very thin powder layers.
D: 19.0-19.1 g.cm- 3 (measured), 18.94g.cm-3 (X-ray; calculated as W3W);
transition temperature: ~-W --+ r:t.-W occurs at 530-800C, depending on the
presence of third elements; it occurs without any significant weight change;
electrical resistivity: 200 ~. cm, CO: gray or black; CR: cubic A15; ~-W type
Pm3n-0\; a = 5.0512 A [4.39].

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

CP:

151

It easily dissolves in diluted or conc. H20 2 . The residual oxygen is either present on
the surfaces or grain boundaries of the commonly nanocrystalline metal powder, or
randomly distributed over all eight positions of the A 15 W3W lattice.

4.2.7.2. Nonstoichiometric Binary Compounds

Crystallographic Shear Structures (CS Structures) [4.30]
STR: When W0 3 is slightly reduced at approximately 730C either in sealed tubes or in
hydrogen atmosphere, the oxygen deficit is taken up by CS planes. In principle, CS
planes can be built up of edge- and plane-sharing structures. This corresponds to a
partial replacement of W+6 by W+ 5 and W+4 ions.
At the lowest degree of reduction, block~ of four edge-sharing octahedra are
present (Fig. 4.6.) The CS planes in this case are disordered. If the degree of
reduction gets higher, more and more ordered structures exist (bands of parallel CS
planes). For equispaced CS planes the composition corresponds to Wn03n-l.
These ordered structures have a composition range between W02 .96-2.933
(W2S 0 74-W 1S 044).
Below W02 .92, the structure changes to six edge-sharing octahedra (Fig. 4.7.)
according to the general formula W n0 3n - 2 where n = 25-18, and this corresponds
to a range W02.920--2.889 (~ W 20 0 58 ).

CS structures have the ability to build-in foreign metal ions, especially in the reduced
form. This seems to play an important role in the dopant uptake of tungsten blue oxide in
non-sag tungsten production (for more information see Section 5.4.5).

FIGURE 4.6. Idealized crystallographic shear (CS) planes in W0 3_ x ; blocks of four edge-sharing octahedra

[4.30].

152

CHAPTER 4

FIGURE 4.7. Idealized crystallographic shear (CS) planes in W0 3_ x ; blocks of six edge-sharing octahedra

[4.30].

4.2.7.3. Stoichiometric Ternary Compounds (Oxidized)

(a) Tungsten Trioxide Hydrates and Monotungstates
Tungsten Trioxide Monohydrate: W03 H20 (W02(OH)~. It is a monomeric, gaseous
compound which fonns above 600C by reaction of water vapor and tungsten oxides. It is
responsible for the Chemical Vapor Transport (CVT) of tungsten during the reduction
process and hence for the grain size of the finally fonned tungsten powder. Another
important aspect of the CVT of tungsten is the overgrowth of dopant compounds by
tungsten during the powder production of non-sag tungsten (see also Sections 3.3.5 and
5.4.5).
Tungstic Acids. In contrast to the compound of the same overall composition
mentioned above, tungstic acids are highly condensed, almost insoluble and amorphous
species with decreasing water content:

W0 3 2H2 0

(W : H20 ratio I : 2)

WOr H 20
2W03 H20

(W : H2 0 ratio 1 : 1)
(W : H2 0 ratio 2 : 1)

Within this group dehydration can be achieved by heating. Heating to approximately

500C leads to the final stage of dehydration, the water-free, crystallized tungsten
trioxide.

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

153

Tungstic Acid: H2 W0 4 (WOr H20)

PR:

PP:
CP:

A:

By reaction of hot alkali tungstate solutions with strong mineral acids, preferably
HCl.
APT is reacted with strong acids (preferably HCI). This product is more pure than
the above, because of the higher purity of APT.
Amorphous, yellow powder.
Nearly insoluble in water or acid solutions, but freshly prepared tungstic acid forms
colloidal solutions when treated with water. It dissolves in strongly alkaline media,
ammonia, and HE It is stable up to 150C.
Tungstic acid in former times was the most important, highly pure intermediate in
the technical process from ore to tungsten powder. It was in most applications
substituted by APT. Today, it serves as starting material for the production of
submicron tungsten powder and of several tungsten compounds. Tungstic acid is
also used as a pigment for bright yellow glazes.

= '" 2)
Precipitation by acidifying tungstate solutions in the cold.
Addition of sodium tungstate solution to 1 molar nitric acid until pH I is reached.
It is a white, amorphous solid.
Converts to tungstic acid on boiling; it is stable up to 55C and loses water in the
temperature range 55 C-75 DC; it seems to be a mixture of hexa- and decatungstate.

White Tungstic Acid: WOJ xH20 (x

PR:
PP:
CP:

Ditungstentrioxide Monohydrate: 2WOJ H20

PR:
PP:

By drying tungstic acid not over 300C.

Orange yellow colored solid.

Monotungstates: WI, W 2 , w3, w4, W 5 , W 6 -tungstates [4.31]

STR: In the tungstates, tungsten is found either in octahedral or in tetrahedral coordina-

PR:

tion. If the radius of the ternary cation is plotted against electronegativity, one may
observe the following properties (Fig. 4.8.):
With high electronegativity and small ionic radius (0.7 ~ 0.95 A) octahedral
coordination prevails (wolframite group).
With low electronegativity and large ionic radius tetrahedral coordination is
observed (scheelite group).
Metals at the border line may show polymorphism or other structural complexity
(for example, under high pressure, CaW04 exhibits wolframite structure).
Nonhydrated sodium tungstate can be prepared by fusion reactions, such as:

If it is prepared by crystallization of an aqueous solution, the dihydrate always

forms Na2W04 . 2H20.
Potassium tungstate can be formed under the s~e conditions.
Ammonium tungstate is produced by reaction of liquid ammonia with tungstic acid
in a nonhydrated form. A hydrated form does not exist, because ammonium ions

154

CHAPTER 4

W oct

2.0

--

> 1.5

CO

0>

Q)

c:

o
~

Q)

W 1.0

K.
0.6

0.8

1.0

1.2

Rb

1.4

Ion radius (A)

FIGURE 4.8. Plot of the octahedral ionic radius versus electronegativity for the fully oxidized tungstates; the
dotted line represents the approximate boundary between structures in which W is tetrahedrally or octahedrally
coordinated [4.31].

PP:
CP:

shift the condensation tendency of tungstate ions to slightly alkaline conditions.

This means that only polytungstates can be obtained from ammonia-containing
solutions by crystallization.
The preparation of insoluble tungstates (divalent or trivalent metal ions with the
exception of Mg) is carried out by reacting aqueous sodium tungstate solutions
with the corresponding soluble metal salt.
Alkali tungstates are white, hygroscopic substances. Some further properties of
various tungstates are listed in Table 4.4.
Alkali tungstates are very soluble in water and crystallize in hydrated form. They
lose the water of hydration nearly quantitatively at 100C. All other tungstates
with the exception of Mg are sparingly soluble in water. They decompose by
addition of hot mineral acids (except phosphoric acid) to tungstic acid and the
corresponding metal salt.

Calcium Tungstate: CaW04

PR: Calcium chloride is added in excess to a sodium tungstate melt.
Addition of calcium chloride solution to sodium tungstate solution. This reaction
was used on a large scale in former times as purification step (precipitation of
artificial scheelite).
PP: CR: tetragonal Dy5Pd2-type; a = 5.25 A, c = 11.38 A; calcium tungstate is fluorescent when exposed to short-wavelength UV radiation.

ISS

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

TABLE 4.4. Physical Properties of Monotungstates

X-ray

Fonnula

Measured

Li2W04
K2W04
SrW04
CdW04
BaW04
PbW04
FeW04

3.17
3.12
6.184

6.372

6.35

6.00

7.391

7.647

MnW04

7.135

7.261

MPCOc)

CO

CR

738
928

white
white
white

1490
1130

white
white

842
1130

gray-green
white

Rhombic a=8.888 A; p= 107.78

Monocline
Tetragonal a = 5.405 A; c = 11.90 A
Tetragonal
Tetragonal a = 5.64 A; c = 12.689 A
Tetragonal a=5.44 A; c= 12.01 A
Moncline p=90
a=4.70 A; b=5.69 A; c=4.93 A
Monoclinic P= 89.17
a=4.84 A; b=5.76 A; c=4.97 A
Monoclinic p= 89.30
a=4.68 A; b=5.73 A; c=4.95 A
Monoclinic p= 90
a=4.66 A; b=5.69 A; c=4.98 A
Monoclinic p= 89.7
a=4.68 A; b=5.66 A; c=4.92 A
Monoclinic p= 89.40
a=4.69 A; b=5.66 A; c=4.93 A
Tetragonal
Monoclinic p= 109.63
a=7.68 A; b= 11.46 A; c= 11.40 A

ZnW0 4
CoW04
MgW04
NiW04
Bi 2(W04h
EU2(W04h
Ce2(W04h

CP:
A:

Only sparingly soluble in water.

It is one of the two most important minerals: scheelite.
As artificial scheelite it was an important intermediate.
For the production of phosphores for lasers, fluorescence lamps, oscilloscopes,
luminescence colors, X-ray observation screens, luminous paints, and scintillation
counters. The property of fluorescence is very important in prospecting and mining
of scheelite ore.

Zirconium Tungstate: ZrW2 0 8 [4.40]

Zirconium tungstate and the isostructural hafuium tungstate (HtW208) show strong
negative thermal expansion from 0.3 K up to their decomposition temperature of

"'1050 K.
PP:

A:

Thermal expansion: -8.9 x 10-6 . K- J for ZrW 2 0 S; decomposes at "'1050 into

tungsten and zirkonuim oxide; CR: cubic, P2 J3; a = 9.1575 A; order-disorder
transition at 428 K.
For composite materials with an overall zero bulk expansion (electronic devices).

The importance of tungstates is manifold. We find nearly all naturally occurring

tungsten minerals in the group of monotungstates. Barium and zinc tungstates are used for
bright white pigments. Sodium tungstate in aqueous solution is the most important
intermediate between ore and APT or tungstic acid in the technical production.

156

CHAPTER 4

(b) Isopolytungstates and their Acids. The transition of monotungstate ions to tungstic acid
by acidification was already discussed earlier. It is accompanied by the formation of
several species of preferential condensation grades. Evidently due to structural reasons,
they show a high degree of stability at some pH value. Some important and typical
examples are given in Table 4.5. The respective salts of these polyions are not easy to
synthesize, because in aqueous solutions equilibria always exist between the different
species and, moreover, kinetic effects of variously rapid reactions make the situation even
worse.
The stability and optimal H30+ (W04 2- ratio of certain species can be further
influenced by other ions, such as ammonium which shifts the polycondensation to higher
pH values.
The most important isopolytungstates are listed below.

Ammonium Metatungstate: [(NH4)6(H2W12040)' 3H20]

PR:
CP:
A:

Ion exchange procedure.

Electrodialysis.
The advantage of AMT is its high solubility in water, giving the possibility to
prepare highly concentrated, alkali-free tungsten solutions.
AMT solutions are in use to treat supports (like alumnia pellets) for catalyst
production. AMT is further used in nuclear shielding, as corrosion inhibitor, and
as a precursor for other tungsten chemicals.

Sodium Metatungstate: [NadH2W12 0 40). 3H20]

PR:

PP:

A:

Formation of the monotungstate and subsequent removal of the sodium ions from
the solution by ion exchange to form a dilute sodium metatungstate solution, and
subsequent concentration of the solution.
High concentrations of sodium metatungstate provide solutions with densities up to
3.1 g. cm- 1; viscosity: > 40 cP (1 cP = 10-3 N s m- 2) at a specific gravity of
3 gml- 1
It is used as high-density liquid in crude oil drilling, as important intermediate for
the production of heteropolytungstates, as a fuel cell electrode material, in cigarette
filters, and for fireproofing textiles.

TABLE 4.5. Some Typical Polytungstate Ions

H3O+jW04 2-

Formula

Name

0.333
0.667

W120 46 20W30 11 44~W3013

10H lOW 12046
HW60 21 5W120 40 8H2W12 0 40 6H3W60 21 3W0 3 H2O

Para Z
Tritungstate

1.167
1.33
1.50
2.00

Para B
Para A
Meta
Pseudo-meta
Tungstic acid

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

157

Lithium Metatungstate: (Li6(H2W12040)' 3H20] [4.41]

PR: See above.
PP: Solutions of lithium tungstate are even more dense than sodium tungstate solutions,
although lithium is the least dense alkali metal; maximum density is about
3.5 g. cm-I; viscosity: ""'30 cP at a specific gmvity of 3 g. ml- I .
Ammonium Paratungstate: (NH4) loH2W12042' 4H20 [4.42]
PR: Its preparation will be treated in detail in Chapter 5.
PP: D: 4.61 g. cm- I (measured), 4.639 g. cm- I (X-ray); CR: monoclinic, P2t!n;
besides the technically important tetrahydmte, depending on crystallization conditions, a triclinic hexahydmte and an orthorhombic decahydmte can be formed.
A:
It is today the most common, highly pure intermediate for most tungsten products
(see Chapter 5.3).

Ion associate complexes of isopolytungstates with secondary and tertiary alkyl

amines play an important role in the technical solvent extmction process (see Section 5.3.)
(c) Heteropolytungstates and their Acids [4.43-4.44]. Heteropoly tungstates may be
classified according to the ratio of hetero to tungsten atoms. Some typical examples are
given in Table 4.6. More than 30 elements are known to function as heteroatom in those
compounds with sometimes varying stoichiometric ratios of hetero to tungsten atoms
(Table 4.7).
TABLE 4.6. Some Typical Species of Heteropolytungstates
Atom ratio (hetero:W)

Central atom

I: 12

pH, AsH, Si H , GeH , TiH, Co3+,

Fe3+, AI3+, cr1+, Ga3+, Te4 +, B3+
Si4+, PtH
Be2+
Series A, Te6+, C+
Series B, NiH, Ga3+
p5+, AsH
pH, AsH

1:10
1:9
1:6

2:18
2:17

Typical fonnula

[Xn+ (W IOOx)]<2x-60-n)-

[X2+(W90 31 )]6[xn+(W6024WI2-n)[xn+(w 60 24H 6)](6-n)[X2 n+(W 1S 0 62)](l2-2)[X2 n+(w 170xW2x-102-2n)-

TABLE 4.7. Ions Capable of Acting as Central Ion in Tungsten Heteropoly Compounds
Group

II
III
IV
V
VI
VII
VIII
Lanthanoid series

New ruPAC designation 1985

1+11
2 + 12
13

14
5 + 15
6 + 16
7 + 17
8,9, 10

Ion
H+, Cu2+
Be2+, Zn2+
B3+, AI3+, Ga3+
Si4+, Ge4+, SnH
N5+, p3+, pH, AsH, AsH, Sb3+, V4 +, V5+
Cr1+, SH, SeH , Te4+, Te6+
Mn2+, Mn4+, 17+
Fe3+, Co2+, C0 3+, Ni 2+, NiH, Rh3+, Pt4+
Ce3+, CeH

CHAPTER 4

158

FIGURE 4.9. Cubooctahedron diagram of the [PW I2 0 40]3- anion. The W atoms are at the comers of the
polyhedron and the oxygen atoms at the midpoints of the edges.

STR:

W0 6 octahedra sharing comers and edges surround the tetrahedrally or octahedrally coordinated heteroatom. Also, the central polyhedron shares comers and
edges (or both) with the W06 octahedra. Thus each of the Watoms is connected to
the central heteroatom via an oxygen bridge. Schematic structures are given in Fig.
4.9. The actual structures are frequently distorted.
In crystals, the relatively large heteropoly anions cause interstices between each
other which are occupied by cations or by water molecules. Therefore, there exist
no strong bondings between the individual heteropoly tungstate ions. Instead, they
are joined by hydrogen bondings through water of hydration molecules.
In case of large cations, they frequently crystallize as acid salt even although an
excess of the cation is present in the mother liquor. There is not so much space
available which would be required for the cations to form a neutral salt. Those salts
are often less hydrated than those of smaller cations. The large cations take so
much space that there is not so much left for the water molecules.
CP:
Heteropolytungstates can be characterized as follows:
High molecular weight <3000.
High degree of hydration.
Unusually high solubility in water and also in some organic solvents.
Strong oxidizing action in aqueous solution.
Strong acidity in free acid form.
Decomposition into monotungstates in strongly basic aqueous solution.
Highly colored anions or colored reaction products.
A:
They were produced in large quantities for:
Precipitants for basic dyes forming lakes and toners.
Catalysts.
Passivation of steel.
Ion associate complexes of (Si, P) heteropolytungstates with alkylamines may disturb
the technical solvent extraction process. In particular, if tertiary amines are in use, the
complexes are insoluble in the solvent and form third phases (see Section 5.3).

Sodium-12- Tungstophosphate
A:

This is a very important compound and is employed for many applications:

Catalyst in petroleum and organic chemical industry (isomerization, polymerzation,

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

159

oil bodying, nitrile synthesis, aromatization, desulfurization, dehydrochlorination,

dehydration, ring closure, keton synthesis).
Manufacture of organic pigments (precipitating agent for synthetic dye stuffs, forms
brilliant colored pigments, highly stable to heat, light and moisture; printing inks,
paper tinting, paper coloring, nontoxic paints, wax pigmentation.
Carroting (surface treating of furs).
Antistatic agent for treatment of acryl-based synthetic fibers.
The oxidizing power makes it useful in various photographic processes, manufacture
of photosensitive systems, image producing, or printout systems.
Leather tanning, waterproofing of plastic films.
Additive to chrome-plating baths.
In cements and adhesives to impart water resistivity.
(d) Peroxotungstates and Acids
PR: Peroxotungstic acid forms when W0 3 is reacted with hydrogen peroxide.
Peroxotungstic acid is also generated when tungsten powder is dissolved in
hydrogen peroxide.
Reaction of tungstates with hydrogen peroxide yields in peroxotungstates.
PP: CO: bright yellow.
CP: Hydrogen peroxide may be regarded as a bidentate ligand which substitutes
1he oxygen ligands at 1he tungstate ion. The 3-membered ring configuration

H20
+

0,

pH

I".'w is thermodynamically more stable than the W=OorW,

OH

bonding,

presuming the hydrogen peroxide concentration is high enough.

Depending on the pH value, peroxotungstate species with different W:0 2 ratio are
formed. A survey is presented in Table 4.8. In basic solutions the peroxo groups
occupy all eight coordination positions at the W ion. By lowering the pH the W:0 2
ratio decreases. As soon as oxygen ligands are present at the W ion dimerization
becomes possible, as is shown in the following equation:

This species is sometimes written in the hydrated form

TABLE 4.8. Peroxotungstate Species

W:O ratio

1:4
1:3
1:2

I: I

Fonnula
WO s 2- [W(02)4fW07 2- [WO(02h1 2W06 2- [W02(02hfW20 11 2- [(02hOWOWO(02hfWO s 2- [W0 3(02)12-

Fonnation
In basic solutions by an excess of hydrogen
Thennal decomposition of 1:4 peroxotungstates
By slight acidification of alkalitungstate solutions containing
excess of hydrogen peroxide
By neutralization of peroxotungstic acid solutions increase

CHAPTER 4

160

and this corresponds to

A:

In general, the presence of hydrogen peroxide prevents the polycondensation

tendency of the tungstate ions. All peroxo tungsten compounds are unstable. The
degree of instability increases with increasing W:02 ratio. The 1:4 compound even
may explode.
Peroxotungstic acids like H2W06 or H2 W2011 can be extracted from nitrate
solutions into trialkyl-(C7-C9)-benzyl ammonium nitrate dissolved in trichloroethelyne.
Peroxotungstates of the following cations have been prepared: Na, K, Ca, Sr, Ba,
Co, Sc, and Y.
Hydrogen peroxide is used to dissolve tungsten metal powder samples for analytical
purposes. The advantage over other solvents is that no ,further anion is present which
may disturb in subsequent reactions, such as fluoride. In this procedure hydrogen
peroxide not only functions as solvent, but also as chelating agent preventing the
tungsten from precipitating.

4.2.7.4. Nonstoichiometric Ternary Compounds (Reduced); Tungsten Bronzes [4.30,

4.45]. Tungsten bronzes (TBs) are well-defined nonstoichiometric compounds of the
general formula [4.45]

where y can be less than or greater than z x/2; z is the valency of the bronze-forming
atom.
The ideal composition is y = z . x/2 ("true" tungsten bronzes; Mx W03 ).
Not only y but also x influences the physical as well as the chemical properties. For
example,
x < 0.3: bronzes are semiconducting;
x> 0.3: bronzes are conducting (metallic bonding), because x has an influence on the
Fermi levels.
The name bronzes derives from the metallic luster and the intense color. The stability
of the TBs depends on the size of "M". A decrease in the ionic radius lowers the stability.
A survey ofTB-forming elements is given in Table 4.9. Very small ions, like Si or Ge, do
not form bronzes. TBs shows a high degree of chemical inertness, especially toward
nonoxidizing acids (including HF).
According to the above general formula, a phase diagram could be constructed
describing the regions for the existence of the following groups of compounds (Fig. 4.10):
monotungstates and polytungstates;
oxides and bronzes of ideal composition;
oxygen-deficient oxides;
reduced polytungstates;
bronzes with cation excess;
bronzes with cation deficiency;
oxygen-deficient oxides with cation contamination.

161

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

TABLE 4.9. Tungsten Bronze Forming Elements [4.30]

GroupQ

Element

Structureb

Li
Na

0-{).5
o-{).I1
0.26-0.38
0.41-0.95
0.06-0.10
0.19-0.33
0.06-0.10
0.19-0.33
o-{).I3
0.03-0.05
0.14-0.16
0.20-{).21
0-{).08
o-{).Ol
0-{).03
0-{).04
0-{).02
0-{).12
0-{).03
0.20-{).33
?-O.10
0.13-0.33
0.04-0.06
0.15-0.18
0.21-0.29
0.03-0.05
0.18-0.35
0-{).07
0-{).07
0.12-0.20
0-{).05
0.05-0.07
0.02-0.19
0.Q7-?
0.Q7-?

P
P
T
P
10
H
10
H
P
Is
PC
T
P
P
P
P
P
P
P
H
Is
H
Is
10
T
Is
T
P
P
Is
P
Is

Rb
Cs
2

Ca
Ba

4
7
8
9
10
12

Zr
Mn
Fe
Co
Ni
Cd
In

13

Tl

14

Sn

Pb
15

As
Sb
Bi

Lanthanoid series
Actinoid series

La
U
Th

P
P

New IUPAC proposal 1985.

Abbreviations: P, perovskite; T, tetragonal; H, hexagonal; I, intergrowth; Is, single row of tunnels in parent W03 matrix; 10. double
tunnels; PC, pentagonal column bronzes related to tetragonal TB
type.

STR:

A variety of different structures are known:

The most frequent is the cubic perovskite structure. The ternary metal atoms
occupy the cage sites present in the W03 host framework (Fig. 4.11).
Tetragonal tungsten bronzes are constructed of pentagonal bipyramides forming
pentagonal columns (PCs); they can be described as an ordered arrangement of
PCs within the W0 3 parent structure. The foreign metal atoms are located in the
pentagonal tunnels and sometimes also in the square tunnels (Fig. 4.12).

162

CHAPTER 4

A
(z . x)

...

0.3
y<

0.2

oxidation
reduction
~

.2

"0

o
c
o

z.x

Y>-2-

0.1

W20 0 S8

(W n0 3n - 2 sheared structures)
0.04

0.08

0.12

0.16

0.20

FIGURE 4.10. Idealized phase diagram of the M-W-O system [4.45]: (A) monotungstates and polytungstates;
(B) oxide and bronzes of ideal composition; (C) oxygen-deficient oxides; (D) reduced polytungstates; (E) bronzes
with cation excess; (F) bronzes with cation deficiency; (G) oxygen-deficient oxides with cation contamination.

PR:

In hexagonal tungsten bronzes comer-shared W06 octahedra are arranged to fonn

hexagonal tunnels (Fig. 4.13.). The metal atoms are interpolated into the tunnels.
Their composition range varies between x=0.15 to 0.33. They fonn when larger
metal atoms are reacted with W03 at high temperature.
Intergrowth tungsten bronzes exist when hexagonal tungsten bronzes coherently
intergrow with the parent W03 structure (Figs. 4.14 and 4.15).
Chemical reduction in acidic solution, either by metal amalgams or electrolytically
(mainly for hydrogen bronzes).
Thennal reduction during many different reactions (solid/solid, molten/molten,
solid/gaseous state reaction, etc.).
Electrochemical reduction of molten mixtures.

Sodium Tungsten Bronzes

They are the most extensively investigated tungsten bronzes.
PR:

Electrolytic reduction.
Vapor-phase deposition.
Fusion.
Solid state reaction at 500-850 C in vacuum.

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

~NaI
\;;;iI

163

EDW 00
I

FIGURE 4.11. Perovskite structure: (a) Single cell; NaW03. (b) Idealized: the large metal atoms are interpolated
into the cage sites in the W03 parent matrix of W03 octahedra; the perovskite bronzes have a partial filling of the
structure to produce a nonstoichiometric phase [4.30].

FIGURE 4.12. Tetragonal tungsten bronze structure: the circles represent possible metal atom sites and the
shaded squares W06 octahedra [4.30).

164

CHAPTER 4

FIGURE 4.13. Hexagonal tungsten bronze structure: the circles represent the positions occupied by a variable
population of interpolated metal atoms [4.30).

PP:

A:

CO: intensely colored from golden yellow to bluish black, depending on x, metallic
sheen; CR: cubic E2), perovskite type.
x

a (A)

Color [4.4]

0.93
0.64
0.46
0.32

3.850
3.834
3.825
3.813

Golden-yellow
Red-yellow
Red-violet
Blue-violet

temperature coefficient of resistivity: x> 0.3 -+ positive; x < 0.3 -+ negative.

Promotes the catalytic oxidation of CO or reformer gas in fuel cells.

Ammonium and Hydrogen Tungsten Bronzes

Besides the above-mentioned elements, hydronium and ammonium ions also form
TB compounds.
Ammonium tungsten bronzes correspond to x = 0.06-0.33; y = 0.030-0.165.
With the exception of the lowest x and y: <NH4)O. 06 W03 (H20)o. 11 , which is
tetragonal, they belong to the hexagonal system. The dominance of the hexagonal
symmetry can be understood as a consequence of the large size of the ammonium ion.
They are partially contained in the industrial tungsten blue oxide, especially at low APT
decomposition temperatures or short exposure times.
In hydrogen tungsten bronzes x values between 0.03 and 0.53 and y values from
0.015 to 0.3 can be found. They have quite different structures (tetragonal, orthorhombic,
monocline, hexagonal, and cubic) and can also be detected in tungsten blue oxides.
We refer the reader to a complete literature survey on tungsten bronzes presented
elsewhere [4.32].

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

165

FIGURE 4.14. Intergrowth tungsten bronze structure oftbe pyrochlore type (single tunnel) [4.30].

4.2.8. Tungsten and Sulfur [4.4, 4.7, 4.8, 4.11, 4.25, 4.46]

The binary compounds WS 2 and WS 3 exist.

Tungsten Disulfide: WS2
PR: Direct synthesis of W and S at 800-900 c in pure nitrogen atmosphere (f3-WS2).
Sulfur vapor over W03 at 1400 DC; 2W03 + 4S ~ 2WS2 + 302 (cx-WS 2).
Thermal decomposition of WS 3
PP: D: 7.73 g.cm- 3 ; MP: 1250 c (decomposition); CO: black.
CR: f3-WS~: hexagonal C7; MoS2(h)-type, a =3.145-3.165 A, c=12.25-12.35 A;
a.-WS2' rhombohedral, MoS2(r)-type; a=3.162 A, c= 18.50 A.
WS2 has a layer structure, each metal atom being surrounded by a trigonal, prismatic
array of S atoms; WS 2 is a diamagnetic semiconductor.
CP: It is insoluble in water, diluted nitric and sulfuric acid. It dissolves in HF-aqua regia
mixture, molten alkalis, and alkali carbonates. It decomposes by heating in air, but
its thermal stability is 90C higher than that of MoS 2

166

CHAPTER 4

FIGURE 4.15. Intergrowth tungsten bronze structure of the pyrochlore type (double tunnel) [4.30].

A:

Fonns adherent, soft, continuous films on different surfaces and exhibits good
lubricant properties, similar to MoS2 Is used as lubricant.

Tunsten Trisulfide: WS3

PR:

It can only be produced indirectly by decomposing sodium thiotungstate with hot

sulfuric or hydrochloric acid.

PP:

Reddish brown, amorphous powder; diamagnetic and semiconducting.

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

CP:

167

It is slightly soluble in cold water and colloidal in hot. It dissolves in alkali

carbonates or hydroxides. Decomposition to WS 2 and W occurs at 170C.

4.2.9. Tungsten and Selenium [4.4,4.7, 4.8, 4.11, 4.25, 4.46]

As with sulfur, tungsten forms two binary compounds: WSe2 and WSe3.
Tungsten Diselenide: WSe2
PR: By direct reaction of the elements at elevated temperature in an inert atmosphere.
By thermal dissociation of WSe3'
PP: D: 9.22-9.32 g. cm- 3 ; MP: 800C (decomposition); CO: black; CR: hexagonal C7 ,
MoSez-type; a = 3.28 A, c = 12.95 A.
CP: Only barely dissociated between 700C and 800 c in vacuo, but at higher
temperature rapid dissociation.
A:
It is in use as lubricant, like the analogous sulfur compound.
Tungsten Triselenide: WSe3
PR: Sodium tungstate solution is saturated with H2Se, whereby selenotungstate ions
[WSe4]2- are formed which are subsequently decomposed to WSe3 by acidification.
PP: Black, amorphous powder.
CP: It is soluble in concentrated HCl and decomposes at 220C to WSe2'

4.2.10. Tungsten and Tellurium [4.4,4.8,4.11]

The only known binary compound is WTe2'

Tungsten Ditelluride: WTe2
PR: By direct combination of the elements below 750C under vacuum.
PP: D: 9.49 g. cm- 3; MP: 972 c (decomposition); CO: yellow, crystalline needles.
CP: It is insoluble in water, ammonia, hydrochloric, and sulfuric acid, but is decomposed
by nitric acid. Reaction with oxygen occurs at 650-700 C and thermal decomposition in vacuum starts at 600-620 0c.

4.2.11. Tungsten and Fluorine [4.1, 4.4, 4.7, 4.11, 4.25, 4.46]

Tungsten halide compounds in general and the fluorides and chlorides in particular
are the basis for sytheses of a vast variety of substitution, mixed ligands, and coordination
compounds, because they are stable compounds and, especially, halide atoms are easy to
substitute. Halide atoms also may be substituted by each other, such as in fluoro chloro
compounds of the general formula WF n CI6 _ n Besides the hexavalent species WX6,
WO~, W02X 2, the lower state compounds are also of interest as precursors. The special
importance of the oxo and dioxo halides is that they are coordinatively unsaturated.
Therefore, they easily form coordination compounds by addition. Analogously to the
oxohalides, thio-, seleno-, and nitridohalides are known.

168

CHAPTER 4

Tungsten fonns with fluorine the binary compounds WF 6, WF s, and WF4 while the
following ternary compounds are fonned with fluorine and oxygen: WOF4 , W02F2, and
WOF 2

Tungsten Hexafluoride: WF6

The direct florination of tungsten in a flowing system at 350-400 DC is the most
convenient procedure: W + 3F2 --+ WF 6 .
By reaction of tungsten hexachloride with HF: WCI 6 + 6HF --+ WF6 + 6HCI; or
with arsenic trifluoride: WCI 6 + 2AsF3 --+ WF6 + 2AsC13; or with antimony pentafluoride: WCI6 + 3SbFs --+ WF6 + 3SbF3C12 .
Tungsten also reacts with CIF, CIF3, and BrF3 to fonn WF 6 . W0 3 reacts with HF,
BrF3, and SF4 to WF 6.
PP: :::::2.0 DC white solid; CR: 0 DC cubic, a = 6.28 A; - 30C orthorhombic;
a=2.68 A, b=18.81 A, c=5.09A; between 2.0 DC and 19.9C pale yellow
liquid; specific gravity: 3.44 g. cm- 1 (at 15 DC); ::: 17.5 DC, colorless gas; D: related
to hydrogen 149; related to air 12.9; it is one of the heaviest gases known.
CP: It is unstable in the presence of moisture: WF6 + 4H 20 --+ H2 W04 + 6HF.
Soluble in benzene and cyciohexane (bright red), in dioxane (pale red), and in ether
(violet-brown). It is very reactive, attacks most metals, erodes glass slowly in the
absence of moisture, but rapidly in the presence of moisture.
A:
WF6 is today widely used by the semiconductor industry in the manufacture of
integrated circuits as metallization source (chemical vapor deposition of W; see
Sections 5.6.3 and 5.7.5). Due to the "explosion" of the electronics industry the
demand of WF6 during the past decades has been rapidly growing and remains
presently at approximately 200 t/year worldwide [4.47]. This corresponds to 127 t
of W or 0.3--0.4% of the world total production. WF6 for that purpose must be of
highest purity ranging from 99.9995 to 99.98%, depending on application.
PR:

Tungsten Pentafluoride: WFj

PR: By reduction of WF6 on a hot tungsten filament.
STR It has a tetrameric structure [(WFS)4]; the four Watoms fonn a square plane bridged
by four F atoms; each Watom is octahedrallY coordinated to four tenninal fluorides.
CO: Yellow.

Tungsten Tetrafluoride: WF4

Reduction of WF6 with PF3 in presence of liquid anhydrous H2F2.
Nonvolatile, hygroscopic, reddish brown solid; CR: orthorhombic.

PR:
PP:

Tungsten Oxytetrajiuoride: WOF4

PR: By sustitution of CI: WOCI 4 + 4HF --+ WOF 4 + 4HCl.
By reaction of W0 3 with lead fluoride at 700 DC: 2PbF 2 + W0 3 --+
WOF 4 +2PbO.
A mixture of oxygen and fluorine reacts with W at elevated temperature to WOF 4.
PP: It is a crystalline, white substance; MP: 1l0-1l9 DC; BP: 187.5-190 DC.
STR: The structure is tetrameric [(WOF4)4] with W-F ... W bridges and tenninal W=O
groups.

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

169

CP: It is extremely hygroscopic (decomposition to H2 W04 and HF) and it is soluble in

CHCh and slightly soluble in CS 2.
Tungsten Dioxydifluoride: W0 2F2
PR: Hydrolysis of WOF4.
PP: White solid.
Tungsten Oxydifluoride: WOF2
PR: Reaction of tungsten dioxide and HF at 600C: W0 2 + 2HF -+ WOF 2 + H20.
PP: Inert, gray solid substance; polymeric with 0 and F bridges.

4.2.12. Tungsten and Chlorine [4.l, 4.4, 4.7, 4.l1, 4.25, 4.46]
The binary compounds WCI 6 , WCI 5, WCI4, and WClz and the ternary oxychlorides
WOCI4, W0 2Clz, WOCh, and WOClz exist.
Tungsten Hexachloride: WCl6
PR: By direct chlorination of tungsten in a flowing system at 600C. On cooling
between 168 C and 170C a violent, explosion-like expansion of the solid occurs
due to an Cl2-Cll transition.
By reaction of compounds like S2Clz, PCI5, HC1, COCI 2, or CC14 with tungsten or
tungsten trioxide.
PP: D: 3.52 g. cm- 3 ; MP: 275C; BP: 346.7 C; CO: blue black solid; CR: hexagonal;
a=6.100 A, c= 16.70 A.
CP: It is very soluble in CS 2, alcohol, CCI3, CCI4, ether, benzene, acetone, and
ammonia. Decomposition with water starts above 60C. Gaseous WC16 will be
stepwise reduced by hydrogen to lower chlorides and finally to W.
A:
It is commercially available and is used as starting material for the chemical vapor
deposition of tungsten (see Sections 5.6.3. and 5.7.5) and for the preparation of
metathesis catalysts which form double and triple bonds with carbon (see Chapter
10).
Tungsten Pentachloride: WCl 5
PR: Reduction of WC1 6 by red phosphorus at 280C or by hydrogen at 400 0C.
Reaction of WC16 with tetrachlorethene at 100C: 2WC16 + C2C4 -+
2WC15 + C 2Ck
Disproportion of WCI 4.
PP: Crystalline, deliquescent solid; D: 3.875 g. cm- 3 ; MP: 243-248 C; BP: 276C;
CO: black or dark green.
CP: Hygroscopic, decomposes in water to W20058, slightly soluble in CS 2, disproportionates into WC4 and WCI 6 . It is dimeric (WCI5h- The long W-W distance of
3.814 A and a magnetic momentum of 1.0 J.1B (per atom at room temperature)
indicate metal bonding.
Tungsten Tetrachloride: WCl 4
PR: Reaction of WC1 6 with W(CO)4 or P4 in a sealed tube.
Thermal gradient reduction of WCl6 by Al or hydrogen.

170

CHAPTER 4

PP:

Nonvolatile, crystalline, deliquescent solid; diamagnetic; D: 4.624 g. cm-3; CO:

black; CR: orthorhombic; a = 8.07 A, b = 8.89 A, c = 6.85 A.
It is hygroscopic and hydrolyzes, sublimes, and decomposes on heating. Consists of
infinite linear chains of octahedra sharing a common edge. The alternating short and
long metal distances imply metal-metal bonding.
It is a useful precursor for many coordination compounds.

CP:

A:

Tungsten Trichloride: WCl3

PR:
CP:

By oxidation of [(W6CIs)CI4] with chlorine.

Consists of clusters [(W6{JlrCI}I21CI6; CO: red.

Tungsten Dichloride: WCl2

PR:
PP:
CP:

Reduction of WCl 6 with Al in a NaAICl4 melt or with sodium amalgam.

Disproportionation of WC4.
Not volatile, not fusible, amorphous powder; D: 5.436 g. cm- 3; CO: gray or yellow.
It is a cluster compound containing the face-bridged octahedral core tetracation
[W6{Jl3-CI }g]4+; [W6CIs]CI4.

Tungsten Oxytetrachloride: WOCl4

PR:

PP:
CP:

Refluxing SOCh on W0 3 (purification by sublimation).

Direct reaction of chlorine with W powder in presence of oxygen or water vapor.
Reaction of WCl6 and W0 3 : W0 3 + 2WCl6 --* 3WOC4.
2CChF 2 + 2W02 --* WOCl4 + WOF4 + CO.
Crystalline solid; D: 11.92 g. cm-3; MP: 209-211 DC; BP: 327C; CO: red.
Soluble in CS 2, benzene, sulfur chloride; reacts with water to form tungstic acid;
carbon --* WCI6; oxygen --* W02 Ch; dry NH3 --* W nitrides; PCl5 --* WCI6;
HF --* WOF4; it is a polymer containing slightly distorted OWOCl4 octahedra
connected via C4W =0 --* (W =O)C1 4 bonding.

Tungsten Dioxydichloride: W02 Cl2

PR:
PP:
CP:

Reaction of W02 and CCI4 in a bomb.

Chlorination of W0 3 by chlorine at 900C or by HCI at 300C.
Crystalline solid; MP: 266C (sublimes); CO: pale yellow.
Soluble in cold water and in alkaline solutions; decomposes in hot water and is
insoluble in common organic solvents. It is polymeric and consists of [{W0 2CI 2 }n]
with W-0 ...W bridges.

Tungsten Oxytrichloride: WOCl3

PR:

PP:
CP:

Heating of a mixture of WCl 5 and Sb20 3 at 100C up to 170 DC.

3WCl5 + Sb20 3 --* 3WOCI3 + 2SbC13
Reduction of WOCl4 by Al in sealed tubes at 100-140 DC.
Crystalline solid; CO: green.
Infinite W-O-W chains.

Tungsten Oxydichloride: WOCl2

PR:

Heating of a mixture of W, W0 3, and WCl6 at 250-450 DC.

W + W0 3 + WCl6 --* 3WOCI2

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

PP:
CP:

171

Thennal decomposition of WOCl3 results in a blue-black polymorph.

A golden brown polymorph fonns by reduction of WOCl4 with SnCh.
Needle-like crystals with copper luster; D: 5.92 g. cm- 3; CR: monoclinic;
a= 12.87 A, b=3.76 A, c=6.46 A, ~= 104.2.
On heating in air it fonns W0 2Ch and finally W0 3. It does not react with acids at
room temperature, but is soluble in hot nitric acid and hydrogen peroxide. It is a
polymer with 0 and CI bridges.

4.2.13. Tungsten and Bromine [4.1,4.4,4.7,4.11,4.25,4.46]

As in the case of chlorine, there exist five binary compounds from WBr6 to WBrz,
and two ternary compounds with bromine and oxygen, WOBr4 and W02Br2.

Tungsten Hexabromide: WBr6

PR: Metathetical exchange reaction of BBr3 with WCk
Reaction of bromine vapor and tungsten powder.
PP: Crystallized substance; MP: 232C; CO: bluish black.
CP: Decomposes on heating, hydrolyzes easily, and dissolves in ammonia.
Tungsten Pentabromide: WBrs
PR: Reaction of bromine vapor with tungsten powder at 450-500 DC.
Reduction of WCl6 with HBr.
PP: Crystallizes in needles; MP: 276C; BP: 333C; CO: brownish black to violet
brown.
CP: Extremely sensitive to moisture, unstable in air, soluble in ether, CCl4, CHCI3,
CHBr3, and benzene.
Tungsten Tetrabromide: WBr4
PR: Reduction of WBrs by Al or W
Decomposition ofWBrs at 180C in vacuum.
PP: CR: orthorhombic; a = 8.49 A, b = 9.29 A, c = 7.25 A; CO: black.
CP: Decomposes at 400C; has the same structure as WC14.
Tungsten Tribromide: WBrJ
PR: WBr2 will be soaked in liquid bromine at 50C.
PP: Black powder.
CP: Disproportionates at 450-500 C into WBrs and WB 2, and is soluble in water.
Tungsten Dibromide: WBr2
PR: Reduction of WBrs by hydrogen in presence of ZnCl2 at 400C.
Disproportion of WCl4 at 450C: 3WBr4 --+ WBr2 + 2WBrs.
PP: Black solid.
CP: Decomposes at 400C.
Tungsten Oxytribromide: WOBrJ
PR: Heating a mixture of WBrs and Sb20 3 under vacuum at 100C for 1 hour and at
150C for 24 hours: 3WBrs + Sb20 3 --+ 3WOBr3 + 2SbBr3.

172

PP:
CP:
SIR:

CHAPTER 4

Brown fibrous structure.

Decomposes at 250C: 3WOBr3 -+ WOBr4
It consists of infinite W-O-W chains.

+ W02Br2 + WBr3'

Tungsten Oxytetrabromide: WOBr4

PR: By reaction of bromine vapor with a mixture of W03, C and W.
By reaction of CBr4 with W0 2 at 250C.
2W + W03 + 6Br2 -+ 3WOBr4.
PP: Deliquescent needles; MP: 277C; BP: 327C; CO: black.
CP: Very hygroscopic, decomposes rapidly in air, the polymeric structure is analogous to
WOC4.
Tungsten Dioxydibromide: W02Br2
PR: Passing a mixture of oxygen and bromine over tungsten at 300C or higher.
PP: Infusible red or brown substance.
CP: Less volatile than WOCI4, decomposes on heating: 2W02Br2 -+ W0 3 + WOBr4;
the polymeric structure is analogous to W02CI2.

Bromine additions play an important role in tungsten-halogen lamps (see Chapter 7).
It is commonly added as alkylbromides.
4.2.14. Tungsten and Iodine [4.1, 4.4, 4.7, 4.11, 4.25, 4.46]

The following binary tungsten-iodine and ternary tungsten-oxygen-iodine

compounds are known: W14, W13, WI2(W61\2), W6115 , W0I2, W021z, W021, and W0I4.
Tungsten Tetraiodide: WI4
PR: Reaction of concentrated hydroiodic acid with WCl 6 at 100C.
WC16 with HI gas at 110 DC.
W0 2 and A1I 3.
PP: Black crystallized powder; D: 5.2 g. cm- 3.
CP: Will be decomposed by air to WI2 and 12, is soluble in water, but hydrolyzes;
insoluble in ether and CHCh, slowly soluble in cold alcohol; chlorine replaces
iodine at 18C, and bromine replaces it at 100C. Dissolution and iodine release
occurs with KOH solutions, molten alkali carbonates, molten potassium bisulfide,
and liquid potassium.
Tungsten Triiodide: WIj
PR: Reaction of iodine gas and W(CO)6 in sealed tubes at 120C.
Is also one of the products of the reaction of W61 \2 with iodine gas.
PP: Gray needles.
Tungsten Diiodide: Wh

PR:
PP:

Passing iodine gas over tungsten powder at 800C.

Reaction of iodine and W(CO)6 in nitrogen.
Brownish crystalline substance; hexamer structure: [W6Ig]1214/2'

173

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

Hexatungsten Pentadecaniodide: WJl5

PR:
PP:
CP:

By reaction of iodine gas and W6I\2 at 350C in vacuum (coexists with WI3).
Brown.
Stable in air, decomposes in KOH solutions and in concentrated sulfuric acid, and
also has a hexameric structure [W6Is]4[I3h/2'

Tungsten Oxydiodide: WOh

PR:

PP:
CP:

By heating a mixture of tungsten, tungsten trioxide, and iodine powder under

vacuum in a two-zone furnace (500C and 700C): 2W + W0 3 + 3h -+ 3WOh
followed by purification via benzene extraction.
Diamagnetic glittering plates; CO: black; CR: a= 18.01 A, c=7.58 A.
Slightly soluble in water and organic solvents and very soluble in NaOH and KOH
solutions.

Tungsten Dioxydiiodide: W02h

PR:
PP:
CP:
A:

Direct reaction of W0 2 and iodine gas.

D: 6.39 g. cm-3; MP: 6.40C; BP: 200C; CO: dark brown; CR: monocline,
a= 17.095 A, b=3.899 A, c=7.492 A, ~= 102.66.
Decomposes at 277 c to W02I and iodine. As a side product also WOh is formed.
Formation of gaseous W0 2I2 is considered to effect the removal of tungsten from
lamp walls in halogene lamps (Chapter 7).

4.3. MIXED LIGANDS AND COORDINATION COMPOUNDS OF TUNGSTEN

[4.1, 4.25, 4.46]

4.3.1. Mixed Ligand Compounds

Some compounds belonging to this group have already been treated in the foregoing
subsection, namely, the oxide hydrates and oxohalides. Mixed ligand compounds may
contain only inorganic, or mixed inorganic and organic, or merely organic ligands. A large
number of these compounds are coordination compounds (complexes) and a high
percentage are organometallics. Therefore, it is not easy to draw a borderline between
these groups and in various cases some overlapping occurs. For example, carbonyl,
thiocarbonyl, cyanide, and isocyanide ligands are counted with the organometallics in
some publications while others regard them as pure inorganic ligands.
Some typical mixed ligand compounds are substituted carbonyl or halide compounds
of the general formula:

x = halide or carbonyl
L = different ligands
Typical
examples
are
[W(CO)sCl],
[W(CO)4Br2],
[W(COhCI3(..,5 -CsHs)], and [W(CO)(..,s -CsHs)].

[W(CO)3CI(..,5-C5H5)],

174

CHAPTER 4

4.3.2. Coordination Compounds

The chemistry of tungsten coordination compounds is quite complicated. The reasons

are:
Tungsten forms complexes in the valence states -2, -1,0, +2, +3, +4, +5, +6.
The coordination number is variable (max 13).
Tungsten has the tendency to form clusters and polynuclear complexes with
varying number of atoms. In these compounds the tungsten-to-tungsten bond
varies between single and quadruple. The lower the valence state the higher the
degree of W-W bonding.
A short survey of the possible donor atoms and corresponding ligand ions or groups is
given in Table 4.10.
Hexavalent tungsten can be found in the form of hexahalides, which give rise to a
variety of substitution products containing the structural unit W6+. Hexavalent tungsten in
addition shows a strong tendency to form bonds of higher order than one with donor atoms
like 0, S, Se, or N. Consequently, one must distinguish between complexes of the
structural units W6+, WOH, and W02H and the analogous complexes which contain =S,
=Se, or, =NR instead of oxygen. Thiotungstate ions like WS4 2-, WOS 3 2-, and W02S22act as bidentate ligands for other metal ions.
In pentavalent tungsten complexes the units W5+, WOH, and WOH also exist, but in
addition dimeric structures can be found containing the W204 H unit. They are not as
common as in pentavalent molybdenum chemistry.
Tetravalent tungsten complexes contain the structural units WH, WOH, or dimeric
configuration having W=W double bonds, as well as trinuclear clusters with three W
atoms bound together in a triangular configuration.
Most complexes of trivalent tungsten are dinuclear such as, for example, in the
[W2C19]6 3- ion with some of them having a triple bond between the two tungsten atoms.
In divalent tungsten complexes besides monomer configurations, also dimeric
configurations having a quadruple W-W bond are known.
TABLE 4.10. Donor Atoms and Corresponding Ions or Groups in Tungsten Coordination
Compounds
Donor atom
C
N
P
As

S
Se
Te
F

Cl
Br

Ion/group
Alkylidine, carbonyl, isocyanide, cyanide
Nitrido, imido, amino, thiocyanate, nitrilo, pyridine, dinitrogen, diazenido, hydrazido, diazene,
nitrene, hydroxylamino
Phosphine derivatives
Arsine derivatives
Aqua, oxaqua, oxo, peroxo, alkoxide, aryloxide, carboxylate, p-diketonato
Sulfido, persulfide, thiolato, dithiocarbamate, dithiol, xanthate
Selenido
Tellurido
Fluoride
Chloride
Bromide

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

175

As already mentioned above, 0, - I, and - 2 valent tungsten complexes also exist.

The ligands can be inorganic or/and organic ions and groups, and can be mono- or
polydentate. A huge number of mixed ligand complexes as well as chelate complexes have
been prepared. Further details can be found in other compilations [4.25,4.46,4.50,4.52].
Some typical examples are shown below.

TABLE 4.11. Typical Examples for Organometallic Tungsten Compounds

Ligand type

Typical examples

W-C-

Alkyl

(CH3)WCI s

rOt

Aryl

(CJis)WCI 3

Alkene-(1)yl

("s-CsHs)(COhW-CH=C<

Alkyne-( l)y I

K3[W(-C=CHh(COhl

Designation

Interacting atoms

,,1

I
I

W, r!
/c= "
W-C=C-

Acyl

'c=0
Alkylidene

W=r!

"

,,2

CI(COhW=C-R

Alkene

[W(,,2-ethyleneh(CO)41

Alkyne

[W(,,2-alkynehCOI

Allyl

[W(,,3-allyl)X(CO)41

Diene

[W(,,4-dieneh(COhl

Cyclopentadienyl

[W("s-CsHshCI21

I(3)W

Arene

[W(,,6-arene)(COhl

@W

Cycloheptatrienyl

[W(,,7-C7H7)(COhl+

,
,

"c=<
/
W

-CaC-

,,3

,,4

w:D
I

C=C~W
\
C=C

,,5

,,6

,,7

YV

",-CH3
(" 5-CsHs)(COhLW-C~
0
R
(COhW=C<
OR

Alkylidyne

WEC-

176

CHAPTER 4

Carbonyl complexes and derivatives: [W(CO)4]4-, [W(co)sf-, [W 2(CO) 10]2-,

[W(CO)4Br]-, [W(COh(11 5-C 5H5)]-, [W(COM11 5-C 5H5)]+'
Cyanide and Isocyanide complexes and derivatives: [W(CN)s]4-, [W(CN)sP-,
[W(CN)6]4-, [W(CNRh]2-, [W(CNRh(11 5-C5H5)]'
Thiocarbonyl complexes: [W(CS)(CO)4I], [W(CS)(COh(11 5-C5H5)].

4.4. ORGANOMETALLIC TUNGSTEN COMPOUNDS [4.1,4.48,4.49]

Organometallic tungsten compounds can be characterized by having one or several

bondings between a tungsten atom and one or more carbon atoms belonging to one organic
ligand. The bonding between the tungsten and one carbon atom can be single, double, or
triple.
The number of compounds synthesized so far is extremely large. The organic ligands
can be classified by the number of carbon atoms (1-7) interacting with tungsten, as will be
shown in Table 4.11.
The most important precursors for the preparation of organometallic tungsten
compounds are the halides and oxohalides, as well as hexacarbonyl.
In lower valence states stabilization is gained by n-bonding. The preferential
oxidation state is 2+, but also lower and higher valent compounds are known. The
most important class of these compounds is cr-organyl-tricarbonyl-(cyclopentadienyl)tungsten and their derivatives (especially the pentamethyl-cyclopentadienyl) or analogous
compounds having six-membered ring systems. The general formula corresponds to
W(COh(115-C5H5)R where R can be alkyl, acyl, alkenyl, or aryl. These compounds are
used in catalytic systems of diverse organic syntheses (see Chapter 10). For more details
reference is made to the compilations [4.48,4.49, and 4.51].

REFERENCES FOR CHAPTER 4

4.1.

N. N. Greenwood and A. Earnshaw, The Chemistry of the Elements, p. 1291, Pergamon Press, Oxford
(1984).
4.2. I. Barin, Thermochemical Data of Pure Substances, 3rd edition, Vol. II, VCH, Weinheim (1995).
4.3. O. Knacke, O. Kubaschewski, and K. Hesselmann, eds., Thermochemical Properties of Inorganic
Substances II" 2nd edition, Springer-Verlag, Heidelberg (1991).
4.4. S. W. H. Yih and C. T. Wang, Tungsten, Plenum Press, New York (1979).
4.5. Gmelin Handbook of Inorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. A6a, SpringerVerlag, Heidelberg (1991).
4.6. Gmelin Handbook of Inorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. A6b, SpringerVerlag, Heidelberg (1988).
4.7. 1. A. Mullendore, in: Kirk-Otmer: Encyclopedia of Chemical Technology, Vol. 23, 3rd ed., Tungsten and
Tungsten Alloys, Wiley, Chichester (1983).
4.8. S. V. Nagender Naidu and P. Rarna Rao, Phase Diagrams of Binary Tungsten Alloys, Indian Institute of
Metals, Calcutta (1991).
4.9. E. Rudy and S. Windisch, Ternary Phase Equilibria in Transition Metal-B-C-Si Systems, Part I, Vol. 3,
Technical Report AFML-TR-65-2 (1965).
4.10. H. Duschanek and P. Rogl, J. Phase Equilibria 16 (1995), 150-161.
4.11. Gmelin Handbook of Inorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. A5b, SpringerVerlag, Heidelberg (1993).

TUNGSTEN COMPOUNDS AND THEIR APPLICATION

4.12.
4.13.
4.14.
4.15.
4.16.
4.17.
4.18.
4.19.
4.20.
4.21.
4.22.
4.23.
4.24.
4.25.
4.26.
4.27.
4.28.
4.29.

4.30.
4.31.
4.32.
4.33.
4.34.
4.35.

4.36.
4.37.
4.38.
4.39.
4.40.
4.41.
4.42.
4.43.
4.44.
4.45.
4.46.
4.47.
4.48.
4.49.
4.50.

4.51.
4.52.

177

F. Binder, Radex Rundsch. 4 (1975), 531-557.

E. Rudy, Report AFML-TR-65-2, Part V, Aerojet-General Corp., Sacramento, California (1969).
E. Rudy, Report AFML-TR-2, Part II, Vol. 18, Wright-Patterson Airforce Base, Ohio (1968).
T. B. Massalski, Binary Alloy Phase Diagrams, 2nd edition, Vol. I, p. 896, ASM International (1990).
P. Gustafson, Mater. Sci. Technol. 2 (1986), 653--658.
E. Rudy and 1. R. Hoffman, Planseeber. Pulvermetall. 15 (1967),174-178.
K. Yvon, H. Nowotny, and F. Benesovsky, Monatsh. Chem. 99 (1968),726-729.
D. N. French and D. A. Thomas, Trans. AIME 233 (1965), 950.
M. Lee, Metall. Trans. 14A (1983), 1625.
L. Pons, in: Anisotropy in Single-Crystal Refractory Compounds (F. V. Vahldiek and 1. Mersol, eds.), Vol. 2,
pp. 393-444, Plenum Press, New York (1968).
B. Roebuck, Int. J Refract. Met. Hard Mater. 13 (1995), 265-279.
Y. Ishizawaand T. Tanaka,lnst. Phys. Con! Ser. No. 75, pp. 29-43,AdamHilger Ltd., BristolandBoston(1986).
L. F. Mattheis and D. R. Hamann, Phys. Rev. B 30 (1984), 1731-1738.
1. A. McCleverty, "Tungsten; Inorganic and Coordination Chemistry," in: Enyclopedia of Inorganic
Chemistry (R. B. King, ed.), Vol. 8, pp. 4240-4268, Wiley, Chichester (1994).
V. A. Maksimov and P. T. Shamrai, Inorg. Mater. 5 (1969), 965-966.
H. A. Wriedt, Bull. Alloy Phase Diagrams 10 (1989), 358-367.
A. F. Guillermet and S. Jonsson, Z. Metallkd. 84 (1993), 106-117.
Gmelin Handbuch der Anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. B2, SpringerVerlag, Heidelberg (1979).
R. 1. D. Tilley, Int. J Refract. Met. Hard Mater. 13 (1995), 93-109.
T. Ekstrom and R. 1. D. Tilley, Chem. Scr. 16 (1980), 1-23.
Gmelin Handbuch der Anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten, SUppl. Vol. B3, SpringerVerlag, Heidelberg (1979).
Gmelin Handuch der Anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. B4, SpringerVerlag, Heidelberg (1980).
Gmelin Handbook ofInorganic ChemiStry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. B5, Springer-Verlag,
Heidelberg (1984).
Gmelin Handbook ofInorganic Chemistry, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. B6, Springer-Verlag,
Heidelberg (1984).
M. A. Sobhan, R. T. Kivaisi, B. Stjema, and C. G. Granqvist, Sol. Energy Mater. Sol. Cells 44 (1996), 451-455.
T. Millner, A. J. Hegediis, K. Sasvari, and 1. Neugebauer, Z. Anorg. Allg. Chem. 289 (1957), 288-312.
W R. Morcom, W L. Worrell, H. G. Sell, and H. 1. Kaplan, Metall. Trans. 5 (1974), 155-161.
A. Bartl, Doctoral Thesis, TV Vienna, Austria (1997).
1. S. O. Evans, T. A. Mary, T. Vogt, M. A. Subramanian, and A. W Sleight, Chern. Mater. 8 (1996),2809-2823.
W P. C. Duyvesteyn, H. Liu, N. L. Labao, and P. L. Shrestha, US Patent 5,178,848 (1993).
1. W van Put, Doctoral Thesis, Delft University, The Netherlands, Delft University Press (1991).
Sylvania GTE, Tech. Inf Bull. CM-9003-(9j80).
G. A. Tsigdinos, "Heteropoly Compounds of Molybdenum and Tungsten," Climax Molybdenum Co. Bull.
Cdb-12a (Sept. 1966).
L. Bartha, A. B. Kiss, and T. Szalay, Int. J Refract. Met. Hard Mater. 13 (1995), 77-91.
Z. Dori, Prog. Inorg. Chern. 28 (1981), 239-299.
Company brochure, Air Products and Chemicals, Inc., Allentown; PA, USA (1994).
A. Segnitz: "Organo-Wolframverbindungen," in: Houben-Weyl; Methoden der Organischen Chemie (E.
Mii1ler, O. Bayer, H. Meerwein, und K. Ziegler, eds.), Bd. 13/7, pp. 489--520, Georg Thieme, Stuttgart (1975).
A. Mayr, "Tungsten; Organometallic Chemistry," in: Encyclopedia of Inorganic Chemistry (R. B. King,
ed.), Vol. 8, pp. 4268-4284, Wiley, Chichester (1994).
R. Piischel and E. Lassner, "Chelates and Chelating Agents in the Analytical Chemistry of Molybdenum
and Tungsten," in: Chelates in Analytical Chemistry, Vol. 1 (H. A. Flaschka and A. 1. Bamard, eds.), Marcel
Dekker, New York (1967).
G. Wilkinson, F. G. A. Stone, and E. W Abel, eds., Comprehensive Organometallic Chemistry, Pergamon
Press, Oxford (1981).
G. Wilkinson, D. Gillard, and 1. A. McCleverty, eds., Comprhensive Coordination Chemistry, Pergamon
Press, Oxford (1987).

5
Industrial Production
S.l. MINING AND ORE BENEFICIATION [S.1-S.5]

5.1.1. Mining
Tungsten mines are relatively small and rarely produce more than 2000 t of ore per day.
Mining is mainly limited by the size of the ore bodies, which are not very large. Open pitmining is the exception. These mines are short-lived and soon convert to underground
operations.
Mining methods for tungsten ore are not at all exceptional and usually are adapted to
the geology of the ore deposit.

5.1.2. Ore Beneficiation

5.1.2.1. General. The majority of tungsten deposits only contains some tenths of a
percent of W0 3 On the other hand, ore concentrates in international trading require 6S75% W03. Therefore, a very high amount of gangue material must be separated. This is
why ore dressing plants are always located near the mine (to save transportation costs).
Companies which process their own concentrates produce low-grade concentrates (6-40%
W0 3) in order to minimize the loss of tungsten minerals which increases with increasing
concentration grade.
Another important aspect of the beneficiation process today is the disposal of the
separated gangue materials, which in case of flotation also contain chemicals. Especially in
areas with rigid environmental restrictions, a deposition near the plant is sometimes not
possible and transportation over longer distances is therefore necessary. Depending on the
mine conditions, a total or partial refill of the tailings into the mine is possible.
In regard to the beneficiation of ore, the positive properties of the tungsten minerals
are the high specific gravity (scheelite and wolframite) and the ferromagnetism (wolframite). A negative property is their brittleness, leading to a partial loss by too fine particles
formed during the disintegration steps.
In principle, two important properties of the ore determine the flow sheet of an ore
dressing plant:
179

180

CHAPTER 5

1. The particle size of the tungsten ore which detennines the degree of disintegration
necessary to liberate the tungsten mineral (liberation size).
2. The type and concentration of the accompanying minerals (gangue) which have to
be separated dictate the mode and number of separation steps.
Ore beneficiation consists of two main steps: comminution and concentration.
Comminution is first perfonned by crushing. Equipment in use comprises jaw, cone, or
impact crushers working mostly in closed circuits with vibratory screens. The second step
in comminution is grinding, which is undertaken in rod or ball mills working in closed
circuits with classifiers.
For concentration (separation of gangue minerals) several methods can be applied,
depending mainly on the composition of the ore. They include ore sorting, gravity
methods, flotation, magnetic, and electrostatic separation.
Beneficiation of tungsten ores by gravity was the classical method, followed by a
"cleaning" step (Fig. 5.1). The recovery depends on the ore characteristics (mainly
liberation size) and ranges typically between 60 and 85%. The main loss is in the
slimes, because the tungsten minerals are the most friable ones present in ores.
As regards "cleaning," for example, a roast process can be applied to convert pyrite
to a magnetic fonn, followed by its magnetic separation together with garnet and pyroxene.
Another cleaning step in the presence of sulfide minerals would be a sulfide flotation.
Gravity methods can also be applied for both scheelite and wolframite. The usual
equipment consists of spirals, cones, tables, and a sink-float.
In order to optimize the yield, modem technology includes the following additional
steps or combinations:
Preconcentration. This could be accomplished by sorting, use of jigs, or heavy media
separation.
Whole Flotation. If tungsten ore mineralization is too fine, the total amount of mined
ore can be subject to flotation. A corresponding flow sheet of the Mittersill ore dressing
plant (Wolfram Bergbau und Hiittengesellschaft m.b.H.) is presented in Fig. 5.2.

Ore from mine

Crushing

Classification

Grinding

Gravity concentration

Cleaning

Ore concentrate

FIGURE 5.1. Classical gravity beneficiation of tungsten ore.

INDUSTRIAL PRODUCTION

181

heap ore

@
railing
pond
Felbcrtal

slurry
pipeline

ruhlfe lden

hydraulic
backfilling

mine

scheelite concentrate

FIGURE 5.2. Mittersill ore dressing plant (Courtesy of Wolfram Bergau- und Hiittenges. m.b.H., Austria): (1)
impact crusher, (2) elevator, (3) screen, (4) ore bins, (5) ball mill, (6) cone classifier, (7) dewatering screen, (8) low
intensity magnetic separator, (9) rough flotation, (10) conditioner, (11) I. Cleaner, (12) 2. Cleaner, (13) conditioner,
(14) 3. Cleaner, (15) concentrate thickener, (16) drum filter, (17) hydrocyclone, (18) backfilling thickener.

Scavenging Circuits. These are combinations of gravity separation and flotation to

recover the loss via slimes occurring in the classical procedure. A simplified flow sheet of
the Cantung Mine beneficiation plant (Canada Tungsten Mining Corporation) is given in
Fig. 5.3.
A flow sheet can be quite complicated especially in the case of complex ores
containing wolframite and scheelite besides other valuable minerals. As an example, ore
dressing at Xihuashan Mine is presented in Fig. 5.4.
5.1.2.2. Ore Sorting

Ore sorting is a modem preconcentrating stage ahead of the other conventional

dressing methods. In former times, sorting was done by hand. Today, mechanical sorting is
possible.

182

CHAPTER 5

Ore from mine

3 Step crushing

Rod mill

Screens and cyclones

Cu and Bi byproduct plant

Sulfide flotation

Cyclone

Scheelite flotation

70% W03 concentrate

Rougher tables

Cleaner tables

Magnetic separation

1----

72% W0 3 concentrate

~----------------~

FIGURE 5.3. Simplified flow sheet of a combined gravity-flotation procedure as applied at Canada Tungsten
Mining Corporation (before closing down) [5.1].

Successful application of an ore sorter depends on two important criteria:

The valuable fraction must be liberated from gangue at a size range which can be
handled by the machine.
The valuable fraction must be identifiable within the time available for examination
by the machine.
The principle of an ore sorting process is shown in Fig. 5.5. It consists of 4 main parts
which are presentation, sensing, air blast, and data processing. In the presentation step, a
monolayer of particles spread from each other is achieved. In the sensing stage, the
location of valuables is determined, and in the separation step the valuable particles are
blown off, leaving the gangue on the conveyer belt.

INDUSTRIAL PRODUCTION

183
rough concentrate from gravity concenlratlon

Table Iiolalion and flotation

rough concentrates

non.mag~e~c minerals

I I

Electrostatic separation

+--

Rod mill

Wolframite flotallon

flotation

.~

8allmill

Mo-Sf flotation

SUI~ldes

JI

Sulphides Ilotalion

Scheehle

High intensity magnetic separation

magnetic mlOerals

sulphides

Mo-Sf separation

I I

Cu Ilotalion

Slime tables

mid~llng

WO~I~~mile
casSreri,e

Magnetic scp,ualor

Cass1terile
concentrate

Wollramlte
concentrates

Rare earlh
rnlddllngs

Scheellte

concenllales

Molybdenite Bismuth
COIlC.
Cone.

Copper
Cone.

FIGURE 5.4. Flow sheet of the cleaning and comprehensive utilization of Xihuashan tungsten mill [5.4].

Photometric ore sorting can be applied to wolframite. It can also be used for scheelite,
if contained in white quarzite veins. Fluorescence ore sorting is a suitable method for
scheelite.
5.1.2.3. Scheelite Flotation. The existing literature on scheelite flotation IS quite
extensive, because the gangue minerals vary in quality and quantity not only from mine to
mine, but also within the same deposit with progress in mining. In principle, each ore
dressing plant has its own recipe, and this will always be adjusted to the variations in ore
constitution. Although patents have been granted, the detailed procedures are kept secret
by most companies.
The following chemicals and conditions are generally in use:
Collectors. Oleic acid, mixtures of oleic and linolic acid, sodium oleat, tall oil,
saponified tall oil and mixtures of these substances, coriander oil, rape oil, mustard oil,
pyridine base, water-soluble carbohydrates of the green oil type, also in saponated form,
fatty acids and their salts, fish oil, hydroxy-ethylene-cellulose, carboxy-methyl-cellulose.
Frothers. Alcohols, cresylic acid, aerosol, oil of terpentine, ketones.
Depressants. Quebracho or tannin for calcite, formic acid for apatite, lactic acid for
mica, and sodium salts of hydrolyzed polyacrylnitrile.
The optimal pH value is between 9.0 and 10.5 and must be controlled to a tolerance
of 0.1 units, depending on the ore composition and the reagents in use. Sodium
carbonate or sodium hydroxide solutions are used for pH regulation.
Sodium silicate serves as a dispersant for calcite. The addition of polyacrylamide or
flocculents increases the flotation yield.

184

CHAPTER 5

Sized ru n of mine material

Data
processing

---

----;- --

-.

~ ---

Presentation

Randommonolayer
Single file
Bulkpresentalion

Sensing

Photometric (optical)
Radiometric
Conductivity
Ultraviolet
Fluorescence

Separation

Waste

Air blast

Ore

FIGURE 5.5. The ore sorting process (scheme) [5.5].

5_1.2.4. Wolframite Flotation. Wolframite flotation is perfonned similarly to the

above-described scheelite flotation, but is not pH sensitive and can therefore be undertaken
in both acidic and alkaline solutions. Generally, flotation is rarely applied to wolframite.
The reason is that wolframite occurs mainly in vein-type deposits whose mineralization is
much coarser than to most scheelite ores. Therefore, gravity and magnetic methods are
more desirable than flotation.
5.1.2.5. Magnetic Separation. Magnetic separation can be applied for two possible
reasons: to clean scheelite by removing magnetic minerals like garnet, magnetite, and
roasted pyrite; to concentrate wolframite by separation from nonmagnetic minerals like
cassiterite or unroasted pyrite. Dry and wet magnetic separators are employed.
5.1.2.6. High Tension Separation. Electrodynamic or electrostatic separators are
used only for scheelite-cassiterite mixtures_ In contrast to scheelite, cassiterite is
conducting.

5.2. HYDROMETALLURGY
5.2.1. Introduction including Ecological and Economical Considerations [5.6]

This section deals exclusively with the modem process of tungsten hydrometallurgy.
In this introduction the classical and modem concepts are compared, outlining the
differences as well as the advantages and disadvantages. In Figs. 5.6 and 5.7, simplified
flow sheets of the classical and modem processes are given.

W0 3

Calcination

H2W04

W0 3

Calcination

W0 3

Calcination

Scheelite

H2W04

Decomposition by HCI

FIGURE 5.6. Tungsten hydrometallurgy-classical methods.

Decomposition by HCI

CaW04

Precipitation

Purification steps

Sodium tungstate solution

Alkali roast

Wolframite

H2W04

APT

....

186

CHAPTER 5

Raw materials
Scheelite ore concentrates
Wolframite ore concentrates
Oxidized scrap
Alkali digestion

Sodium tungstate solution

Purification steps

Acidic sodium
isopolytungstate solution
Ion exchange
(liquid or resin)

Ammonium
isopolytungstate solution
Crystallization

Ammonium
paratungstate (APT)
Calcination

Tungsten trioxide
or

Tungsten blue oxide (TBO)

FIGURE 5.7. Tungsten hydrometallurgy-Modem Process.

In the classical procedure, scheelite was mainly processed by acid digestion while
wolframite was dissolved by an alkaline digestion. The important solid, pure intermediate
was, in both cases, tungstic acid, derived either directly or after several precipitation steps.
Tungstic acid is formed by a precipitation, which is a more or less instantaneous process.
Foreign ions present in solution during the precipitation are partly entrapped or coprecipitated and contaminate the tungstic acid.
In the modem process, APT is the intermediate which, in contrast to tungstic acid, is
gained by crystallization. Crystallization compared to precipitation is a much slower
process, consequently less of the impurities present in the mother liquor will be contained
in the crystallized product.

INDUSTRIAL PRODUCTION

187

The second important difference between classical and modem technology is the
method used to separate the large amount of sodium ions. The old method used a calcium
tungstate precipitation and subsequent hydrochloric acid treatment for that purpose, while
in the modem process solvent extraction or ion exchange converts sodium into ammonium
isopolytungstate solution.
The modem technology offers the following further advantages:

Different raw materials can be handled in the same equipment,

The energy efficiency is better,
The labor force is less,
The process can be easily monitored,
The efficiency is high,
The final product APT is of high purity and consistency.

Although recycling of chemicals is performed in many cases, such as sulfide recycling in

molybdenum precipitation, ammonia recycling in APT crystallization, and solvent
circulation in solvent extraction, still a certain waste of chemicals and also of energy
exists which add to the high cost and to environmental pollution. The main example is the
sodium carbonate pressure leaching process to digest scheelite. This procedure affords at
least a threefold stoichiometric excess of sodium carbonate, and in addition, the
corresponding amount of sulfuric acid in the subsequent process to neutralize it. Finally,
the equivalent tonnage of sodium sulfate leaves the plant in the solvent extraction raffinate.
Depending in which country the plant is located and at which site, the following
possibilities exist to dispose of this undesired by-product:
Disposal to a nearby river is allowed (lowest cost, but environmental pollution).
Disposal at a distant place (enhanced cost by preconcentration and transport,
environmental pollution).
Disposal not allowed at all (highest cost mainly caused by crystallization, which is
an energy-intensive operation, profitable sale of the product not possible, but
saving of the environment).
Research tendencies to solve this and similar problems can be found in the
literature and will also be discussed in this section.
A simplified flow sheet of a conversion plant which can deal with all types of raw
materials is presented in Fig. 5.8.

5.2.2. Raw Materials and Their Treatment

5.2.2.1. Raw Materials [5.1,5.2,5.6]. There exist three different types of tungstenbearing materials which are currently in use as raw materials for APT production: scheelite
ore concentrates and wolframite ore concentrates are the primary sources, and various
tungsten-containing scrap materials act as secondary sources.
Primary Tungsten Sources. Marketable ore concentrates of both types usually
contain between 65 and 75% W0 3 . This guarantees a low percentage of foreign material
in contrast to low-grade concentrates, which are in use in fully integrated companies (from
mine to APT). Those contain typically between 6 and 40% W0 3 and a correspondingly

CHAPTER 5

188

Scheelite
concentrates

Wolframite
concentrates

Roasting
Grinding

Oxidation
Grinding
Screening

Calcination
Leaching
Grinding

Tungsten
soft scrap

~.

Na2C03
pressure leach

NaOH
pressure leach

Tungsten
hard scrap

Alkali fusion
or

Electrolytic
dissolution

Impure sodium tungstate solution

containing gangue and/or precipitates
Filtration

Silica precipitation

Molybdenum precipitation

Purified sodium polytungstate solution

Solvent extraction

Ammonium polytungstate solution

Crystallization

Ammonium paratungstate
FIGURE 5.8. Simplified flow sheet of a modem tungsten conversion plant.

high percentage of gangue minerals, which sometimes create difficulties in processing

when compared to high-grade concentrates. Typical ranges for contents of the main
components and of impurities in high-grade scheelite and wolframite concentrates are
given in Table 5.1.

189

INDUSTRIAL PRODUCTION

Secondary Tungsten Sources [5.7]. Tungsten scrap materials typically range from 40
to 95% W, but can reach up to nearly lOO%. Therefore, they are high-grade tungsten raw
materials of high value and have been in use as tungsten sources for many decades. One
distinguishes between "hard" scrap and "soft" scrap (cutting or grinding sludges,
sweepings, powders, turnings, etc.), which consists of more or less fine particles.
Tungsten scrap is not only recycled via the chemical conversion, as will be described
here, but also by other methods. Therefore a special chapter is devoted to these methods
(see Chapter 11).
Tungsten scrap with only few exceptions is easier to convert to APT than ore
concentrates, because usually it does not contain the "harmful" elements like phosphorus,
arsenic, or silicon.
5.2.2.2. Pretreatment of Raw Materials. The following reasons may necessitate
pretreatment steps:
The separation of elements or compounds which have not been removed by the
concentration and which disturb the further processing. This can be done either by
selective dissolution or by roasting processes. In the latter, the disturbing elements
are at least partially volatilized.
The decrease in particle size enabling a quantitative dissolution during digestion.
This is especially necessary for concentrates gained by gravity methods.
The oxidation of tungsten in scrap material to the hexavalent state by heating in air
or oxygen-enriched air, because only hexavalent tungsten dissolves in an alkaline
leach process.
TABLE 5.1. Composition of High-grade Tungsten Ore Concentrates [5.6]
Constituent
W03
CaO
Mg
B
Al
Si
Sn
Pb
P
As
Bi
S
Ti
V
Cr
Mo
MnO
FeO
Co
Cu
Ni

Scheelite
(%)

Wolframite
(%)

66-72 (theor. 80.52)

10.9-20.3 (theor. 19.48)
0.03-D.84
<0.001-<0.01
0.22-2.0
0.77-2.94
<0.01-3.0
0.025-D.25
0.008-D.17
0.0066-D.17
O.OO6-D.IO
0.056-D.41
< 0.00 I-D.075
< 0.004-D.0 I
<0.001-D.008
<0.01-2.0
0.052-D.654
0.28-3.74
<0.005
<0.025-D.lO
0.003-D.Ol

60.5-65.8 (theor. 76.53 Fe, 76.57 Mn)

2.14-2.30
0.04
0.005
0.40-1.17
1.36-2.73
1.00
0.066
0.014-D.22
0.058-1.00
0.045-D.20
0.50-1.10
0.23
0.015
<0.005-D.008
<0.003
2.53-10.8 (theor. 23.43)
13.2-17.0 (theor. 23.66)
0.05
0.039-D.75
0.01

190

CHAPTER 5

Leaching of Scheelite with Hydrochloric Acid [5.1]. The process must be performed
at room temperature, otherwise scheelite can be dissolved. It reduces phosphorus, arsenic,
and sulfur values to less than 200 J.1g/ g P, 100 J.1g/ g As, and 1% S by dissolution of apatite,
sulfide, and arsenide minerals.
Roasting [5.1,5.2,5.8,5.9]. This process can be applied to scheelite and wolframite.
Treatment of ore concentrates at 660 to 1000 C in air in rotary furnaces has been proposed
to reduce the content of sulfides and/or arsenides by volatilization as sulfur dioxide and
As20 3, respectively. The addition of powdered coke leads to the evaporation of tin as SnO
(900-1000C). If CO-C02 or water vapor is added to the roasting atmosphere, arsenic and
antimony volatilize.
Calcination is also applied to mixtures of concentrates rich in sulfides and those rich
in calcite. The S02 and S03 evolved react with the calcite to form CaS04.
The roast treatment not only has a positive aspect but, especially at higher
temperature, can bear negative consequences. For example, a change in solubility of
gangue minerals may occur. MoS 2, which is insoluble dwjng digestion, will be converted
at higher temperature to easily soluble Mo03; also, insoluble arsenides are oxidized to
As20 3 and in case the temperature is not high enough for its sublimation, it remains and
will be dissolved.
Calcination of flotation concentrates at 500-600 C for a short time is used to oxidize
organic reagents stemming from the flotation process. This is important, because as
surface-active substances they cause excessive foaming during the further processing steps
and interfere in solvent extraction.
A general procedure cannot be given. The individual composition of the ore
concentrate together with environmental considerations as well as economical calculations
dictate the optimal direction.
Grinding of Ore Concentrates [5.6]. Flotation-derived concentrates in most cases are
fine enough to provide a complete dissolution during digestion. In contrast, concentrates
prepared by gravity procedures are much too coarse and must be disintegrated by grinding
in ball mills mostly working in closed circuits with a screen. Nearly all wolframite and a
certain percentage of scheeiite concentrates have to be ground.
Particle size plays a crucial role, especially in wolframite digestion. Mostly < 325
mesh (44 J.1m) is recommended and seldom < 200 mesh. Forscheelite, < 200 mesh is sufficient.
Oxidation ofSoft Scrap [5.7]. By heating tungsten bearing fine grained scrap in air or
oxygen-enriched air, the tungsten is oxidized to the hexavalent state. Other constituents of
the scrap, like Co, TiC, TaC, NbC from cemented carbides or Fe, Ni from heavy metal
alloys, are oxidized to the respective oxides. W0 3 partly reacts with these oxides to
tungstates. Alumina or silicon carbide remain unattacked during the oxidation process
while diamond is burned off. These materials are contained in grinding or cutting sludges.
In addition to W03, the tungstates contained in the oxidized scrap dissolve easily during
the alkaline digestion.
Scrap oxidation is performed either in push-type or in rotary furnaces. Also,
multihearth roasting furnaces are employed.
The oxidation is followed by grinding and screening. Coarser particles are recycled to
the oxidation.

191

INDUSTRIAL PRODUCTION

5.2.3. Digestion-Dissolution
The aim of this step is to dissolve the W0 3 or the tungstates and convert them into
water-soluble sodium tungstate. Besides, a partial or total separation of gangue minerals
can be achieved, depending on the type of compounds present. Raw materials consisting of
fine grains can be treated by aqueous solutions of sodium hydroxide or sodium carbonate.
The latter is more selective (it does not dissolve so much of gangue minerals) compared to
sodium hydroxide. Compact, bigger pieces of scrap, which cannot be oxidized in a
reasonable time, can be dissolved either in an oxidizing, alkaline melt (sodium carbonate
with additions of sodium nitrate or sodium nitrite), or by electrolysis.
The usual equipment for pressure digestion consists of horizontal roll autoclaves or
vertical autoclaves equipped with a stirrer. For NaOH leaching, Inconel-lined vessels are
necessary, while mild steel can be used in case of sodium carbonate. Heating is performed
by direct steam injection.
Digestion of Wolframite [5.1, 5.10-5.15]. Conventionally, wolframite is attacked by
sodium hydroxide either under atmospheric pressure in higher concentrated solution or
under elevated pressure in a more diluted solution. Today, the less concentrated solution is
preferred, because it dissolves less gangue minerals. Table 5.2 describes digest conditions.
(Fe, Mn)W04

+ 2NaOH -+ Na2 W04 + (Fe, Mn)(OH)2

Concentrates low in manganese may be dissolved also by a pressure leach as used for
scheelite (Table 5.2).
Usual yields range between 97 and 99%.
Digestion ofScheelite [5.1,5.16-5.22,5.24-5.26]. The method mostly applied today
in western world countries is dissolution with sodium carbonate under increased pressure.
Costs are higher compared to the above-described sodium hydroxide digestion of
wolframite due to the fact that a high stoichiometric excess of the reagent must be applied.

The equilibrium constant K = [WO~-]/[CO~-] is increased by temperature. Therefore, the

digest is performed at elevated pressure and temperature. An increase in carbonate
concentration decreases the equilibrium constant, and when it exceeds a certain limit
the formation of an insoluble scale at the surface of the scheelite crystals slows down the
further dissolution. The scale consists of Na2C03 CaC03 and/or Na2C03 2CaC03.
Accordingly, maintaining the sodium carbonate concentration within a certain limit is
crucial for a good yield, which should be around 99% or better. There are proposals in the
literature to improve the process:
Countercurrent leaching scheme to save chemicals.
Continuous process with feedback control.
Use of ball mill autoclave.
Within the last ten years, in many plants in China, a change was made from sodium
carbonate to sodium hydroxide leaching with simultaneous mechanical activation. By
mechanical activation the density of lattice defects is increased, and correspondingly the
activity and decomposition rate is also increased. Mechanical activation favors the
decomposition not only kinetically but also thermodynamically. The advantages of this

1:1.0S

NaOH
7-10%

Na2C03
16.S%

Na2C03
10-18%

Na2C03
12-18%
+NaOH
NaOH
NaOH
20%

Wolframite

Wolframite low in Mn

Scheelite and
scheelite/wolframite mixtures

Scheelite/Wolfrarnite mixtures

Scheelite and
scheelite/wolfrarnite mixtures

Oxidized scrap
1:1.4

1:3-7

Weight ratio
Na2C03/W03
I: 1.0-1.6

I :2.S-S.0

1:3.3

1:1.0-1.5

Molar ratio W0 3:agent

NaOH
40-50%
ISM

Agent
concentration

Wolframite

Raw material

IS0-200 C
2--4h

180C
I.5h

10-12

10

Mechanical activation in a ball

mill autoclave

NaOH addition stoichiometric to

Fe + Mn content

IS--40

200-2S0C
4h

The lower the W0 3 conc., the

higher the stoichiometric excess of
Na2C03 must be

12-26

190-23SoC
I.S-4h

200-2S0 g solids/I

22

10-12

17S-190C
4h

After cooling, addition of water to

4.5-S.S M NaOH

Remarks

210C
3h

Atmospheric

Pressure
(bar)

100-150C
O.S-4h

Temperature
time

TABLE 5.2. Digest Conditions [5.6]

U>

(J

...~

193

INDUSTRIAL PRODUCTION

method are: less chemical consumption (0.7-0.8 kg NaOH/kg W03 compared to 1.351.40 kg Na2C03 /kg W0 3); very good yield (>98%); and flexibility in regard to the raw
material. The method can be applied to high and low scheelite concentrates as well as to all
kinds of scheelite and wolframite mixed ore concentrates.
Table 5.2 describes digest conditions.
Digestion oj Wolframite-Scheelite Mixtures [5.1,5.23-5.26]. As long as the wolframite percentage is low, sodium carbonate pressure leach as applied to scheelite can be
successful. Another method uses mixtures of sodium carbonate and sodium hydroxide
because, as opposed to the digestion of scheelite, NaHC0 3 is formed in wolframite
digestion by hydrolysis of FeC03.
FeW04 + Na2C03 --+ Na2 W04 + FeC03
FeC03 + Na2C03

+ 2H20

--+ Fe(OH)2

+ 2NaHC03

This means that an increased sodium carbonate consumption is slowing down the
dissolution of the tungstates. The addition of sodium hydroxide converts the primary to
secondary sodium carbonate:

Finally wolframite-scheelite mixtures can be successfully digested by a pressure

leach with sodium hydroxide under instantaneous mechanical activation (ball mill
autoclave). As already pointed out, under scheelite digestion this method is frequently
applied in China.
Digestion oJOxidized Scrap [5.6]. Dissolution of these materials can be achieved by
sodium hydroxide pressure digestion, similar to that used for wolframite (Table 5.2).
Depending on the scrap type, yields are between 97 and 99%.
Fusion or Roast Technique Jor Hard Scrap [5.27-5.29]. Fusion is performed at high
temperature with sodium nitrite or sodium nitrate as oxidizing agent and sodium carbonate
as dilutant and must be carefully controlled. Roasting is done in rotary kilns at 800C
using a 1:3 mixture of oxidizer and diluent. After cooling, the cake must be crushed prior
to dissolution in water.
Electrolysis [5.30]. The principle of the method is to have the tungsten scrap as
anode in an electrolyte like sodium hydroxide or sodium carbonate aqueous solution. The
tungsten atoms are oxidized electrolytically to the hexavalent state. The resulting solution
contains sodium tungstate besides msoluble oxides or oxide hydrates of the other
constituents. Rotating drum or disk electrolytic cells are the applied equipment.
During rotation, the scrap pieces are constantly rearranged, enabling scale and metal
skeleton removal. The surface is continuously renewed (cleaned) during the operation.
Scale at the surface is formed by oxidation of other scrap constituents to insoluble oxides.
Metal skeleton is formed if metallic constituents of the scrap are not oxidized, as in the
case of heavy metal. The Ni-Fe binder phase is not attacked; only the tungsten grains
dissolve. Both scale and skeleton are diffusion barriers and slow down the process.
Solutions containing 100 g W0 3/1 can be produced by the process. Typical parameters are given in Table 5.3.

194

CHAPTER 5

TABLE 5.3. Typical Energetic and Dissolution Parameters of the Electrolytic

Dissolution Process for Different Tungsten Scraps [5.30]
Scrap type
W rods or plates
W-Cu rods and plates
Heavy metal pieces
low iron
higher iron

Specific energy consumption

(W.h.g- I )

Dissolution rate
(g.A - I . h - I )

II
6.0

1.12
0.86

3.8
4.3

0.81
0.83

5.2.4. Purification Steps

These includes the following main steps: separation of the undissolved digestion
residue by filtration, and separation of the impurities dissolved in the sodium tungstate
solution. Sometimes, the first step can be combined with the first precipitation step (silica
precipitation).
Depending on the raw material composition and digestion conditions, the sodium
tungstate solution after filtration may contain more or less impurity ions in variable
concentrations. Typical contaminating ions are silicate, phosphate, arsenate, molybdate,
fluoride, lead, bismuth, and aluminum. Their main origins are the gangue minerals
(aluminum may be contained as [AIF6]3- ion as a consequence of a fluoride content
after digestion and subsequent silica precipitation by addition of aluminum sulfate). If their
concentrations are too high, they cause disturbances during further processing, or they
contaminate the final product APT.
The first separation step is called silica precipitation but, besides silica, other ions like
phosphate and fluoride are at least partially separated.
The second step is a precipitation of sulfide-forming cations mainly applied to
separate molybdenum, but others like arsenic, antimony, bismuth, cobalt, etc. are
precipitated, too.
The silica precipitation is performed in alkaline solution, while the sulfide precipitation is effective at pH 2.5-3.0. This change in pH is usually achieved by addition of
sulfuric acid. An interruption at pH 6 offers the possibility to separate the above-mentioned
[AIF 6]3- ion (minimum solubility of Na3[AIF3]). The acidification transforms all excess
Na + into soluble sodium sulfate.
Acidification can also be achieved by electrodialysis and offers in addition the
possibility to recover NaOH solution, which can be used for digestion. This is a modem
process until now used only in a pilot plant scale. However, economic as well as ecologic
considerations will dictate such processes in the future.
It is interesting to note that the majority of APT plants, including those which have
been installed recently, still make use of these precipitation reactions, although in the
literature other methods have also been proposed. Basically, separations of this type
always involve an increased loss of the valuable metal with difficulties in filtration and
washing.
The only exception in this regard is the PR of China. For many years more modem
processes have been applied in several plants and are still being improved, based on ion

195

INDUSTRIAL PRODUCTION

exchange and adsorption. One of the main advantages of the ion exchange process is the
fact that separation of disturbing ions like silicate, arsenate, phosphate, and molybdate can
be combined with the step consisting of conversion of sodium to ammonium tungstate, so
reducing considerably investment as well as working costs.
5.2.4.1. Separation of Undissolved Residues. The residue in the case of ore
concentrates consists of undissolved gangue minerals and insoluble reaction compounds
such as CaC03 or (Fe,Mn)(OHh, or, in the case of grinding scrap, insoluble metal oxides.
The latter contain valuable components like Co, Ta, or Ni and can be subject to metal
recovery.
The equipment used in this step can be a disk, plate, and frame, multistage-drum or
belt filters with counterflow washing with water. Additional repulping and application of a
high-pressure filter press reduces the loss of dissolved tungsten.
5.2.4.2. Silica Precipitation [5.1,5.31,5.32]. Silicate ions may be precipitated by

addition of aluminum sulfate or magnesium sulfate solution to the sodium tungstate

solution. Also, mixtures of both reagents are used. The addition is made to the slightly
alkaline or neutral solution. The respective chemical reactions are complex but can be
described in a simplified manner by the following equations:
2Na2Si03 + AI2(S04)3 + Na2C03
Na2Si03

Na20 . Al 20 3 . 2Si02 + 2C02 + 3Na2S04

+ 2MgS04 + 2NaS04 ~ 2MgO . Si02 + 2Na2S04 + 2C02

Several procedures were proposed. Especially, in the case of higher silica concentrations, a double precipitation is recommended. Typically 0.08 kg AI2(S04)' 18H20 and
0.03 kg MgS04 . 7H20 per kg W03 are added to the 70-80 C hot solution after digestion.
The pH value should be between 9.0 and 9.5. Filtration can be started after I hour.
Another procedure is stepwise addition of the two chemicals to precipitate silicate,
phosphate, and fluoride ions:
Step 1: addition of MgS04 to a final concentration of 0.5-2.0 g Mg/l at 60-110 C

and pH 10-11, followed by filtration.

Step 2: addition of Ah(S04)3 to a final concentration of 0.2-3.0 g AI/I at 20-60C

and pH 7-8.
In this way, original concentrations of20-10,000mgSi/l, 10-120mg P/I, and 5004000 mg F/1 can be reduced to ~ 10 mg Si/I, ~ 5 mg P/1, and ~ 200 mg F/1.
The precipitates have very poor filtration behavior. Therefore, cellulose is often added
as a filter aid.
5.2.4.3. Combination of Gangue Removal and Silica Precipitation [5.1,5.6,5.31].
Because of the aforementioned bad filtration behavior of the silica precipitate, the
undissolved gangue itself will be used as a filter aid. This comprises the addition of the
precipitating agents, as described above, to the digested slurry. In case of higher silica
concentration (several g/l) a second precipitation must be applied, reducing the level to
~20mg/1.

5.2.4.4. Molybdenum Separation [5.1,5.31,5.33,5.35]. Ifsodium sulfide is added to

the neutral or slightly alkaline sodium tungstate solution, the molybdate ion is converted to

CHAPTER 5

196

thiomolybdate:

The reaction is completed after one hour at pH 10 and a temperature of 80-85 0c. The
addition of sulfuric acid to change the pH to 2.5-3.0 converts the thiomolybdate to
insoluble molybdenum trisulfide.
MoS~-

+ 2H30+ -+ MoS3 + H20 + H2S

The stoichiometric excess for a complete reaction depends on the Mo--W ratio and
should be at least 3. Lower Mo concentration affords a higher excess. The precipitate is
filtered off (mostly by plate and frame filter), while the remaining molybdenum concentration in solution is :::: 10 mg/I. As already mentioned, other insoluble sulfide-forming
cations are coprecipitated.
The H2 S evolved during acidification has to be scrubbed by NaOH. In case of a
solution derived from a sodium carbonate leach during the acidification, a high amount of
CO 2 is generated due to the neutralization by sulfuric acid up to pH about 6. This CO2 is
vented, but should not pass the scrubber so as not to consume NaOH (H2S is only
generated at lower pH).
A continuous operation using a multistage stirred-tank reactor system and hydrogen
sulfide recycling by absorption in the sodium tungstate solution substantially reduces the
amount of sulfuric acid.

5.2.4.5. Molybdenum Removal by Adsorption on Activated Carbon [5.34]. The

principle of the method is the selective adsorption of thiomolybdate by activated carbon
while tungstate will not be retained.
The neutral sodium tungstate solution (PH 7.2-7.4) is treated with sodium sulfide to
convert the molybdate to thiomolybdate and then adjusted to pH 8.2-8.4. Activated carbon
is added and mixed for about 30 minutes. The loss of tungsten by coadsorption is less than
0.1%. Specific adsorption ofMo varies between 4.5 and 6.1 mg Mo/g carbon, depending
on the carbon quality. A multistage crossflow adsorption is proposed for industrial
application, leading to higher specific adsorption in the earlier stages.
5.2.4.6. Electrodialysis [5.30]. The sodium tungstate solution after silica precipitation contains sodium tungstate as well as an excess of sodium ions stemming from the
surplus necessary for an optimal yield in digestion (1.5 times excess of NaOH for
wolframite and 3.3 and more times excess of Na2C03 for scheelite digest). The
electrodialysis offers the possibility to recover the sodium ions as NaOH. The solution
is fed into the anode compartment of the electrodialyzer. In the electric field, the Na ions
pass through the cation exchange membrane into the cation compartment (NaOH). An
equivalent amount of H+ ions is generated in the tungstate solution. Hence the sodium ion
concentration is decreased in the tungstate solution and at the same time increased in the
NaOH solution. When the decrease exceeds the stoichiometric ratio for Na2 W04 ,
polycondensation starts and isopolytungstates form. The acidification of the tungstate
solution by electrolytically generated protons cannot be compared with the usual
acidification in the technical process, which is performed by addition of concentrated
sulfuric acid. The electrodialysis protons are homogeneously generated in the solution and

INDUSTRIAL PRODUCTION

197

the pH gradually decreases. During the classical process, locally and intermediately higher
sulfuric acid concentrations arise and were equalized briefly by stirring. However, alreadyformed higher condensed tungstate species will not be dissociated completely. Therefore,
the condensation grade of the tungstate species prepared by electrodialysis is more
homogeneous, and solutions containing 200 g W0 3 /1 are stable at pH 1.7-1.8, which is
not the case for solutions derived from the classical procedure. A further reason for the
different properties is the absence of excessive sodium sulfate in the electrodialyzed
solution. High sodium sulfate concentration also influences the condensation. Those
solutions behave differently in the next step, namely, solvent extraction.
5.2.4.7. Ion Exchange. As already pointed out in the Introduction, the methods have
been evaluated in China and are in industrial large-scale use. The ion exchange method
combines the separation of impurities with the sodium-to-ammonium tungstate conversion. The sodium ions pass the anion exchanger. Owing to the chemical similarity of
tungstate and molybdate, the latter must be converted first to thiomolybdate, which has a
quite different desorption behavior. Silicate, phosphate, a,nd arsenate are also not retained
by the resin while tungstate and thiomolybdate are adsorbed. Selective desorption of
tungstate is achieved by eluation with a mixture of aqueous ammonia and ammonium
chloride and, subsequently, thiomolybdate is eluated by a basic hypochlorite solution. For
more details see Section 5.2.5.2.
5.2.5. Conversion of Sodium into Ammonium Isopolytungstate Solution
The goal of this step-in the classical process achieved by tungstic acid precipitation
and successive dissolution by ammonia-is almost total separation of sodium ions, which
have been necessary for dissolution but which interfere in the further production steps to
tungsten metal powder. Their concentration will be decreased quite effectively from
approx. 70 gil to less than 10 mg/I.
Today, two industrial methods are applied to meet this target: solvent extraction
(liquid ion exchange) and ion exchange by resins. Plants using the latter process can be
found exclusively in China (smaller works). All others are equipped with liquid ion
exchange facilities.
5.2.5.1. Solvent Extraction [5.9,5.15,5.36-5.44]. Generally, SX in hydrometallurgy
serves to concentrate diluted solutions, to separate, and to purify. In tungsten hydrometallurgy, as already explained, separation is the main target and, to a certain extent,
purification from ions which do not condense with tungstate and from those which are not
extracted by the solvent can be achieved.
The basic principle is to extract isopolytungstate species from the slightly acidic
aqueous solution by the reagent dissolved in the organic phase, thus separating the
tungstate from sodium ions. Re-extraction (stripping) from the organic into the aqueous
phase is performed by aqueous ammonia solution. The organic solvent circulates from the
extraction to scrubbing, stripping, regeneration, and back to extraction. Losses are only
due to the very low solubility of the organic compounds in aqueous solutions.
A simplified flow sheet is presented in Fig. 5.9. One distinguishes between 4
steps: extraction, scrubbing, stripping, and solvent regeneration. The last step in some
plants is a separated one, as indicated in Fig. 5.9, but may be combined with the extrac-

CHAPTER 5

198

Extraction

Stripping

Scrubbing

aqueous
sodium
sulfate
solution
(raffinate)

deionized
water

scrub
raffinate

aqueous
ammonia
solution

aqueous
ammonium
polytungstate
solution

Solvent regeneration
stripped solvent

regenerated solvent

organic
aqueous
- - - W-containing

aqueous
sulfuric acid
solution

regeneration
raffinate

FIGURE 5.9. Solvent extraction process to convert Na in

N~-tungstate

(simplified flow sheet).

tion step in others, by feeding an appropriate amount of sulfuric acid to the aqueous
phase.
The equipment used for the different steps consists of mixer-settlers in a two-stage or
multistage mode comprising countercurrent flow of aqueous and organic phases, and with
partial internal circulation in each step to achieve the retention time required. Also, static
mixers combined with mixing pumps and settlers can be used in extraction and scrubbing,
where only short retention time is necessary.
Organic Solvent. The organic phase consists of two or three components: reagent,
modifier, and solvent. A basic study of possible chemicals for these components dates back
to 1960 [5.33].
As reagent, long-chained primary, secondary, tertiary amines, and quaternary ammonium compounds (chain length Cs to C 10) like trioctylamine or tridecylamine are used.
Commercial products are Primene JMC (primary amine), Amberlite LA-lor Adogen 283
(secondary amine), Trioctylamine (tertiary amine), and Aliquat 336 (quaternary ammonium compound). Depending on the process conditions, either secondary or tertiary
amines are preferred, the latter more seldom. Primary amine is advantageous for higher
condensed tungstate solutions as derived from the electrodialytic process. The concentration of the reagent varies in the literature between 1 and 22 vol%.
As solvent, either kerosene or mixtures of alkyl benzenes like toluene or xylene have
been proposed. Due to the toxicity of the latter, kerosene is widely preferred.
The modifier is only necessary in case of kerosene as solvent, because of the low
solubility of the isopolytungstate-amine complex which is hence increased. Compounds
used as modifiers are isodecanol (5%) or tributyl-phosphate (12%).

INDUSTRIAL PRODUCTION

199

Extraction. The slightly acidic polytungstate solution is contacted with the regenerated organic phase, whereby the isopolytungstate reacts with the protonated cationic amine
to form an ion associate complex. In the following equation, a tertiary amine and, as the
tungsten species, a paratungstate A is assumed:
S(R3NHt

+ ~SO~- + HW60~1 + SNa+

--+ (R3NH)5HW6021

+ ~SO~- + SNa+

The equation is idealized, because in the slightly acidic solution a variety of condensed
tungsten species can be contained depending mainly on pH and time of standing after
acidification [S.34]. Due to equilibria at the same pH, several polytungstates are present at
the same time. Therefore, not only one species is extracted. Higher condensed polytungstates should be avoided if tertiary amine is the reagent because of the formation of the
third phase. The complex of higher condensed species and tertiary amines is much less
soluble and "precipitates" cause worse disengagement. In this case, primary amines are
advantageous.
In regard to the throughput of heteropolytungstates, secondary and tertiary amines
behave differently. Secondary amine allows tungstophosphate and silicophosphate to pass
solvent extraction, while in the case of tertiary amine a honey-like tertiary phase is formed
which, due to its high density, settles to the bottom. In both cases, severe disturbances
result. If heteropolytungstates pass the solvent extraction in the following step (APT
crystallization), the tungsten solubility in the mother liquor will be enhanced (see Section
S.2.6). The formation of the third phase not only consumes amine and separates it from the
process, but may also lead to a plug-up of pipings, etc. Therefore, proper separation of Si
and P impurities in advance is important.
The equivalent amounts of sulfate and sodium ions, as well as any surplus of these
ions applied in excess during digestion and acidification, remain in the aqueous phase and
leave the extraction as raffinate. Depending on conditions, it can be discarded or must be
treated further (see the introduction).
Indications about extraction conditions in the literature vary due to different amines
and differences in concentration, loading, and pH values:

W03 concentration of the aqueous phase, 10-180 gil.

pH of aqueous phase, 1.8-2.6.
W0 3 concentration of the loaded organic phase, 2S-120 gil (usually 30-60 gil).
organic:aqueous ratio mainly 1: 1.
Temperature, approximately SO C.

Scrubbing. In a technical scale, some entrainment of the organic phase after

extraction can not be avoided. In scrubbing, the main portion will be removed. The
scrubbing raffinate contains low concentrations of sodium sulfate. A typical aqueous:
organic ratio is 1:20. A combination of coalescer and settler can be used for entrainment
removal besides mixer settlers.
Stripping. In this stage, the loaded organic is contacted with aqueous ammonia
solution. The ion associate complex is decomposed, and the isopolytungstate is reextracted into the aqueous phase as ammonium salt according to the following idealized
equation:
(R3NH)5HW60Z1

+ SOH-

--+ HW60~1

Conditions for this step also differ in the literature:

+ S(R3NH)OH

200

CHAPTER 5

pH 8-13.
Ammonia concentration 1.2-10%.
Retention time of some minutes, due to slow reaction rate.
Aqueous:organic ratio must be adjusted to the concentration of the loaded organic,
between 1:1 and 1:5. High concentrated ammonium isopolytungstate solutions are
desirable. The higher the concentration, the less water has to be evaporated during
the subsequent production step (crystallization) and that means a saving of energy
and a reduction in cost.

The entrainment of the ammonium tungstate solution can be minimized by applying a

direct current voltage of 15 V to promote coalescence [5.39].
Solvent Regeneration. In this stage, the stripped organic phase is contacted with an
aqueous sulfuric acid solution. Hence the amine is again protonated and at the same time
the organic phase is cleaned from the entrainment of the ammoniacal solution. As already
mentioned, this stage is substituted in some plants by combining protonation with
extraction. Entrainment of the unloaded organic phase can be decreased by contacting it
with activated carbon [5.38].
5.2.5.2. Ion Exchange [5.45-5.47]. The process comprises the following steps:
Preconditioning o/the sodium tungstate solution. The pH is adjusted to 7.2-7.4 by
electrodialysis (see Section 5.2.4.7.). Electrodialysis is preferred to the addition of acids,
because additional anions affect the sorption capacity of the resin and furthermore result in
stable tungstate solutions. Subsequently, molybdate is converted into thiomolybdate by
addition of sodium sulfide solution.
Sorption on the anion exchange resin.
RCl2 + Na2 W04 --+ RW04 + 2NaCI

RCl2 + Na2MoS4 --+ RMoS 4 + 2NaCI

The resin in the chloride form exchanges chloride against tungstate and thiomolybdate. In
the case of tungstate the equation is idealized, because in reality isopolytungstates are
present. Sodium ions as well as the anions silicate, phosphate, and arsenate are not retained
and pass the column. The pH is too high for heteropolytungstate formation with these
anions.
China-made anion exchange resins 930, 931, 934, and 937 are employed.
Desorption. Tungstate is first removed from the resin by a mixture of aqueous
ammonia solution and ammonium chloride and exchanged by chloride.

The equation is also idealized. Thiomolybdate remains combined with the resin and can be
subsequently stripped by a basic hypochlorite solution.
In a series column operation, thiomolybdate is sorbed in the first column (resin 930)
and tungstate in the second (resin 937). The molybdenum removal is better than 99%. The
impurity to W03 ratios are:
Mo
As

1.2-1.4
0.8-1.5

X
X

10-5
10-5

1.2-3.0
2.5-2.7

10-4
10-5

INDUSTRIAL PRODUCTION

201

As already pointed out, the majority of the pertinent literature is in Chinese while the
rare contributions in English are not at all detailed. Therefore, we have included in the next
subsection (5.2.6) a summary of modem Chinese APT technology written by an expert
from China.
5.2.6. Modern Methods in Chinese Plants [5.48-5.50].
Ion Exchange Technology in Processing Impure Wolframite Concentrates to APT, by
Zhao Quinsheng (Central South University of Technology, Changsha, Hunan, 410083,
China).
Hot sodium hydroxide digestion is widely used in China for treating wolframite
concentrates. Recent developments show that NaOH digestion can also be used for treating
impure wolframite concentrates such as scheelite-wolframite mixed ore concentrates or
low-grade wolframite concentrates containing unusually high contents of P, As, Si, and Ca
by adding some additives to accelerate the decomposition process of the tungsten minerals
and to restrict the dissolution of some impurities. As a result of NaOH digestion, the raw
sodium tungstate solution is formed. Impurity elements like Mo, P, As, and Si are also
partly dissolved by NaOH, which must be removed. The classical process combines
precipitation reactions with liquid solvent extraction and is adopted by most manufacturers
worldwide at present.
The ion exchange technology for purifying tungstate solution has been used in China
for nearly 15 years [5.48]. More than 50% of the APT produced in China is manufactured
by using ion exchange technology. Figure 5.10 shows a pertinent simplified flow sheet.
Normally, the raw sodium tungstate solution after NaOH digestion contains 200 g
WO~- /1, which is too high for the ion exchange process. The ion exchange capacity
decreases with increasing W0 3 concentration, as is shown in Fig. 5.11.
In the ion exchange process, ions with higher affinity to the resin can replace those
with lower affinity. A strong basic anion exchange resin (type 201 x 7) is used. There are
several anions like WO~-, MoO~-, HAsO~-, HPO~-, SiO~-, OH-, etc. in the sodium
tungstate solution. Beside OH-, all others have higher affinity than Cl- and can be
adsorbed by the resin as follows:

The affinities of these ions can be arranged in a decreasing series:

WO~- > MoO~- > HAsO~- > HPO~- > SiO~- > Cl- > OH-

During the early stage of adsorption, all ions with higher affinities than that of Cl- can be
adsorbed on the resin. However, when the diluted tungstate solution continuously passes
through the ion exchange column, anions with smaller affinities, adsorbed already on the
resin, are replaced by tungstate and removed by the effluent solution, as described by the
following equation:

202

CHAPTER 5

Slurry from NaOH Digestion of

Tungsten Ore Concentrates

Dilution and Filtration

---t

woi-

on Ion
Adsorption of
Exchange Resin

Residue to
Diposal

Liquid Waste to
Disposal

Stripping

---t

Ammonium Tungstate
Solution
FIGURE 5.10. Flow sheet of the single ion exchange process for purifying and converting impure sodium
tungstate solution.

80

75

l70
~

I
I!

c::

Ia

c::
=
.2
.c::

45
40
0

40

80

120

160

Concentration (gil)

FIGURE 5.11. Relationship between W04 concentration and ion exchange capacity [5.48].

203

INDUSTRIAL PRODUCTION

By this substitution reaction, removal of P, As, and Si up to 90% is possible; however, the
difference in affinities of molybdate and tungstate is too small to separate them. Obviously,
all cation impurities can be removed by this process.
The ion exchange technology is not only used for purification of the tungstate
solution, but also for converting it into ammonium tungstate solution. The conversion is
done while stripping the tungstate from the loaded resin, using ammonium salts as
stripping agents. The stripping effects of various ammonium salts is compared in Fig. 5.12.
It is seen that ammonium chloride is the most suitable agent. In addition, it regenerates the
resin to the desired form:

In order to avoid precipitation of APT crystals during stripping, an overstoichiometric

amount of NH4 0H is applied.
Wolframite concentrates, generally used as raw material in a single ion exchange
processing, are first-grade No.2 quality according to Chinese National Standard GB-282581. The corresponding analysis is: W0 3 > 65% and impurities (%) < 0.7 S, 0.1 P, 0.1 As,
0.05 Mo, 3.0 Ca, 0.25 Cu, 0.2 Sn, and 5.0 Si.
The effect of a single ion exchange treatment of raw sodium tungstate solutions is
shown in Table 5.4. As indicated before, the single ion exchange technology is not suitable
for removing molybdenum. In order to treat impure wolframite concentrates and low-grade
wolframite-scheelite concentrates containing more than 0.1 % Mo and higher concentrations of other disturbing impurities, a double ion exchange technology was recently

.B

200

:;,

"0
UJ

"C
Q)

a.
a.

''::

iil
c:

.2
a;....

100

c:
Q)

o
o
o

c:

11

13

Sample number
FIGURE 5.12. Stripping effect of various ammonium salts [5.48].

204

CHAPTER 5

TABLE 5.4. Composition of Solutions of a Single Ion Exchange Process [5.49]

Solution

W0 3
(g. 1-1)

Raw sodium tungstate

Adsorption effluent
Ammonium tungstate

20.08
0.1
203.2

37.4
68
3.5

As

Si

(~g. g-l W03 )

104.6
190
10.3

1245
2400
73.8

Mo
1594
1750
738

developed and put into production in the Four Rings Metallurgical Factory, Yunan,
Guangdong Province.
A process for separating molybdate from tungstate using ion exchange was patented
by Chen Zhouxi et at. in China [5.50]. The principal flow sheet of the technology is
presented in Fig. 5.13.
The feed solution can be either a sodium tungstate solution containing Mo, P, As, and
Si or an ammonium tungstate solution containing Mo as the main impurity element
stemming from a single exchange technique. After pH adjustment, NaHS or (NH4 hS is
added to convert molybdate into thiomolybdate. The solution passes a column of strongly
basic anion exchange resin (Type D-20l). Thiomolybdate is adsorbed and the effluent
contains the tungstate purified from Mo. In order to regenerate the anion exchange resin, a
solution of an oxidizer is used to desorbe the thiomolybdate.
The original Mo/W0 3 concentration ratio in the feed solution of 0.06--0.84%
decreases to 0.001-0.015% in the effluent. The recovery of the operation is better than
99%. It was used in 1992 commercially to produce sodium and ammonium tungstate of
high purity. The purification effect is demonstrated by the analyses in Table 5.5.

Tungstate solution
containing molybdate

!
pH adjustment and
conversion to thiomolybdate

!
Ion exchange - sorption
of thiomolybdate
~

Adsorption effluent

Stripping effluent

1-1

Mo free tungstate solution

Stripping of tungstate
~

1 Stripping of thiomolybdate 1
~

Eluate to disposal

FIGURE 5.l3. Flow sheet of molybdate removal from tungstate solution by ion exchange [5.50].

205

INDUSTRIAL PRODUCTION

TABLE 5.5. Composition of Solutions of a Double Ion Exchange

Process [5.49]
Solution
Ammonium tungstate feed
Adsorptioin effluent
W stripping effluent
Mo stripping effluent

W03
(g .1- 1)

Mo
(g. I-I)

Mo/W0 3

200.5
195.6
53.5
4.4

0.592
0.005
0.04
4.07

0.283
0.0025
0.8
92

(%)

If the ammonium tungstate solution after the Mo removal is evaporated for APT
crystallization, a 98.4% evaporation can be applied. At 92% ~vaporation the APT contains
< 20 Ilg/g Mo, and at 98.4% evaporation 20-80llg/g Mo.
With this basic knowledge, a new double ion exchange technology for processing
impure wolframite concentrates was created by combining the above-described double ion
exchange processes. The resulting APT is of utmost purity also in regard to Mo. The
complete flow sheet for treating primary tungsten raw materials using the double ion
exchange technology for commercial APT production is given in Fig. 5.14. Typical
impurity concentrations (Ilg/ g) in the resulting APT are: A12, As 1, Ca 5, Cr 5, Fe 5, K 10,
Mg3, Mnl, M07, Na9, Ni5, P7, Si5, Sn2.
The APT quality is very consistent and either comparable or even better than that of
other producers. The main feature of the technology is the lower production cost and the
ability to produce APT according to the customers' requirements (e.g., Mo content).

5.2.7. Ammonium Paratungstate Crystallization [5.51-5.58]

The ammonium isopolytungstate solution from solvent extraction is evaporated,

whereby ammonia and water are volatilized. The ammonia concentration in the mother
liquor decreases (PH drop) and at the same time the W03 concentration increases. Para
tungstate B [H2 W 12 0 42 ]10- is formed whose ammonium salt has low solubility. Crystallization of (N~)lOH2 W 12042. 4H20 is the result. The degree of evaporation depends on
the purity of the feed solution and the required purity of the APT. Usually, if proper
purification was performed in the foregoing steps it ranges somewhere between 90 and
99%.
The physical properties of the crystallized APT can be influenced by crystallization
conditions to some extent. They are of importance for the further processing to metal
powder by hydrogen reduction.
The crystallization is not only a conversion of the dissolved to the solid salt, but the
last purification step in the hydrometallurgical process. Most of the impurity elements still
present in the process solution have much higher solubility than APT and are consequently
enriched in the mother liquor.
Impurity concentrations of the feed solution to the crystallizer, of the mother liquor,
and of the crystallized APT are related. For a given required APT, purity limits in impUrity
concentrations of the mother liquor are allowed, and these limits in combination with the
impurity level of the feed solution determine the degree of evaporation. Upper impurity

206

CHAPTER 5

Primary tungsten raw materials

Vibratory milling
Alkali digestion in ball mill autoclave
Slurry dilution
Press filtering and washing

First ion exchange

MoO/' ~ MoS/' conversion
Second ion exchange
First crystallization
Filtering and washing

Wash water

Mother liquor

APT

Second crystallization

Mother liquor

APT---'

1 0 - - - - - - - - CaO
Filtrate to disposal
FIGURE 5.14. Flow sheet for treating primary tungsten raw materials by using double ion exchange technology.
By courtesy of Four Rings Metallurgical Factory, Yunan, Guangdong, China.

levels in the feed solution and corresponding concentrations in the mother liquor and the
crystallized APT are given in Table 5.6.
The crystal slurry is separated from the mother liquor by filtration and washing with
deionized water. Filtrate and wash water are recycled and, finally, the APT is dried.
On a technical scale, batch and continuous operation are used. Batch crystallization
leads to better consistency of the product, but the continuous process is more economic.

207

INDUSTRIAL PRODUCTION

TABLE 5.6. APT crystallization: Impurity Concentrations [5.6]

Element

Feed solution

Concentration of
mother liquor

(gTI)

(gI-I)

(Ilg.g- I)

< 100a
<3000b
< 2000b
<3000
<200
<60
<100
<400
<200
<1200

<10

Al

<10

As
F
Fe
Mo
Na
P
Si

<50
<250
<10
<10
<10
<50
<10
<100

V
a
b

APT

<20
<10
<10
<20
<10
<20
<20
<20

In the absence of fluoride.

At fluoride concentrations of about 4000 mg/I.

The presence of foreign element impurities not only affects the impurity level of APT,
but also in some cases the APT crystal size. The higher the concentration of these
impurities, the smaller the average crystal size. Heteropolytungstate-forming ions like
silicate, phosphate, and arsenate lead to a considerable increase in the solubility of
tungsten in the mother liquor. The consequences are worse crystallization yield and higher
cost due to increased recycling of tungsten. A precipitation by aluminum salt solution to
separate phosphorus and silicon prior to crystallization is proposed. However, it is
preferable to make all advanced purification steps more effective.
Ammonia Recovery. The ammonia evaporating together with water during crystallization is condensed and can be reused in the stripping stage of solvent extraction.
Sometimes, a concentration by distillation is applied to maintain the water balance.
Mother Liquor Recycling. In order to maintain the impurity concentration in the APT
low and consistent, the evaporation can never be 100%. This means that a certain amount
of mother liquor must be recycled to the former steps of the conversion plant. Usually, an
amount of mother liquor corresponding to 5-10% of the volume fed into the crystallizer is
pumped back (90-95% evaporation). This can be done continuously or batchwise. There
exist several possibilities to treat the recycled mother liquor, which is high in impurities.
In some factories it is added to the feed make-up prior to digestion. The disadvantage
of this method is that, during heating, an ammonia pressure is built up that prevents one to
reach the desired temperature by steam injection. An intermediate venting of the autoclave
is necessary, resulting in loss of time and in increased steam demand. The advantage is the
proper separation of all impurities.
A second possibility is to add the mother liquor immediately after the digestion but
prior to the purification steps. The disadvantage here is the presence of ammonium ions.
As already pointed out, ammonium ions favor the condensation of tungsten ions, and this
changes the properties and behavior of the isopolytungstate solution in the solvent
extraction process. Therefore, this type of mother liquor recycling affords a separate
pretreatment to expel the ammonia. This can be done by addition ofNaOH and subsequent
heating. The ammonia evolved can be fed into the ammonia recovery system of crystallization. The treatment produces a precipitate which can be separated by filtration.

208

CHAPTER 5

A third possibility is to further crystallize the mother liquor yielding a more impure
APT [5.55]. Also, treatment of mother liquor with hexamethylene tetramine to precipitate
tungsten was proposed [5.56].

5.3. THE HIGHLY PURE INTERMEDIATES

5.3.1. Ammonium Paratungstate (APT) [5.51-5.53,5.59]

Today, APT is the most important and almost exclusively used precursor for tungsten
products. Only in tungsten melting metallurgy (Chapter 8; production of steel, stellites,
etc.) and for producing WC directly from ore concentrates (Chapter 9; Menstruum WC)
are other starting materials used (ferrotungsten, scrap, melting base, tungsten ores).
Intermediates, such as tungsten trioxide, tungsten blue oxide, tungstic acid, and ammonium metatungstate can be derived from APT as shown in Fig. 5.15, either by partial or
complete thermal decomposition or by chemical attack.
Although a hexahydrate and a decahydrate exist, only the tetrahydrate (NH4)10
[H2 W1242] 4H20 forms under industrial conditions, since the hexahydrate is unstable
at temperatures exceeding 96C, while the decahydrate crystallyzes only from solutions at
room temperature.
Characterization. The important physical and chemical properties have been already
treated in subsection (b) of Section 4.2.7.3. Here, only parameters relevant for the further
processing are discussed.
APT is a white, crystallized powder. The average crystal size of commercial products
ranges between 30 and 100 ~m. The SEM image in Fig. 5.16 reveals mainly faceted

APT

Calcination
oxidizing conditions
(air)

f---+

Tungsten
trioxide

Calcination
reducing conditions
~
(exclusion of air or in hydrogen)

Tungsten
blue oxide

Decomposition
byHCI

r-+

Tungstic
acid

Partial thermal
decomposition

f---+

Ammonium
metatungstate

FIGURE 5.15. APT-derived important intermediates.

209

INDUSTRIAL PRODUCTION

FIGURE 5.16. SEM image of APT crystals.

crystals and only few intergrown or agglomerates. A typical grain size distribution of a
crystallized APT is shown in Fig. 5.17. For special purposes, classified APT is also
available. Specification of some physical properties and impurity concentrations as
common today are given in Table 5.7. They reflect the high standard of the technical
APT production.
Specially purified APT (by multiple-step liquid extraction of selected, very pure
batches under clean-room conditions) for the production of 4N and 6N tungsten sputter
targets shows a much lower impurity content (Table 5.8).
Solubility. The solubility in water is low: 20 g W0 3 /1 at 20C; 60 g W0 3 /1 at 90C.
Thermal Decomposition. On heating APT to 300-800C, ammonia and water are
evolved. The decomposition becomes more complete as the temperature and duration
of heating are increased. The final decomposition product is determined by the

:1------

~ :j
!

IL

15
10
5

20

40

60

80

100

120

140

Size (11m)

FIGURE 5.17. APT grain size distribution.

160

160

200

210

CHAPTER 5

TABLE 5.7. APT Trace Impurities (upper limits in ~g.

Element
Al
As
Bi
Ca
Cd
Co
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
P
Pb
S
Sb
Si
Sn
Ti
V
F
Li
Nb

2
5
10
I
10
10
10
3
10
10
7
10
20
10
7
7
7
8
10
1
10
10

4
0.1

0.1
3
10
1
7
5
1
5
0.5
7
1
3
0.3
3
3

g-'t

5
10

5
7

7
5
0.5
5

5
5
10
10
5
10
10
5
5
30
10
5
7
5
7
10
10
10
10
25
5
10

1
2
1
6
5
7

5
5
I
10
9
5
5
20
9
5
5
5
2
8

1
6
7
7
14
8
7
0.5

6
0.5

ai, Chinese National Standard GB 10116-88. 2, Excellent Chinese Grade (Hengdu

Tungsten). 3, Specification Wolfram Bergba- und Hiittenges.m.b.H. 4, Osram Sylvania. 5, Fujian Xiamen Tungsten Products Plant. 6, Chinese Production, Jiangxi
Province.

decomposition temperature, the retention time in the furnace, and the reduction potential of
the decomposition atmosphere.
If heating takes place with exclusion of air or in an inert gas atmosphere (e.g.,
nitrogen), part of the ammonia evolved is cracked and the hydrogen formed can cause a
slight reduction of the hexavalent tungsten matrix ("autoreduction capability of APT";
partial formation of pentavalent tungsten). The degree of reduction and the formation of
compounds is determined by the decomposition conditions. The product of this type of
calcination is greenish blue to flashing dark blue in color and is called Tungsten Blue
Oxide (TBO).
If the decomposition is carried out under oxidizing conditions, a slight reduction can
occur intermediately at low decomposition temperatures, but the final product is always
tungsten trioxide (W0 3).
Besides temperature, time, and decomposition atmosphere, the amount of APT plays
an important role since the mass of APT itself is responsible for producing a certain
amount of ammonia and water as well as hydrogen and nitrogen. The powder layer,
depending on its thickness and porosity (which increases during decomposition due to an

INDUSTRIAL PRODUCTION

211

TABLE 5.8. APT Specification for

UHP Tungsten According to the
Upper Permissible Limits in UHP
W Powder
Element
Al
B

Na
K

Ca
Mg
Fe
Co
Cr
Ni
Cu
U
Th

Concentration (ng . g -I)

140
35
35
35
70
70
70
15
70
15
15
0.1
0.1

increase in density), retains the gases released for some time. This fact explains the
occasional contradictory literature on the decomposition of APT. For example, it is easy to
understand that in a boat of comparable size 109 or 1000 g will produce different
atmospheres, especially inside the powder layers whose heights are also quite different.
Under oxidizing conditions, the APT decomposition path is as follows [5.55]:
Between 20 and 100 e only dehydration occurs and the product is crystallized,
dehydrated APT:

In the temperature range 180-225e, ammonia is released and the APT converts to
amorphous ammonium meta tungstate (AMT):

Between 230 and 325 e, ammonia as well as water vapor are evolved. The product is also
amorphous:

By increasing the temperature to 400-500 e, all residual ammonia and water is released,
and the reaction product is tungsten trioxide:

212

CHAPTER 5

Under a slightly reducing atmosphere between 220 and 325 C amorphous, and above
325 C, crystallized, ammonium tungsten bronzes form: (NH4)x W0 3 . Under stronger
reducing conditions, conversion to lower tungsten oxides takes place.
5.3.2. Tungsten Trioxide [5.60]

In a technical scale, W0 3 is almost exclusively produced by calcination of APT under

oxidizing conditions (in air). Usual equipment consists of rotary furnaces operating at
500-700C. Sufficient air supply must be provided to suppress any reducing reaction by
the partly cracked ammonia. The ammonia evolved can be recovered by absorption in cold
water and concentrated by subsequent distillation.
The W0 3 particles are pseudomorphous to APT. This means that the shape and size
of the particles are the same as the APT crystals, but they are built of very small W0 3
grains (Fig. 5.18) forming a large oxide sponge with a high degree of microporosity
(specific surface area). Their grain size and agglomerate structure depend on the
calcination conditions (heating rate, temperature, and time). Higher temperature and low
heating rates result in coarser grains. Above 700 C, coarse, faceted W0 3 single crystals
form due to enhanced chemical vapor transport of the oxide.
Low-temperature calcined W0 3 (approximately 500-550 0c) [5.55] is highly reactive
and dissolves easily in water, which is not the case for higher-temperature calcined W0 3 .
For special purposes, especially in the case where a high specific surface area is
necessary and APT pseudomorphology is undesirable, W0 3 can be produced also by
calcination of tungstic acid.
As precursor for the Wand WC powder production, W0 3 lost its importance mainly
to tungsten blue oxide. W0 3 is also used as a yellow pigment.
5.3.3. Tungsten Blue Oxide (TBO) [5.59-5.65]
Industrial Production. TBO is formed by calcination of APT under slightly reducing
conditions. The conversion can be performed either in multitube push-type furnaces or in

FIGURE 5.18. SEM image of W0 3 . The size and shape of the APT crystals are maintained, but the
pseudomorphs consist of agglomerates of very small oxide crystals with sizes between 0.1 and 0.5 J.1m.

INDUSTRIAL PRODUCTION

213

rotary kilns. Various atmospheres are used. Generally, in push-type furnaces a flow of
hydrogen or hydrogen-nitrogen mixtures is applied. In rotary furnaces one usually takes
advantage of the reducing capacity of the gases evolved during the decomposition (H, H2,
NH 3), leading to the desired formation of reduced tungsten species.
Temperature may vary between 400 and 900C. Literature values can be misleading,
because some are related to the real temperature of the powder layer while others are
furnace temperatures measured at the wall of the heating compartment or tube. These
temperatures can differ considerably, due to the overall endothermic behavior of the
APT -+ TBO decomposition reaction. The exposure time in a rotary kiln is usually much
shorter than in the pusher, and the decomposition temperature is therefore higher for
obtaining a similar degree of thermal decomposition.
Chemical Composi~on. TBO is not a defined chemical compound but is a mixture of
different constituents, such as ammonium, hydrogen and hydronium tungsten bronze
phases, tungsten trioxide, tungsten-~-oxide (W0 2.9 or W20058), and tungsten-y-oxide
(W0 2.72 or WI 80 49)' Under more reductive conditions, even traces ofW0 2 and ~-tungsten
can be present. The relative amounts of the various compounds in the TBO depend on the
calcination parameters:

temperature,
heating time,
composition and pressure of atmosphere,
mass of APT flow with time,
gas flow,
layer height in the boat (pusher furnace),
slope and rotation rate (rotary furnace).

The oxygen index (molar ratio OfW) is commonly used to characterize the degree of
reduction of TBO. However, since most TBOs also contain ammonia and water in addition
to Wand 0, a more complete description is given by x(NH3) . y(H 20) . WOn'
A series of analyzed industrial samples gave the following ranges for the coefficients
x and y and the index n: x = 0.02-0.09,y = 0.02-0.14, and n = 2.82-2.99.
Qualitative and quantitative X-ray analyses of the same samples revealed quite a large
scatter in composition: tungsten bronzes, 0--45%; W0 3 , 0--45%; W02.9 , 5-20%; W0 2.72 ,
0-25%, and amorphous, 30-55%.
Amorphous species form by dehydration which, on further heating, convert into
crystalline binary tungsten oxides as well as tungsten bronzes. The conversion from
amorphous to crystalline is a slow process. Therefore, if the heating period is short, as in
rotary furnaces, the time available for overall crystallization is insufficient. This is why
rotary-furnace-derived TBO can be high in amorphous oxide when compared to TBO from
pushers.
Homogeneity. The product of a rotary furnace is more homogeneous in regard to
differences between individual particles, since the powder is constantly being mixed by the
turning motion. The TBO particles are composed of different compounds but are similar in
composition, while pusher-furnace-derived TBO particles can also differ in composition,
depending on their position in the powder layer.
Physical Properties. The physical properties of the TBO, such as particle size, size
distribution, and internal porosity, are of importance for the subsequent hydrogen

214

CHAPTER 5

reduction process. Size, size distribution, and shape of the particles are related to the
starting APT as a consequence of the crystallization process and do not alter significantly
during the decomposition process, while the internal porosity is a result of the blueing
process. The degree of porosity is influenced by the calcination parameters (time,
temperature). The specific surface area in commercial TBOs can vary between 1 and

15 m 2/g.

Both macromorphology and microporosity of the TBO particles determine the

permeability and thus the diffusion resistance of a TBO powder layer, which is a
determining factor for the material exchange (H2 -+ H20) during reduction.
Characterization. Industrially produced TBO is characterized chemically by the
residual ammonia content and the oxygen index, leaving open the composition in
regard to the different compounds and the amount of amprphous species. These
determinations would be much too tedious and expensive for routine purposes. As long
as APT quality and calcination conditions are kept constant, the composition will be
reproducible. Consequently, in many companies it is preferred not to buy TBO but APT,
and to perform the blueing in-house.
Physical characterization of TBO includes particle size and size distribution measurement by laser diffraction (macroporosity) as well as specific surface area measurement
(microporosity). Particle size measurement by FSSS (Fisher Sub-Sieve Sizer), as also
sometimes used, is misleading, because of the porosity of the TBO particles. An empirical
relationship between FSSS and particle size measured by laser scattering can, however, be
detected if the microporosity of the samples is uniform (constant blueing conditions).
5.3.4. Tungstic Acid

In order to make use of the high APT purity in modern processing, tungstic acid is
produced today by treating an aqueous APT crystal slurry with hydrochloric acid. In this
way APT is decomposed and H2 W0 4 is precipitated. After filtration, it must be thoroughly
washed to remove ammonium chloride. The earlier process of precipitation from sodium
tungstate solutions by addition of acids no longer has industrial importance.
Tungstic acid has a high active surface. The former, most important intermediate
today is only used in smaller quantities for special purposes:
production of ultrafine tungsten and tungsten carbide powders in order to
circumvent the sometimes disturbing pseudomorphology of APT-derived products,
and
for tungsten chemicals.

5.3.5 Ammonium Metatungstate (AMT) [5.66-5.71]

AMT corresponds to the formula (NH4)6H2 W12040. xH 20. The amount of combined
water is variable. Usual technical products contain 3 to 4 molecules. Further dehydration
may lead to insoluble compounds.
The importance of AMT lies in its extremely high water solubility. This is why AMT
has gained increasing usage as an intermediate in a variety of applications like chemicals
and catalysts.

INDUSTRIAL PRODUCTION

215

Several methods of preparation exist:

Acidification of an ammonium tungstate solution and subsequent evaporation.
Partial thermal decomposition (see Section 5.3.1).
Addition of tungstic acid to an ammonium tungstate solution to reach a 1:2
NH 3 /W0 3 ratio, followed by evaporation.
Partial replacement of ammonium ions by hydrogen ions using selective ion
exchange and subsequent evaporation.
The latter three methods are preferred today for industrial production, since acidification
and thermal decomposition result in lower-quality products (less purity and amorphous
structures respectively).
Acidification by Electrodialysis. Ammonium polytungstate solution from the solvent
extraction process, possesses a pH of approximately 9 and is circulated through an
electrodialyzer as anolyte. In the electrical field, ammonium ions pass through the cationic
exchange membrane forming ammonium sulfate in the circulating catholyte. The pH drops
due to the decrease in the ammonium ion concentration (generation of H+), and the
reaction is stopped at pH 3. The following overall equation represents the process:

The resulting solution is evaporated to a syrup from which white crystals are recovered.
The yield is better than 99%.
Properties. AMT is a white crystallized powder having a density of 4g/cm3 .
Decomposition starts at 100C and, between 200 and 300 C, it converts to the anhydrous
form. Further decomposition leads to W0 3 .
The solubility in water at 22C is 1635 gil corresponding to 1500 g W0 3 /1 and, at
80C 2200 gil are dissolved. The pH of the aqueous solution depends on the AMT
concentration and varies at 25C between 2.5 and 5, indicating that different species are
present simultaneously.
High solubility is the prerequisite for the production of highly concentrated, alkalifree tungsten solutions. They are necessary to impregnate suitable carrier materials like
y-A12 0 3 in catalyst production.

5.4. TUNGSTEN METAL POWDER PRODUCTION

5.4.1. General

The manufacture of tungsten metal powder is a crucial step in tungsten metal and
alloy production, since the powder properties significantly affect the properties in
subsequent operations, such as pressing, sintering, and metalworking. Between 70 to
80% of tungsten worldwide is produced via powder metallurgy and thus passes through
this important stage. In the past, advances in powder technology have greatly contributed
to the development of tungsten and its alloys, as well as today's high standard of product
quality. Powder grades are tailormade for the subsequent applications, and the powder

216

CHAPTER 5

industry is facing a competitive market where the stringent fulfillment of exacting demands
is an important part of business success.
The powder is characterized by chemical (purity), physical (grain size, size distribution, shape, agglomeration, etc.), and technological properties (fluidity, compaction
density, etc.), which are influenced by the production process and which can be
controlled-to a certain extent-by the process parameters.
Today, production of tungsten metal powder is accomplished almost exclusively by
the hydrogen reduction of high-purity tungsten oxides. Reduction of the oxides by carbon,
common in the early years of metal production, is presently used only for the production of
tungsten carbide (direct carburization). The hydrogen reduction of tungsten halides (Axel
Johnson process) has not become established on a large scale.
The common starting materials are tungsten trioxide (W0 3) and tungsten blue oxide
(W0 3 _ x ), the latter being the most widely used material. Tungstic acid (H2 W04 ) is used
only for selected metal grades.
In principle, APT can also be directly reduced without any prior calcination step. The
disadvantage of direct reduction is the formation of ammonia which has to be scrubbed,
but a certain amount of ammonia cracks and dilutes the hydrogen by nitrogen. Consequently, from time to time, part of the contaminated, circulating hydrogen must be vented,
thus increasing costs.
Reduction is carried out in pusher furnaces in which the powder passes through the
furnace in boats or in rotary furnaces (see below). Walking beam furnaces or furnaces with
internal band conveyors are less often used. Fluidized-bed reactors are still not in
commercial use, except for the production ofnanophase Wor WC/Co powder precursors
(see Chapter 9). Furnaces are provided with several temperature zones controlled between
600 and 1100 C. A large excess of hydrogen is used, which is recycled to the furnace after
purification. The flow of hydrogen is usually in a countercurrent direction, more rarely
cocurrent. The hydrogen acts not only as a reducing agent but also carries away the water
formed.
The reduction of tungsten oxides by hydrogen to tungsten metal is, in some respect, a
unique process. It offers the possibility to produce tungsten powder of any desired average
grain size between 0.1 and 10 ~m (and, in the case of doped oxides, even up to 100 ~m),
starting from the same oxide precursor. Individual tungsten particles form during reduction
as a result of chemical vapor transport of tungsten (vaporization/deposition process),
which is responsible for the final powder characteristics. For principles of the reactions
occurring during reduction as well as thermodynamic, kinetic, and morphological aspects,
see Section 3.3.
By changing the reduction parameters, powder characteristics like average grain size,
grain size distribution, etc. can be regulated. Temperature and humidity (i.e., the water
vapor partial pressure prevalent during reduction) are the two main parameters in steering
the average grain size of the W powder, the latter being related to a number of oxide and
process-related variables as indicated in Fig. 5.19 and discussed briefly below. The reason
for the strong influence of the humidity on powder grain size originates in the strong
dependence of humidity on the nucleation rate of the metal phase and the high mobility of
tungsten due to the presence of a volatile tungsten compound ([(W02 (OH)zD. The lower
the humidity, the higher the nucleation rate (under isothermal conditions) and the smaller
the grain size.

217

INDUSTRIAL PRODUCTION

Tungsten powder
average grain size

Temperature

Water vapor concentration

Tungsten oxide related

Hydrogen related

Tungsten mass flow

Flow rate

Powder layer height

Flow direction

Porosity of powder layer

Dew point

FIGURE 5.19. Average grain influencing reduction parameters.

Temperature. Temperature influences the rate of all reactions occurring during

reduction, hence the dynamic humidity and partial pressure of the volatile [W02(OHh]
which forms during reduction and which is responsible for the chemical vapor transport
(CVT) of tungsten. Temperature and tungsten particle size are directly proportional while
temperature and time required for final reduction are inversely proportional.
Oxide Feed. The tungsten mass flow determines the amount of H20 liberated during
the entire reduction process. The higher the flow, the larger the grain size.
Tungsten Powder Layer Height. During reduction and accompanying water formation, the powder layer exerts a considerable diffusion resistance against water removal
from the layer. The higher the layer, the greater the diffusion resistance and the more
slowly the reaction water will be removed. The local humidity is higher at the bottom of
the layer than at the top. It is this local humidity which determines the nucleation and
growth conditions of the metal particles formed at a particular temperature. The layer
height is directly proportional to powder grain size.
Porosity of the Powder Layer. The porosity of the powder layer, and thus its
permeability, is determined by the macroporosity (intermediate space between the oxide
particles) and by the microporosity (porosity of the individual oxide particles). For
parameters influencing porosity, see Section 5.3. The higher the porosity of the powder
layer, the better the material exchange H2 0 -+ H2 during reduction and the less the grains
of the tungsten particles will grow, resulting ina smaller particle size.
Hydrogen Flow Rate. A higher hydrogen flow enhances the material exchange due to
the more rapid removal of water vapor. Therefore, the flow is inversely proportional to the
average grain size.
Hydrogen Flow Direction. Cocurrent hydrogen flow with respect to the tungsten flow
generates a higher dynamic humidity at the later part of the reduction, while counter-

218

CHAPTER 5

current flow (which is the standard condition) provides higher humidity during the early
reduction stages.
Hydrogen Dew Point. The dew point of the incoming hydrogen influences the overall
humidity during reduction. More "wet" hydrogen enhances the tungsten particle growth.
It is important to understand that the final average grain size of the tungsten powder is
a consequence of combining all the aforementioned parameters. The general rules are:
Small grain size: low temperature, dry hydrogen, high hydrogen flow rate, low dew point,
small powder layer height, high porosity, low oxide feed.
Large grain size: high temperature, wet hydrogen, low hydrogen flow rate, high dew
point, large powder layer height, low porosity, high oxide feed.

Empirically based equations for calculating the reduction time and average particle
size have been derived for rotary furnaces [5.73] but have never been applied industrially.
The main difficulty is that the properties of the raw material play a crucial role in the
reduction process, and these characteristics are not well enough represented by corresponding equations. This is particularly true for smaller grain sizes. For larger grain sizes,
the influence of the oxide precursor is less pronounced.
In industrial practice, the choice of proper reduction parameters is based exclusively
on empirical experience. Besides the average grain size, the reduction parameters also
influence the grain size distribution, agglomeration, apparent density, and grain morphology.
Grain Size Distribution. Grain size distribution is to a great extent the consequence
of powder layer height. The growth conditions for the individual particles are different and
depend on their position within the powder layer. The humidity is higher in the interior and
decreases as one approaches the surface. This gradient results in larger grain-sized
particles inside and smaller grain-sized particles at the surface or surface-neighboring
areas. It is easy to understand that the distribution is broader for high powder layers and
closer for lower layers. In any case, the distribution can be improved (made closer) by
using "wet" hydrogen, since the water vapor gradient from inside to outside the layer will
be decreased.
Agglomeration. One may distinguish between three different types of agglomeration:
Reduction performed at low temperature and dry conditions results in the
formation of metal sponges, which are pseudomorphous to the oxide precursor.
The very fine crystals (some tenths of a micron or less) stick together in a more or
less loose manner. Usually, this type of agglomeration is called pseudomorphology
(Fig. 5.20). The finer the crystal size, the higher the strength of these agglomerates.
Bigger tungsten crystals (from approximately I to several ~m), grown individually
grown by CVT, are welded together. This type of agglomeration is looser than the
above-described pseudomorphology and occurs mainly under dry reduction
conditions (Fig. 5.21). It can be influenced by the dew point of the incoming
hydrogen. Higher dew point results in less agglomeration. The strength of this
agglomeration type increases with temperature.
Reduction under wet conditions and high temperature (high concentration of
[W02 (OHh] in the vapor phase) results in strongly intergrown crystals (Fig. 5.22).

INDUSTRIAL PRODUCTION

219

FIGURE 5.20. Fine tungsten powder agglomerated in APT pseudomorphs; average grain size of the powder is
below 0.5 11m.

FIGURE 5.21. Individually grown and loosely agglomerated tungsten crystals.

FIGURE 5.22. Coarse intergrown tungsten crystals.

220

CHAPTER 5

Agglomeration is closely related (inversely proportional) to the apparent density of

the tungsten powder. Correspondingly, the apparent density can be influenced within
certain limits by the hydrogen dew point. Agglomeration is a prerequisite for good
compactability of the tungsten powder.
Morphology. As noted earlier, low temperature and dry conditions largely suppress
any CVT of tungsten and lead to the formation of metal sponges, which are pseudomorphous to the oxide precursor (APT, H2 W04 ). They consist of very fine, polygonal and
polycrystalline metal particles. With increasing temperature and humidity, individual
tungsten grains form by CVT over comparably large distances. The particles are faceted
and commonly exhibit the characteristic shape of the cubic metal (Fig. 5.23). Well-faceted
crystals, showing growth steps and being partly intergrown, are characteristic for very
humid conditions (high temperature, large powder layer height; Fig. 5.24).

FIGURE 5.23. Small cubic tungsten particles obtained at a reduction temperature of 800 DC (powder layer height,
2mm).

FIGURE 5.24. Well-faceted tungsten crystals obtained at a reduction temperature of 1000C (powder layer
height, 20 mm).

221

INDUSTRIAL PRODUCTION

5.4.2. Push-Type Furnace

Metal boats are charged with oxide to a height ranging from a few mm up to several
cm and are pushed in stages through the furnace in corrosion-resistant steel tubes at
specific time intervals. By introducing a new boat into the tube, the row in front is pushed
forward by the length of a boat. Hydrogen in excess flows either co- or countercurrent to
the tungsten flow direction. The hydrogen is not only responsible for the reduction process
itself but serves also to remove the water vapor formed and also acts as protecting
atmosphere in the cooling zone. The "wetted" hydrogen leaving the furnace is dried to a
desired dew point and recycled to the furnace (Fig. 5.25). As indicated, hydrogen having
higher dew points can also be fed into the furnace.
Hydrogen has to be applied in large excess, which guarantees a fast flow over the
powder layer. The excess depends on the desired grain size (smaller for coarse and higher
for fine powder). The range is somewhere between 2.5 and 40 times stoichiometric.
Multitube furnaces (14 to 18 tubes arranged in two rows) are frequently in use today.
A photograph of an industrial aggregate is shown in Fig. 5.26. The boat material, in most
cases, is an iron alloy high in Ni and Cr (Inconel). More seldom, because of the high price,
boats are made of TZM (molybdenum alloy with Ti, Zr, and C) or pure tungsten.
The big disadvantage of the iron alloys is that diffusion of the elements occurs into
the contacting tungsten powder layer. In this respect, Ni is the most dangerous element
although widely used. Ni rapidly diffuses over the tungsten grains, thereby weakening the
surface of the bottom and wall of the boats. With time, a Ni, Fe, Cr, and W containing
scale is formed. This scale sticks more or less firmly to the boat. After several travels
through the furnace, it gets thicker and partly breaks off, contaminating heterogeneously
the tungsten powder. Bigger-scale particles can be separated by the always applied
screening process following the reduction, but the smaller particles remain in the tungsten
powder. The higher the temperature and humidity, the more pronounced the scale
formation. Cast alloy material (coarse microstructure) shows enhanced scale formation
compared to boats made of rolled sheet. Alloys containing Co instead of Ni are more
resistant, but the high price of Co makes them unacceptable for boats. Co containing alloys
are only in use as tubes in rotary furnaces.
The furnaces are either gas fired or electrically heated in three or four separate zones.
Furnace temperatures range between 600 and 1100 C. For smaller and medium W grain

H2 from
reduction furnace

Scrubbing

Gas holder ----"T"""""----'

Virgin H2 - - - - - '
H2to
reduction furnace

Dryer
FIGURE 5.25. Hydrogen regeneration.

2nd cooling stage

(-10C)

222

CHAPTER 5

FIGURE 5.26. Push-type reduction furnaces. By courtesy of Wolfram Bergbau- und Hiittenges. m.b.H., Austria.

sizes, a temperature profile is preferred in order to decrease the time necessary for the last
reduction step from W02 to W (slow reduction rate). For larger grain sizes (>6I1m),
isothermal reduction conditions are applied.
The reduction is commonly carried out in one stage. Alternatively, a two-stage
reduction sequence can be applied instead. In this case, the first reduction stage takes place
at lower temperature (500-700C; formation of brown oxide, W02 ) and the second stage
at 600-1100 C (formation of tungsten metal).
In industrial practice, the boats are loaded with a certain oxide weight (layer height)
and pushed through the furnace with a given temperature profile and hydrogen throughput.
After dynamic equilibrium is reached, the particle size of the metal powder is measured. If
the powder does not meet the requirements, parameter adjustments such as change in
temperature, boat load, hydrogen throughput, or push time are introduced.
Typical reduction conditions and furnace capacities for various grain sizes are
summarized in Table 5.9.
Subsequent to reduction, the powders are screened on 60 mesh (sometimes also on
200 mesh) to eliminate contaminants stemming from furnace or boat materials and are
blended to form a homogeneous powder batch. No special atmosphere is necessary for
handling, since the powder surfaces are rapidly saturated with oxygen and water vapor.
However, below 111m, the powders may be pyrophoric and precautions are necessary, in
particular below 0.5 11m. Reduction under cocurrent hydrogen flow is the most effective
method to avoid burning of the fine powders. Already, during the cooling stage in the
furnace, the powder is contacted to "wet" hydrogen, and the surface is saturated when the

223

INDUSTRIAL PRODUCTION
TABLE

5.9. Push-type

Furnace Reduction Conditions

Capacities for Different Grain Sizes

and

FSSS grain size (11m)

Temperature Cc)

Capacity (kg. W h- 1)

0.7
1.5
2.5
5.0
12.0

600-800
700--900
800--950
900-1000
900-1000

40--50
90--110
100--120
130--150
140--160

By courtesy of Wolfram Bergbau- und Hiittenges.m.b.H., St. Martin i.S. A-85452.

powder leaves the furnace. Under countercurrent flow conditions, the powder has to be
slowly saturated with oxygen. This can be achieved either by an inert gas storage (nitrogen
or argon containing small amounts of oxygen) or by exposing the powder to the
atmosphere in small portions in order to omit local overheating. This can be done by
leaving the powder in the boat for approximately 30 min.
It is obvious that the furnace capacity for smaller grain sizes, especially for submicron
tungsten powder, is low. Only very thin powder layers can be applied to retain the grains
from growing. In order to improve the capacity, a double or triple boat technique was
invented. The reduction boat is topped with one or two upper boats in a way permitting
hydrogen flow between the boats, so that the capacity also for smaller grain sizes could be
increased considerably.
Modem furnaces are fully automated, which means that all variables can be set and
controlled. Loading, pushing, and discharge of the boats is done by machine.
The advantage of the push-type furnace in comparison to rotary kilns is its flexibility
in switching from one condition (grain size) to the next and in its high capacity, especially
for finer powder qualities. Disadvantages are higher energy consumption, broader grain
size distribution, more contamination by the scale from the boats, and higher maintenance
costs.

5.4.3. Rotary Furnace [5.72-5.74]

In contrast to the pusher furnace, where a static powder layer passes the furnace, a
dynamic powder flow exists in the rotary kiln. Furnace rotation and the incline
continuously move the powder through the hot zone. Feed rate, rotational speed, incline,
and lifters inside the tube determine the depths of the dynamic powder layer. High oxide
feed, longer retention time (low rotation rate and low inclination), and higher temperature
result in coarser tungsten grains.
The powder layer is constantly disturbed by the rotating motion and powder from
inside moves to the surface, and vice versa. Therefore, the water-vapor retaining capability
of the layer is less compared to a static layer, and the material exchange rate H2 0 -+ H2 is
enhanced. The humidity in a rotary furnace powder layer is thus lower and reduction
proceeds under drier conditions as compared to the pusher. Lifters inside the tube
combined with stepwise turning motion make sure that a powder bed is maintained for
a certain time giving the chance to build up higher humidity both temporarily and locally.

224

CHAPTER 5

In order to maintain the humidity for a certain grain growth, only a small excess of
hydrogen is necessary (between 2 and 3 times stoichiometric). This corresponds to the
lowest values in pusher furnaces, applied only for very coarse powders. Moreover, the
influence of the H2 /W0 3 molar ratio on grain growth is less pronounced.
The hydrogen is cleaned and dried and recycled to the furnace after preheating. Due
to the constant motion of the powder bed, fine particles are swept away by the hydrogen
flow and form a dust, which must be removed by filtering devices.
Very coarse blue oxide (derived from classified APT) has improved flow characteristics and helps to prevent powder from sticking to the tube wall.
Due to the "dry" conditions, rotary-furnace-produced tungsten powder shows a high
degree of agglomeration, lower apparent density, and closer grain size distribution. The
upper size limit is approx. 6 ~m.
A photograph of a modem rotary furnace is shown in Fig. 5.27.
Furnaces in use possess different sizes: inner diameter between 300 and 750 mm and
heated drum length from 3000 to 8000 mm. The most common equipment has an inner

FIGURE 5.27. Rotary reduction furnace. By courtesy of Wolfram Bergbau- und Hiittenges. m.b.H., Austria.

225

INDUSTRIAL PRODUCTION

TABLE 5.10. Rotary Furnace Conditions and Capacities for

Different Grain Sizesa
FSSS grain size
(1JlII)

0.7
I.3
4.0
a

eC)

Capacity
(kg.W.h- l )

600--800
600--850
800--920

18
27
58

Temperature

By courtesy of Elino Industrie-Ofenbau, Carl Hanf GMBH

+ CO, Diiren 0-52355.

diameter of 600 mm and a heated length of 6500 mm. Typical conditions and furnace
capacities for various tungsten powder grain sizes are presented in Table 5.10.
Advantages of the rotary furnace compared to a pusher are less specific energy
consumption (valid only for median grain sizes), while automation is less complicated and
consequently maintenance is cheaper. Finally, the product is more homogeneous and less
contaminated, but the particle morphology may differ significantly from powders produced
in pusher furnaces.

5.4.4. Tungsten Powder

Technical tungsten powder grades prepared by hydrogen reduction are commercially
available in average grain sizes from O.lllm (100 nm) up to 100 Ilm. The whole palette of
grain sizes find their application in cemented carbides. The main portion of tungsten
powder produced (80%) is directed into that production.
The starting powder for the powder metallurgical production of pure tungsten (ductile
tungsten) and sintered tungsten alloys commonly covers grain sizes between 2 and 61lm.
Finer or coarser powder would be unsuitable. Fine powder has too high a sintering activity,
not permitting the evaporation of trace impurities during the sintering process, while
coarser powder leads to incomplete sintering under usual conditions.
Extremely coarse powder gained by classification (to separate any finer particles)
exhibits excellent flow characteristics and is used in plasma spraying.
Chemical Properties: Purity. The purity of the tungsten powder is of particular
importance in PM manufacturing of tungsten metal, since during subsequent sintering
further purification through evaporation is only possible to a certain extent. The demand
for purity of tungsten powder has increased steadily during the last three decades.
Considerable improvements in hydrometallurgy have led to concentrations fairly below
10 Ilg/ g for most of the elements. This trend with time can be demonstrated by comparing
today's usual specifications with those given in the last book on "Tungsten" by Yih and
Wang, published in 1979 (Table 5.11).
The reason for this enhanced demand for powder purity originates in the fact that
remaining impurities after sintering greatly affect the workability (grain boundary strength,
recrystallization temperature, etc.) and properties of the final product.
By considering the impurity-to-tungsten ratio (disregarding oxygen and ammonia) on
the long path from ore concentrates to compact tungsten metal, one observes a constant
increase in purity up to the stage of APT crystallization. At this stage, the maximum purity

226

CHAPTER 5

TABLE 5.11. Tungsten Powder

Specifications 1975-1995 (upper
limits in Ilg. g - I b

Elements
Al
Fe
Ni
Mo
Si
Ca
Co

Cr
Cu
Mg
Mn
Pb
Sn
a

10
50
30
100
10
20
10
10
10
10
10
10
10

3
8
5
16
5
3
5
8
2
5
3

10
10
10
20
10
10
2
10
2
2
2
2
10

I, Yih and Wang, 1975.2, Chinese Specification, 1995. 3, Data Sheet, 1993, Wolfram
Bergbau- und Hiittenges.m.b.H.
Lower impurity levels in recent years are
partly a consequence of improved production techniques and partly due to the higher
sensitivity of advanced analytical methods.

is more or less reached. Consequently, the purity of the tungsten powder depends mainly
on the APT cleanliness. During APT processing to tungsten powder, the purity already
decreases again. Sources of contamination are contacts with metallic tubes or boats in the
respective furnaces. Slightly enhanced, overall concentrations of elements, such as Fe, Ni,
Cr, and Co, are the consequence. This type of impurity occurs heterogeneously and
represents small areas of locally high concentrations of foreign elements. If they are big
enough, they might act as the origin of sintering defects.
Moreover, volatile elements or compounds present in the hydrogen atmosphere can
be adsorbed by the tungsten powder during reduction. A typical example are alkali metals.
Physical Properties. The relevant physical properties are average particle size,
particle size distribution, apparent, tap, and compact or green density, specific surface
area, degree of agglomeration, and morphology. All of them are closely related to each
other and can be influenced by the oxide properties and reduction conditions. They
represent the important criteria for further processing and are responsible for the
compactability, sintering behavior, dissolution reactions during liquid-phase sintering,
and carburization reactions. Physical properties, methods of determination, and standardized procedures are outlined in Table 5.12.
Average particle size and particle size distribution: "As supplied" tungsten powder
is always more or less agglomerated, depending on reduction conditions. Therefore, the
measured value for the average particle size in this stage does not always correspond to the
real particle size and can even be misleading. For example, certain submicron powders
show "as supplied" grain sizes of 1-2 Jlm, but after deagglomeration these values drop to
0.4-0.5 Jlm.

INDUSTRIAL PRODUCTION

227

TABLE 5.12. Important Physical Properties of Tungsten Powder

Parameter

Method of determination

Average grain size

as supplied

Fisher Subsieve Sizer

lab milled

Fisher Subsieve Sizer

Grain size distribution

Apparent density
Tap density
Press density

Specific surface

SEM
Gravity sedimentation turbidimentry
Gravity sedimentation
X-ray absorption
Laser scattering
Scott volumeter
By measuring the height of a cylindrical
compact prepared under standardized
conditions (powder weight, compacting
pressure and time)
BET

Strength of compacts
Morphology

Remarks
ASTM No. B 330-88
(reapproved 1993)
after standardized milling
according to ASTM No. B 430-90
for average grain size ~ 0.5 11m
ASTM No. B 430-90
ASTM No. B 761-90
ASTM No. B 822-92
ASTM No. B 329-76
ASTM No. B 527-93
ASTM No. B 331-85
(reapproved 1990)

ASTM No. C 1096--86

(reapproved 1992)
ASTM No. B 312-82
(reapproved 1988)

SEM

For production control in the range of 1-10 11m, the "as supplied" value mostly
suffices, because production conditions are kept constant within close limits. However, for
submicron-sized powders as well as for a correct characterization of coarser powders, only
the "lab milled" powder should be used.
Normally, the particle size distribution is a function of particle size. The bigger the
particle size, the broader the distribution. For a given particle size, the distribution can be
made narrower during production by application of wet hydrogen and/or doping the oxide
with alkali compounds. Typical particle size distributions are shown in Fig. 5.28.
Particle size distribution measurements should also be made from "lab milled"
samples only. The particle size distribution influences the compacting behavior and green
density as well as the sintering properties.
Agglomeration. The difference between "as supplied" and "lab milled" (deagglomerated) particle sizes is a measure of the degree of agglomeration. Agglomeration is very
important for the strength of the green compacts and is therefore a necessary property for
powder going into ductile tungsten production.
Much effort was expended in the past to elucidate the dependence of the compactability of tungsten powders on grain size and grain size distribution. Low grain sizes result
in too low green densities due to high friction between the particles. The closer the grain
size distribution, the poorer the particle packing. So it is evident that broader grain size
distributions, or even blends of powders having different average grain sizes, result in
better packing and higher strength of the green compacts.

228

CHAPTER 5

01

!i

J.

.. "

Diameter

(I'm)

"arla.lon In hrciC'le Size Dlslrlbulion

~ubOlicton

Il ~m WPowdcr
Broad Plrllele b.e

" Po.-d.r

12

~OI

" Po"d ..

CloJ(' P.rtlcle

Sl~('

Di.Jlribulton

Distribution

FIGURE 5.28. Particle size distribution of fine and coarse tungsten powders.

Compaction and green strength is also influenced by the particle morphology. More
irregular shapes, which cause interlocking, improve the strength.
Specific surface area. The specific surface area is an important criterion for the
sintering activity (solid state sintering), dissolution processes (liquid phase sintering), and
reaction with gaseous or solid substances during carburization. Commonly, it is in the
range of 0.01 m2 /g (coarse powders) up to 12 m2 /g (very fine powders).
Based on the specific surface area, an average particle size can be calculated.
However, attention must be paid to the surface roughness of the particles. In particular,
very fine powder particles do not exhibit "smooth" but "rough" surfaces, or even exhibit a
certain degree of microporosity, hence increasing the specific surface area by a factor
of 2-4.
Apparent and tap density. Both densities increase with increasing average grain size,
but are additionally influenced by the grain size distribution, particle shape, and degree of
agglomeration.

INDUSTRIAL PRODUCTION

229

Compressibility. This is the ratio of green density to apparent density. The ratio
increases with increasing pressure until a limit is reached. Compressibility of a powder is
an important criterion for press and die design.

5.4.5. Reduction of Doped Tungsten Oxides

Powder properties can be significantly influenced by the addition of a third element
(see also Section 3.3). This can be used industrially for the production of special W grades,
as mentioned below.
Tungsten blue oxide is treated with aqueous solutions of the respective element
compounds.
Sodium Doping. Addition of sodium compounds (concentrations equal to 50200 ~g/g Na) is a usual practice in producing coarse tungsten powder. At lOOO C and
wet conditions, average grain sizes of 10-25 ~m (depending on the layer height) can be
attained. The high temperature guarantees the extensive evaporation of sodium at the end
of the reduction process. Only for special purposes must a subsequent acid or water leach
be applied to decrease the residual sodium concentration.
Lithium Doping. Lithium compounds are preferred to sodium for very coarse
tungsten powder grades (50-100 ~m); Fig. 5.29. Reduction conditions are the same as
above. Due to the high lithium boiling point, it is not volatilized during reduction like

FIGURE 5.29. Coarse W powder obtained by Lithium doping (average particle size 50 ~m). By courtesy of
Wolfram Bergbau- und Hiittenges. m.b.H., Austria.

230

CHAPTER 5

FIGURE 5.30. HF-washed NS-doped tungsten powder. Large dopant phases which were not completely
incorporated during W crystal growth are removed by the acid washing treatment, leaving behind large holes
in the crystal surfaces.

sodium. The powder has to be leached by hydrochloric acid and water subsequent to
reduction. Finer fractions are separated by screening.
Chromium Doping. Additions of chromium compounds can be used to produce very
fine metal powder grades 0.5Ilm). Chromium oxide is formed during reduction and acts
as a grain refiner. Such powders are subsequently carburized to ultrafine We. Special
precautions have to be taken because of the high pyrophoricity of the metal powder.
NS or AKS Doping [5.75,5.76]. For the production of lamp filaments (incandescent
or halide lamps), a special doping procedure is carried out which is commonly called NS
or AKS doping. NS stands for the non-sag properties of the coiled lamp filament, while
AKS represents the doping elements used in this process: AI, K, and Si.
TBO is doped with about 3000-3500 Ilg/g K, 2000-2500 Ilg/g Si, and 400500 Ilg/ g AI in the form of potassium silicate and aluminum chloride or nitrate solutions.
In reduction of the doped oxide, part of the dopants is incorporated during the CVT
growth of the metal particles in the form of silicates. Excess dopant, which remains on the
tungsten crystal surfaces, is removed by subsequent leaching of the powder in HCI and HF
acids, while the dopants which are internally trapped are retained. The amount of
incorporated dopants, the size of the inclusions, and their chemical composition can be
influenced by the reduction conditions within certain limits. About 100-150 ppm K, 60100ppm AI, and 200-300 ppm Si commonly remain in the acid-washed powder.

INDUSTRIAL PRODUCTION

231

During sintering, the silicates dissociate thermally. The smaller AI, Si, and 0 atoms
escape via diffusion through the tungsten crystal lattice while the large K atoms remain in
the form of potassium bubbles. It is this potassium that enables the formation of rows of
smaller potassium bubbles during wire fabrication, and consequently the formation of a
long-grained interlocking microstructure during subsequent filament operation, which is
the key to the non-sag characteristics (see Section 6.2.1).
The grain-growth-enhancing effect ofK makes it possible to conduct the reduction at
a lower temperature than normally applied for W powder of 3-5 11m average grain size
(Le., the optimal grain size for powder-metallurgical processing). The hydrogen gas has to
be kept separated from hydrogen for reduction of undoped powder, because it is
contaminated by K.
NS-W powders exhibit a characteristic morphology, exhibiting holes in the crystal
surfaces (Fig. 5.30), which form during HF-washing of the powders due to the dissolution
of partly intergrown silicate phases.

5.5. POWDER METALLURGY [5.1,5.2,5.77-5.84]

Although the Wollaston process of producing platinum metal from platinum powder
is regarded as the birth of modern powder metallurgy (1808-1815), it was the pioneering
work of W. D. Coolidge on the production of ductile tungsten wires in 1909-1913 which
led to its first commercial application [5.82]. Over the years, only few changes were made
in the industrial production of tungsten metal. Powder metallurgy is still today the main
route in tungsten and tungsten alloy manufacture. Unlike other refractory metals, such as
Zr, Hf, Nb, or Ta, melting technology has assumed no industrial importance for metal
production, since the very high temperature, necessary for melting, and the resulting
coarse microstructure of the "as cast" material make further processing both difficult and
costly.
Powder metallurgy comprises two steps: compaction andsintering.

5.5.1. Compacting
Tungsten powder is consolidated into a compact by two main routes: pressing in rigid
dies (uniaxial pressing) and isostatic pressing in flexible molds (compaction under
hydrostatic pressure). Other techniques, such as powder rolling, cold extrusion, explosive
compaction, slip casting, vibratory compaction, or metal injection molding, have gained
no industrial importance.
Tungsten powder is not easy to compact due to its relatively high hardness and
difficult deformation. Nevertheless, in most cases compaction is performed without
lubricant to avoid any contamination by the additive. The resulting compacts are generally
sufficiently strong so that they can be handled without breaking. For machining the part, it
must be pre-sintered beforehand.

5.5.1.1. Die Pressing. Pressing of powders in rigid dies is carried out either in
mechanical or hydraulic presses. The pressure is applied from the top, or from the top and
bottom (double action presses). Die and punches are made of high-speed tool steel or

232

CHAPTER 5

(more rarely) hardmetal. Mechanical presses (pressing forces up to 1 MN) are used for
small parts and high production rates. They allow a higher degree of dimensional precision
and are well suited for process automation. Hydraulic presses are mainly used for simple
preforms. Large presses with up to 30 MN (3000 t) pressing force are used for pressing of
plates which are to be rolled to sheet metal. The size and shape of the compact are limited
by the capacity of the press and also by the geometry of the part.
Due to the friction between the powder and the die wall and the nature of the load
distribution inside the die, the pressing density is not the same all over the compact. This is
more pronounced for large parts and large part heights and can lead to crack formation
and/ or distortion of the pressed compact during sintering. Large and critical parts are
therefore commonly pressed isostatically.
Typical compaction pressures are in the range of 200-400 MPa (2-4 t/cm 2) but can
reach 1000 MPa (using hardmetal dies and punches). The green density (compact density)
is in the range of 55--65% of the theoretical density (75% at most) and it depends upon the
applied pressure, particle size, size distribution, particle shape, and size of the compact.
There are several theoretical equations relating green density and applied pressure, but in
practice empirical relations are used.
The relationship between the average particle size and the green density as well as the
compressive strength of compacts is shown in Fig. 5.31 for a constant pressure. Although
the green density increases as average particle size varies from I to 9 Jlm, the compressive
strength exhibits a maximum between 3 and 6 Jlm. This maximum corresponds to the
preferred particle size range for most tungsten compacts.

5.5.1.2. Cold Isostatic Pressing. In isostatic pressing, the powder is filled into
flexible molds made of rubber or elastomers and subjected to hydrostatic pressure. The
pressure is commonly in the range of 200 to 400 MPa. As a result of the uniform pressure,
a much higher uniformity in density is achieved. Isostatic pressing has gained much

Green density

co

64

>'iii
c:

CD
"C

a..
~

40

62

CD
c:
....CD

60

iii
30

c:

CD
CD
C)

.... 58

CD

>
'iii
C/)

CD
....
a.

Compressive strength

56
54

20 U
1

7
8
4
5
6
3
Average particle size, FSSS (Il m)

FIGURE 5.31. Relationship between average grain size and green density as well as compressive strength for a
constant compacting pressure (241 MPa) [5.1].

INDUSTRIAL PRODUCTION

233

importance during the last 30 years, because it offers several further important advantages
compared to rigid die pressing:

Lower pressure required for a certain green density.

Higher strength of the compacts.
More free choice in dimension (ratio of diameter to lengths).
Parts with undercuts and reentrant angles can be pressed.
Thin-walled tubes can be produced.
Very large parts can be compacted.

Less precise dimensional control and a much lower rate of production are the two main
disadvantages.
Isostatic pressing is carried out by two different techniques: wetbag pressing or
drybag pressing. In wetbag pressing, the powder is filled into flexible molds which are
sealed outside the pressure vessel. Several molds (either of the same shape or of different
shape) are then immersed in a fluid, most commonly water, and pressure is applied
isostatically. Wetbag pressing is the most common technique for producing forging or
rolling preforms, where an even and high green density is more important than dimensional control. Nevertheless, wetbag pressing is also used for more complex geometries
and even near-net shape parts. The size of the part is limited by the size of the pressure
vessel. Tungsten ingots of up to 1000 kg are produced via wetbag pressing.
In drybag pressing, the e1astomeric mold is fixed into the pressure vessel. The mold is
filled with powder and sealed with a cover plate. Then pressure is applied between the
mold and the vessel wall. After pressure release, the cover is removed and the part
removed. Then the procedure starts again. Pressure vessels with both top and bottom plate
are also in use, and allow more rapid removal of the compact.
Drybag pressing is used for simple shapes, such as plugs, and high production rates
(mass production). However, only one compact (with one specific shape) can be pressed at
a time.
5.5.2. Sintering [5.78-5.84]
5.5.2.1. General. In order to increase the strength of the green compacts, they are
subjected to heat treatment, which is called sintering. The main aim of sintering is
densification in order to provide the metal with the necessary physical and mechanical
properties and a density which is suitable for subsequent thermomechanical processing.
Sintering of tungsten is commonly carried out in a temperature range of 2000 up to
3050 DC under flowing hydrogen, either by direct sintering (self-resistance heating) or
indirect sintering (resistance element heating systems). The density thereby obtained
should be a minimum of 90% of the theoretical density, but is commonly in the range
between 92 to 98%.
The main driving force for sintering is the lowering of free energy, which takes place
when individual particles grow together, pores shrink, and the high surface area of the
compact (i.e., its high excess surface energy) decreases. The decrease in surface area is
accomplished by diffusional flow of matter into the pore volume under the action of
capillary forces (surface tension forces). Besides shrinkage, recovery (change of subgrain
structures and strain relief), recrystallization (formation of strain free crystals low in

234

CHAPTER 5

dislocation density), and grain growth occur during sintering, also contributing to the
minimization of free energy.
Sintering is commonly regarded as taking place in three stages [5.79,5.83]:
During the early stage, necks are formed between individual particles and grow by
diffusion, increasing the interparticle contact area. The powder aggregate shrinks,
involving center to center approach of the particles. In this stage, the degree of
densification is still low and the pore structure is open and fully connected.
With increasing neck formation (intermediate stage), the necks become blunted
and lose their identity. The pores are assumed to be cylindrical. Their radii vary
along their lengths and, with increasing shrinkage, the pore channels break up into
small, still partly interconnected segments. During this stage (channel closure
stage), pronounced densification occurs and significant grain growth occurs
concurrent with shrinkage.
Finally, in the last stage (isolated pore stage), the pore segments further break up
into chains of discrete, isolated pores of more or less spherical symmetry. This
stage occurs when about 90% of the theoretical density is achieved. The sintering
density then approaches asymptotically the practical limit of 92-98%.
Investigations have shown that the densification is controlled by grain boundary
diffusion over most of the densification range, unless at very high densities it becomes
controlled by lattice diffusion [5.83].
Since the motion of grain boundaries, necessary for grain growth, is impeded by the
presence of pores, grain coarsening proceeds at a higher rate above 97% density. Grain
sizes of the as-sintered ingots are commonly in the range of 10 to 30 11m.
Besides temperature and time, several other parameters influence densification, such
as powder particle size, green density, sintering atmosphere, powder purity, compact
size/weight, heating rate, thermal gradients, and the presence of insoluble phases such as
oxides (Th02, La203, ee02, zr02) or metallic potassium (NS-tungsten).
Temperature is the most important parameter affecting densification. Higher sintering
densities can be obtained much more rapidly by increasing the sintering temperature than
by prolonging the sintering time. Below 1900 0 e little densification occurs, unless very
long sintering times are applied. For example, more than 50 h are required to obtain 92%
density for a tungsten powder of 4 11m particle size at 1800 e [5.83]. At 2400 e, the time
necessary to obtain high densities decreases to 1-2 hours and, at 3000 o e, 20-30 minutes.
The finer the starting tungsten powder, the more rapid the densification at a given
temperature [5.1].
The influence of temperature and time on densification can be estimated by using socalled density diagrams (sinteringmaps) [5.85], which are based on approximate sintering
models. Nevertheless, empirical rate equations are used for industrial purposes to calculate
the necessary sintering times at different temperatures.
Tungsten sintering, in practice, is always performed in reducing atmosphere which
removes the oxygen coating of the powder particle surfaces. High-purity dry hydrogen is
commonly used. Under vacuum or in inert gas, sintering is retained by residual oxygen,
and the desired density will not be achieved.
The high temperatures used for sintering also lead to a significant purification of the
metal by volatilization of impurities. Interstitials, such as oxygen and carbon, cannot be
removed completely and have to be held at a low level; otherwise, they significantly affect

INDUSTRIAL PRODUCTION

235

the workability of the parts. Metallic elements, such as Fe, Ni, Cr, Nb, etc., are evaporized
but remain partly in the ingot, forming solid solutions. If Fe or Ni is present in the form of
large, heterogeneous contaminations (such as iron particles stemming from reduction
boats), these can lead to local melt formation and subsequent formation of large voids,
which do not close during sintering. Elements which have practically no solubility in
tungsten, such as alkali or earth-alkali metals, are volatilized. Grain boundary diffusion
and carrier distillation effects most likely play an important role in "cleaning" the grain
boundaries.
Since the ductility of tungsten is very sensitive to most of the impurities, purification
is important. Therefore, special care must be taken so that, during sintering, evaporation
can take place to the desired extent (i.e., as long as there is an open porosity). If the ingot
densifies too quickly, impurities can be trapped. Due to the higher sintering temperature,
direct sintering is more effective in cleaning than indirect sintering.
The sintering of tungsten can be enhanced by the addition of small amounts of
alloying elements (0.5-1 %), such as Ni or Pd. This phenomenon is called "activated
sintering." It is explained by the enhancement of grain-boundary diffusion in tungsten due
to the presence of the respective element in the grain boundary. High densities (up to 99%)
can be obtained even at 11 00 0 C (at this temperature, tungsten compacts are commonly
presintered). However, since the alloys so produced are rather brittle, activated sintering
has never been used in industry.
The sintering of doped tungsten is a peculiar case in sintering of tungsten. This
includes dispersion-strengthened materials such as thoriated tungsten or tungsten with
additions of Ce02, La203, and zr0 2 as well as NS (non-sag) tungsten used for lamp
filaments.
The oxide dispersoids are stable at the high sintering temperatures and do not dissolve
in the tungsten matrix. They pin the grain boundaries during the later stages of sintering,
and in that way they significantly restrict grain coarsening. So, for example, a rod
containing 0.75% of thoria has a grain size of 5000 to 10,000 grains per square millimeter,
as compared to 1500 grains per square millimeter for a similar rod of pure tungsten [5.1].
NS-doped tungsten powder contains small inclusions of potassium aluminosilicates,
which were incorporated during the reduction process. During sintering, the silicates
dissociate thermally and submicron potassium bubbles form in the tungsten ingot. Similar
to the oxides, these bubbles pin the grain boundaries and inhibit grain coarsening during
sintering. Since potassium is gaseous during sintering, the bubbles are under high pressure,
which is balanced by the surface tension of the pore. They can be seen as small pores in the
fracture surfaces of NS-doped tungsten besides the significantly coarser residual sintering
pores, as characteristic for undoped tungsten (see also Section 6.2.1) [5.86]. They
constitute the starting point for subsequent formation of rows of bubbles during thermomechanical processing.
Up to the sixties, the only sintering method used in practice was direct sintering.
Although still in use for the production of doped tungsten, it has lost its importance. From
then on, mainly because of the increasing demand for larger parts and the higher capacity
of the aggregates, indirect sintering furnaces were developed. This technique is used
nowadays as the main route for producing pure tungsten [5.80].

5.5.2.2. Direct Sintering [5.81,5.87]. A schematic drawing of a direct sintering

furnace is given in Fig. 5.32. The tungsten bars are commonly mounted in pairs or several

236
Water outlet

t I

CHAPTER 5
Atmosphere outlet

Water-cooled shell

4+-++-......,10"}-- Upper bar holder

(movable tungsten clip)

1 . - - - - 6 t - - Pressed bar
Lower bar holder
(stationary tungsten clip)

~ Water inlet

Water outlet

tt

Water inlet

Atmosphere inlet

To power supply

FIGURE 5.32. Schematic drawing of a direct sintering aggregate. By courtesy of Metals Handbook, ASM
International, Ohio.

units in a vertical, free-hanging position in a water-cooled chamber. They are clamped at

the top using water cooled copper contacts with tungsten inserts. The bottom clip is made
of tungsten and must be able to move as the bar shrinks. An electric current of several
thousand ampere is passed through it in order to heat it up to about 3000C by Joule's
energy. Sintering is conducted under dry hydrogen. Several such sintering aggregates are
commonly connected in series so that several bars can be sintered simultaneously (Fig.
5.33).
In order to achieve an appreciable strength of the green compact, which is necessary
for handling, it must be presintered. Presintering is done by heating to 1100-1300C in
muffle furnaces under hydrogen atmosphere. By reducing the oxide film covering the
powder particles, the cohesive strength between them is increased.
The electric current is used to control the sintering progress and as an indicator of the
prevailing sintering temperature. The current is kept low at the beginning due to the higher
resistance of the bar (less contact between the grains). It is then gradually raised up to
2700-3000C. Proper hold points are important to allow the outgassing of impurities.
This is particularly important for the sintering of NS-tungsten, because both Al and Si
evaporate rapidly above 2150 C [5.82]. If densification occurs too rapidly, excess dopants
will be trapped and may cause difficulties during subsequent working or even bursting of
the bar during sintering. Commonly, less than 10 ppm Si and Al but 60 to 100 ppm K
remain in the ingot, the latter in the form of desired potassium bubbles.
Depending on the bar size, the heating up time is between 20 and 60 minutes and
the holding time between 30 and 60 minutes. Typical bar dimensions are 15-25 mm x

INDUSTRIAL PRODUCTION

237

FIGURE 5.33. Modem industrial aggregate for direct sintering of tungsten. By courtesy of OSRAM GmbH,
Germany.

15-25 mm X 600-920 mm. Typical weights are between 1.5 and 6 kg. The density of
sintered bars is between 88 and 96% of the theoretical value. It is always lower for NStungsten (88 to 92%) than for pure tungsten, due to the presence of trapped potassium.
Advantages compared to indirect sintering are short sintering times; temperature
inside the bar is higher than on the surface, which favors the diffusion and evaporation of
impurities; higher temperature results in a better cleaning effect; a simpler aggregate with
comparatively low maintenance costs.

238

CHAPTER 5

Disadvantages are the dimensional restrictions (only simple shapes can be sintered);
low capacity; the loss of the "heads" (clipped parts) which did not obtain the temperature
due to the heat sink effect of the contacts.

5.5.2.3. Indirect Sintering. A schematic view of a furnace with resistance-heating

elements is presented in Fig. 5.34. A modem industrial aggregate is shown in Fig. 5.35.
The green compacts (no presintering necessary) are placed inside a cylindrical or basketlike heating element of the furnace (constructed of Mo or preferably W). In a radial
direction to the outside, the furnace is adapted with radiation shields (inner shields made of
W, outer sheets made of Mo), which protect the furnace wall and concentrate the heat to

10

10

FIGURE 5.34. Schematic view of an indirect resistance-heated sintering furnace [5.1]: (I) furnace assembly with
dome, (2) bottom heat shield, (3) top heat shield, (5) heating elements, (6) element support posts, (7) element
clamping posts, (8) heat shields, (9) pyrometer window, (10) cooling inlet, (II) cooling outlet.

INDUSTRIAL PRODUCTION

239

FIGURE 5.35. Modern industrial aggregate for indirect sintering. By courtesy of Plansee AG, Reutte, Austria.

the center. A vacuum system is necessary to empty the furnace prior to hydrogen flooding.
Maximal dimension depends on furnace size. Compacts of any desired shape can be
sintered. In order to achieve even shrinkage, the compacts have to be placed on green
tungsten shims [5.1]. A slow heating rate is essential; otherwise, surface densification will
occur too early, not allowing the outgassing of the interior. Internal stress will be built up,
resulting in cracks. This is particularly important for large parts, such as tungsten billets for
forging or rolling. Holding times are essential. In practice, the sintering schedule is
adjusted to the specific requirements. Proper furnace loading is also important, since the
shielding of parts by others can lead to different densification rates and less uniformity
[5.84].
Common sintering temperatures are between 2000 and 2700C (max). Sintering
times of 8 to 24 hours, or even more, are common. Furnaces having uniform hot zones of
up to 1.2 m and 0.14 m2 are available. Weights in the range of a thousand kg can be
sintered in one batch. For very high loads, deformation of the part can take place due to
gravity.
Advantages are: no loss of material, no dimensional restrictions, and high capacity.
Disadvantages are: longer heating times, less purification, lower efficiency of heating, and
higher maintenance costs.
5.5.3. Fabrication of Tungsten
5.5.3.1. Fabrication of Wrought PIM Tungsten [5.88-5.90]. With only some exceptions, tungsten is used in the form of pore-free preforms ("wrought" P1M tungsten). To

240

CHAPTER 5

obtain a completely dense material, as well as the desired shape and mechnical properties,
a complex, multistage, hot and cold forming process is required. * The most important
forming techniques for tungsten are rolling (for rods and sheet products), round forging
(for large diameter parts), swaging (for rods),forging (for large parts), and drawing (for
wires and tubes). Secondary forming processes include flat rolling of wires, flow turning,
spinning, deep drawing and wire coiling. For a detailed description of the most important
forming processes, we refer the reader to the book, Tungsten, by Yih and Wang [5.1].
In general, plastic forming of tungsten is difficult and needs experience. In the assintered condition, tungsten is brittle except at quite high temperatures, because it is
recrystallized (coarse grained) and not fully dense. Unlike most metals, the low-temperature ductility of tungsten increases with progressive deformation, because embrittlement is
due to grain boundary segregations of interstitial soluble elements, such as oxygen,
carbon, and nitrogen. With the breakdown of the coarse microstructure during deformation, these impurities are distributed over a larger intergranular area, which makes the
material more ductile and less sensitive to cracking during forming at lower temperatures.
The aim of primary working (ingot breakdown) is threefold: (1) to eliminate the
residual porosity, (2) to convert the original massive form (sinter ingot) to the desired
preshape (sheet, rod, tube), and (3) to refine the grain size in order to improve the
formability during subsequent cold forming.
The first forming step is usually carried out at 1500-1700C. As a result of the low
specific heat and high thermal conductivity of tungsten, several reheating stages are
necessary in the first stages of shaping, because the heat is lost rapidly at these
temperatures and the ingot cools down rapidly. If the metal is worked at too low a
temperature, cracks and splits will easily develop. As the forming process continues, the
forming temperature is reduced progressively since the recrystallization temperature
decreases as deformation proceeds. Tungsten is generally worked below its recrystallization temperature, because recrystallization is combined with grain boundary embrittlement. With increasing work hardening during deformation, both the hardness and strength
of the worked part increase significantly, and intermediate stress relief annealings are
necessary to minimize the hazards of cracking (laminating) and to avoid overstraining the
working tool.
With a high degree of deformation, the grains become more fibrous due to the
unidirectional nature of the processing (sheet rolling, rod extrusion, or swaging) and
mechanical anisotropy develops. Strong crystallographic orientations and planar distributions of dispersed, insoluble phases (in the case of doped tungsten) are characteristic of the
stage of heavily worked tungsten. Figure 5.36 shows the wrought structure ofa cold-rolled
tungsten sheet. Table 5.13 gives an example of a rolling schedule for tungsten sheet.
The grain size and grain structure of the final product have a great influence on the
mechanical properties. They can be controlled by the type of deformation, degree of
deformation, and annealing processes both during and after machining.
Forming of tungsten is commonly carried out without protecting atmosphere. In air,
tungsten is readily oxidized. Tungsten trioxide forms on the surfaces of the worked piece

* In a metallurgical sense, hot-working is considered as working above the recrystallization temperature. This is
not the common case for tungsten, where working is usually performed below the recrystallization temperature.
Thus, from this point of view, tungsten is commonly cold-worked, although very high temperatures are applied
for deformation.

241

INDUSTRIAL PRODUCTION

FIGURE 5.36. Typical wrought microstructure of rolled tungsten sheet. By courtesy of Plansee AG, Reutte,
Austria.

TABLE 5.13. Example of Rolling Schedules for Tungsten Sheet [5.lt

Pass number

2
3 (Cross roll)
4

Thickness after pass (mm)

Preheating temperature ("C)

Starting thickness 24.4 mm

20.3
15.9
12.7
10.2

1450
1400
1400
1400

Anneal Smin at 1300-1350C

5
8.26
1350
6
6.86
1300
7 (Cross roll)
5.46
1300
8
4.4S
1300
Anneal 5 min at 1250-1300C; caustic pickling and condition

9
10
II
12
13

3.S1
2.82
2.26
1.80
1.63

12S0
1200
1200
IISO
1150

Anneal 5 min at 1150C; caustic pickling and acid etching

a

Source: G. C. Bodine jr. (1963), Tungsten Sheet Rolling Program, Final Report, Fansteel
Metallurgical Corp.,North Chicago, Illinois.

242

CHAPTER 5

and, above 800C, it volatilizes. The oxide layer acts as a protective layer against
contamination from the working tools and is removed at certain stages of deformation
by pickling and/or machining. Intermediate annealing and stress-relieving annealing is
performed under hydrogen to avoid enhanced oxidation of the metal and sublimation of
W0 3
5.5.3.2. Shaping [5.88-5.90]. Compared to ductile metals and alloys, the fabricability of tungsten is rather poor. Tungsten should always be heated before shaping. The
temperature range for forming has a lower limit, set by the brittle-to-ductile transformation
temperature, and an upper limit, set by the recrystallization temperature. This temperature
is mainly dependent on the purity, the history of deformation, and heat treatment of the
material. As an example, Fig. 5.37 shows the recommended temperature ranges for
bending and punching of tungsten sheet as a function of sheet thickness [5.89]. Highly
deformed products, such as thin tungsten wires, ribbons, or foils, are ductile at room
temperature.
Thin, strongly deformed sheet and foil have a pronounced structure in the longitudinal direction due to elongation of the grains during rolling. The bending properties
along the direction of rolling are therefore different from those across it. Therefore,
tung~ten sheet should always be bent in a way such that the bending edge is perpendicular
to the rolling direction. Ifbending in the longitudinal direction cannot be avoided, owing to

1600 r---------------------~
total recrystallization
after 1 hour
1400
1200

IL---------range of recrystallization

1----------------\ incipient recrystal- ~..rT777

lization

G 1000
~

...
iii
...
(J)

:::J

800

(J)

c.
E

{E. 600
400

200

transition temperature
L--....l---'---i'---'---J...--..I._~..J

1 2 3 456 7
Sheet thickness, d (mm)

FIGURE 5.37. Transition temperature, recrystallization temperature, and recommended temperature range for
bending and stamping of tungsten sheet as a function of thickness [5.89].

INDUSTRIAL PRODUCTION

243

the design, much higher bending temperatures are required. At high temperatures, tungsten
sheet can be stamped, punched, and sheared. Sharp tools are essential to clean cutting
action without sheet cracking or delamination. Tungsten cylinders and cones can be
formed by spinning, flow turning, or forging. The use of stress-relieved tungsten is
suggested for optimum fabrication results.
Tungsten can be drilled, turned, milled, planed and ground. However, machining
operations require experience and close adherence to optimum conditions. Preheating to
about 200C and the use of cemented carbide tools are recommended. Complex shapes
and holes can be formed by spark erosion; the tungsten workpiece forms the anode and the
working electrode the cathode.
Tungsten (pure W, doped W, W-Re) is commercially available as wrought products
(sintered bars, billets, forgings, rods, pins, disks, cylinders, tubes, sheet, strips, ribbons,
wires) as well as in the form of a broad variety of shaped parts (crucibles, springs, heating
elements, rivets, etc.).
Cold-rolled sheet is supplied in thicknesses from 0.025 to 0.4 mm (up to
500mm x 1000mm). Hot-rolled tungsten is available with thicknesses from 0.4 to
8.0mm (maximum lengths, 1500mm). Tungsten wires are manufactured from about
1 mm down to 5 ~m.
5.5.3.3. Bonding of Tungsten [5.89].
Mechanical Joining. Mechanical joining, such as riveting, tacking, or lacing,
comprises the simplest methods of joining tungsten, provided the joint need not be
impermeable to liquids and/or gases. Tungsten and molybdenum parts can be used.
Mechanical fastening is used for constructional parts, such as heating elements, containers,
large shields, etc.
Brazing. The parts to be joined must be free from grease, oils, oxides, or other
impurities. As tungsten is very sensitive to oxidation, brazing must be carried out under a
protective gas, hydrogen, or in vacuum. If maximum strength of the joint is required,
tungsten has to be brazed below its recrystallization temperature. Typical brazing metals
and temperatures are: Rh (1970C), Ni (l430C), or AgCu20Pd15 (700-900C). More
detailed information is available [5.78,5.89]. Brazing of tungsten to ceramics, graphite,
and silicon has gained importance for the fabrication of refractory-metal composites.
Welding. Tungsten possesses only moderate welding properties. Welding must be
carried out under controlled weld atmospheres, preferably in a dry box, since any
contamination by oxygen will reduce the ductility of the joint. Before welding, the
metal has to be degreased and pickled, commonly with a mixture of nitric acid and
hydrofluoric acid (90/10 vol%). The weld seams have a coarse-grained structure in the
hot-fusion zone, owing to recrystallization, and can therefore withstand only low
mechanical stresses. Tungsten can be welded by tungsten inert gas (TIG) welding, laser
beam welding (for thin parts), plasma welding, friction welding and electron beam
welding, the latter being the preferred fusion-welding method. Tungsten can be welded
to W-Re and Mo-Re alloys. W-Re alloys (in particular W-26Re) are recommended as
filler metal for welding tungsten. Tungsten can be diffusion welded to tungsten and other
metals. Despite the high melting point of tungsten, temperatures of 1300-2000C and
pressures of2-20N . mm- 2 give satisfactory joints. Intermediate layers ofNi, Pt, Rh, Ru,
and Pd accelerate the diffusion process. Tungsten wires can be spot-welded to other metals
under a protective gas blanket.

244

CHAPTER 5

5.6. ALTERNATIVE PROCESSES [5.1,5.2]

The following methods have been applied to tungsten production: arc melting,
electron beam melting, hot pressing, isostatic hot pressing, explosion compacting, slip
casting, plasma spraying, chemical vapor deposition, powder rolling, and injection
molding. Below, we restrict ourselves to the methods of technical importance.

5.6.1. Electron-Beam Zone Melting

Electron-beam zone melting is the preferred method to produce larger, single crystals
of tungsten. A polycrystalline tungsten rod is mounted in a vertical position in a stainless
steel bell jar and connected as anode. An electron gun (cathode) traverses along the rod by
means of a screw drive mechanism. On melting in a high vacuum, a narrow, liquid zone is
formed and held between the two solid rod parts by surface tension forces. As the zone
slowly moves through the rod several times (zone floating), impurities are concentrated on
both the starting and terminal end of the rod. Since the molten zone is quite narrow and the
temperature gradient in its vicinity rather steep, nucleation starts essentially at one point
and a single crystal is the result. The concentration of many impurity elements can be
decreased considerably by zone melting. AI, Ca, Co, Cu, Fe, Mg, Mn, Ni, Si, Na, Ti, Ta,
Re, Mo, and Nb concentrations can be reduced below Illg/ g. Interstitials remain at the
following Ilg/ g levels: 0 2 2, N 2 5, C 10, and H 1.
Single crystals produced by this technique represent the most ductile tungsten
currently available. They are free of grain boundaries, which cause brittleness.

5.6.2. Plasma Spraying

Tungsten layers can be sprayed with suitable guns on different base materials. Easily
flowing tungsten powder is used as spraying material. Ar or Ar-N2 mixtures are preferred
as protective atmosphere in closed chambers. Besides coatings, tungsten parts can be
produced by spraying tungsten layers on an adequately formed nucleus until the desired
thickness is reached.
One may distinguish between three basic types of substrates:

expendable substrates are removed after spraying by selective chemical etching,

melting, etc;
reusable substrates are used for mass production;
permanent substrates become an integral part of the article.
Density ranges from 50-95% of the theoretical value. Higher density is attained by
uniform substrate temperature and low deposition rate. Density and strength might be
increased by post-sintering in hydrogen. Low-pressure plasma spraying is mainly used for
producing intricate shapes, which otherwise cannot be manufactured.

5.6.3. Chemical Vapor Deposition

For the deposition of tungsten layers from the vapor phase, an easily evaporable
tungsten compound is mixed with a carrier gas and/or a reducing gas and decomposed or

INDUSTRIAL PRODUCTION

245

reduced at a higher temperature on a suitable substrate. Possible tungsten compounds are

halides (preferably WF6 and WCI6), tungsten hexacarbonyl [W(CO)6], and organometallics. Until now, only the halides are technically important. Extensive investigations in the
direction of organometallics have been undertaken in order to overcome the disadvantages
of the halides: the high-temperature necessary for reduction and the difficulties caused by
corrosion. For more details see Sections 3.5.2 and 3.5.3.
Sheets, produced via CVD, exhibit a quite different microstructure as compared to
wrought P1M products in terms of grain size and grain orientation. Both parameters can be
influenced by the deposition conditions. Very fine grained CVD tungsten can be obtained
by changing the growth conditions during deposition, for example, by brushing, which
increases the nucleation rate and disrupts the columnar grain growth. The method also
offers the possibility to deposit layers of desired crystal orientation or even single crystal
layers.
CVD of tungsten and tungsten-rhenium (co-deposition of WF6 and ReF 6) is of
commercial importance for the production of X-ray targets. W-Re layers of up to 1 mm are
deposited onto graphite disks with diameters between 100 and 150mm [5.78]. Furthermore, CVD can be used to produce cylindrical or conical shells, cones (such as shaped
charge liners), crucibles, and thin-walled tubes.
Low-pressure CVD of tungsten is tested for via plugs, low-resistance contact barriers,
and interconnects for high density silicon integrated circuits. Depending on the substrate
and the deposition conditions, either \1.- or ~-tungsten, or a mixture of both, forms. This is
of interest, because the resistivity of ~- W is about 10 times higher than that of \1.- W.

5.6.4. Physical Vapor Deposition (Sputtering)

Difficulties combined with the elevated temperature in case of CVD can be
eliminated by PVD (sputtering). The substrate in this process remains fairly cold.
An ion beam (preferably noble gas) is generated in a high vacuum chamber by
discharge. It is focused on the water-cooled tungsten sputter target where, by its impact
energy, it frees tungsten particles which are deposited on the substrate.
Improved design of sputtering equipment somewhat allows the substitution of CVD
tungsten. Sputter targets used for thin layers in microelectronics manufacture are made of
high or ultrahigh purity W, W-lO%Ti, and WSix (see Section 5.7.7).

5.7. SPECIAL TUNGSTEN FORMS AND QUALITIES [5.1,5.2,5.90,5.91-5.95]

5.7.1. Globular Tungsten Powder

This substance can be produced from fine powder in a plasma jet (150-200 !lm).
Monodisperse single crystal powder (containing small amounts of Ni) can be obtained by
acid leaching of sintered W-Ni-Cu parts.

5.7.2. Coarse Tungsten Powder

Good flowing tungsten powder can be obtained by:

246

CHAPTER 5

reduction of W0 3 or blue oxide at high temperature, high humidity, and alkali

metal doping, and subsequent classification;
disintegration and classification of presintered parts;
fine tungsten powder +0.2% Ni is compacted and sintered at 1840C for 3 hours.
Afterward, the Ni is dissolved by nitric acid.

5.7.3. Single Crystals

These crystals can be produced by:
Annealing of deformed wires or rods by secondary recrystallization and extreme
crystal growth; thoriated tungsten wires (cross-section, e.g., 100 l.lm) are pulled
through a narrow zone at a temperature of above 2000C at the same time the
crystal grows, forming a single crystal up to several meters long (pintsch procedure
[5.96].
Zone melting.
Single crystals, due to their high cost are used only in special cases like basic
investigations. The outstanding properties are a lack of grain boundaries and a very small
number of lattice defects. They are therefore easy to deform, show great ductility, and
possess no embrittlement up to the highest temperature.

5.7.4. Tungsten Thin Films

The preparation of tungsten thin films is possible by different techniques, such as
electrolysis from salt melts, CVD, PVD and plasma spraying, diffusion, plating, electrophoresis, and metallizing. The technically important techniques are only CVD, PVD, and
plasma spraying.

5.7.5. Porous Tungsten

The production of structural parts from porous tungsten affords special technological
procedures. Starting materials are mainly coarse or globular powders, which are
compressed by special methods, sometimes with the addition of pore-forming materials
which evaporate during sintering. Also in sintering, special procedures such as infiltration,
evaporation, surface treatment, etc. are necessary. Powder grain size, compacting pressure,
sintering temperature, and time determine pore size and permeability. At densities between
50 and 78% all pores are open; at >90% density, all are closed.
Parts with an open coherent porosity of > 20% are used in vacuum tubes and
aeronautics because of refractoriness, shape stability, and chemical resistance. Examples
are porous cathodes impregnated with alkaline earth oxides as electron-emitting sources in
special tubes, thermoionic converters heated by nuclear energy, and ionic propulsion units
with porous tungsten plates as ionic sources for Cs vapor of high temperature. Those
emitter plates are made of globular tungsten powder (7 Jlm) density 75-85%, pore size
2-30 Jlm, and pore number 1.4 x 106-8 x 106 . cm- 3 .

247

INDUSTRIAL PRODUCTION

Porous tungsten can be infiltrated either to machine it to rigid dimensional tolerances,

or to fabricate composite materials. Copper, which is used as infiltrant for machining, can
be completely removed afterward by heating in vacuum by evaporation.

5.7.6. High-Purity Tungsten

High-purity or Ultrahigh-purity tungsten (5N or 6N) is used in the form of thin films
in the electronic industry as the preferred material for gate electrodes, interconnects, and
barrier metal because of its high resistance to electron migration, high temperature
stability, and tendency to form stable silicides. In principle, tungsten thin films can be
produced either by CVD or by PVD. In regard to CVD tungsten thin films, we also refer
the reader to Sections 3.5.2 and 3.5.3 as well as to Section 5.6.4. For PVD-produced
tungsten thin films, sputter targets of high purity (6N) are commercially available today.
It can be seen in Fig. 5.38 that the concentration of trace impurities in sputter targets
has shown a downward trend since 1988. This is partly due to improvements in production
methods, which in fact caused higher purity. But it is also somewhat a consequence of
increased sensitivity of the analytical methods and the equipment used nowadays. For
example, some years ago it was only possible to say that the concentration is less than
1 ~gjg or less than 0.1 ~gjg. Meanwhile, application of more advanced techniques has
changed the situation drastically. In high-purity tungsten, metal levels of trace elements are
in the ngj g level or even below, which affords special analytical requirements such as
working in clean room laboratories, prevention of any other contamination, and correct
calibration. The most important analytical methods applied today for that purpose are glow
discharge mass spectrometry, secondary ion mass spectrometry, isotope dilution, and trace
matrix separation combined with inductively coupled plasma atomic emission spectrometry.
As an example, it was possible by isotope dilution analysis to determine 15 ppq of
uranium in an Ultrapure tungsten sample. This corresponds to 15 x 10- 15 %. These
10

E
a.
a.

0.1

0)

0.01

0.001

.:.:.
.....
0)

Fe. Cr. Ni

0.0001
0.00001

~--~--~----~--~----~--~

1988 1989 1990 1991

1992 1993 1994

FIGURE 5.38. Ultrahigh-purity tungsten: improvement in purity from 1988 to 1994 (target data); production
takes place under clean room conditions [5.96].

248

CHAPTER 5

Commercial APT

!
APT (high in u/Th)

l
APT (low in u/Th)

!
Double or multiple
crystallization

..
AMT

UHPAPT

UHPWPowder

Pressing, sintering

Mixing with Ti or Si

Compacting, sintering
hot rolling

WRod

Hot pressing,
machining

UHP sputter target

..

Electron beam
floating zone refining

TiWorWSix
sputter target

UHPWrod
FIGURE 5.39. Production methods for high-purity tungsten.

249

INDUSTRIAL PRODUCTION

methods are generally very complex and expensive. Comparative analyses often show
significant scatter.
Of all impurity elements, in particular uranium and thorium contents must be
extremely low. These naturally radioactive elements cause "soft errors" in memory
circuits due to a-particle emission.
The diverse production methods for high-purity tungsten are summarized in Fig. 5.39.
The raw material is always APT with different purity depending on the uranium and
thorium content. It depends on the ore deposit whether these two elements are present or
not. If the concentrations are unacceptably high after dissolution, a selective solvent
extraction process removes the main portion of these two elements.
The APT is calcined to AMT at 220C and dissolved in water while adding hydrogen
peroxide and sulfuric acid. The traces of uranium and thorium are extracted selectively by
0.5 M di-2-ethylhexyl phosphoric acid and 0.l25 M tri-n-octyl-phosphine oxide dissolved
in kerosene.
APT, having acceptably low uranium/thorium concentrations, can be further purified
by double or multiple crystallization to remove other impurities.
The W0 3 gained after calcination in both cases is reduced by hydrogen to tungsten
powder of high purity. Typical analyses of Wand WSix powder are shown in Table 5.l4.
This c"an be used for the production of tungsten, tungsten silicide, or TiW sputtering
targets.
If further purification is desired, rods are produced by compacting and sintering and
are subjected to electron beam floating zone melting. This procedure does not further
reduce the concentrations of uranium and thorium, but does reduce the concentrations of
many other elements as shown in Table 5.15. Moreover, residual impurity elements are
distributed homogeneously.
Powders having sufficient purity can be processed to sputter targets directly. Tungsten
sputter targets have usually been manufactured by hot pressing. However, it was not
possible to maintain the high-purity level of the powder. A much better way to achieve
TABLE 5.14. Purity of Ultrahigh-Purity Grade Tungsten and
Tungsten Slicide Powder [5.78]
Trace impurity
Al
B
Na
K
Ca
Mg
Fe
Co
Cr
Ni
Cu
U
Th

Tungsten powder
Ilg/g)

Tungsten silicide powder

Ilg/g)

0.20
0.05
0.05
0.05
0.10
0.10
0.25
0.02
0.25
0.15
0.02
0.0003
0.0003

0.50
0.50
0.10
0.50
0.50
0.10
1.0
0.05
1.0
0.30
0.20
0.0003
0.0003

250

CHAPTER 5

TABLE 5.15. Impurity change from APT via Tungsten Powder

and a Tungsten Rod after Electron Beam Zone Refining [5.77]
Element

APT

Fe
Ni
Al
Na
Mo
K
Cr
Si
Mn
U
Th

6.0

1.3
3.7
7.7
15.8
3.5
2.8
7.1
3.2
0.062
0.140

W powder
concentrations (I-lg . g -1)
8.0
2.0
9.0
7.0
9.0
3.2
3.9
11.0
4.1
0.0011
0.01

Refined rod
2.9
< 1.0
< 1.0

0.013
< 1.0

0.014
0.34
< 1.0
< 1.0
0.0011
O.oJ

high purity in the sputter target is to employ a special sintering process which reduces the
contents of gaseous impurities and provides a homogeneous distribution of all remaining
impurities.
In general, it is important to maintain clean room conditions throughout the
production steps.

REFERENCES FOR CHAPTER 5

5.1.
5.2.
5.3.
5.4.
5.5.
5.6.

5.7.
5.8.
5.9.
5.10.
5.11.
5.12.
5.13.
5.14.
5.15.
5.16.

St. W. H. Yih and C. T. Wang, Tungsten, Plenum Press, New York (1979).
Gmelin Handbuch der anorganischen Chemie, 8th ed., Syst. No. 54, Tungsten, Suppl. Vol. Al (1979).
R. G. Woolery, in: Proceedings of the 2nd Tungsten Symposium, San Francisco, pp. 52--63, Mining
Journals Books Ltd., London (1982).
Wu Weisun, in: Proc. 2nd Tungsten Symposium, San Francisco, pp. 64-70, Mining Journals Books Ltd.,
London (1982).
G.1. Willey, in: Proc. 2nd Tungsten Symposium, San Francisco, pp. 21-28, Mining Journals Books Ltd.,
London, (1982).
E. Lassner, "From Tungsten Concentrates and Scrap to Highly Pure APT," in: The Chemistry of Non Sag
Tungsten (L. Bartha, E. Lassner, W. D. Schubert, and B. (Lux, eds.), pp. 35-44, Pergamon Press, Oxford
(1995).
E. Lassner and B. Kieffer, Conference Proceedings on Advances in Hard Materials Production, London,
MPR Publishing Services Ltd., Shrewsbury (1986).
R. F. Hogsett, D. K. Huggins, and L. B. Beckstead, US Patent 4.338,287 (1982).
E. Lassner, in: Extractive Metallurgy of Refractory Metals (H. Y. Sohn, O. N. Carlson, and 1. T. Smith,
Eds.), Proc. TMS-AIME llOth Annual Meeting, pp. 269-272, Chicago (1981).
L. R. Quatrini, M. B. Terlizzi, and B. E. Martin, US Patent 4.353,878 (1982).
L. R. Quatrini and B. E. Martin, US Patent 4.353,789 (1982).
L. R. Quatrini, M. C. Vogt, and B. E. Martin, US Patent 4.353,881 (1982).
L. R. Quatrini, US Patent 4.353,880 (1982).
1. Zhou and 1. Xue, in: Proc. 5th Int. Tungsten Symp. Budapest, pp. 73-86, MRP Publishing Services Ltd.,
Shrewsbury (1990).
V. Zbranek, Z. Zbranek, and D. A. Burnham, US Patent 4.092,400 (1978).
N. N. Maslenitskii, Tsvetn. Met. 4-5 (1939), 140-143.

INDUSTRlAL PRODUCTION

251

5.17. B. Burwell, US Patent 3.256,057 (1966).

5.18. N. N. Maslenitskii and P. M. Perlov, in: Proc. 5th Int. Mineral Processing Congr., Group VII, p. 839,
Pergamon Press, New York (1960).
5.19. P. Queneau, Thesis, University of Minnesota, MN 687369 (1967).
5.20. A. N. Zelikmann and G. E. Meerson, Metallurgiya, Redkikh Metallov, Ch. I, Metallurgiya, Moscow (1973).
5.21. P. B. Queneau, 1. W Beckstead, and D. K. Huggins, US Patent 4.325,919 (1982).
5.22. P. B. Queneau, 1. W Beckstead, and R. Hogsett, Int. Patent COIG41/00; Int. Publ. No. W082/02540
(1982).
5.23. P. B. Queneau, D. K. Huggins, and 1. W Beckstead, US Patent 4.320,096 (1982).
5.24. H. Li, M. Liu, and Y. Li, in: Proc. of the 1st Int. Can! on the Metallurgy and Materials Science of If, Ti, Re
and Sb, Changsha, Vol. 1, pp. 192-197, International Academic Publishers, Pergamon Press, Oxford
(1988).
5.25. H. Li, M. Liu, P. Sun, and Y. Li, J. Cent. South Univ. Techno!. 2 (1995), 16-20.
5.26. H. Li, M. Liu, and Z. Si, Chinese Patent 85100350.8, 1985-04-01.
5.27. C. 1. Smithells, Tungsten, Chemical Publishing Co., New York (1953).
5.28. Societe Electrometallurgique de Saint Etienne, u.K. Patent 695.843 (1953).
5.29. M. Karczynski, Polish Patent 48.426 (1964).
5.30. K. Vadasdi, "Effluent-Free Manufacture of Ammonium Paratungstate by Recycling the Byproducts," in:
The Chemistry of Non-Sag Tungsten (1. Bartha, E. Lassner, W D. Schubert, and B. Lux, eds.), pp. 45--60,
Elsevier Science Ltd., Oxford (1995).
5.31. E. Lassner, Oesterr. Chem. Z. 80 (1979), 111-115.
5.32. P. B. Queneau, 1. W Beckstead, and D. K. Huggins, US Patent 4.311,679 (1982).
5.33. N. N. Maslenitskii and P. M. Perlov, in: Proc. 5th Int. Mineral Processing Congr., Group VII, p. 839,
Pergamon Press, New York (1960).
5.34. Gong Bofan, Huang Weizhuang, and Zhang Qixiu, Int. J. Refract. Met. Hard Mater. 14 (1996),
319-324.
5.35. A. I. Bellingham, US Patent 3.939,245 (1976).
5.36. C. R. Kurtak, US Patent 3.158,438 (1964).
5.37. T. K. Kim, R. W Mooney, and V Chiola, Sep. Sci. 3 (1968), 467-478.
5.38. T. K. Kim, 1. E. Ritsko, M. B. Macinnes, and M. C. Vogt, US Patent 4.379,126 (1983).
5.39. M. B. Macinnes, R. P. Macclintic, and T. K. Kim, US Patent 4.360,502 (1982).
5.40. C. Burwell, US Patent 3.256,057 (1966).
5.41. T. K. Kim, US Patent 4.175,109 (1970).
5.42. 1. W Beckstead, H. Dale, and D. K. Huggins, US Patent 4.328,190 (1982).
5.43. T. K. Kim, M. B. Macinnes, R. P. Macclintic, and M. C. Vogt, US Patent 4.374.099 (1983).
5.44. M. B. Maccines and T. K. Kim, Int. J. Refract. Met. Hard Mater. 5 (1986), 78-81.
5.45. T. Zuo, Q. Zhao, S. Li, and 1. Wang, in: Proc. of the 1st Int. Can! on the Metallurgy and Materials Science
of W Ti, RE and Sb, Changsha Vol. I, pp. 11-20, International Academic Publishers, Pergamon Press,
(1988).
5.46. 1. Zhou and 1. Xue, in: Proc. 5th Int. Tungsten Symp. Budapest, pp. 73-86 MPR Publishing Services Ltd.,
Shrewsbury (1990).
5.47. Huang Weizhung, Zhang Qixu, Gong Bofang, Huang Zhaoying, and Luo Aiping, Int. J. Refract. Met. Hard
Mater. 13 (1995) 217-220.
5.48. Hu Zhaorui and Yun Lui, in: Application of Ion Exchange in Tungsten Hydrometallurgy in Ion Exchange
for Industry (M. Streat, ed.), p. 385, Ellis Horwood, Chichester (1988).
5.49. Zhao Heng, in: Proc. of the 7th Int. Tungsten Symposium, Goslar, pp. 142-149, ITIA, London (1996).
5.50. Chen Zhouxi et al., Chinese Patent 88105712.6.
5.51. 1. W van Put, T. W Zegers, A. van Sandvijk, and P. 1. M. van der Straten, in: Proc. 2nd Int. Conf on
Separation Science and Technology (E. B. Baird and C. Vijayan, eds.), pp. 387-394, C.S.Ch.E.M.H.I.,
Hamilton, Canada (1989).
5.52. 1. W van Put, P. M. de Konig, A. van Sandvijk, and G. 1. Witkamp, in: Proc. 11th Symp. on Industrial
Crystallization (A. Mersman, ed.), pp. 647-652, Garmisch-Partenkirchen (1990).
5.53. 1. W van Put, W P. C. Duyvesteyn, and 1. van Vliet-Jahnberg, in: Proc. 12th Plansee Seminar (H. Bildstein
and H. M. Ortner, eds.), pp. 433-435, Metallwerk Plansee, Reutte, Austria, (1989).
5.54. A. W Bryson, D. Glassner, and W F. Lutz, Hydrometallurgy 2 (1976), 1985.

252

CHAPTER 5

5.55. 1. W. van Put, Thesis, Delft University Press (1991).

5.56. 1. R. Scheithauer, C. D. Vanderpool, M. B. Macinnes, and 1. H. Miller, US Patent 4.624,844 (1986); Europ.
Patent 0194.346AI.
5.57. 1. B. Goddard, US Patent 4.326,061 (1982).
5.58. 1. B. Goddard, SME-AIME Annual Meeting, Los Angeles, Reprint
5.59. H. 1. Lunk, B. Ziemer, M. Salmen, and D. Heidemann, in: Proc. 13th Int. Plansee Seminar (H. Bildstein
and R. Eck, eds.), Vol. I, pp. 38-56, Metallwerk Plansee, Reutte, Austria (1993).
5.60. AMAX Metals Group, Tungsten Chemicals, W-17 (1984).
5.61. E. Lassner and W. D. Schubert, "Tungsten Blue Oxide," in: The Chemistry ofNon-Sag Tungsten (L. Bartha,
E. Lassner, W. D. Schubert, and B. Lux, eds.), pp. 111-118, Elsevier Science Ltd., Oxford (1995).
5.62. H.1. Lunk, B. Ziemer, B. Salmen, and D. Heidemann, Int. J Refract. Met. Hard Mater. 12 (1993/94),1726.
5.63. W. D. Schubert and E. Lassner, Int. J Refract. Met. Hard Mater. 10 (1991), 171-183.
5.64.1. W. van Put and T. W. Zegers, Int. J Refract. Met. Hard Mater. 10 (1991),115-122.
5.65. Z. Zou, E. T. Wu, and C. Qian, in: Proc. I lth Plansee Seminar (H. Bilstein and H. M. Ortner, eds.), pp.
337-348, Metallwerk Plansee, Reutte, Austria (1985).
5.66. V Chiola, 1. M. Laferty, and C. D. Vanderpool, US Patent 3.175,881 (1965).
5.67. V Chiola and P. R. Dodds, US Patent 3.591,331 (1971).
5.68. 1. O. Hay and R. 1. Grodek, British Patent 1.267,585 (1972).
5.69. T. K. Kim, 1. M. Laferty, M. B. Mcinnis, and 1. C. Patton, US Patent 3.857,928 (1974).
5.70. C. D. Vanderpool, M. B. Macinnes, and 1. C. Patton, US Patent 3.936,362.
5.71. D. E. Collier, C. 1. Couch, and D. N. Hingle, Hydrometallurgy, pp. 1-9, 81/G5.
5.72. G. Schwier, in: Proc. 10th Plansee Seminar (H. M. Ortner, ed.), Vol. 2, pp. 369-383, Metallwerk Plansee,
Reutte, Austria (1981).
5.73. H. Palmour III, PM 94, Refractory Metals, pp. 2025-2028.
5.74. V Reich, PM 94, Refractory Metals, pp. 111-113.
5.75. L. Bartha, E. Lassner, W. D. Schubert, and B. Lux, eds., The Chemistry of Non-Sag Tungsten, Elsevier
Science Ltd., Oxford (1995).
5.76. E. Pink and L. Bartha, eds. The Metallurgy of Doped/Non-Sag Tungsten, Elsevier, New York (1989).
5.77. W. D. Coolidge, JAm. Inst. Elec. Eng. 29 (1909), 953.
5.78. E. Pink and R. Eck, "Refractory Metals and their Alloys," in: Materials Science and Technology (R. W.
Cahn, P. Hassen, and E. 1. Kramer, eds.), Vol. 8, pp. 589-641, VCH, Weinheim (1991).
5.79. F. V Lenel, Powder Metallurgy, Metal Powder Industries Federation, Princeton (1980).
5.80. Metallwerk Plansee GmbH, Refractory Metals, Verlag Moderne Industrie AG & Co., Landsberg/Lech,
Germany.
5.81. 1. A. Mullendore, in: The Metallurgy ofDoped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 61-81,
Elsevier, New York (1989).
5.82. S. Yamazaki, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 47-59,
Elsevier, New York, (1989).
5.83. B. P. Bewlay and C. L. Briant, in: The Chemistry of Non-Sag Tungsten (L. Bartha, E. Lassner, W. D.
Schubert, and B. Lux, eds.), p. 137-159, Elsevier Science Ltd., Oxford (1995).
5.84. R. F. Cheney, "Sintering of Refractory Metals," in: Metals Handbook, 9th ed., Vol. 7, pp. 389-393, ASM
International, Metals Park, Ohio (1984).
5.85. F. B. Swinkels and M. F. Ashby, Acta Metall. 29 (1981), 259-281.
5.86. O. Horacsek, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 175-187,
Elsevier, New York (1989).
5.87. D. J. Jones, Journal of Less-Common Metals 2 (1960), 76-85.
5.88. W. Rostoker, "Conversion of Refractory Metals," in: Refractory Metals and Alloys II (M. Semchyshen and
I. Perlmutter, eds.) pp. 379-394, Interscience Publishers, Wiley, New York (1963).
5.89. Wolfram, Company Brochure, Metallwerk Plansee, Reutte, Austria (1980).
5.90. 1. P. Wittenauer, T. G. Nieh, and 1. Wadsworth, Adv. Mater. Process. 9 (1992), 29-37.
5.91. H. S. Yui, 1. S. Kim, K. 1. Rhee, J. C. Lee, and H. Y. Sohn,Int. J Refract. Met. Hard Mater. 11 (1992), 317332.
5.92. R. Huenert, G. Winter, W. Kiliani, and D. Greifendorf, Int. J Refract. Met. Hard Mater. 11 (1992), 331335.

INDUSTRIAL PRODUCTION

253

5.93. V Glebovsky, in: Proc. 12th Plansee Seminar (H. Bildstein and H. M. Ortner, eds.), Vol. 3, pp. 379-389.
Metallwerk Plansee, Reutle, Austria (1989).
5.94. Yoshiki Doi, Characteristics and Application of High Purity Tungsten and Chemical Vapor Deposited
Tungstein, paper presented at the Annual ITIA Meeting, Huntsville, USA; transcript of papers: lTIA,
London (1994).
5.95. G. Winter, Tungsten and Special Tungsten Compounds-Their Application in Modern Technology, paper
presented at the Annual lTIA Meeting, Huntsville, USA; transcript of papers: lTIA, London (1994).
5.96. G. D. Rieck, Tungsten and its Compounds, Pergamon Press, Oxford (1967).

6
Tungsten Alloys
A large number of tungsten alloys and composites were investigated in the past, but only
some of them achieved technical importance. However, these are of significant commercial
interest and will be discussed in the following sections. Alloys derived from melting
metallurgy, such as ferrotungsten, melting base, steels, superalloys, and stellites, as well as
the large group of cemented carbides, are not discussed here but will be treated in Chapters
8 and 9.
The aim of alloying tungsten is to improve its chemical, physical, and mechanical
properties at both ambient conditions and elevated temperatures. Beyond that, it is possible
to combine the useful properties of tungsten with those of the alloying additives.
Low-temperature brittleness is the most crucial aspect in the manufacture of pure
tungsten metal. Therefore, in the past, much effort has been directed at lowering the
ductile-to-brittle transition temperature (DBTT) and hence improving the fabricability of
the metal. In this regard, tungsten-rhenium alloys have gained outstanding importance.
They exhibit a significantly lower DBTT and are even stronger than unalloyed tungsten at
high temperatures.
To fully realize the potential of tungsten in high-temperature applications, dispersion
strengthening and precipitation hardening turned out to be the most effective way to
increase the high-temperature strength and creep resistance. Non-sag tungsten alloys or
thoriated tungsten, used for lamp filaments, are examples of this group of materials.
Moreover, alloys based on W-Re-Th0 2 and W-Re-HfC are among the strongest alloys
used today for construction parts in high-vacuum and high-temperature technology.
These tungsten alloys are commonly produced by powder metallurgical techniques.
Solid-state sintering, liquid-phase sintering, and infiltration techniques are employed. Fine
particle size and high purity of the powder components are a prerequisite in order to
achieve an even distribution during sintering and a high degree of material purity. Powder
metallurgy offers a simple route to the desired ratio of components, simply by mixing them
properly. Concentration(s) of the alloying element(s) in tungsten base alloys may vary from
traces in the case of microalloys (non-sag tungsten contains less than 100 Ilg/ g K) to more
than 20 wt% (W-Re). Depending on the additives, single-phase or two-phase alloys are
produced.
Applications of the alloys described here are presented in Chapter 7. More information on the metal-physical principles can be found elsewhere [6.1-6.4].
255

256

CHAPTER 6

6.1. SUBSTITUTIONAL ALLOYS (SOLID-SOLUTION ALLOYS)

In the case of solid-solution alloys, the size and valence of the solute atoms determine
the degree of solid-solution strengthening. Differences in atomic size of about 5% between
the solvent and the solute are necessary to produce a significant strain-hardening effect
[6.1 ].
Molybdenum, niobium, and tantalum (bcc crystal structure) form a continuous series
of solid solutions with tungsten, but only Nb and Ta additions lead to a strong straining
effect. Higher additions of Nb and Ta raise the recrystallization temperature but also
increase the DBTT and thus decrease the workability. Therefore, none of these binary or
ternary tungsten base alloys developed in the past [6.2] has attained commercial
importance. The only important tungsten-base solid-solution alloy today is tungstenrhenium.
Tungsten additions to Mo, Nb, and Ta, on the other hand, improve the hightemperature strength and creep properties of these metals. Niobium-based alloys, such
as, Nb-l0W-2.5Zr, Nb-2STa-IOW-IZr, or Nb-l0W-IOHf-O.1Y, are potential candidates for nuclear engineering and space applications, due to their high strength and good
strength-to-specific weight ratio [6.5]. Tantalum-based alloys are used owing to their high
corrosion resistance (see below). Molybdenum-tungsten alloys (Mo--l OW and Mo--30W),
which were used earlier in the zinc-producing industry because of their long-term
corrosion resistance to molten zinc, lost their importance to Sialon (SiN/AI 2 0 3). Mo-5W, formerly applied in rotating X-ray anodes as support for the W-Re layer, was
substituted by TZM (Mo-O.5Ti-O.OSZr-O.05C). Recently, molybdenum-tungsten alloys
have been used as sputter targets for the manufacture of flat panel displays (thin-film
transistor technique).

6.1.1. Tungsten-Rhenium Alloys [6.1-6.4]

Rhenium is the most important alloying element for tungsten. Rhenium additions
increase the ductility of wrought products at low temperatures and also improve their hightemperature strength and plasticity. Furthermore, Re additions stabilize the grain structure,
increase the recrystallization temperature, reduce the degree of recrystallization embrittlement, and significantly improve the weldability. Finally, W-Re alloys also exhibit
superior corrosion behavior than unalloyed tungsten. The only drawback of rhenium is its
high price (1000 US$/kg as compared to 13.5 US$/kg for tungsten).
The improved fabricability and low-temperature ductility of W-Re alloys (the socalled rhenium ductilizing effect) is attributed to the increased solubility of interstitials (N,
C, 0, S, P) in the substitutional alloys. Interstitials are scavenged from low-energy
positions (dislocations, grain boundaries) and distributed in the form of clusters within
the lattice, so increasing the grain-boundary rupture strength. Rhenium also promotes
excessive twinning, which might act as a supplementary deformation mechanism to
slipping for stress relaxation, particularly at lower deformation temperatures [6.3].
The rhenium ductilizing effect is more pronounced at higher Re additions (Fig. 6.1)
and reaches a maximum at the solid solubility limit (approximately 27wt% Re). Beyond
that, embrittlement occurs due to the formation ofa hard and brittle a-phase. The hardness
ofW-26 Re alloy is 20% larger than that of unalloyed tungsten and the hot tensile strength

257

TUNGSTEN ALLOYS

Temperature, of

100

200

300

400

600

500

700

Ductile

4
I-

vl

:J

:0(1)
....

"0

C
CIl

.0

E 12
::I
E
c

16

20~

__

~~

______

100

______

200

300

____

400

Temperature,OC
FIGURE 6.1. Effect of alloying additions of rhenium on ductility [6.3,6.5].

is more than twice as high (Fig. 6.2). Recrystallized W-26Re exhibits a fine and even
microstructure and behaves in a ductile manner at room temperature.
Compositions. The most important alloy compositions are W-(3-'5)Re, W-lORe,
and W-(25-26)Re.
Alloys with 5% Re exhibit the hardness minimum and highest creep strength; alloys
with >8% Re show good workability (minimum forging temperature 1560C, annealing
temp. 1600-1800C) and have significantly better welding properties. The 3-10% Re is
extremely resistant against alternating thermal stresses which occur in X-ray anodes with
high-energy electron beams.
Thermocouples based on W-Re alloys can be used to measure temperatures up to
2300C and, for short times, up to 2700C (W-3RejW-25Re; W-5RejW-26Re).
Production. The classical method is to mix Re and W powders in the desired ratio
prior to compaction and sintering. In order to achieve a more even W-Re distribution,
restricting the danger of local a-phase formation, mixtures of tungstic acid or tungsten
trioxide or tungsten blue oxide with ammonium perrhenate can be used as raw materials.
These mixtures are co-reduced by hydrogen to metal powder.
Tungsten-rhenium alloys are hot-swaged and hot-rolled in air and are well suited for
shaping, including tube fabrication.

258

CHAPTER 6

1500
(J]

Q.

en
~

:J

cP.
~

1000

500

0
60

.Q
(J]

CJ)

40

(I)

Q)

20

:::J

()

a:l
~
U-

500

1000

Test temperature

1500
(OC)

FIGURE 6.2. Tensile properties as a function of test temperature ofW and W-Re disks forged by 75%, measured
in the radial direction [6.1].

CVD ofW-Re alloys has gained commercial importance for the production of X-ray
rotating anodes, cones, tubes, and other shapes. Layer thickness for X-ray targets is 1 mm.
As in CVD of pure tungsten, fluorides or chlorides of W and Re are also used as
precursors. Densities over 99% of the theoretical can be obtained [6.1].

6.1.2. Tungsten-Tantalum Alloys [6.1,6.2]

In the binary system, tungsten-tantalum unlimited solubility exists. However, only
the Ta-rich side is of technical importance. Tantal1.lm-tungsten alloys combine the good
corrosion resistance and high elasticity of tantalum with the better hot strength of tungsten.
Typical alloy compositions are Ta-2.5W, Ta-7.SW, and Ta-lOW They are used for
construction materials in strongly corrosive media and in cases where the hot strength
of pure tantalum is inadequate.
Due to the high affinity of tantalum to oxygen, nitrogen, and hydrogen, sintering can
only be performed in high vacuum as is usual for pure tantalum.

TUNGSTEN ALLOYS

259

Tantalum can be further strengthened by additions of Hf (solution-strengthening) or

hafnium nitrides, carbides, or oxides (precipitation/dispersion-strengthening). Typical
alloy compositions of this type are Ta-8W-2Hf and Ta-lOW-2.5Hf-O.Ol C.
6.1.3. Tungsten-Titanium Alloys [6.6]

The alloy W-lOTi is used as a sputtering target in the manufacture of microelectronics devices, such as VLSI, ULSI (very large resp. ultralarge-scale integration), and
DRAM (dynamic random access memory) chips. Thin W-Ti layers are sputtered onto
silicon substrates and act as a diffusion barrier against aluminum (interconnect).
W-lOTi alloy targets are produced by blending ultrapure tungsten (see also Section
5.7.6) and titanium powders (obtained by vacuum melting and subsequent pulverizing),
followed by pressure sintering. The impurity level has to be very low, in particular, with
respect to radioactive elements (U, Th) and "mobile" alkaline metals (Na, K).
6.2 DISPERSION-STRENGTHENED AND PRECIPITATION-HARDENING ALLOYS
At half of the solidus temperature on the absolute temperature scale (i.e., '" 1850 C
for tungsten), solid-solution alloys lose much of their strength, and dispersion-strengthened or precipitation-hardened alloys are significantly stronger and creep resistant. This is
caused by the interaction between the dispersoids and dislocations, as well. as subgrains
and grain boundaries.
6.2.1. Non-Sag Tungsten

Tungsten wires are commonly used as filaments in incandescent lamps. However, due
to the high working temperatures (up to 3000C in halogen lamps), pure tungsten wire
would fail rapidly as a result of grain boundary sliding ("offsetting"). In order to increase
the lifetime of such filaments even under severe conditions (e.g., in automotive applications), so-called non-sag tungsten wires are used. The term non-sag refers to the resistance
of the wire against deformation ("sagging") under its own weight at incandescent
temperatures (Fig. 6.3).
Non-sag tungsten, or, as it is frequently called, doped tungsten, is a remarkable kind
of dispersion-strengthened material. It owes its outstanding creep resistance to rows of
potassium-filled bubbles (gaseous at. working temperatures) which are aligned in rows
parallel to the wire axis. These bubbles ("soft" dispersoids) act as barriers for grain
boundary migration in the radial direction and effect the recrystallization process in such a
way as to lead to an interlocking long-grained microstructure, which gives the wire its nonsag properties (see Chapter 1; Fig. 1.9).
The discovery of NS-W was a fluke. In the early days of wire fabrication (19101925), it was realized that the source of the tungsten oxide used for metal powder
fabrication played an important role in determining the wire quality [6.7]. Tungsten
filaments, originating from material which was treated in clay crucibles, preferably those
manufactured by the Battersea Company near London, were shown to be much more stable
in shape at incandescent temperatures. Subsequent chemical analysis indicated that a

260

CHAPTER 6

Ibl

FIGURE 6.3. Coil of a 230Vj60W lamp before use (a); sagging and defonnation of the coil after several
hundred hours of operation due to bad NS quality (b).

certain pick-up of potassium, aluminum, and silicon occurred during powder processing,
but only potassium remained in the sintered ingot in significant amounts.
It is not clear today whether this contamination occurred during calcination of
tungstic acid prior to reduction [6.8] or during the reduction process [6.7], and when the
dopants were finally added intentionally in the various tungsten plants. Most of the doping
and reduction technology was and still is based on empirical knowledge. Even today,
different producers have their own ideas about the action of the dopants and their behavior
during subsequent processing.
Although it has been known for a long time that the long-grained and interlocking
microstructure was the reason for the high-temperature creep resistance of the doped
tungsten wires, no explanation could at first be given for the action of the dopants. The
formation of thin "tubes" of dope aligned parallel to the wire axis and the occurrence of
rows of pores, which hinder the movement of dislocations and thus increase the resistance
to off-setting, was first stated in the late fifties [6.9]. Later, the formation of potassiumfilled bubbles was demonstrated by TEM investigations, and metallic potassium was
identified as the bubble-forming material [6.10, 6.11].
Today, the desired NS properties of tungsten wires are achieved by doping the starting
tungsten blue oxide with aqueous solutions of potassium silicate and aluminum chloride or
nitrate (AKS doping). Only when all three elements are present during hydrogen reduction,
a sufficient amount of dopant is incorporated into the metal grains as potassium
aluminosilicates and cannot be removed by a subsequent acid washing. During sintering
of the powder compacts by direct electrical resistance heating under hydrogen, the
potassium aluminosilicates dissociate and both silicon and aluminum are removed by
diffusion and volatilization. Potassium, which is insoluble in the tungsten matrix, remains
in the sintered ingot forming bubbles. The potassium-filled bubbles are elongated during
swaging and wire drawing and finally form the rows of tiny bubbles (typically in the range
of 20-40 nm), which are essential for generating the NS microstructure.
A schematic representation of the non-sag tungsten wire fabrication process is shown
in Fig. 6.4.

261

TUNGSTEN ALLOYS

- formation of W-crystals
- incorporation of DOPANTS

in the WllrainS
H, -REDUCTION

ACIDWASHING

remova l of
unincorporated
K. AI , Si

I w-powderl

- rows of K-bubbles
- high sag-resistance
evaporation
of AI . 5i , (K)

FIGURE 6.4. From tungsten blue oxide to non-sag wire (schematic presentation).

Non-Sag Tungsten Wire Fabrication

Metal Powder Production [6.12]. The production of non sag tungsten powder starts
with the careful selection of an appropriate ammonium paratungstate (APT) in terms of
crystal size, size distribution, particle agglomeration, chemical purity, and the subsequent
"blueing" of the APT under slightly reducing conditions at 400-900 C, in rotary furnaces
or in push-type kilns. The resulting tungsten blue oxide (characterized by its residual
ammonia and water content, its specific surface area, as well as its w:o ratio) is then doped
in batches with aqueous solutions of potassium silicate and aluminum chloride or nitrate.
Typical dopant compositions are between 2000-3000llg/g K, 200-500llm/g AI, and
2000-3000llg/g Si. After drying the oxide, several batches are blended into one
homogeneous lot of some 1000 kg.
Reduction is performed in multitube push-type furnaces as described in Section 5.4.2.
Typical furnace temperature profiles start with low reduction temperatures of 500-700 C
and end with maximum temperatures of 800-900 0c. Alternatively, a two-stage reduction
is carried out instead. In the first stage, the reduction takes place at 400-700 C (formation
of W0 2 ) and the second stage at 600-900C (formation of tungsten metal). Between the
two stages, the powder is blended.
After reduction, nearly all the dopants added to the TBO appear to remain in the
tungsten powder. However, only 10% are actually incorporated within the tungsten
particles, and the rest is only on the particle surfaces. During subsequent washing of the
powders with distilled or deionized water and concentrated HCl and HF acids, only the
dopants which are entrapped internally are retained. Leaching is commonly performed in
small batches, around 100 kg. The powder is then dried at 100-120C to remove water,
screened to eliminate boat scale, and blended to form a larger, homogeneous powder lot.
The specific goal of the NS powder production is to produce a tungsten powder with
an average particle size in the range 3-5 11m and typical dopant levels of 100-160 Ilm/g K,
20-60llg/g AI, and 200-250llg/g Si (analyzed after the acid-wash).

262

CHAPTER 6

Sintering [6.2,6.13]. The powder is pressed into compacts, usually of size (1228) x (12-28) x (600-800) mm, and then presintered under hydrogen at 1100-1300 C for
30--60 minutes to increase the strength of the compacts so they can be handled for the final
direct sintering process. Sintering is carried out under hydrogen in bell-jar-type furnaces
until the density reaches about 90% of the theoretical value (see also Section 5.5.2.2). A
typical sintering schedule uses two intermediate soak points prior to the final hold at 27003000 DC, with carefully controlled heating rates between the holds to render the outgassing
of impurities and of parts of the dopants as long as there is open porosity [6.13]. Between
1800 and 2150C all of the silicon, and above 2150C most of the aluminum, is
evaporated as a result of the progressive dissociation and reduction of the incorporated
potassium aluminosilicates [6.14]. About 60 to lOOllg/g K, but less than 10 Ilg/g Si and
AI, remain in the sintered bar. It is commonly agreed today that a minimum of 50 Ilg/ g K
is necessary to produce a moderate non-sagging wire quality.
In the sintered bar, potassium bubbles appear as tiny pores of about 0.1 11m diameter
in the fracture surface, besides some larger, residual sintering pores (Fig. 6.5).
Metal Working. Today, the first step in metal working is commonly hot rolling (16001650 DC; Kocks rolling mills) of the sintered bar down to rods of about 7-10 mm diameter.
Rolling is then followed by a multiple swaging process, in which the temperature is
gradually reduced as the working proceeds (1600C --+ 1200C). Reduction in area pass
is 10% at the start of swaging and increases to 35-40% for the latter stages. Work
hardening of the material requires intermediate recrystallization anneals, which are carried
out between 2000-2300 C [6.13]. Prior to annealing, the oxide scale must be removed by
pickling to avoid embrittlement.
Wire drawing usually starts at a diameter of about 0.1 inch (2.54 mm) [6.2]. During
the first few drawings, a 40% reduction per pass can be applied at about 1000C. The
drawing temperature then decreases with decreasing diameter, and the reduction per pass is

FIGURE 6.5. Fracture surface of a NS-doped tungsten ingot showing tiny potassium-filled pores besides some
larger residual sintering pores. The potassium-filled pores are in the order of O.llllll and were formed during
sintering by dissociation and reduction of aluminosilicates. By courtesy of O. Horacsek, Hungarian Academy of
Science.

263

TUNGSTEN ALLOYS

lowered to 10%. Intermediate in-line anneals serve only for stress relief. A graphite
suspension in water is commonly used as lubricant. Cemented carbide dies are used for
large-size drawing, but below 0.25 mm diamond dies are used. A typical non-sag tungsten
wire processing schedule is shown in Fig. 6.6 [6.2].
Theoretically, a sintered ingot of 3 kg will yield 365,000 m of a 24-!lm filament,
corresponding to 500,000 coils for a 40-W incandescent lamp [6.15]. Wires for common
household lamps are in the range of 17 to 45 !lm, but wires can be drawn down to 10 !lm.
Even smaller diameters can be produced by subsequent etching.

Doped tungsten powder

Press to 55% theoretical density
Compact (14 mm x 14 mm x 600-800 mm)
Presinter and sinter; swage at 1650 C, 20% red./pass
Sintered bar (7.6 mm diameter)
Pickle and condition, anneal at 2150 C, 300 mm/min
Swage at 1350 C, 25% red./pass (4 passes)
Swaged rod (4.3 mm diameter)
Anneal at 2000 C, 250 mmlmin
Tandem (2-dle) swage at 1300 C, 28% red./pass, 3 mlmin
Swaged rod (2.9 mm diameter)
Tandem (2-die) swage at 1250 C, 28% (ad./pass, 3 mlmin
Swaged rod (2.0 mm diameter)
Draw at 1050 C (3 passes) 40% red./pass, 4.8-8.5 mlmln
Wire (1.19 mm diameter)
Draw at 1000 C (1" pass), drop 50 C every pass (4 passes)
35% red./pass, 8.5-11.3 m/mln
Wire (0.58 mm diameter)
Draw at 450 C (14 passes)
15% red./pass,12.5-27.7 m/min
Wire (100 mgl200 mm length)
In-llne anneal (constant current)
Draw at 350 C (16 passes), 15% red./pass, 40-68 mlmln
Wire (7.0 mgl200 mm length)
In-line anneal (gas), no die in series
Draw at 300 C (14 passes); 13% red./pass, 91-99 mlmin
Wire (1.0 mg/200 mm length)
Draw at room temperature (10 passes)
10% red./pass, 120 m/mln
Wire (0.340 mgl200 mm length)

FIGURE 6.6. Typical non-sag tungsten wire processing schedule [6.2].

264

CHAPTER 6

FIGURE 6.7. SEM image of a non-sag tungsten filament: (a) before use, (b) after several hundred hours of
operation in a 60-W incandescent lamp. Facetting of the originally round filament occurred due to evaporation of
tungsten at the high operation temperature (2400- 2500 C).

Most incandescent lamps today exhibit a coiled-coil geometry of the filament (Fig.
6.7). Coiled-coils are made by winding the filaments on mandrels of molybdenum and
subsequent winding of the coil on a larger mandrel to form a secondary coil [6.2]. The coil
is then heated to about 1700C to relieve the stresses formed during coiling and to
stabilize the geometry [6.7]. The mandrels are then selectively dissolved in nitric acid or
mixtures of nitric and sulfuric acid.
The last stage in filament fabrication is a heat treatment to recrystallize the wire at
temperatures greater than 2200 C* to form the characteristic NS structure. This procedure
is either performed after mounting of the filament (in situ recrystallization) or in a separate

* Compared

to unalloyed tungsten the recrystallization temperature of doped tungsten is shifted from 10001200 C up to 2000 C or even higher [6.1].

TUNGSTEN ALLOYS

265

furnace unit prior to assembly, where the filaments are positioned in tungsten boats. The
latter procedure is only used for high-performance halogen lamps [6.7].
Microstructural Evolution. As the sintered ingot is worked down into rods and wires,
the residual sinter porosity is eliminated. However, the potassium-filled bubbles remain
and will be elongated with increasing deformation, forming tube-like structures with an
aspect ratio of 20 or even higher (Fig. 6.8). On annealing, these tubes break-up into strings
of bubbles, driven by surface-energy minimization (Fig. 6.9) [6.16]. The number, size, and
dispersion of the bubbles finally present in a non-sag tungsten wire is determined by the
thermomechanical history of the wire. Usually, a large number of small-diameter bubbles
is desired for obtaining a high NS-W wire quality.
Drawn tungsten wire has a fibrous microstructure. Under incandescent temperatures,
the highly cold worked structure recrystallizes, and the fibrous microstructure is transformed into an interlocking grain structure (swallow-tail type), with grain lengths many
times the diameter of the wire, which give the wire its non-sag properties. A comparison of
a pure and NS-doped tungsten wire in the as-deformed condition and after recrystallization
is shown in Fig. 1.9 (Chapter 1).
For vibration-resistant applications, such as traffic signal lamps, non-sag wires with
(1-5 wt%) Re are used. Rhenium improves the ductility of the recrystallized wire and also
increases its electrical resistance, thus allowing a thicker and mechanically more stable
wire [6.17].
More than 90% of non-sag tungsten is used for incandescent lamps. Small amounts
are used as defroster heating wires in automobile windshields and as heating wire coils for
aluminum evaporation in metallizing applications. For the latter application, thick wires
are used (0.5-1 mm). The service life of these wires is not only determined by their
mechanical properties but also by their corrosion resistance, which is higher for coarsergrained doped materials with a minimum of grain-boundary volume [6.18].

FIGURE 6.8. SEM micrograph showing cylindrical, potassium-containing tubes in the as-deformed tungsten. By
courtesy of O. Horacsek, Hungarian Academy of Science.

266

CHAPTER 6

FIGURE 6.9. On annealing, the tubes break-up into strings of potassium-filled bubbles. By courtesy of O.
Horacsek, Hungarian Academy of Science.

6.2.2. Alloys with Oxide Dispersoids

The addition of small amounts of finely dispersed oxides to metals, for increasing
their high-temperature strength, is a common practice in physical metallurgy (ODS alloys).
In the case of tungsten, this technology can be traced back to its early roots. In 1913 (i.e.,
long before the mechanisms of dispersion-strengthening were known), impact and
vibration-resistant filaments were produced based on thoriated tungsten [6.8], basically
as is still done today for rough service uses.
Thoria is insoluble in tungsten and has a very high thermal stability, which makes it
an ideal dispersoid. Additions of 0.5-4 wt% significantly restrict grain coarsening during
sintering, which leads to a fine-grained sintered microstructure. The thoria particles are
located at the grain boundaries, thus impeding their mobility (Fig. 6.10). Thoria additions
increase the recrystallization temperature and improve the hot strength and creep resistance
of tungsten. Strength and fracture elongation of W-2Th02 are shown in Fig. 6.11 as
dependent on the degree of working [6.1, 6.19]. Thoria hardening in W-Re is effective up
to 2200 K ('" 1930 0c) [6.20].
During swaging and wire drawing, the particles are aligned in strings parallel to the
drawing direction of the wire. On recrystallization, an elongated grain structure forms, but
the resistance against sagging at incandescent temperatures is comparatively low. Therefore, thoriated tungsten filaments are used only where the hot strength is more important
than the sagging characteristics [6.2].
Thoria additions lower the electronic work function (which is the electrical energy
that an electron needs to escape from a tungsten surface) and enhances the thermionic

267

TUNGSTEN ALLOYS

FIGURE 6.10. Fracture surface of a sintered W-2% Th0 2 ingot; SEM micrograph showing rounded thoria
particles located at the grain boundaries. By courtesy of Plansee AG, Reutte, Austria.

True stra in
0 .5 0.7

n
I (So/S)
1. 0

1 .5 2 34

500

2000

co
a...

300

(f)

()
~

1000

I::J

100 l-

I--

a en
0
a
c

-1 00

15

o
(;J

OJ

10

..Q
aJ
aJ

'-

:::J
( ,)

CO
LL

'-

0 .4
02
0 .6
0 .8
Engineer ing strain (So - S) /S o

1.0

FIGURE 6.11. Ultimate tensile strength and fracture elongation at different temperature and ductile-to-brittle
transition temperature as dependent on the degree of working for W-2%Th0 2 [6.1,6.19].

268

CHAPTER 6

emission. This is important for all types of tubes, such as radio valves and X-ray tubes, as
well as for discharge lamps and welding electrodes (TIG welding*), where it improves the
ignition behavior of the electrodes and gives a constant and stable arc.
Under working conditions, thoria is reduced and the released thorium atoms act as an
emitter [6.21]. The reason for the formation of a monoatomic layer of metallic thorium is
still not yet fully understood. It is assumed that the reduction of thoria by tungsten
according to
2W + xTh0 2 -+ 2WOx

+ xTh

is favored by the simultaneous evaporation of the tungsten oxide, because of its high vapor
pressure [6.20].
Electrodes for high electron emission must be activated before use. This is done by
rapid heating above 2000C and subsequent activation annealing at l500-l800C [6.2,
6.22]. Alternatively, the activation can be performed in a carburizing atmosphere, which
leads to a slight carburization of the surface. Both procedures enhance the formation of
metallic thorium. A steady transport of thorium by grain boundary diffusion to the surface
and an appropriate microstructure on the electrode tip are of particular importance for a
good welding performance [6.23].
Production. Thorium is either added to the oxide as a solution of thorium nitrate or
chloride, or as colloidal hydroxide, or it is mixed in the form of very fine thoria particles
with the tungsten powder. Sintering is commonly carried out between 2600-2800 C [6.2].
Thoriated tungsten is stronger than unalloyed tungsten and also has a lower DBTT.
Therefore, working has to be carried out at comparatively higher temperatures and inprocess annealing is required more often; otherwise, cracking and particle/matrix
decohesion can occur [6.2].
Thorium is a radioactive element and a natural (X-emitter (specific activity 4250 Bq/ g
and half-life 1.4 x 1010 a). Health and environmental considerations (evaporation and
inhalation of thorium during welding; disposal of grinding dust) have therefore in the past
led to a partial replacement of thoriated tungsten for welding operations and in the lighting
industry by thoria-free ODS tungsten alloys [6.24].
Fine oxide dispersions of yttrium, lanthanum, zirconium, hafnium, cerium, and
erbium behave similarly to thoria, although they possess different emission characteristics,
depending on temperature. This affords an adaptation of the operating conditions
compared to W-Th0 2. Electrodes, based on W-(1-2wt%)Ce02 and W-(I-2
wt%)La203 have demonstrated good arc striking, excellent arc stability, and low electrode
consumption and have even outperformed thoriated tungsten in welding applications
[6.23,6.24]. They can be used for both AC- and DC-TIG welding. Electrodes with 0.8
wt% zr0 2 are used for special welding operations [6.22].
6.2.3. Alloys with Carbide Dispersoids
Hafnium carbide (HiC) is the most potent strengthener for tungsten at elevated
temperatures [6.4]. Particle size and interparticle spacing must be small in order to lead to

* Tungsten-Inert Gas-Welding.

TUNGSTEN ALLOYS

269

an effective strengthening. This occurs as a result of dislocation pinning through the fine
(50-100nm) HfC particles, which inhibits the slip of atom planes and the migration of
grain boundaries [6.25]. A uniform distribution of fine HfC particles can be obtained from
supersaturated solid solutions ("precipitation hardening"). Other carbides, such as TaC,
NbC, and zrC, were also tested but proved to be less effective, especially at temperatures
above 1900C [6.2]. The outstanding action of HfC is attributed to its high thermodynamic stability and its comparatively low solubility and diffusivity in tungsten at high
temperatures.
400

---b- W-3.6Re

co
a..

---0-- W-3.6Re-1Th02

300

W-3.6Re-O.26HfC

..c::
+-'

Cl

c::

...

200

(])

+-'

en

(])

:;

100

o
1200

1600

2000

2400

2800

3200

Temperature (K)
600 r---------------------------------~

co

- - b - - W-3.6Re

500

a..
~

- - 0 - - W-3.6Re-1Th02
~

400

W-3.6Re-O.26HfC

..c::

+-'

Cl

c::

...

300

(])

200

(])

+-'

en

'00
c::

100
0
1200

1600

2000

2400

Temperature (K)

2800

3200

FIGURE 6.12. Yield strength and ultimate tensile strength of a W-3.6Re-O.26HfC alloy as a function of
temperature, in comparison to unalloyed W, W-3.6Re, and W-3.6Re-lTh02 [6.20].

270

CHAPTER 6

Significant strengthening is observed up to 2700 K (2427 QC). Above this, rapid

particle coarsening occurs, and the strengthening effect fades. Additions of rhenium
provide ductility at low temperatures by decreasing the DBTT.
W-Re-HfC alloys cannot be produced by powder metallurgical techniques owing to
the high oxygen affinity of hafuium. They are produced by arc melting.
Yield strength and ultimate tensile strength of a W-3.6Re-0.26HfC alloy as a
function of temperature are shown in Fig. 6.12, in comparison to pure W, W-3.6Re,
and W-3.6Re-lTh0 2 .
W-Re-HfC alloys are the strongest man-made metallic materials at temperatures
above 2000 K [6.25].

6.3. TUNGSTEN COMPOSITES

Tungsten metal exhibits outstanding thermal properties, which makes it attractive for
a broad range of applications. However, for certain applications, its electrical and thermal
conductivity, sensitivity toward oxidation, and poor workability are unsatisfactory. These
limitations have led to the development of two-phase alloys, in which the useful properties
of tungsten are combined with those of the additive.
Two important alloy systems belong to this group and are described below:
The so-called heavy metal alloys, based on W-Ni-Fe and W-Ni-Cu-(Fe). They
are used wherever high density, excellent mechanical properties, and good workability are required .
W-Cu and W-Ag alloys, in which the high electrical and thermal conductivity of
copper or silver is combined with the high hardness and wear resistance of
tungsten.

6.3.1. Tungsten Heavy Metal Alloys [6.1,6.2,6.26,6.27,6.29]

The term tungsten heavy metal or heavy metal is used for a group of two-phase
composites, based on W-Ni-Fe and W-Ni-Cu. They are characterized by a high density
and a novel (unique) combination of strength and ductility.
Tungsten is the main component of the alloys (typically present in the range of 90 to
98 wt%) and the reason for their high density (between 17 and 18.5 g/cm\ Nickel, iron,
and copper serve as a binder matrix, which holds the brittle tungsten grains together and
which makes the alloys ductile and easy to machine. A typical microstructure is shown in
Fig. 6.13. It consists of spherical tungsten grains (20 to 60 11m in diameter), which are
embedded in a tough, metallic matrix. While the grains are nearly pure bcc tungsten, the
fcc binder matrix contains about 20 wt% W in solid solution.
Heavy metals are used for applications, where the high specific weight of the material
plays an important role. They are used as counterweights, rotating inertia members, X-ray
and y-radiation shields, as rigid tools for machining, as well as for defense purposes
(kinetic energy penetrators, fragmentation devices, etc.).
Fabrication [6.26,6.27]. Heavy metals are produced by conventional P1M techniques. A flow chart of the fabrication process is shown in Fig. 6.14. Elemental powders(W,

TUNGSTEN ALLOYS

271

FIGURE 6.13. Optical micrograph of a 95 wt% tungsten heavy metal alloy with 3.2 wt% Ni and 1.8 wt% Cu. By
courtesy of Plansee AG, Reutte, Austria.

Fe, Ni, eu) are blended in mixers or ball mills to the desired ratio, compacted to form a
green body, and subsequently liquid-phase sintered. Assuming proper manufacturing
conditions, they exhibit full or near-theoretical density in the as-sintered condition.
Powder particle sizes are in the range of 2 to 6 11m. Both die pressing and isostatic
pressing (dry- and wet-bag pressing) are in use. No lubricant is commonly added, since the
green strength is high enough to handle the compacts. Powder injection molding (PIM) is
used for applications where net shaping is desired and large quantities of complex parts are
produced [6.28].
Sintering is commonly carried out in molybdenum-wire resistance-heated furnaces
under hydrogen or nitrogen mixtures (dissociated ammonia) but can also be performed in
vacuum units. The use of wet hydrogen has become industrial practice to suppress
hydrogen embrittlement (water vapor porosity) [6.29]. The temperature/time program of
the sintering cycle must be adjusted to the composition and size of the sintered parts. A

FIGURE 6.14. Flowchart of the fabrication process of tungsten heavy metal penetrators [6.26].

272

CHAPTER 6

cleaning step in hydrogen at 1000C is commonly performed to render the outgassing of

volatile compounds. High-purity powder grades must be used for sintering. Otherwise,
blistering will occur on liquid-phase sintering, and interface precipitations will occur on
cooling.
Isothermal sintering is carried out above the eutectic temperature, typically between
1440 and 1500 C, but can be as high as 1600C. The higher the tungsten content of the
alloy, the higher the temperature. Tungsten-nickel-copper alloys are sintered at somewhat
lower temperatures than tungsten-nickel-iron alloys. Sintering times are between 30
minutes and two hours. Prolonged sintering under dry hydrogen leads to swelling of the
parts and significant embrittlement due to pore coarsening by Ostwald ripening and
coalescence. Residual porosities larger than 0.5% drastically reduce the ductility
[6.30,6.31].
On cooling, a significant amount of tungsten remains in solid solution, depending on
the binder composition and the cooling rate. The solubility is highest in binary W-Ni
alloys (up to 40 wt%; resulting in low ductility), but additions of Fe and Cu depress it to
lower values (typically 20 to 25 wt% W), providing a tough and ductile binder matrix.
Shrinkage porosity can form on cooling, in particular in W-Ni-Cu alloys and at high
cooling rates; furthermore, impurities (P, S, C) segregate at the tungsten matrix interface.
Both effects must be controlled by proper manufacturing conditions, since they significantly lower the ductility.
Optimal mechanical properties require subsequent heat treatment of the alloys.
Solution annealing at 900 to 1300C and subsequent quenching to avoid impurity
segregation and formation of intermetallic phases significantly improves ductility [6.26].
Heavy metals can be cold-worked (swaged, rolled) to increase hardness and strengths at
the expense of elongation and toughness. Aging at 500C after 25% deformation (cold
work) is a compromise to achieve both high ultimate tensile strengths and elongation.
Conventional procedure is a double swaging-heat treatment cycle (deformation 25-30%)
[6.32]. The structure shows that the original spheroids formed during liquid-phase
sintering have been transformed to ellipsoids on deformation (Fig. 6.15). Recently, the
advantage of large deformation levels (up to 95%) yielding in a fibrous microstructure
(Fig. 6.16) has been demonstrated [6.32]. Heavy metals exhibit a ductile-to-brittle
transition temperature. In comparison to pure tungsten it is, however, not a sharp transition
but spreads over several hundred degrees [6.1].
Although a considerable amount of densification already occurs during solid-state
heating, isothermal liquid-phase sintering is a prerequisite to obtain near-theoretical
density and a high degree of microstructural homogeneity. Rapid final densification
occurs on formation of the liquid phase under the action of capillary forces. Particle
rearrangements, solutionjreprecipitation processes (Ostwald ripening), and coalescence
contribute to a higher packing density and a significant grain coarsening during sintering.
The original tungsten powder grain size of 3-5 ~m is transformed during sintering to
rounded tungsten particles (spheroids) of at least 10 times the original grain diameter.
Shape accommodation (formation ofpolyhedras with rounded comers) plays an important
role in high tungsten alloys, where only a small quantity ofliquid is available for achieving
full density.
Depending on the binder composition, heavy metals can be classified into two main
groups [6.2,6.27]:

TUNGSTEN ALLOYS

273

FIGURE 6.15. Microstructure of a W-Ni-Fe-Co alloy after swaging by 30% [6.32]. By courtesy of A. R.
Bentley, M. C. Hogwood, and M. Power, Defence Research Agency, Kent, England.

(a) W-Ni-Fe (Ni: 1-7 wt%, Fe: 0.8-3 wt%, Mo: 0-4 wt%)
This group is ferromagnetic. Typical nickel-to-iron ratios range between 1: 1 and 4: 1.
The preferred ratio is 7:3, since this composition avoids the formation of intermetallic
phases [6.26]. W-Ni-Fe alloys exhibit excellent strength/ductility combinations and can
be cold-worked to a reduction of 60% without intermediate annealing. Molybdenum acts
as grain refiner. Higher additions of iron and/or additions of molybdenum cause a
significant matrix-strengthening effect and improve high-temperature strength (Gyromet
grades) [6.2]. Other additives, such as Co, Ta, and Re, act as grain refiners and increase
hardness and strength, but lower the ductility [6.29]. Rhenium additions are of interest in
net-shape production, since their high strength in the as-sintered condition (1180 MPa)
does not require any post-sintering treatment [6.33].

FIGURE 6.16. Microstructure of a W-Ni-Fe-Co alloy after defromation by 80% [6.32]. By courtesy of A. R.
Bentley, M. C. Hogwood, and M. Power, Defence Research Agency, Kent, England.

CHAPTER 6

274

(b) W-Ni-Cu (Ni: 1-7wfllo, eu: 0.5-3 wt%, Fe: 0-7 wt%)
Members of this group are nonmagnetic and exhibit a higher electrical conductivity.
The nickel-to-copper ratio ranges from 3:2 to 4:1. W-Ni-Cu alloys exhibit lower strength
and ductility than comparable W-Ni-Fe alloys. Due to the low melting point of copper,
low heating rates are required to obtain full density .
Properties: [6.1,6.34,6.35]. Some properties of tungsten heavy metal alloys are
summarized in Table 6.1. Ultimate tensile strengths (660-1350 MPa), yield strengths
(565-300 MPa), and fracture elongations (5-30%) can vary widely. In general, mechanical
properties are very sensitive to processing conditions, impurities, and microstructure.
Problems with controlling porosity, impurities, and microstructural homogeneity are
therefore common in heavy alloy fabrication [6.31]. Residual sinter porosity and the
formation of interface precipitates are the main reason for inferior material properties.
With increasing tungsten content, the contiguity of the tungsten grains increases (i.e.,
the W-W interfacial area as a fraction of the total interface area) and both strength and
ductility decrease [6.36]. This "microstructural limitation" of the properties is demonstrated in Fig. 6.17 for alloys of different tungsten contents [6.1, 6.31, 6.3 7]. While the

TABLE 6.1. Properties of Tungsten Heavy Metal Alloysa

Properties
Density (g. cm - 3)
Young's modulus (GPa)
Shear modulus (GPa)
Poisson's ratio
Hardness
HVIO

17.0--18.5
350-400
0.28-0.29
270--360

HB30

17.0--18.5
320--380
125-160
0.28---{).29

17.1-18.6

270-470
250-450

RC
Tensile strength (MPa)
20C
100C
500C
1000C

Yield strength (0.2%) (MPa)

Fracture elongation (%)
Compressive strength (MPa)
Kjc (MPamm- l )
Compressibility (% of length)
Charpy value (notched)
Specific electrical resistivity (Jill. m)
Mean linear coefficient of thermal expansion
at 20--800C (mm - I . K -I)
Half-value thickness
against 60Co
against 137Cs
a

28-40
870--1000
340-650
220--260
600--700
10--30
3500-4500
800-6300

660--1350
560--730
340-610
90--260
565-1300

880--1320

750--1240
5-30

3500--5500
45-60

0.9-2.8
0.10-0.18
5.2-6.5 x 10- 6

0.10-0.18
5.2-6.5 x 10- 6

8.5-9.7
4.5-5.2

1. Pink and R. Eck, "Refractory Metals and their Alloys," in: Materials Science and Technology (R. W Cahn, P. Haasen, and
E. 1. Kramer, eds.), Vol. 8, pp. 591-638, VCH, Weinheim, (1996).
2. Metallwerk Plansee, Densimet Schweremetall Legierungen, Company Brochure, Reutle, Austria (1982).
3. Ashot Ashkelon, Tungsten-Based Products, Company Brochure.

275

TUNGSTEN ALLOYS

1.0

40

1200

--

*.....

--a..

1000

30 "
C

(.)

til

0.8

:;:

til

0)

20 c
0
Qi

800

C/)

0.6

...
Q)

10

600

400
90

0
94
96
98
92
Tungsten content (wt.%)

...til
LL

0)

:;:

...:::I

::>

...:>.

0.4

0.2

100

FIGURE 6.17. Optimum values of room-temperature strength and fracture elongation and their correlation with
the contiguity factor for tungsten heavy alloys [6.1,6.31].

trend is most pronounced for the fracture elongation, the tensile strength exhibits a
maximum at about 85 vol% W (93 wtOlo) before it significantly decreases. At 99 wt% W,
the elongation is zero and the tensile strength is about 400 MPa [6.38].
Recently, it was demonstrated that large deformation levels yield exceptionally high
tensile strength levels (up to 1700 MPa), combined with fracture elongations above 15%
[6.32].
Besides the high density and the unique combination of high strength and ductility,
there are other attributes which make heavy metals a versatile product:

the high modulus of elasticity (much higher than steel),

excellent vibration damping characteristics (for chatter-free heavy machining),
its good machinability,
the high absorption ability for X-rays and y-rays,
good thermal and electrical conductivity,
low electrical erosion and welding tendency,
good corrosion resistance.

Research in tungsten heavy alloys was previously boosted by the ballistic application
as anti-amor kinetic energy penetrator. Numerous papers have appeared in the last years
and reviews have recently been published [6.29,6.39,6.40]. Although the basic requirements are the same for civil as for defense applications (i.e., high density, strength, and
elongation), there are two more important factors which must be considered for this
specific application: the behavior of the material under high strain rate conditions and their
ballistic performance (i.e., their penetration ability). In particular, the latter aspect is of
critical importance and, to a certain extent, still a weak point of heavy metals. Their
ballistic performance is inferior compared to depleted uranium, which is still used as a

276

CHAPTER 6

standard penetrator material [6.29]. Nevertheless, recent environmental considerations

have put a strong emphasis on substituting depleted uranium by heavy metals because of
its radioactivity. Efforts to improve the ballistic performance through proper processing
and compositional modifications have failed [6.29,6.39]. Recent research has therefore
focused on alternative matrix alloys, such as tungsten-hafuium, tungsten-uranium
composites [6.29], or heavy metal alloys with a spiculating core of we [6.41].

6.3.2. Tungsten Copper and Tungsten Silver [6.1,6.2,6.42].

Tungsten--copper and tungsten-silver composites are widely used in mechanical and
electrical engineering. Typical applications include high-, medium-, and low-voltage
circuit breakers (W-Cu, W-Ag), resistance welding electrodes, electrode materials for
electrical discharge machining, and heat sink materials for microelectronic packaging (Weu). More recently, W-Cu composites have been tested as heat flux components in
experimental fusion reactors [6.43] and as materials in MHD (magnetohydrodynamics)
power generation systems [6.44].
They combine the high hardness, hot strength, and wear resistance of tungsten with
the outstanding electrical and thermal conductivity of the two high-conductivity metals.
Furthermore tungsten increases the resistance of the materials against spark and arc
erosion (bum-off) and lowers the sticking and welding tendency, which are both important
criteria for heavy-duty electrical contacts, where switching currents can be up to 100 kA
and arc temperatures of a few 10,000 K can occur within a few milliseconds [6.45]. Under
arcing conditions, the contacts are cooled through the melting and evaporation of silver or
copper ("transpiration" cooling), an effect which was earlier used for rocket nozzle throat
liners (for example for underwater-launched ballistic missiles) made of W-Ag [6.2].
w-eu and W-Ag composites are not real alloys in the strict meaning of the word,
because the mutual solubility of the components is practically zero. Therefore, they are
also sometimes called "pseudoalloys."
Production. Powder metallurgy is the only viable way to produce composites of high
quality. The method of production depends on the composition ratio. Materials with 10 to
40 wt% copper (20 to 50 wt% silver) are commonly produced by infiltration, while at
higher copper or silver contents the powders are blended, pressed, and subsequently solidstate sintered [6.42,6.46].
Infiltration of porous structures. A porous tungsten skeleton is first produced by
pressing either pure tungsten powder or a mixture of tungsten powder and small amounts
of the additive element with subsequent sintering in a reducing atmosphere at comparably
low temperatures (l200-1800C). The porosity of the compact can be varied by the
pressing and sintering parameters, as well as by the grain size distribution of the starting
tungsten powder. It determines the percentage of infiltrate. For higher copper and silver
contents, loose tungsten powder can also be infiltrated.
Infiltration is accomplished by immersion, dipping, or flooding, either in vacuum,
inert gas, or, preferably, in reducing atmosphere at 1150-1250 0c. Oxide layers disturb the
infiltration process and must be removed prior to infiltration by reduction with hydrogen.
Trace impurities can impair the wetting between tungsten and the melt (for example,
silicon) but can also improve it (nickel, cobalt, iron) [6.47]. Densities of 96% up to near-

TUNGSTEN ALLOYS

277

theoretical can be obtained. The presence of some closed pores in the sintered skeleton
prevents full density.
Machining is required to finish the infiltrated parts.
Sintering. Liquid-phase sintering oflow-copper (silver) composites results in inferior
product qualities, due to a higher degree of residual porosity. Shrinkage during sintering is
difficult to control, and large distortions of the parts commonly occur. However, very fine
powder grades, which render a better particle rearrangement during sintering, can be
sintered to full density [6.48]. Additions of nickel, cobalt, or iron improve the sintering
behavior, but lower the conductivity. Powder injection molding of complex shapes for
large volume fabrication of W-Cu parts used for microelectronic packaging and subsequent liquid-phase sintering might be an interesting alternative to infiltration [6.28].
Sintering of high-copper and silver alloys is carried out below the melting point of the
two metals. In such alloys, tungsten does not form a contiguous skeleton, as in the case of
infiltration. About 92% of the theoretical density can be obtained. Parts are repressed after
sintering or can be hot-worked (rolled, swaged, extruded) to achieve higher densities.
A typical microstructure of an infiltrated W-20Cu alloy is shown in Fig. 6.l8. It
consists of tungsten particles which are embedded in a soft copper matrix. The particles
form a ridged skeleton, which is an important prerequisite for obtaining optimal
mechanical properties. As the components Wand Cu have no mutual solubility, no
appreciable change in the particle size and morphology of tungsten takes place during
infiltration, i.e., the particle size and morphology of the tungsten powder approximately
matches the particle size and morphology of tungsten in the W-Cu alloy, though in
densified form.
Composition and Properties. The tungsten content in contact materials varies
between 10 and 90 wt%, but composites between 70 and 90 wt% W, and W particle

FIGURE 6.18. Microstructure of tungsten infiltrated with 20 wt% copper. By courtesy of Plansee AG, Reutte,
Austria].

278

CHAPTER 6

sizes between 5 and 50 11m are the standard materials for arcing contacts [6.42,6.45].
Compositions between 25 and 40 wt% Cu are preferably used for electrical discharge
machining [6.49]. For special switches, WC-(20-65) Ag and WC-(30-50)Cu composites
are used.
W-Ag exhibits better oxidation and corrosion resistance than W-Cu and has higher
current-carrying capacity. They are used in switching devices operated in air. However, the
costs are significantly higher. W-Cu composites are preferably used in vacuum, inert gas,
and oil-switching devices.
Some important electrical and mechanical properties of W-Cu and W-Ag alloys are
summarized in Table 6.2 (see page 279).
At room temperature the composites behave in a brittle manner. However, they are
readily machinable using cemented carbide tools (ISO KlO). The arcing contacts can be
clamped, brazed, or directly infiltrated onto the circuit-breaker components (commonly
pure copper or copper alloys).
6.3.3. Tungsten-Fiber-Reinforced Composites
Tungsten metal and tungsten alloy wires have attracted interest for fiber reinforcement
of high-temperature materials, in particular of superalloys. However, until now, they have
remained "exotic" materials with only a few specific applications or for model studies of
mechanical behavior and are not produced on a large scale. This is at least partially due to
the difficulties encountered during composite fabrication and the resulting high cost of the
material.
When a ductile and oxidation-resistant matrix is used, it protects the tungsten from
oxidation. The specific strength of the composite is much higher than that of the matrix
alloy, especially at elevated temperature. Strength and ductilities of typical tungsten-wire
reinforced superalloys, measured in the wire direction, are shown in Fig. 6.19 for the
temperature range of 20 to 1000 C, together with a micrograph of a composite sectioned
perpendicular to the fiber axis. Other than in the wire direction, the properties of the
materials are dominated by the respective matrix [6.50]. Typical fiber diameters are in the
range of 100 to 500 11m.
Most studies on wire reinforcement were performed on superalloy matrices (Incoloya,
Waspaloy, FeCrAIY), in particular at temperatures between 800 and 1200C, where the
usual strengthening mechanisms in superalloys begin to fail [6.50]. The most promising
research today is on the fabrication of high-pressure turbopump blades for the space shuttle
main engine [6.51,6.52]. Reinforcement by NS tungsten fibers was shown to be superior
in regard to stress-rupture properties as compared to undoped tungsten, but even higher
strength was obtained using a W-Re-HfC wire [6.52]. This wire has a tensile strength of
more than 6 times the strength of the strongest nickel- and cobalt-base superalloys at
rv 1100 C [6.51]. A Waspaloy /W-24Re-HfC wire reinforced composite was shown to be
most promising [6.52].
Other interesting matrices are niobium alloys (investigated for long-term, hightemperature applications in space power systems) and copper and silver for electrical
applications and heat exchangers [6.50,6.51].
Tungsten fiber composites can be produced either by infiltration of the fiber bundle by
the molten matrix, or by bringing together the fibers and matrix in the solid state and

180
180

Hardness HB 10

0.050
0.050
35
134
134

Electrical conductivity (% IACSt

Thermal conductivity
(W m-IK- I)

138

36

0.042

160

390

225

14.5

W25Cu

147

43

0.040

150

390

155
155

47

0.037
0.038

150
130

340
340-380

200
185

13.5
13.5

14.0
230

W35Cu

W30Cu

176

51

0.033

140

290

180

12.7

W40Cu

195

58

0.026

120

250

150

11.6

W50Cu

138
138

35

0.050
0.050

8.7

180
180

340
240-470

159
154

47

0.037
0.040

110
100

290
300

132
135

14.0
14.0

15.6
15.6
205
210

W35Ag

W20Ag

First Line: Metallwerk Plansee Elmet Kontaktwerksto/fe, Company Brochure, Reutte, Austria (1982).
Second Line: E. Pink and R. Eck, Refractory Metals and their Alloys, in: Materials Science and Technology (R. W. Cahn, P. Haasen, and E. 1. Kramer, eds.), Vol. 8. pp. 591--{)38, VCH, Weinheim
(1996).
% lACS = % International Annealed Copper Standard.
100% IACS=58.00j.lQ-lm- 1; 1/% lACS = 1.7241 j.lQm.

(Jill. m)

Specific electrical resistivity

8.5

440
400-460

Tensile strength at 20C

(MPa)

Mean linear coefficient of thermal

expansion (mm- I . K- I)

225
230

15.4
15.4

Density
(g. cm- 3 )

Young's modulus
(GPa)

W20Cu

PropertiesU

TABLE 6.2. Important Properties of W-Cu and W-Ag Pseudoalloys

-l

N
-..I

[/J

:
r

tn
Z

-l

[/J

CI

a..

10

r==-___

11111111

500
TEMPERATURE (-C)

embrittled
fibers

1000

FIGURE 6.19. Micrographs of a tungsten-wire reinforced superalloy sectioned perpendicular to the wire direction (left); strength and ductilities of typical tungsten-wire
reinforced superalloys with 40 vol% fibers (right) [6.50]. By courtesy of R. Warren, Lulea University of Technology, Sweden.

u..w

I-<.!)

~~
Uz
~9

wo

~ 20
!!...
Z

I-

z
w

en

....J

en

I-

lLJ
Il::'

<.!)

I-

::r:

<.!)

...

0\

-l

."

n
::c:
;.>

QC

TUNGSTEN ALLOYS

281

subsequent solid-state consolidation at comparatively low temperatures (approximately

1100 QC). Excessive fiber/matrix interactions are the main disadvantages of the liquidphase route, leading to the partial dissolution of tungsten in the matrix, the formation of
brittle intermetallic phases, and a partial recrystallization of the tungsten wire, making it
weak and brittle. Diffusion barrier coatings (refractory oxides, carbides, or nitrides) have
therefore been developed, but further increase the cost of the material. For solid state
processing, the danger of reaction between matrix and fibers is less, due to the lower
temperatures. It is essential to have the matrix either in wire or in sheet, foil, or powder
form. Consolidation is achieved by mechanical pressing and subsequent solid state
sintering or by hot pressing.

REFERENCES FOR CHAPTER 6

6.1.
6.2.
6.3.
6.4.
6.5.
6.6.
6.7.
6.8.
6.9.
6.10.
6.11.
6.12.
6.13.
6.14.
6.15.
6.16.
6.17.
6.18.
6.19.
6.20.
6.21.
6.22.
6.23.
6.24.

E. Pink and R. Eck, "Refractory Metals and Their Alloys," in: Materials Science and Technology (R. W.
Cahn and E. J. Kramer, eds.), Vol. 8, pp. 591-638, VCH, Weinheim (1996).
S. W. H. Yih and C. T. Wang, Tungsten, Plenum Press, New York (1979).
G. T. Hahn, A. Gilbert, and R. I. Jaffee, in: Refractory Metals and Alloys II (M. Semchyshen and I.
Perlmutter, eds.), Vol. 17, pp. 23-63, Wiley, New York (1962).
W. D. Klopp, J Less-Common Met. 42 (1975), 261-278.
S. Geraldi, Metals Handbook, 10th edition, Vol. 2, pp. 565-571, ASMI, Materials Park, Ohio (1990).
O. Kanou, in: Proc. 7th Int. Tungsten Symposium, pp. 410--418, lIlA, London SW90QT, UK (1996).
C. L. Briant and B. P. Bewlay, MRS Bulletin, pp. 67-73 (August 1995).
T. Millner and J. Neugebauer, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha,
eds.), pp. 1-13, Elsevier, London (1989).
G. D. Rieck, Tungsten and its Compounds, Pergamon Press, Oxford (1967).
D. M. Moon, R. Stickler, and A. L. Wolfe, in: Proc. 6th Int. Plansee Seminar (F. Benesovsky, ed.), p. 67,
Springer-Verlag, Vienna-New York (1969).
D. B. Snow, Metall. Trans. 3 (1972), 2553.
W. D. Schubert, B. Lux, and B. Zeiler, Int. J Refract. Met. Hard Mater. 13 (1995), 119-135.
J. A. Mullendore, in: The Metallurgy ofDoped/Non-Sag Tungsten (E. Pink and B. Bartha, eds.), pp. 61-80,
Elsevier, London (1989).
Sh. Yamazaki, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and B. Bartha, eds.), pp. 47-59,
Elsevier, London (1989).
O. Schob, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 83-111,
Elsevier, London (1989).
O. Horacsek, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 175-187,
Elsevier, London (1989).
S. Leber, R. J. Arena, and D. L. Biy, Metals Handbook, 9th edition, Vol. 7, pp. 629-631, ASMI, Materials
Park, Ohio (1984).
R. Eck, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 285-291,
Elsevier, London (1989).
R. Eck and E. Pink, in: Proc. 13th Plansee Seminar (H. Bildstein and R. Eck, eds.), Vol. I, pp. 16-25,
Metallwerk Plansee, Reutte, Austria (1993).
A. Luo and D. L. Jacobson, in: Proc. 13th Plansee Seminar (H. Bildstein and R. Eck, eds.), Vol. I, pp. 263277, Metallwerk Plansee, Reutte, Austria (1993).
W. J. van den Hoek and A. G. Jack, "Lamps," in: Ullmann's Encyclopedia of Industrial Chemistry, Vol.
A15, pp. 115-150, VCH, Weinhein (1996).
Electrodes for TIG Welding, company brochure, Metallwerk Plansee, Reutte, Austria (1986).
O. Prause, in: Proc. 7th Int. Tungsten Symposium, pp. 289-293, lIlA, London SW90QT, UK (1996).
J. Resch and G. Leichfried, in: Proc. 7th Int. Tungsten Symposium, pp. 294-311, lIlA, London SW90QT,
UK (1996).

282
6.25.
6.26.
6.27.
6.28.
6.29.
6.30.
6.31.
6.32.
6.33.
6.34.
6.35.
6.36.
6.37.
6.38.
6.39.
6.40.
6.41.
6.42.
6.43.
6.44.
6.45.
6.46.
6.47.
6.48.

6.49.
6.50.
6.51.
6.52.

CHAPTER 6
K. S. Shin, A. Luo, and D. L. Jacobson, JOM, pp. 12-15 (August 1990).
T. W Pernice, Metals Handbook, 9th edition, Vol. 7, pp. 688-691, ASMI, Materials Park, Ohi<> (1984).
R. F. Cheney, Metals Handbook, 10th edition, Vol. 2, pp. 392-393, ASMI, Metals Park, Ohio (1990).
R. M. German, in: Proc. 14th Plansee Seminar (G. Kneringer, P. Rodhammer, and P. Willhartitz, eds.), Vol.
I, pp. 194-206, Plansee AG, Reutte, Austria (1997).
W D. Cai, Y. Li, R. J. Dowding, F. A. Mohamed, and E. J. Lavemia, Rev. Particulate Mater. 3 (1995), 71132.
R. M. German and K. S. Chum, Metall. Trans. 15A (1984), 747-752.
R. M. German, L. L. Bourguignon, and B. H. Rabin, J. Met. 37, No.8 (1985), 36-39.
M. C. Hogwood and A. R. Bentey, in: Tungsten and Refractory Metals 2 (A. Bose and R. J. Dowding, eds.),
pp. 37-45, MPIF, Princeton, N.J. 1995).
A. Bose, G. Jennan, and R. M. German, Powder Metall. Int. 21, No.3 (1989), 9-13.
Densimet Heavy Metal Alloys, company brochure, Metallwerk Plansee, Reutte, Austria (1982).
Tungsten-Based Products, company brochure, Ashot Ashkelon, Israel.
H. Danninger and B. Lux, in: Proc. 14th Plansee Seminar (G. Kneringer, P. Rodharnmer, and P. Wilhartitz,
eds.), Vol. I, pp. 315-328, Metallwerk Plansee, Reutte (1997).
B. H. Rabin, A. Bose, and R. M. German, Int. J. Powder Metall. 25, No. I (1989),22-27.
R. M. Germain, in: Tungsten and Tungsten Alloys-1992 (A. Bose and R. J. Dowding, eds.), pp. 37-45,
MPIF, Princeton, N.J. (1992).
R. J. Dowding, M. C. Hogwood, L. Wong, and R. L. Woodward, in: Tungsten and Refractory Metals 2 (A.
Bose and R. J. Dowding, eds.), pp. 3-9, MPIF, Princeton, N.J. (1995).
L. S. Magness and D. Kapoor, in: Tungsten and Refractory Metals 2 (A. Bose and R. J. Dowding, eds.), pp.
11-20, MPIF, Princeton, N.J. (1995).
L. Ekbom, L. Holmberg, and A. Persson, in: Tungsten and Tungsten Alloys-1992 (A. Bose and R. J.
Dowding, eds.), pp. 551-558, MPIF, Princeton, N.J. (1992).
Y. Shen, P. Lattari, J. Gardner, and H. Weigard, Metals Handbook, 10th edition, Vol. 2, pp. 840-868, ASMI,
Materials Park Ohio (1990).
Y. !toh, M. Takahasi, and H. Takano, Fusion Eng. Design 31 (1996), 279-289.
L. C. Farrar and J. A. Shields, Jr., JOM, pp. 30-35 (August 1992).
L. Zehnder, in: Proc. 7th Int. Tungsten Symposium, pp. 317-335, ITIA, London SW90QT, UK (1996).
H. Mayer, in: Pulvermetallurgie und Sinterwerkstoffe (F. Benesovsky, ed.), pp. 179-189, Metallwerk
Plansee, Reutte (1982).
H. Danninger and B. Lux, in: Proc. 13th Plansee Seminar (H. Bildstein and R. Eck, eds.) Vol. I, pp. 486500, Metallwerk Plansee, Reutte (1993).
J. L. Johnson, K. F. Hens, and R. M. German, in: Tungsten and Refractory Metals 2 (A. Bose and R.
Dowding, eds.), pp. 245-252, MPIF, Princeton, N.J. (1994).
Electrode Materials for Electrical Discharge Machining, company brochure, Metallwerk Plansee, Reutte,
Austrialia (1985).
R. Warren, in: The Metallurgy of Doped/Non-Sag Tungsten (E. Pink and L. Bartha, eds.), pp. 293-301,
Elsevier, London (1989).
T. Grobstein and D. W Petrasek, Metals Handbook, 10th edition, Vol. 2, pp. 582-584, ASMI, Materials
Park, Ohio (1990).
F. J. Ritzert and R. L. Dreshfield, in: Tungsten and Tungsten Alloys-J992 (A. Bose and R. J. Dowding,
eds.), pp. 455-462, MPIF, Princeton, N.J. (1992).

7
Tungsten and Tungsten Alloy
Products
Besides tungsten metal, only those alloys which were described in Chapter 6 are treated in
this chapter. Alloys produced by melting metallurgy as well as the cemented carbides
(hardmetals) and their application are treated in Chapters 8 and 9, respectively.
The application of tungsten metal which, at the beginning of the 20th century, was
used only as a filament in incandescent bulbs and as alloying element in steel, as well as of
its alloys, is nowadays very widespread and covers quite different fields such as lighting,
electronics, high-temperature technology, medicine, aviation, military uses, sports, and so
on.
In order to facilitate use of this survey, the application of tungsten and of its alloys is
divided into 13 different fields. The material used for the application(s) is given in
parentheses. (Note: HM is used for Heavy Metal).

7.1. LIGHTING APPLICATION [7.1-7.7]

About 4-5% of annual worldwide tungsten production is consumed by the lighting

industry. The largest market is still for incandescent lamps (worldwide production in 1995:
~ 10.5 billion lamps) [7.8]. However, more than 70% of artificial lighting is today
generated by fluorescent lamps [7.1] and this portion is steadily increasing. Modem
production units can produce more than 6000 lamps per hour [7.1]. Tungsten is used in the
form of wires, coils, and coiled coils in incandescent lamps, and as electrode in low- and
high-pressure discharge lamps.
The most important artificial light sources are summarized in Table. 7.1 in terms of
their main lighting characteristics and applications. From this table it is clear that the
efficiency of the light sources (efficacy in 1m . W- 1) has significantly improved over the
years, in particular through the development of discharge lamps. Modem fluorescent lamps
consume 80% less energy than an incandescent lamp of similar light output. Their cost
effectiveness can save more in energy and replacement than their initial purchase price.
Calculations have indicated that more than 40% of the energy used worldwide for lighting
283

CHAPTER 7

284

TABLE 7.1. Standard Lamp Types, Lamp Efficacy, and Typical Applications G

Type

Maximum
lamp
efficacy
(lm W- I )

Burning
life
(hours)

Edison lamp
Incandescent
lamps
Halogen

1.4
22

40
1,000

27

2,000

Fluorescent

104

7,500

63

12,000

94
125

6,000
12,000

200

10,000

Gas discharge
High-pressure
mercury
Metal-halide
High-pressure
sodium
Low-pressure
sodium

Main characteristics

Typical applications

Carbon filament/1879
Easy to install, wide
variety; low cost price
Compact; high light
output; longer life
Wide choice of colors;
economical in use

General lighting in the home;

decorative lighting
Accent lighting; floodlighting;
background lighting
Lighting in public buildings;
streetlighting; home lighting

High efficacy; reasonable

color quality
Excellent color rendering
Very high efficacy; good
color rendering
Extremely high efficacy;
poor color rendering;
monochromatic light

Sports grounds; industry;

residential areas
Floodlighting; plant irradiation
Public lighting; floodlighting;
industrial lighting
Wherever energy/cost
effectiveness is important, i.e.,
highway lighting

By courtesy of Philips Lighting B.v., The Netherlands [7.5].

could be saved simply by changing from standard to more advanced light sources, without
losing light quality [7.5].
Filaments in incandescent lamps (NS-W). Incandescent bulbs of various shapes and
colors dominate the market worldwide in regard to general lighting. They are used as
household lamps, automotive lamps, traffic signal lamps, reflector lamps, etc., but also for
many thousands of specialty lamps. Wire diameters are commonly in the range of 17 to
50 ~m, but can be as low as 6 ~m or as high as 1 mm. Power ratings are from a few watts
(watch lighting) to several thousand watts (lighthouse lamps; up to lO kW).
The overwhelming majority of filaments are made of NS-W (sag-resistant tungsten).
Only for special shock and vibration resistant lamps are W- Th0 2 or W-Re wires used.
Every year, about 20 billion meters of lamp wire are drawn, a length which corresponds to
about 50 times the earth-moon distance [7.8].
Large bulbs are used for conventional incandescent lamps to spread the evaporating
tungsten over a large area in order to minimize wall blackening. Operating temperatures
are between 1700 and 2500 DC. Household lamps have inert gas fillings (Ar-N2) and
operating pressures of about 1 bar.
Filaments in halogen lamps (NS-W of highest non-sagging quality). Halogen lamps
are used more and more for indoor and outdoor applications due to their higher luminous
efficiency as compared to conventional lamps. Bulb wall blackening is avoided by the
addition of halogen compounds (commonly mono-, di-, or tribromomethane) and a higher
inert gas pressure. Therefore, halogen lamps can be operated at higher temperatures (up to
3000 DC) and have about twice the lifetime (Table 7.1) of conventional incandescent
lamps.

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

285

The principle of the halogen cycle is shown schematically in Fig. 7.l [7.9]. In the
vicinity of the filament, where the temperature is above 1200 C (area C), only atoms of
tungsten and halogens are present. Near the quartz bulb wall (area A in Fig. 7.l) a
tungsten-halogen compound forms, which diffuses into the hot zone, where it dissociates.
The "brought back" tungsten is deposited on the filament. A minimum bulb temperature
of 250C and a maximum bulb temperature of about 900 C is necessary to maintain the
tungsten regenerative cycle.
In commercial bulbs tungsten--oxygen-halogen compounds will form rather than
tungsten-halogen compounds, because traces of oxygen are always present. Nevertheless,
the principle of regeneration is the same.
In case of iodine or bromine additions, the deposition of tungsten does not occur at
the place where it was evaporated, so hot spots will form by thinning in time, finally
resulting in failure of the lamp. Computer calculations have demonstrated that fluorine
additions would result in a selective deposition of tungsten: directly on the hot spot.
However, such additions are not practicable, because they would destroy the glass bulb
[7.1].
The variety of coils used in incandescent and halogen lamps is shown in Fig. 7.2.
Electrodes in fluorescent lamps (coated Wor NS-W). Fluorescent lamps are lowpressure mercury discharge lamps which radiate in the UV region. The ultraviolet radiation
is converted into light by means of a fluorescent layer (Phosphor). Tungsten coils, coated
with a mixture of Ca, Ba, and Sr compounds, are used as electrodes (emitter).
Electrodes in high-intensity discharge lamps (coated NS-W, W- Th0 2, NS-W- Th0 2,
W-Ce02' W-La203, and porous W). These lamps are light sources with high output, high
luminous efficiency, and long operating time (Table 7.1). Their main applications are
lighting of roads, outdoor areas, halls, shopping windows, agriculture, horticulture, and in
photography and medical technology. They were only introduced recently for automotive
lighting (Xe microdischarge lamps).
The following lamp types are in use:
High-pressure mercury lamps (mercury vapor pressure during operation:
105 Pa).
Low-pressure sodium lamps (sodium vapor pressure rv 0.5 Pa).
High-pressure sodium lamps (sodium vapor pressure 104 Pa).
Metal halide gas discharge lamps.

WXn -

W nX

WXn

4X

(high temperature)

00"'

W nX -

rv

W filament
I'-::;LJ--I_gas nX
filament 3000 o
1200 0
W nX _ WXn
600 0 e
- _ - W nX - WXn
min. 250 0 e --"';;;;:'_L__ bulb wall

e,TT
e

(low temperature)

FIGURE 7.1. Schematic presentation of the "halogen cycle" in a halogen lamp [7.9]. (W ... tungsten;
nX .. .halogen; WXn .. .tungsten-halogen compound). The cycle is described in the text.

286

CHAPTER 7

FIGURE 7.2. Tungsten filaments of different shape used in (a) traffic signal lamps, (b) projection/overhead
lamps, (c) entertainment lighting/studio/film lamps, (d) video/photo lamps. By courtesy of Philips Lighting B.V,
The Netherlands.

Coiled tungsten electrodes (Fig. 7.3) [7.l0] or tungsten rod electrodes, coiled with a
thinner tungsten wire or coil, all coated with emissive material, are used as electrodes.
Exceptions are metal-halide lamps where thoriated tungsten or W-Ce02 are used.
Emitters usually contain Ba, or mixtures of Ba, Ca, and Sr compounds (tungstates,
aluminates, tantalates) to lower the work function. Electrode temperatures are below
2000 C. All lamps have noble gas fillings (either neon/argon or xenon) to facilitate
ignition.
Figure 7.4 shows a section through a metal-halide lamp [7.11]. A short description of
the principle is given below.
Electrodes for Xe arc lamps and mercury Xe arc lamps (NS-W, W - ThOz, W-La203,
porous W). Xe arc lamps are used as white light source for film and video projection as
well as in scientific instruments (Fig. 7.5). Mercury-xenon arc lamps are used as an UV
light source in photolithography.
Lamp support wire or rod (w, NS-W).

7.2. ELECTRICAL ENGINEERING [7.7, 7.12, 7.13]

Rods as electrical contacts (W).

vacuum-tight conducting connections (W).
Contact rods to Hg switches (W).
Electrical contacts for high-frequency switching (W). Tungsten is an ideal contact
material for switches with high operation frequency, such as car hom contacts, ignition

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

287

FIGURE 7.3. Triple-coil tungsten electrodes coated with emissive material, used in low-pressure sodium lamps
[7.10]. By courtesy of Philips Lighting B.V, The Netherlands.

contact breakers, voltage regulators for automotive generators, etc. Optimal properties may
be achieved by a fine grained microstructure.
Electrical contacts in high-performance switches (W-Cu, W- Ag, WC-Ag). Tungsten-containing contact materials are used for high-voltage (72.5-800 kV), mediumvoltage (5-38 kV), and low-voltage (Sl kV) switches. The magnitude of switching
currents can attain 100 kA, and arc temperatures in excess of 10,000 K can prevail for
milliseconds during switching [7.13]. Evaporated copper (W-Cu) or silver (W-Ag, WCAg) cool the tungsten matrix ("sweat" cooling).
Typical shapes used in high-voltage switches are shown in Fig. 7.6.
Heating filaments for defrosting automobile shields (NS-W).

288

CHAPTER 7

FIGURE 7.4. Section through a metal-halide lamp: (I) mechanically fitted E40 screw base, (2) lead-in
wire/support, (3) sleeve protecting the support, (4) quartz discharge tube, (5) internal phosphor coating, (6)
elliptical hard glass outer bulb, (7) getter ring for maintaining a clean gas atmosphere. By courtesy of Philips
Lighting B. v., The Netherlands.
Light is obtained from an electrical discharge occurring between two spiral-like tungsten electrodes inside a
quartz tube. Within the tube is a mixture of two gases (neon and argon) to facilitate ignition. Also contained
within the tube are metallic mercury and three metals in iodide form (indium, thallium, sodium). After ignition,
the temperature increases and the three iodides as well as the mercury evaporate. After dissociation of the iodides,
the metals alone are excited and produce specific radiation mainly in the visible part of the spectrum. The internal
fluorescent coating converts UV radiation into visible light which further improves the color impression [7.11].

FIGURE 7.5. Thoriated tungsten cathode (left) and thoriated tungsten anode (right) for xenon short-arc lamps
used for wide-angle cinema projection. By courtesy of Plansee AG, Austria.

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

289

FIGURE 7.6. Contacts for high-voltage switch gear. By courtesy of Plansee AG, Austtia.

7.3. ELECTRONICS [7.7]

Electron emitter (w, W-Th02, W-La203, porous W). Other elements are better
emitters, but W has an extremely low vapor pressure and can be used at higher
temperatures (it keeps the filament free of gas contamination, which would decrease
emission). Therefore, W is practically the only material used in this field. Tungsten can be
used as cold (field emission) and as hot emitter (thermionic emission). The principles are
explained in Section 1.2.7. Field emission cathodes are used in field emission electron
microscopy (FESEM). Examples of thermionic emission are manifold:
Heater wires of oxide-coated metal cathodes in electronic tubes (W, W-Th0 2,
porous W infiltrated with barium-aluminum compounds). BaO coating reduces the
work function and thus increases the emission.
Directly heated cathodes for transmitter tubes, rectifier and electrometer tubes (WTh0 2, infiltrated porous W).
Cathodes of electron guns (W, W-Th02); for example, for electron beam
evaporation.
Filaments for magnetrons (NS-W-Th0 2) for microwave ovens.
Heavy-duty cathodes of high-power pulsed microwave tubes (impregnated porous
W); cathodes of this type are shown in Fig. 7.7.
Filaments of transmission and scanning electron microscopes (w, W- Th0 2).
Grid wires in vacuum tubes (NS-W-Th02' W-Th02)'
1ilcuum tight conducting connections (W). Ground pins for glass seals in high
vacuum tubes.
Tungsten charger wires (Corona wires) for laser printers, photocopiers, FAX
machines, etc. (w, gold-plated, Pt-clad W).
Ni or Au coated rounds (W). They serve as supports for silicon single-crystal plates
used in transistors, diodes, thyristors, and rectifiers.

290

CHAPTER 7

FIGURE 7.7. Porous tungsten cathodes with large surfaces for high-power microwave tubes. By courtesy of
Plansee AG, Austria.

Heat sinks (W-Cu) [7.14]. Downscaling and higher integration of electronic devices
(LSI via VLSI to ULSI) generates very high and localized heat release. An effective and
fast heat transfer to heat sinks of W-Cu composite materials of different W /Cu ratios is
necessary. The following properties are of importance:
High thermal conductivity and high thermal stability enabling a co-firing with the
ceramic substrates.
Low thermal expansion coefficient matching that of the co-fired ceramic capsule.
Good machinability and smooth surface finishing.
Different heat-sink geometries are shown in Fig. 7.8. They are used for submounting
chips, as substrates for IC, as base metal for H/F power transistors, as motherboard for
multichip LSI, and housing for H/F power amplifiers.
Thin films in microelectronic and optoelectronic devices (w, WSix , W-Ti). The
important properties are listed below.
Examples are source, drain, and gate metallization in transistors and diffusion
barriers, ohmic contacts, and interconnections (lines, vias, plugs) in integrated circuits.
The favorable electrical, mechanical, and chemical properties are:
high activation barrier for self diffusion,
high resistance to electron migration under high current densities,
low thermal expansion coefficient only marginally changing with temperature,
matching very well that of Si,
formation of ohmic contacts with Si and III/V semiconductors (GaAs or InP),
the only metal not reacting with III/V semiconductors up to 600-700 C,
does not form silicides <600C in contact with pure or n+ /p+ doped Si layers.
Thin films are deposited by CVD or sputtering.

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

291

FIGURE 7.8. W-Cu heat sinks for electronic devices. By courtesy of Sumitomo Electric Industries, Ltd., Japan.

7.4. HIGH TEMPERATURE TECHNIQUE [7.7, 7.15]

Heating element in high-temperature furnaces (W, W-La203)' The elements are

fabricated from wire, mesh, sheet, strip, or open seam tube for temperatures up to 2800 C
(for typical examples, see Figs. 7.9 and 7.10). Due to the high electrical conductivity of
tungsten, large currents and low voltages must be applied. Furnaces operate either in a
high vacuum or in a protective atmosphere.
Radiation shields in high-temperature furnaces (W). Shields closest to the tungsten
heating element are made of tungsten sheets.

FIGURE 7.9. Spacer parts, heating elements, and heat radiation shields of tungsten and molybdenum for a hightemperature furnace; height of the heating zone, approximately 440 mm. By courtesy of Plansee AG, Austria.

292

CHAPTER 7

FIGURE 7.10. Tungsten/molybdenum furnace for heat treatment of lamp components; charge carrier and heating
elements are made of tungsten. By courtesy of Plansee AG, Austria.

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

293

Charge carriers for high-temperature fornaces (W). Other furnace components are
rivets, pins, washers, spacers, and annealing boats (W).
Susceptors in high-frequency fornaces (W, W-La203)
Thermocouples (W/W-Re). These are employed for temperature measurement in
high-temperature furnaces in hydrogen or inert atmosphere or in vacuum, and in
combustion chambers of turbine engines or reactors (up to 2600 QC). Couples mostly
used are W-5Re/W-26Re and W-3Re/W-25Re.
Ionization chambers (W) for implantation of ions in the doping of silicon wafers.
Applications which might gain more importance in the future are Construction
material for the plasma chamber in magnetohydrodynamic power generation (Wand WCu) and target plates in fosion reactors (w, W-La203) [7.16, 7.17].
Recent plasma experiments and theoretical and numerical studies show that tungsten
may be the best, if not the only, material to withstand the extraordinary operating
conditions in a nuclear fusion reactor divertor. The divertor, being that part of the
vacuum vessel where the plasma particles interact with the first wall, and where a large
fraction of the fusion heat is removed, consists of water-cooled copper heat-exchanger
elements covered with a plasma facing armor. The plasma particles (electrons, protons, and
ex-particles) are directed by the magnetic field toward the divertor target plates, where they
are neutralized and pumped. The convective heat flux reaches 20 MW m- 2 and the
attendant surface temperature more than 3000 QC. Therefore, a suitable armor material
must have a high thermal conductivity (in order to transfer high heat fluxes), low thermal
expansion coefficient and low Young's modulus (in order to keep thermal stresses low),
and a high melting point and low sputtering yield (in order to keep erosion low). Although
tungsten does not have as high a thermal conductivity and as Iowa Young's modulus as
carbon--carbon composite materials, which are foreseen for the sections of the divertor
with the highest heat flux, many experts believe that, in the long run, reasonable lifetimes
will only be achieved by tungsten divertor plates, which have the lowest erosion rates of all
materials in sections of the divertor with relatively low plasma temperature but high
particle density.
In the current ITER (International Thermonuclear Experimental Reactor) divertor
design (November 1997; Fig. 7.11), almost the whole divertor surface is covered by
tungsten, representing more than 100 tons in the case of the ITER construction. During
ITER operation, the tungsten target plates will have to be replaced between 5 to 8 times
due to sputtering erosion [7.16].

7.5. WELDING, CUTTING, PLASMA SPRAYING, SPARK EROSION, AND VAPOR

DEPOSITION [7.7, 7.18]

Welding electrodes (W, W-Th02' W-La203, W-Ce02, and w-zr02). Their application is in inert gas welding (TIG) and plasma welding. W-Th02 or W-Ce02 carry 50%
more current than W, start more easily, and produce a more stable arc.
Hot cathode emitter for electron beam welding and cutting (W-Th02).
Electrodes (W-Th02' W-Cu) and nozzles in plasma guns (W-Th02). The guns are
used in plasma melting, arc cutting, and plasma spraying.
Electrodes for resistance welding (W-Cu, W-Ce02' W-La203).
Electrodes for electrical discharge machining, arc erosion, and wire eroding (w,

294

CHAPTER 7

FIGURE 7.11. Schematic view of the International Thennonuclear Experimental Reactor (ITER); 100 t of
tungsten are used for the ITER construction. By courtesy of Plansee AG, Austria.

W-Cu, W-Th02 ). Electrical discharge machining allows close dimensional tolerances,

accuracy of form, and a high-quality surface finish to be achieved. Both hardened steels
and hardmetals can be machined (Fig. 7.12). Tungsten in wire eroding is not only used as a
guide and contact for the wire electrodes, but also as erosion wire.
Heating coils or boats for vacuum metallizing (Wand NS- W). Tungsten is an
excellent material for evaporation sources. Coils (Fig. 7.13) and boats (Fig. 7.14) of
different size and shape are used. The main application is to evaporate aluminum, but a
variety of other metals and metal compounds can be evaporated too, as listed in Table 7.2.
Evaporization is performed for metallizing different products such as reflector lamps, head
lamps, television picture tubes, electronic components, plastic, etc.
Disks as sputtering cathodes (high purity W, W-Ti, W- Six, MoW). Sputter targets
are used to produce thin layers in integrated circuits (VLSI) by PVD. Recently, W-Mo
sputter targets were used for the manufacture of flat panel displays [7.19].

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

295

FIGURE 7.12. Electrical discharge machining of a compacting die used in the production of gear parts. By
courtesy of Plansee AG, Austria.

FIGURE 7.13. Tungsten evaporation coils for thin-film technology. By courtesy of Plansee AG, Austria.

296

CHAPTER 7

FIGURE 7.14. Tungsten, molybdenum, and tantalum evaporation boats. By courtesy of Plansee AG, Austria.

TABLE 7.2. Chemical Compounds which can be

Evaporated with Tungsten Evaporation Sourceso
Ag
Al
AI/O.I-2% Si
AI/l-4% Cu
AI/4% Cu/I% Si
Au
BaF 2
BaTi03
BeO
Bi
Bi20 3
CaF2
Cd
CdS
Ce02
Co
Cr
Cr-SiO
Cr2C3

Cr203
CsF
Cu
Er203
EU203
Fe
Fe203
Ga
GaAs
GaP
Gd20 3
Ge
Ge02
In
In/Sn
In203
La203
LU203
LuF3
Mg

By courtesy of Plansee AG, Austria.

MgO
Mn
MnS
NaF
Nb
Nb20 S
Nd20 3
Ni
Ni/Cr
Pb
PbF2
PbO
PbS
PbSe
Pd
Pr203
Pt
Sb
Sb20 3
SC20 3

Se
Si
Sm20 3
Sn
Ta20s
Tb20 3
Te
Ti
Ti20 3
TiO
Ti02
V
W0 3
Y
Yb20 3
Zn
Zn/Se
Zr
zrO
zr0 2

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

297

7.6. X-RAYS, RADIATION, MEDICAL ENGINEERING [7.7]

Static and rotating anodes in X-ray tubes (W and W-Re). For high performance
tubes W-Re is preferred. These anodes are used in modem medical technology (e.g.,
computerized tomographic scanning equipment). Only the surface layer consists of W /3lORe while the support body can be a Mo-Ti-Zr-C microalloy, called TZM, or graphite.
The W-Re layer can be produced either by powder metallurgical techniques or by
CVD. Compared to pure tungsten, which was employed in former times, the W-Re
alloy allows higher electron currents yielding in shorter irradiation times. The important
property is the increased stability to extreme alternating thermal stresses. Roughening of
the heating zone is much retarded in comparison to W, hence increasing the lifetime of
the tube.
Rotating anodes and a complete X-ray tube are presented in Fig. 7.15.
Directly heated cathodes in X-ray tubes (W).
Components for X-ray and radiation shielding (W, HM, HM-polymer) [7.207.23]. Typical examples are containers or flasks for radioactive materials and liquids,
shielded syringes (Fig. 7.16), shielding construction parts such as collimators in computer
tomographic scanners (Fig. 7.17), radiation therapy instruments, and containers and
shielding for oil prospecting using radioactive sources (Fig. 7.18).
Infrared reflector lamps (NS-W) used in medicine and health promotion (for
example, treatment of hyperbilirubinaemia in a baby).

FIGURE 7.15. Rotating X-ray target with graphite body and tungsten-rhenuim focal track. By courtesy of
Plansee AG, Austria.

298

CHAPTER 7

FIGURE 7.16. Heavy metal shielded syringes for applications in nuclear medicine. By courtesy of Plan see AG,
Austria.

7.7. MECHANICAL AND ENGINE ENGINEERING [7.20, 7.21]

Construction parts (W) for high-temperature nuclear reactors, fusion reactors, plasma
technique, and ion propulsion.
Combustion chamber of turbo engines (WRe).
Gyroscope rotors (HM) (Fig. 7.19).
Actuator in self-winding watches (HM).
Floats in flow meters (HM).
Governor balance weights (HM).
Anti-vibration tool holders (HM). Heavy metal is an excellent material for damping
torsional vibrations. Tools with longer shafts are typical applications. Examples are turning
shanks, boring bars, endmills, grinding quills, pinion type cutter supports, etc. (Fig. 7.20).
Counter balance weights (HM).
Flywheel weights (HM).

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

299

FIGURE 7.17. Collimator platelets for computerized tomography. By courtesy of Plansee AG, Austria.

FIGURE 7.18. Heavy metal containers and shieldings for oil prospection using radioactive sources. By courtesy
of Plansee AG, Austria.

300

CHAPTER 7

FIGURE 7.19. Heavy metal rings for gyroscope systems. By courtesy of Plansee AG, Austria.

FIGURE 7.20. Heavy metal boring bars with simple and multiple cutter heads. By courtesy of Plansee AG,
Austria.

7.8. LEISURE TIME AND SPORTS EQUIPMENT (FIG. 7.21)

Professional darts (HM).
Hunting cartridge shots (HM).
Diving weights (HM).
Golf club components (HM, W-Cu).
Hand weights for joggers (HM).

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

301

FIGURE 7.21. Professional darts and hunting cartridge shots. By courtesy of Ashot Ashkelon, Israel.

FIGURE 7.22. "Safeweight" nontoxic angler weights produced from tungsten-loaded polymers using injection
molding or extrusion. By courtesy of Royal Ordnance Speciality Metals Ltd. , England.

Tennis racket balance weights (HM).

Ski balance weights (HM).
Saddle weights (HM).
Fishing weights (Fig. 7.22).
Counter balance weights in Formula J racing cars (HM).

7.9. CHEMICAL INDUSTRY AND METALLURGY [7.23, 7.24]

Components for the production of quartz glass, transparent alumina tubes, and
ceramic single crystals (W, W-La203). Continuous production of quartz glass tubes is
conducted at temperatures higher than 2000 C in order to keep the viscosity of Si02
sufficiently low. Tungsten crucibles, drawing dies, mandrels, and heating elements are

302

CHAPTER 7

FIGURE 7.23. Tungsten-thoria guide nozzles. By courtesy of Plansee AG, Austria.

applied. Melting crucibles in seamless version are among the largest parts of tungsten
produced in one part (weight about 800 kg) [7.25].
Electrodes for melting of caolinit (W).
Spinning nozzles for ceramic wool production (W).
Crucibles for processing of used nuclear fuels and for handling of metal melts (W,
NS-W).
Tooling material for brass and aluminum die casting.
Spring material in corrosive media (Ta-2.5W, Ta-lOW). Preferably for use in
mineral acids and halide vapor.
7.10. SPACE AVIATION [7.7]

Ignition tubes for the main rocket engine (W-Re).

Guide nozzles of satellites for reliable positioning and stabilizing (W-ThOz, Ta2.5W; Fig. 7.23).
Thermionic emitter in spacecrafls (W, CVD-W, W-Re). The important properties are
high-temperature resistance and stability against cesium vapor.
Ion propulsion jet (porous W). Emitters of electrostatic ion propulsion systems.
Construction parts in space nuclear reactors (Wand tungsten fiber reinforced Nb
matrix alloys).
7.11. MILITARY APPLICATIONS [7.22, 7.26]

High kinetic energy penetrators in tank ammunition (HM, W-Hf). A schematic

drawing of a whole cartridge is shown in Fig. 7.24. Upon firing, when the projectile exits
the muzzle, gas and aerodynamic forces cause the sabot segments to separate from the
subprojectile. Muzzle velocity is '" 1700 m/ s. The fin-stabilized subprojectile continues
on its extremely flat trajectory to the target. Penetration is achieved by the kinetic energy of
the high-density penetrator. Differently shaped and sized penetrators can be seen in Fig.
7.25.

303

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

FIGURE 7.24. Schematic drawing of a heavy metal tank ammunition cartridge: (I) windshield, (2) tungsten
heavy-metal penetrator (subprojectile), (3) three-section sabot, (4) stabilizing fin, (5) propellant. By courtesy of
TAAS-Israel Industries Ltd., Israel.

FIGURE 7.25. Selection of large-caliber kinetic energy penetrators; medium-caliber kinetic energy penetrators
and ammunition, as well as tungsten alloy cones for shaped charges. By courtesy of Royal Ordnance Speciality
Metals, England.

304

CHAPTER 7

Armor plating (HM).

Scatter grenades (HM) (fragmentation devices).
Shaped charge liners (W); Fig. 7.25.
Counterweights in tanks (HM).
Nozzles in air to air rockets and jet vanes (W, W-Cu, W-Th02 , Ta-2.5W).
7.12. AVIATION [7.24]

Rim components offlywheels or rotating members in governors (HM).

Counterweights in static and dynamic balancing application in airplanes and
helicopers (aircraft balance weights) (HM). For typical shapes, see Fig. 7.26.
Instrument balance weights (HM).
Airborne antenna bases (HM).
7.13. LASER TECHNIQUE [7.7]

Construction parts in gas lasers (W-Cu).

Cathodes (porous W impregnated with Ba aluminate). Coiled cathodes for laser
resonators are shown in Fig. 7.27.

FIGURE 7.26. Heavy metal parts used in aviation and radiation shielding: (I) passenger aircraft elevator weight,
(2) passenger aircraft rudder weight, (3) helicopter vibration dampener weight, (4) helicopter rotor. vibration
dampening weight, (5) rotor weight, (6) missile component, (7) nuclear medicine shielding, (8) radiation
shielding; rest: miscellaneous parts. By courtesy of Teledyne Advanced Materials, USA.

TUNGSTEN AND TUNGSTEN ALLOY PRODUCTS

305

FIGURE 7.27. Coiled cathodes of porous impregnated tungsten for laser resonators. By courtesy of Plansee AG,
Austria.

REFERENCES FOR CHAPTER 7

7.1.
7.2.
7.3.
7.4.
7.5.
7.6.
7.7.
7.8.
7.9.
7.10.
7.1 1.
7.12.
7.13.
7.14.
7.15.
7.16.
7.17.

7.18.
7.19.
7.20.
7.21.
7.22.
7.23.
7.24.
7.25.
7.26.

W 1. van den Hoek and A. G. Jack, "Lamps," in: Ullmann's Encyclopedia of Industrial Chemistry, Vol.
A15, pp. liS-ISO, VCH, Weinheim (1990).
C~ L. Briant and B. P. Bewlay, MRS Bulletin, pp. 67-73 (August 1995).
F. Mertens, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 277-293, lIlA, London (1997).
L. Bartha, in: Mater. Res. Soc. Symp. Proc. 322 (1994), 531-536.
Philips Light ... More Than Meets The Eye, company brochure, Philips Lighting, The Netherlands (3/87).
Lighting Technology, company brochure, Plansee AG, Austria (9/95).
Tungsten, company brochure, Plansee AG, Austria (3/95).
F. Mertens, Philips Lighting, private communication.
Philips Lighting, product information, The Netherlands (9/88).
SOX-E Low Pressure Sodium Lamps, company brochure, Philips Lighting, The Netherlands (4/86).
Metal Halide Gas-Discharge Lamps, company brochure, Philips Lighting, The Netherlands (4/86).
H. Kippenberg, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 312-316, lIlA, London (1997).
L. Zehnder, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 317-335, lIlA, London (1997).
CMSH, Sumitomo Electric Industries, Ltd., Japan (12/91).
High-Temperature Furnace Construction, company brochure, Plansee AG, Austria (11/96).
L. Plochl, Plansee AG, private communication.
D. Naujoks, K. Asmussen, M. Bessenrodt-Weberpals, S. Deschka, R. Dux, W Engelhardt, A. R. Field, G.
Fussman, 1. C. Fuchs, C. Garcia-Rosales, S. Hirsch, P. Ignacz, G. Lieder, K. F. Mast, R. Neu, R. Radtke, 1.
Roth, and U. Wenzel, ASDEX Upgrade Team, Nucl. Fusion 36 (1996), 671.
Electrode Materialsfor Electrical Discharge Machining, company brochure, Plansee AG, Austria (9/85).
Thin Film Technology, company brochure, Plansee AG, Austria (7/97).
Densimet W + WS, company brochure, Plansee AG, Austria (6/95).
Tungsten-Based Products, company brochure, Ashot Ashkelon, Israel (11/92).
Kinetic Energy Technology, company brochure, Royal Ordnance Speciality Metals Ltd., England.
Chemical Apparatus Engineering, company brochure, Plansee AG, Austria (2/97).
E. Okorn and G. Leichtfried, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 361-370, lIlA, London
(1997).
G. Leichtfried, Plansee AG, private communication.
Tank Ammunition, company brochure, TAAS-Israel Industries Ltd. (8/94).

8
Tungsten in Melting
Metallurgy
8.1. TUNGSTEN IN STEEL

8.1.1. Introduction
At the beginning of the 20th century steel was the biggest tungsten consumer. Since the
development of cemented carbides in 1927, the proportion of total tungsten consumed in
steelmaking has fallen progressively to the current value of 22%. Nevertheless, steel is
today still the second largest tungsten consumer.
The early tungsten-bearing steels were produced by adding tungsten powder to the
steel melt-a rather expensive procedure. In 1914 the master alloy Jerrotungsten was
introduced to the market. Ferrotungsten could be produced in a much cheaper way directly
from ore concentrates, and provided a lower melting point for faster dissolution and higher
tungsten yield compared with the use of tungsten powder. In 1940 a second master alloy
produced from secondary raw materials~alled tungsten melting base~ame into use.
Tungsten can also be added to the steel melt in the form of tungsten scrap or direct
scheelite concentrates.
Tungsten alloying is important mainly for the following steel grades:
cold work tool steels (0.5 to 3 wt'110 W),
hot work tool steels (1.5 to 9 wt% W),
high speed tool steels (2 to 18 wt% W),
and to a lesser extent
stainless steels and constructional steels (<1 wt% resp. 1 wt%).
High speed steels (HSS) consume by far the most tungsten in steelmaking, although
they account for less than 0.02% of the total steel production (1997: about 700 million
metric tons). In fact the tungsten story in special steel making is the story of high speed
steel [8.8].
By shifting the eutectoid point to lower carbon concentrations, tungsten increases the
amount of undissolved and excess carbide in the hardened steel. Tungsten is a carbideforming element and binds the carbon to form straight tungsten carbide or combines with
307

308

CHAPTER 8

other alloying elements to form complex carbides (M6C, M23C6)' The chemistry of these
carbides depends on the concentration of carbon and the alloying elements present as well
as on the heat treatment applied. For the same level of carbon in the steel tungsten will
produce a larger carbide volume than other alloying elements [8.3].
In addition to improving hardness, tungsten and complex carbides also subscribe to
retaining a fine-grained structure in steels as the excess carbides retard the grain growth
during austenizing. The resulting structure can be defined as fine grained with evenly
distributed, very fine carbide precipitations. The influence of tungsten on carbide formation is shown in Fig. 8.1.
The effects of tungsten on steel properties in terms of carbide volume, shape and
distribution, and fine-grained structure are:
more carbides and harder carbides increase hardness and wear resistance;
fine carbides and a small grained structure result in higher toughness for a certain
hardness level.
Good wear resistance is the most important property of tool steels, and tungsten, next
to vanadium, is the most effective carbide-forming element for increasing wear resistance.
Furthermore, tungsten significantly improves the hot hardness and hot strength of steel,
and leads to increased yield strength and tensile strength without detriment to ductility
(elongation and area reduction on breaking) and notched bar fracture toughness. The high
hot and red hardness (which is the maintenance of hot hardness over long-term exposure to
high temperatures) and the high hot strength are the result of a microstructure containing
numerous high-alloy carbides, and a composition heavily alloyed with diffusion-slowing

C 0.95%

annealed

C 0.97% 0.8 W

10.urn
f-----1

FIGURE 8.1. Influence of tungsten on carbide fonnation in steel [8.3].

10.urn
H

TUNGSTEN IN MELTING METALLURGY

309

elements and carbon. Hardening is achieved by high-temperature austemtIzation to

dissolve the alloying elements, followed by transformation during quenching and tempering to martensite with precipitated fine carbides. This produces a material which will retain
most of its hardness up to about 550 DC, and regain it again on cooling to room
temperature.
Tungsten supply in the past was sometimes uncertain, and prices were dictated by the
supply / demand balance. In times of shortage, as in World War II, prices increased
considerably. This led to the introduction of molybdenum as a substitute for tungsten in
high speed steels (most of the early grades contained up to 18 wt% tungsten), with the
advantage that only half the amount is needed compared to tungsten. However, molybdenum-bearing steels generally require more sophisticated and expensive heat-treatment
facilities. The most common grades of high speed steels (HSS) currently used in the
western world contain both tungsten and molybdenum together with vanadium. The US
market, and also Japan, mainly asks for the low-tungsten-high molybdenum grades
(typically 2 wt% W; 9 wt% Mo, and 1 wt% V; M1-type); while in Europe the 6 wt%
W, 5 wt% Mo, and 2 wt% V grade (M2-type) is predominant [8.8]. In the eastern
countries, especially in the former Soviet Union and the Peoples Republic of Chinacountries which have traditionally been self-sufficient in tungsten-little tungsten substitution has occurred.
While tungsten will continue to play an important role in high speed steels, these
materials are under constant attack from the extending range of application of cemented
carbides, particularly in the area of cutting tools. This trend will accelerate with the
modernization of machining plants in developing countries, and with the improvements in
performance and tool life provided by the coating of high speed steels with hard layers
such as TiN [8.8, 8.9]. Already a considerable percentage of drills sold today in the western
world are coated.
From an environmental standpoint, cemented carbide tools are preferable to high
speed steel because any tungsten in steel is lost by dilution (see Section 11.1) in contrast to
cemented carbide scrap, which is a valuable source for recycling [8.10].

8.1.2. Master Alloys

Tungsten in most cases is added to the steel melt as master alloy. The two usual types
are ferrotungsten and melting base. Due to their lower melting point, master alloys dissolve
and distribute more readily in the steel melt as pure tungsten. Moreover, they are cheaper
because they are produced either directly from ore concentrates in case of ferrotungsten or
from scrap materials in case of melting base alloys.
8.1.2.1. Ferrotungsten. Properties. Commercial ferrotungsten contains between 75 and
80 wt% W It has a steel gray appearance and a fine grained structure consisting of FeW (8phase) and Fe2 W (A-phase). It is supplied in 80-100 mm lumps. Standard compositions are
given in Table 8.1.
At 80 wt% W, the liquidus temperature is >2500 DC and the solidus temperature
about 1640 DC. The density is 15.4 g . cm- 3 .
Production. The raw materials for ferrotungsten production are either rich ores or ore
concentrates of the minerals wolframite, huebnerite, ferberite, and scheelite. Also, artificial
scheelite or soft scrap can be used.

CHAPTER 8

310

TABLE 8.1. Composition of Ferrotungsten According

to ISO 5450 [8.2]
FeW 80
W (wt%)
Ranges (wt%)
Si
Mn
Cu
S
P
Tolerable maximum (wt%)
C
Al
Sn
As
Mo

70-85

FeW 80LC
70-85

0.6-1.0
0.6-1.0
0.20-0.25
0.05-0.06
0.05-0.06

0.6-1.0
0.6-1.0
0.20-0.25
0.05-0.06
0.05-0.06

1.0
0.10
0.10
0.10
1.0

0.1
1.0
0.10
0.10
0.5

Tungsten is reduced in preference to iron so that ferrotungsten with high W content

can be produced even if the ore is high in iron. The W0 3 contained in these compounds
can be reduced either carbothermically in electric arc furnaces or metallothermically by
silicon and/or aluminum. The carbothermic or silico-carbothermic method is preferred for
cost reasons and, moreover, the tolerance level for impurities like As and Sn in the raw
materials is higher.
Although tungsten forms carbides with carbon, the formation of WC can be largely
prevented by controlling the CO/C02 ratio. Thus a low-carbon ferrotungsten with
maximum carbon content of 1% is produced.
Carbothermic Production. Because of the high melting temperature tapping off is
not possible. Therefore, the so-called solid block process is normally used. The process
operates in two stages. Both stages are performed in electric arc furnaces. In the refining
stage an ingot of high-grade ferrotungsten and a WOrrich slag is produced. The ingot is
removed and cleaned. In the reduction stage the W0 3 -rich slag is processed to ferrotungsten of 50% Wand a slag having < 1% W0 3 . The 50% ferrotungsten together with the
outer parts of the cleaned, high-grade ingot are recycled to the refining stage. A flow sheet
of the process is given in Fig. 8.2. The tungsten yield is between 97 and 98%.
Typical consumption figures per metric ton of 80% FeW are:
Tungsten concentrate (65% W0 3)
Charcoal
Coke
Fluorspar
Quarzite / calcite
Steel scrap
Bauxite
Electrodes
Electric energy

1600 kg
150 kg
20 kg
80 kg
50 kg
75 kg
80 kg
120 kg
7500-8500 kWh

TUNGSTEN IN MELTING METALLURGY

311

Tungsten ore concentrate

Flux

Charcoal (or coke)

Refining stage, electric arc furnace, block process

Coke
breeze

Steel
scrap

Slag

Flux

(25"1oW)

Ferrotungsten
Ingot

Crushing, sorting

Ferrotungsten
(ca. 50"10 W, slag metal")

Slag

1"1oW)

Tip

Ferrotungsten

(80"10 W)

Recycle metal
(remelt, ca. 25"10 W)

Sale

FIGURE 8.2. Scheme of carbothermic ferrotungsten production [8.2].

Carbothermic-silicothermic production. The process is carried out in three successive stages in an electric arc furnace. Sixty percent of the oxygen in W0 3 are reduced by
carbon and 40% by silicon added as 75% FeSi.
Stage 1: Reduction of tungsten ore concentrates by carbon under a W0 3-rich slag
(10-16% W0 3), and scooping out ferrotungsten (75% W).
Stage 2: Reduction of the W0 3-rich slag from Stage I by FeSi75 and addition of
scrap iron. Tapping off the low-W0 3 slag 1% W0 3). The metal phase
(50-70% W) remains in the furnace.
Stage 3: Refining of the low-W metal phase by addition of ore concentrates. A W0 3rich slag is formed (18-25% W0 3), and the W content of the metal phase
increases.
Advantages are:

continuous operation responsible for prolonged lifetime of the furnace lining;

lower electric energy consumption.
Disadvantages are:

heavy labor of the scooping operation;

accumulation of some ferrotungsten in a "furnace sow" or "salamander" which
can be recovered only after furnace relining.
Typical consumption figures per metric ton of 80% FeW are:
Ore concentrate (60% W0 3)
FeSi75
Scrap iron
Coke
Electrodes
Electric energy

1527 kg
130 kg
80 kg
120 kg
85 kg
4000-4200 kWh

Metal/othermic Production. Metallothermic reduction requires purer raw materials,

because it proceeds very rapidly and some of the impurity elements are reduced as well as

CHAPTER 8

312

tungsten. For example, about 50% of the Sn and As contained in the raw materials end up
in the ferrotungsten, most of which would be evaporized during the longer-lasting
carbothermic process.
Concentrates are mixed with coarsely powdered Al and Si (ratio AI/Si 70:30), and are
charged into a refractory-lined furnace vessel. The vessel is preheated to 400-500 C or,
alternatively, an Al booster is added. The reaction is started by igniting initiators (e.g.,
Ba02 and AI), and is completed in terms of several minutes. After cooling, the vessel is
removed and the metal block (700 to 2000 kg) is separated. The yield is about 96%.
The metallothermic reduction process has lost much of its importance in recent years
owing to the increased costs of Al and Si powders and the necessity for using pure and
expensive raw materials.
8.1.2.2. Melting Base. Properties. This type of master alloy can have various compositions, depending on the scrap material used. Typical examples are listed in Table 8.2.
Production. Different types of tungsten-containing scrap are mixed to meet the
desired composition. Reductive melting is done in electric arc furnaces as usual in steel
melting. The product is finally granulated 10 mm) by casting on a rotating disc and
quenching in water.

8.1.3. Tungsten Alloyed Steels. Tungsten addition to the steel melt.

Tungsten sources which can be added to the steel melt are:
Tungsten metal scrap
Ferrotungsten
Scheelite ore concentrates
Melting base

90
75
55
up

to
to
to
to

98
80
70
36

wt%
wt%
wt%
wt%

W
W
W
W

Tungsten metal scrap and ferrotungsten, due to their high percentage in tungsten,
show poor solubility and this may cause bath inhomogeneities. The tungsten recovery
using this type of tungsten addition lies at approximately 95% in case of low carbon and
slag levels.
Tungsten addition as scheelite permits rapid attainment of alloy homogeneity at a
recovery of 94%. The costs are significantly lower but the addition is limited to 5% W in
steel. Higher amounts increase the slag volume and thus decrease the recovery.
Addition of melting base offers the advantages of known chemistry, easy charge, and
good solubility.
TABLE 8.2. Typical Analyses of Melting Base Alloys (wflt-ot
Designation

Mo

Cr

TMB
TMB+Co
FeWMo
FeWMoCo
Darby Melt

28-38
24-30
7
7
10

0.5
0.6
5.5
5.5
0.4-1.5

0.2
0.5
0.5
0.5
3

0.2
0.2
2
2
1.2

Co

Ni

max 0.6

0.4
0.4
0.4
0.4

0.04
0.04
0.04
0.04
0.03

0.04
0.04
0.04
0.04
0.02

1.2
1.4
1.2
1.2
0.4-1.3

5-10
0.6
5-10

Treibacher Chemische Werke, Lieferprogramm Wolfram-Melting Base, Treibach, Austria (1991).

313

TUNGSTEN IN MELTING METALLURGY

In general, a high melting temperature followed by a low teeming temperature results

in a homogeneous ingot [8.7].
The quality of tool steels has greatly improved in recent years by new melting and
processing routes. Most of the special steels are made by vacuum arc melting (ingot
casting) and subsequent hot working (rolling, forging). However, electro slag refining
(ESR) and, to a lesser extent, vacuum arc remelting (VAR) are increasingly important for
producing high-quality steels with significantly lower nonmetallic inclusion contents.
Hardening and tempering. Tungsten-containing steel affords a certain carbon level
and furthermore a careful heat-treatment practice. One disadvantage of tungsten steel is
that annealing prior to hardening, if not done with utmost care, will cause formation of
tungsten carbides, rich in carbon, which do not dissolve during hardening. Hence the
carbon level in the austenite is reduced and the martensite does not present the desired
hardness. If hardening is performed from a sufficiently high temperature the tungsten
carbides fully or partly dissolve in the austenite prior to quenching. After quenching, they
remain in supersaturated solution in the martensite. During tempering below 400C
tungsten does not diffuse and prevents diffusion of other elements, too. Very fine and
evenly distributed tungsten alloy carbides precipitate from the martensite matrix above
400 C causing a precipitation hardening effect. In this way high-temperature hardness
(around 500C) is increased. The effect of tungsten on hardness during tempering is
shown schematically in Fig. 8.3. At higher temperatures the carbides coalesce and the
effect gets lost.
In very hard tool steels, for example, cold work steels tempered at low temperature,
the wear resistance is originated mainly by hypereutectoid carbide which is not affected by
hardening. These carbides are very small and evenly distributed and very hard (heavy-duty

improvement
by tungsten

rJ)
rJ)

Q)

1J
L-

eo

C-steel

1000

500

Tempering temperature

(OC)

FIGURE 8.3. Effect of tungsten on hardness after tempering (schematic) [8.3].

CHAPTER S

314

cold work tools with high hardness, good wear resistance and toughness at ambient or
slightly higher temperature). Fine grained martensite adds to this effect.
High Speed Tool Steels (HSS). Tungsten is the most prominent alloying element in
HSS on a worldwide basis, although its originally large contribution has reduced in
western countries. About 95% of HSS is used for machining tools, the rest for chipless
forming (punching tools).
The basic analysis of the three main groups of HSS is:
wt%
Group I
Group II
Group III

Cr

0.7
O.S
O.S

4
4
4

IS

Mo

9
6

Experience showed that all three groups behave similarly, but they differ slightly in
respect of susceptibility to decarburization, carbide formation, and heat treatment. They
can be hardened and are austenized at about 80C below their melting temperature prior
to air cooling for hardening. All show appreciable secondary hardening after tempering
(550-600 0C). This is their most important feature, responsible for the retention of
hardness at high temperature occurring during cutting and machining.
The first group (molybdenum free) has been in service for more than 80 years,
possessing all the characteristic properties of a HSS, such as red hardness, high toughness,
and good wear resistance. The advantage of this group of steels as compared to the others
is that the heat treatment is much more tolerable in regard to temperature control and
equipment. It is still the most popular group in the eastern countries, but not in the west.
Table 8.3, row 1, shows that in practice a wide variation of tungsten containing HSS
without Mo exists.
The second group containing only Mo and no W is used more rarely.
Group III is most popular in western countries and has proved most satisfactory in
practical service. Also, within this group a broad variation in the alloying elements W, Mo,
V, and Co is in use as can be seen in Table 8.3, row 2.
Also low-carbon, precipitation-hardened steel can be produced. It owes its substantially higher hardness to the precipitation of intermetallic phases at annealing temperatures
above 600C. Figure 8.4 compares the hardness to tempering-temperature dependence of
high- and low-carbon HSS. However, costs as well as machining and forming difficulties
have prevented its use on a large scale. The analysis is given in Table 8.3, row 3.
In producing HSS, the maximum hardness after tempering is obtained if the
hardening temperature is sufficiently high and the holding time long enough to insure
uniform heating. Higher hardening causes an increase in hardness after tempering, as will
be shown in Fig. 8.5. It must be chosen in such a way that sufficient carbides dissolve, but
no overheating and grain coarsening may occur.
About 8% of HSS are produced today by powder metallurgy (P1M). This amount is
steadily increasing, in particular in the field of cutting tools (gear cutter, endmills, taps, and
broaches) [8.10]. More carbides can be added compared to conventional HSS and a finer
particle size and more uniform distribution of the carbides can be achieved. However, the
costs are still higher than for conventional tool steels, which retards a more general
applicability [8.10].

12
13
14

II

2
3
4
5
6
7
8
9
10

Row No.

0.7-1.5
0.8-1.5
0.2
0.35-D.55
0.25-0.50
0.60-0.65
0.22
0.10-0.12
0.12- 0.83
0.9-1.2
2.25
1.00-1.25
0.70-0.75
0.50

12-18
1.5-6

Tungsten high speed tool steels

Molybdenum high speed tool steels
Low-carbon high speed tool steels
Chromium hot work tool steels
Tungsten hot work tool steels
Molybdenum hot work tool steels
High creep resistant steels (martensitic)
High creep resistant steels (austenitic)
Valve steels
Oil hardening cold work tool steels
High-C high-Cr cold work tool steels
Carbon-tungsten tool steels
Permanent magnet steels
Shock-resisting tool steels

1.3-3.5
5-7
1-2.5

1.5-7
9-18
1.5-6
0.5
2.5-2.8
1-2.5
0.5-1.8

II

C (%)

W (%)

Steel type

0.75-1.5

5-7
2-12
4
12
16-21
14-21
0.5-D.8
12
0-0.75

4
4

Cr (%)

0-0.4
0-1
1-2

1-5
1-5

V (%)

0-3
0-3

5-8

3.5-9.5
7
0-1.5

Mo (%)

0-20
0.25 opt.

5-9
0-12
23

Co (%)

TABLE 8.3. Analyses of Tungsten-Containing Steels (wt"Io) [8.2, 8.4]

0.8
13-20
0.7-20

Ni (%)

0.5 Ti
0-9 Mn, 0-1 Nb, 0-0.5 N2

Others (%)

....

Ul

Y>

Cl

""-<

g;

ti

g;

[FJ

CHAPTER 8

316

70

0::

Mo
Co

65

c: 60

en
en

Q)

c: 55

....

"C

ttl

50

518-0-1
0.75%
4.1 %
1.1 %
18.0%

45

C
Cr
V

40+----,r----r----,----.----.---~

300

400
500
600
700
Tempering temperature

800

900

(oC)

FIGURE 8.4. Effect of tempering on hardness ofa low and high carbon HSS [8.3].

hardening
temperature

(OC)

67
0
0::
I

c: 66

en
en
Q)

c:

.... 65

"C

ttl

64

631------L~----~------,_~~~

460

500
540
580
Tempering temperature (oC)

620

FIGURE 8.5. Influence of hardening temperature on tempering hardness [8.3].

TUNGSTEN IN MELTING METALLURGY

317

Hot Work Tool Steels. This steel type is mainly used for producing casting and
pressing tools in light and heavy machinery. It was important in the development of metal
extruding and die casting. For application where very high temperature and severe wear
exists, tungsten steels are favorable. Analyses of three different hot work steel groups are
given Table 8.3, rows 4-6.
High-Temperature Steels. For applications affording superior high-temperature
strength and high-temperature creep resistance, austenitic and Cr-alloyed tungsten steels
containing also Mo and V are in use (see Table 8.3, rows 7 and 8).
Valve steels, which have to meet stringent requirements regarding hot hardness,
fatigue strength, and wear resistance, also belong to this group. Cr and Ni account for the
matrix (martensite or austenite), carbide and nitride for additional hardness, and tungsten
for wear resistance. Analyses are given in Table 8.3, row 9.
Cold rfOrk Tool Steels. Historically, this was the first group of tungsten-bearing
steels. They are hypoeutectoid, hypereutectoid high Cr, high C steels. The tungsten
function is to increase the carbon proportion and carbide hardness, to result in a fine
structure and improved toughness. Here, too, a carefully controlled heat treatment is
necessary. Moreover, tungsten accounts for the secondary hardening effect on tempering.
Typical alloy compositions are given in Table 8.3, rows 10 and 11.
Some further steels containing tungsten are:
carbon tungsten tool steel (Table 8.3, row 12),
permanent magnet steels (Table 8.3, row l3),
shock-resistant tool steels (Table 8.3, row 14).

8.2. SUPERALLOYS
Superalloys are complex alloys containing up to 16 and more elements. There exist
three basic types: iron, nickel, and cobalt base alloys. Their important properties are:

high-temperature strength;
high creep strength at high temperature;
high thermal fatigue resistance;
good oxidation resistance;
excellent hot corrosion resistance;
air melting capability;
air or argon remelting capability;
good welding properties;
ease of casting.

The group of iron-based alloys are of minor importance. The composition ranges for
Ni- and Co-based alloys (cast and wrought) are summarized in Table 8.4. It can be seen
that the carbon content is in the range of 0.25-1.0% for cast Co-based alloys and 0.050.4% for hot worked.
The hardness is a result of carbide precipitation and solid solution strengthening. The
presence of carbides at the grain boundaries is important for the mechanical properties,
because they retard grain boundary migration.

CHAPTER 8

318

TABLE 8.4. Composition Ranges ofNi- and Co-Based Superalloys (concentrations in wt%)
[8.4, 8.5, 8.6]
Nickel base alloys
Element
Ni
Cr
Co
Mo
W
Ta
Nb
Al
Ti
C
8

Zr
Fe
Mn
Si

Cobalt base alloys

Cast

Wrought

58-73
0--16
0--20
0--6.5
2-20
0--9
0--2.5
3--6.5
0--5
0.12--0.20
0.010--0.020
0--1.5

45--68
12-25.5
0-15
3-9
1--6
0-3.5
0-3
0-4.6
0-5
0.05--0.32
0--0.020
0--0.10
0--1.5
0--1.5
0--1.2

Re

Cast
0-10.5
3-29
52--68
4.5-25
0-9
0-2
0-4.3
0-1
0.18-1.00
0--0.010
0--2.25
0--2.5
0--0.75
0--0.75
0--0.1
0--2.0

Wrought
10--22
20--22
39-52
0--4
4-15
0--4
0--0.5
0--1.3
0.05--0.4

0-4
0--1.5
0--0.50

Tungsten accounts for

solid solution strengthening;
strengthening by formation of intermetallic compounds (e.g., C03 W);
formation of carbides.
Superalloys are produced via master alloys. In case of nickel-based alloys they
include NiW and more complex nickel-based alloys containing Cr, Ta, and Mo. Some
examples are given in Table 8.5. All raw materials have to be of high purity. Melting
(performed in induction furnaces), casting and cooling takes place in vacuum. Master
alloys are supplied in small lumps or as powder.
The most important applications of superalloys are in aircraft engines, marine
vehicles, and stationary powder units as turbine blades and vanes, and in sheet form for
exhaust case assemblies and burner liners. Moreover, they are used in furnace combustion
tubes, muffles, cracking and reformer tubes.

8.3. STELLITE ALLOYS

Stellites are hard-facing and wear-resistant alloys. They have been developed since
1920 and today more than 20 different qualities are on the market. There are two main
groups:
alloys of Co, Cr, W, and C,
alloys of Co, Cr, W, Mo, Ni, Fe, and C.

319

TUNGSTEN IN MELTING METALLURGY

TABLE 8.5. Examples of Master Alloys Used in Superalloy

Production (concentrations in wt%) [8.6]
Element

NiW

NiCrAl g WTi4.4Ta3sMo

55
45

W
Ni
Mo
Cr
Co
Al
Ti
Ta
S
p

7
36
4
40

7
46.6

3.5
26
I
8
4.4

12

3.5
0.001
0.002
0.002

0.001
0.015
0.1

Special hard-facing qualities contain in addition small percentages of Si and B.

Typical analyses of tough, medium, and hard alloys are given in Table 8.6 including
hardness at ambient and high temperature.
The microstructure is defined by hard carbide particles in a Co-rich solid solution
matrix. The wear characteristics of the matrix are influenced by a phase change from a
strain-induced cubic to a hexagonal form. Toughness and abrasion resistance are due to the
volume fraction of carbides and their morphology.
They are prepared from Co-Cr-W master alloys by addition of the other components
in pure quality.
For hard facing, they are used either as welding rods or as powders atomized from a
homogeneous melt. Hard facing is applied to valves, valve seats, bearings, elevators, rock
crushers, mixing components, marine propeller shafts, automotive water or fuel pump
shafts, knives, and blades.
Massive forms are in use as cast only. Finished shape can be attained by grinding,
since they are not amenable to hot rolling or forging. The properties of stellite metal
cutting tools are in the middle between HSS and cemented carbides. Other applications are
wear and corrosion resistant parts, like pump sleeves, valves, valve seats, extrusion dies,
and mill guides.
TABLE 8.6. Analyses and Hardness of Stellite Alloys (Co content is balance) [8.5]

Alloy

Cr (wt%)

W (wt%)

Group I
Examples
tough
medium hard
hard grade

30

4--17

30
30
30

4
9

Group II

25-30

18
0-19

C (wt%)

900 DC

Others

1-3.2
I
1.5
2.5

0.1-2

375
470
600
3.22 Ni, 3-10 Mo,
5-20 Fe

100
150
200

320

CHAPTER 8

REFERENCES FOR CHAPTER 8

8.1.

8.2.

8.3.
8.4.
8.5.
8.6.
8.7.
8.8.
8.9.
8.10.

E. Lassner, H. Ortner, R. M. Fichte, and H. U. Wolf, "Wolfram und Wolframlegierungen und

-verbindungen," in: Ullmanns Encyk/opedie der Technischen Chemie, Band 24, WOlfram, pp. 457-488,
Verlag Chemie GmbH, Weinheim (1983).
E. Lassner, W D. Schubert, E. Loderitz, and H. U. Wolf, "Tungsten, Tungsten Alloys and Tungsten
Compounds," in: Ullmann s Encyclopedia 0/ Industrial Chemistry, 5th ed., Vol. A27, pp. 229-266, VCH,
Weinheim (1996).
M. Kroneis, "Tungsten in Steel," in: Proc. 1st Int. Tungsten Symposium, Stockholm, pp. 96-107, Mining
Journals Books Ltd., London (1979).
S. W H. Yih and Ch. T. Wang, Tungsten, Plenum Press, New York (1979).
1. D. Donaldson, "Cobalt and Cobalt Compounds," in: Ullmann sEncyclopedia o/Industrial Chemistry, 5th
ed. (W Gerhartz, ed.), Vol. A7, VCH, Weinheim (1986).
F. W Strassburg, "Nickel Alloys," in: Ullmann s Encyclopedia 0/ Industrial Chemistry, 5th ed. (B. Evers,
St. Hawkins, and G. Schulz, eds.), Vol. A17, VCH, Weinheim (1991).
F. B. Pickering, "Constitution and Properties of Steels," in: Materials Science and Technology, Vol. 7 (R.
W Cahn, P. Haasen, and E. 1. Kramer, eds.), VCH, Weinheim (1995),
H. 1. Fleischer, Proc. 4th Int. Tungsten Symposium, vancouver, pp. 40-49, MPR Publishing Services Ltd.,
Shrewsbury, England (1988).
G. Kientopfand R. Anderson, Proc. 7th Int. Tungsten Symposium, Goslar, pp. 254-274, lIlA, London, UK
(1996).
A. Egami, Proc. 7th Int. Tungsten Symposium, Goslar, pp. 13-23, lIlA, London, UK (1996).

9
Tungsten in Hardmetals

9.1. INTRODUCTION [9.1-9.6]

Cemented carbides represent a group of hard and wear-resistant refractory composites in
which hard carbide particles are bound together or are "cemented" by a ductile and tough
binder Jllatrix. Although the term "cemented carbide" is still widely used, mainly in
English-speaking countries, and well describes the nature of the composite, they are even
better known internationally as "hardmetals." The latter term will be used in the following.
Hardmetals combine the high hardness and strength of the covalent carbide(s) (WC,
TiC, TaC) with the toughness and plasticity of the metallic binder (Co, Ni, Fe). This unique
combination of hardness and toughness makes them outstanding as tool materials in the
manufacturing industry. Their application is extremely widespread and includes metal
cutting, machining of wood, plastics, composites, and soft ceramics, chipless forming (hot
and cold), mining, construction, rock drilling, structural parts, wear parts, and military
components. The metal cutting group accounts for about 67% of the total use, followed by
mining (13%), machining of wood and plastics (11%), and construction (9%) [9.1].
Straight grades. Tungsten carbide (WC) is the basic and most widely used hard
component. Cobalt was found to be the best binder metal. WC-Co grades, consisting of
two phases (WC and Co; Fig. 9.1), are often referred to as unalloyed grades, straight
tungsten carbide grades, cast iron cutting grades, or edge wear-resistant grades. The cobalt
content varies from 3-13 wt% in cutting applications and goes up to 30 wt% for wear
resistant parts. The average WC grain size varies from submicron to approximately 30 11m
[9.2]. Regarding consumption and broad applicability, this group is the most important
(more than 80%). Alternative binder metals in straight grades can be nickel (corrosion
resistant alloys; woodcutting alloys), steel, or combinations of the iron binder metals.
Tungsten-(titanium-tantalum-niobium)-carbides. WC-Co alloys have been modified in the past to produce a variety of hardmetals with improved high-temperature
properties, such as improved oxidation resistance, hot hardness, hot strength, and diffusion
resistance against iron alloys, by introduction of additional carbides of titanium, tantalum,
niobium, chromium, vanadium, and molybdenum. The ideal microstructure of this second
large group of hardmetals consists of three phases: hexagonal WC, cubic mixed crystal
carbides, and binder (Fig. 9.2) These grades are also referred to as steel cutting grades, or
321

322

CHAPTER 9

FIGURE 9.1. Microstructures of straight WC-(9-10)wt% Co grades; all samples etched with Murakami's
reagent; 1000 x. (a) Submicron grade for finish machining and chipless forming. (b) Fine/medium grade for wear
parts and chipless forming. (c) Coarse grade for impact drills in medium/hard rock drilling. (d) Very coarse grade
for chisels in coal and vein mining. By courtesy of B6hlerit Ges.m.b.H & Co. KG., Austria.

FIGURE 9.2. Microstructure of a 68.5WC-21(Ti,Ta,Nb)C-9.5Co steel cutting grade (ISO P25); etched with
Murakami's reagent; 1000x. The gray faceted grains are WC; the dark gray and globular grains are the
(W,Ta, Ti,Nb}-solid-solution carbides.

TUNGSTEN IN HARD METALS

323

crater-resistant grades. Their basic composition contains 3-12 wt% Co, 60-85 wt% WC,
4--25 wt% TiC, and up to 25 wt% TaC.
Besides WC-based hardmetals, Ti(C,N)-based grades are also used, particularly in
Japan. These so-called cermets contain no, or only minor additions of, We. They will not
be discussed in this book.
Mechanical properties. The mechanical properties of hardmetals depend primarily
on composition, binder content, carbide grain size, and grain size distribution. Variations
in the carbide-to-binder ratio, as well as in the average grain size of the carbide phase(s),
allow adjustment of the desired properties of the composite in terms of hardness, strength,
rigidity, toughness, etc. They form the basis for its widespread applications. One example
may facilitate easy comprehension! Simply by changing the average WC grain size in a
straight WC-6 wt% Co alloy, the Vickers hardness can be increased from 1170 HV30
(6 11m WC) to 1600 HV30 (1.7 11m WC), and even to 2300 HV30 in the case of ultrafine
alloys (0.2 11m WC).
Generally speaking, a decrease in WC grain size is connected with an increase in
hardness, compressive strength, and also bending strength, but gives a decrease in impact
strength, rupture strength, and fracture toughness. An increase in cobalt content yields
lower hardness, lower modulus of elasticity, and compressive strength, but it increases
bending strength and fracture toughness.
As a result of the strong adhesion between the WC grains and the Co binder, straight
WC-Co grades exhibit the best hardness-to-toughness ratio of all hard materials produced
today. This explains their outstanding role in the manufacturing industry.
Hardmetals today. Changes in composition or grain size of hardmetals, in order to
improve the performance of tools, are of minor importance today (with the exception of
ultrafine grades). Most of the recent advancements in the manufacturing industry were
made either by coatings (CVD, MT-CVD, PVD) or by optimization of the tool-design and
tool-holder systems. This is easy to understand. Development of coatings which increase
tool lifetime by a factor of 2-4, or tools with multiple cutting functions, based on already
available ("well known") alloys, will save more time and money in manufacturing than
would be possible by further (limited) alloy improvements.
The machining industry is the best example. The dominant tool material today is the
coated hardmetal which, in combination with indexable insert technology, has had an
enormous impact on the machining industry since its introduction in the late 1960s [9.3].
However, despite these changes in the tool industry, hardmetals have not lost their
importance. On the contrary, larger amounts are produced than ever before and, in their
various areas of application, their proportion is steadily increasing. The reason for this
situation is manifold:
The availability of high-quality raw materials (WC, Co, WC-TiC, TaC, etc.).
The high standard of manufacturing (spray drying, pressure sintering, etc.) and
high product reliability (important for automized manufacture).
Their high strength and rigidity (two to three times that of steel), excellent thermal
conductivity, and low thermal expansion, making them the ideal coating substrate
(excellent adhesion).
The possibility of tailormaking the substrate simply by variations in powders
and/or in sintering technology (recent examples: gradiented structures and WC
plate reinforcement).

324

CHAPTER 9

Unlimited possibilities in tool design by CAD/CAM and subsequent realization by

P /M techniques (near net shape production, powder injection molding, etc.).
Last, but not least, the possibility of material reclamation through well-developed
recycling processes (zinc reclaim, bloating, etc.).

Manufacture ofhardmetals [9.4]. The manufacture ofhardrnetals is based on powder

metallurgical techniques, which include several steps. Each step must be carefully
controlled to achieve a final product with the desired properties. These steps are:

Preparation of WC powder.
Preparation of other carbide powders.
Production of grade powders (blending, powder milling).
Powder consolidation.
Liquid-phase sintering.
Postsinter operations.

This chapter will follow the different stages of powder metallurgical manufacture. More
emphasis will be put on the description of WC powder production methods and qualities,
and the preparation of graded powders. Less emphasis will be put on sintering, hardrnetal
qualities, and applications. In this context, we refer to several excellent books and review
articles dealing particularly with hardmetal technology, properties, and applications [9.1,
9.2, 9.4, 9.7-9.9].
Economic importance. About 23,000 tons/year of hardmetal are currently used in
the western world, which correspond to more than 60% of tungsten demand [9.1]. The
world market volume for hardrnetal cutting tools in 1995 was estimated to be roughly 9
billion DM [9.10]. The price of hardmetals, on average, equals that of silver. For special
tools, like coated indexable inserts, the price is about double that of silver [9.11].

9.2. RAW MATERIALS

9.2.1. Tungsten Carbide Powder

Tungsten carbide powder can be prepared from different raw materials and by various
production methods, as outlined in Fig. 9.3. Only methods used today on a technical or
pilot plant scale are considered. The conventional procedure covers the widest range of
powder qualities in regard to particle size (0.15 ~ 12 Ilm), and consequently the biggest
percentage ofWC presently produced, while all other methods yield special coarse or finegrained powders.

9.2.1.1. Conventional Production (Carburization) [9.7]. Tungsten powder is reacted

with carbon at temperatures between 1300 and 1700 C in a hydrogen atmosphere. The
average particle size and size distribution of the starting W powder determine the particle
size and size distribution of the WC (Fig. 9.4). Only a slight increase in size occurs due to
the change in density from 19.3 g. cm- 3 (W) to 15.7 g . cm- 3 (WC). In addition a certain
agglomeration (local sintering), always occurs, especially at higher temperatures.
As carbon source, high-purity carbon black is mostly used and, more rarely, graphite,
due to its higher price and lower reactivity. Thermal carbon black is preferred to flame

325

TUNGSTEN IN HARDMETALS
Raw malaria. a .. raactecI with each _
but the main percantaga
Raw malarlals In
solution ara convertad
of WC Is formaci by the solid state - gsa reaction
to solid mlldU.. and
W+CH.-->WC+2H..
carburlzed by solid
_-gsa reaction.

carbon

Tungstan

Oxt

Blending

,I

Double staga
nztsryfumaca
1350'C N,
1650'C H,

raducUon

doubles.

ID ..ggI .....

Rapid
~I

1SOO-2000'C

Sub~icron

WCpowder

"r

Tr

r1

Direct

carburlzallon

Aqusous solutions of

Tungsten

B'ik

Blending

carbu_

1800-2000 'C
Hlr Y8CUUm

-'1 0......,..

Submlcron
to 12 11m
WCpowder

High
temperatura

carIIurIzaIIon
onvenaonsl
1300-1700'C

H.~

II

-.1

\
Coarse high
temperature
WCpowder
(12 50 11m)

COjlts
Mixing

Spray drying

Co, CO..
CH.. H.

.~.

lFIUld1zed bs~ I
carburIzIItIon

Nanoph';'Wc+CO
precursor

WC formation by
melllng metallurgy

Tungsten ore

F"~

AI

I~UmotharmICI
reduction
~ Wscrap

r- Fe

Extramly fine WC
formaci In the vapor .
ph...

G........ compounds
CH.
Hz

we..

~
wall_

Nanophasa
WCpowder

W containing melt

I with
Satu..uon
carbon I

Cooling

iAcldIc dissOlutiOni
ofF.
~

Menstruum WC
coarse

FIGURE 9.3. Routes of WC powder production; conventional manufacture and alternative processes, including
WC/Co precursors.

carbon black because of its higher purity. Pelletized qualities have a smaller apparent
density and are therefore easier to handle.
In any case, carbon black is always more impure than tungsten powder, particularly in
regard to the alkali metals, Ca, Si, Mg, and S. Part of these trace elements are volatilized
during carburization (the percentage depending on temperature). This is why finer WC
powders (lower carburization temperature) are usually more impure than coarser powders.
The two components (Wand C) must be blended thoroughly prior to carburization.
This is done in different types of equipment, like Vor double cone blenders, mixing ball
mills, or high-energy mixers. An even blend is of importance because, during carburization, carbon atoms can only move via diffusion or as methane molecules over very short
distances. Pelletizing or compacting enhances diffusion and increases the furnace capacity.
The carbon balance is of utmost importance and depends on several parameters and
properties:
The particle size of the starting tungsten powder; the smaller the size, the higher
the oxygen content. Part of this oxygen is responsible (especially in case of
submicron powder) for a certain carbon consumption during heating, although
hydrogen as a reducing agent is present.
Content of volatile substances in the carbon black.
Temperature.
Hydrogen flow rate and its moisture content.
Type and duration of mixing.

326

CHAPTER 9

submicron

we

coarse

l/ t m

FIGURE 9.4. SEM images of three different WC powders (right), and of the respective W powders (left), from
which they were made. Note that the particle size of the carbide corresponds well with the particle size of the
metal.

In any case, the difference between the percentage of carbon added and the percentage in
the final product can only be determined by practical experience. Nevertheless, it is
possible to produce we powders with a desired carbon content to an accuracy of
+0.01 wt% e by empirical knowledge of the carbon loss or uptake. It is easy to
understand that all raw material properties, as well as production parameters, must be
kept strictly constant in order to meet the above accuracy.
The desired final carbpn content of the we powder depends on the further production
mode of the hardmetal producer (mainly powder milling and sintering conditions), and
varies between slightly understoichiometric to stoichiometric (6.13 wt%) to slightly
overstoichiometric.

TUNGSTEN IN HARD METALS

327

In producing submicron WC powder, small amounts of grain-growth-inhibiting

substances (inhibiting WC grain growth during carburization and especially during
hardmetal sintering) are sometimes added to the W + C charge prior to blending. The
usual chromium or vanadium are added either as oxide or as carbide. The addition of oxide
must be considered in calculating the carbon balance, because additional carbon will be
consumed for reduction and carburization of the metal oxide.
Carbon vessels are filled with the powder blend. Depending on the furnace type,
either boats or boxes made of dense graphite are used. The vessels are covered with a
graphite cover to avoid any contamination, and pass through the furnace.
Push-type furnaces equipped with heated tubes or channels are mainly used (Fig.
9.5). Construction material for the tubes and channels is either alumina or graphite, and the
heating elements are made of molybdenum wire or graphite. Both materials have
advantages, but also significant disadvantages which shorten their lifetime. The advantage
of graphite is its high chemical stability against trace elements which evaporize during
carburization; its disadvantage is the slow but constant reaction with hydrogen and water
vapor. In contrast, alumina is very stable against hydrogen and water vapor, but reacts with
alkali metals (evaporating from the powder blend), which finally weaken the ceramic by
decreasing the melting point.
Molybdenum heating elements, in combination with alumina tubing or muffles, are
limited in temperature to a maximum of 1650 dc.
All this taken together illustrates that a carburization furnace will in any case, be a
compromise in regard to lifetime.
Induction-heated muffle furnaces with graphite susceptors are used more seldom.
The furnace tubes and the heating elements are swept with dry hydrogen, which acts
as protective atmosphere for the product as well as for the sensitive furnace parts.

FIGURE 9.5. Push-type carburization furnaces used for conventional powder carburization (maximum temperature 1650 0C). By courtesy of Wolfram Bergbau und Hiittenges.m.b.H, Austria.

328

CHAPTER 9

Moreover, it carries away a certain amount of the impurities which evaporate from the
product and leads to a purification. Finally, it favors the carburization reaction by
intermediately forming methaQ.e molecules (see also Section 3.6.1). The latter is of special
importance in carburizing coarser tungsten powder.
earburization temperatures vary between 1300 and 1700 e, mainly depending on
the average particle size of the powder. The smaller the particle size, the lower can the
temperature be maintained. Lower carburization temperature leads to a higher degree of
lattice defects, and consequently to a higher reactivity during sintering, which is
undesirable especially for submicron grades. On the other hand, very fine powders tend
to grow already during carburization at higher temperature. Therefore, a compromise has
to be made in carburizing submicron powder.
Usual retention times in the hot furnace zone are between 1 and 2 hours. The
exothermic heat of reaction can be used in the rear part of the hot zone to maintain the
temperature without any heating.
After heating is completed, the vessels pass a cooling zone, still under hydrogen, and
are discharged at room temperature. More modem furnaces are equipped with locks,and
charging and discharging is done automatically. By using locks, no air is allowed to enter
during loading and unloading, thus avoiding reactions with oxygen or moisture. For we
powders with average particle sizes below 0.5 Ilm (ultrafine grades), special care must be
taken due to their pyrophorictiy, and handling is commonly done under inert gas.
During the carburization, the single we particles agglomerate. The higher the
temperature and the finer the powder, the stronger the agglomeration.
Further treatment of the we powder depends on whether or not the powder is
processed in the same factory. In case of in-house processing no further treatment is
necessary. Deagglomeration is carried out during the graded powder manufacture.
In case the we powder is processed elsewhere, a rigid specification must be observed
not only in regard to the carbon content as already described, but also in regard to a series
of physical properties, such as:

Average particle size.

Particle size distribution.
Apparent (bulk) density.
Homogeneity.

To meet these demands, the powder can be deagglomerated or milled. A more or less loose
deagglomeration can be performed by little jaw or impact crushers. A more rigid
deagglomeration consists of short-time (some minutes) milling in a ball mill. Procedures
of these types will suffice for powder grain sizes between 1 and 12 Ilm, unless higher
apparent density values are unnecessary.
For submicron we powders, an extended milling process is applied, especially in
cases where the subsequent wet milling procedure for graded powder preparation is not
very intense (attritor milling). The we milling can be performed either in optimized ball
mills (hardmetal balls; optimal ball milling conditions avoid any contamination from the
steel walls and keep the abrasion of the hardmetal balls at a minimum) or in jet mills in
combination with a sifter. The main reason for this milling is to destroy any coarser we
particle 2 Ilm) which might be responsible for coarse we crystals in the sintered
structure. Furthermore, heterogeneous impurity particles (graphite from the carburizing

TUNGSTEN IN HARDMETALS

329

vessel and Fe-Ni-Cr containing particles from the reduction boat scale) are finely divided
and distributed. This type of milling does not effectively influence the WC average particle
size.
The lower limit in average WC particle size is about 0.l5 /lm (150 nm). To achieve
this limit, very dry reduction conditions must be applied during W powder manufacture in
order to avoid particle coarsening by CVT (see Chapter 3). This decreases the product
capacity and significantly increases the product cost. However, to date, there is still limited
demand for such extreme grades.
The physical parameters are not only responsible for the microstructure after
sintering, but also for the shrinkage behavior during the sintering period. Therefore,
they have to be kept constant within very close limits.
By far, the biggest percentage of WC is produced by this method.
9.2.1.2. High Temperature Carburization [9.l2]. Preparation of the W+C blend is
the same for this type of carburization as for the conventional type. The main difference
with respect to the foregoing method is the carburization temperature ranging between
1800 and 2200 DC. Furnace types in use are push-type graphite tube furnaces, working
semicontinuously with the powder in graphite vessels, and batch-type furnaces (powder in
trays) either inductively or resistance heated. A modem semicontinuous aggregate is
shown in Fig. 9.6.
High-temperature carburized powders are usually coarse (10-50 /lm). Sometimes,
5-10 /lm powder is also treated that way.
At temperatures :::: 1900 DC, extensive grain (crystallite) coarsening takes place, most
likely by coalescence, yielding in WC crystallite sizes up to 50 /lm. Smaller, average-sized
powder (5-10 /lm) can partly consist of single-crystal particles.

FIGURE 9.6. High-temperature carburization furnace with graphite resistance heating system; capable of 1500 to
2400 C continuous operation; automated introduction and discharge ofthe powdered product loaded in graphite
carriers. By courtesy of Harper International, USA.

330

CHAPTER 9

The push-type furnaces work in a flowing hydrogen atmosphere, while in batch-type

furnaces a reduced pressure hydrogen atmosphere is commonly used. In case of batch
carburization and especially for coarse powder, it is advantageous to precarburize the
material first at 1600 c under flowing hydrogen, then to crush and blend it, and finally to
carburize at high temperature. The precarburization is mainly a cleaning step, and the
subsequent crushing and blending results in an improved carbon distribution.
The percentage of high-temperature carburized WC is low. Main applications are in
road planning tools and soft rock drilling tools. Part of the coarse crystallized WC for these
applications is also supplied by Menstruum WC.
9.2.1.3. Direct Carburization [9.13-9.16]. An alternative for producing WC
powders on an industrial scale is the direct carburization process, which was patented
in Japan. It has been exclusively used there for several years to produce submicron and
ultrafine WC powders of high quality. Other than the conventional process, where steering
of the WC particle size is mainly conducted via the metal powder process, this is done
through the quality of the oxide and carbon source as well as the process parameters.
The necessary equipment consists of mixer, pelletizer, drier, and two rotary furnaces,
as shown schematically in Fig. 9.7.
Raw materials such as W0 3 derived from tungstic acid or low-temperature calcined
W0 3 from APT (via AMT, see Section 5.3) can be used. The procedure is as follows.
Eighty-four parts per weight W0 3 and 16.5 parts carbon are mixed for 1 hour in a
Hensell mixer. Twenty-three parts water are added and the mixture is extruded (1.2 mm
diameter) and cut every 3rd mm and dried ( <0.1 % H2 0). The pellets are fed into the first
rotary furnace and flow automatically from there into the second by gravity.
Typical furnace conditions are: stage I-nitrogen atmosphere, 1350 C (reduction);
stage 2-hydrogen atmosphere, 1650 C (carburization).
The exclusion of hydrogen in the reduction step avoids any tungsten crystal growth
via the volatile W0 2 (OHh. The reaction path follows the sequence: W0 3 ~
(W0 2.9 ) ~ W0 2.72 ~ W0 2 ~ W ~ W 2 C ~ We. The intermediately formed, needlelike W02 .72 decomposes to ultrafine W0 2 nuclei, which are further transformed to WC of
starting materials
rotary carbur lzation
furnace

Mixing:
Time: 1 hour

Granulating:
Pellet size: 1.2 x 3 mm

Carburization:
N. furnace : 1000 - 1600 0 C
H. furnace : 1400 - 20000C

FIGURE 9.7. Schematic presentation of the direct carburization process [9.13-9.16]. By courtesy of Tokio
Tungsten Co., Ltd., Japan.

TUNGSTEN IN HARD METALS

331

approximately the same size. To meet the carbon balance (around 6.13 wt% C), precise
control of temperature and atmosphere is necessary throughout the two-stage procedure.
The finest powders available (Cr3C2 doped or undoped) are characterized by a close
particle size distribution and a specific surface area between 3.0 and 3.5 m2 . g-l. The real
particle size ranges around 0.15 11m, as it can be seen from Fig. 9.S.

FIGURE 9.8. SEM image of an ultrafine WC powder grade produced by the direct carburization process;
20,OOOx. By courtesy of Tokio Tungsten Co., Ltd., Japan.

9.2.1.4. Alternative Processes.

Menstruum tungsten carbide production (Kennametal Inc.) [9.17, 9.1S]. The process
is also called a "macro process" and is unique in that it produces WC directly from ore
concentrates, either scheelite, ferberite, or wolframite. A mixture of ore concentrate with
Fe304, AI, CaC2, and/or C reacts exothermally after initiation (by using a starter) to form
WC, Fe, CaO, and Ah03' The metallic components gather in the bottom part of the melt
and the oxidic in the less dense slag at the top. The process represents an aluminothermic
reduction of iron and tungsten oxides. The tungsten metal formed intermediately in the
molten iron is carburized and, due to the limited solubility ofWC in Fe, it crystallizes. The
following chemical reactions proceed (Me can be Fe or Mn):

+ SAl -+ 9Fe + 4Al20 3

MeW0 4 + 2AI + CaC 2 -+ Me + WC + CaO + Al 20 3
6CaW0 4 + 3CaC 2 + IOAI -+ 6WC + 9CaO + 5Al20 3
3Fe304

The reactants are proportioned to develop a self-sustained exothermic reaction at

about 2500 0c. A typical charge of 72 t of reactants is divided into smaller portions
contained in sheet aluminum bags for progressively charging and maintaining the reaction.
The reaction is performed in circular kilns. The furnace has to withstand the high
temperature and attack of the reactants. It is designed to minimize heat losses by using

332

CHAPTER 9

graphite slabs for the inner wall backed by insulating carbon. Complete reduction and
carburization is finished within 60 minutes and yields about 22 t of WC.
Gangue elements, like Ca, Mg, Si, etc. which remain in the oxidic state, as well as the
Alz0 3 formed are contained in the slag. Carbide-forming elements, like Ti, Nb, Ta, and Mo,
appear in the crystal mass and should be contained in the ore concentrates only in low concentration. Elements which are reduced but do not form carbides are dissolved in the iron.
Part of the tungsten can also be supplied by scrap. In this way the amount of Al
necessary for reduction can be decreased, because scrap contains the tungsten already in
the elemental state and dissolves easily in the Fe melt.
The cooled mass consists of the upper slag layer and the metallic part further down.
The latter contains about 65% WC (Fig. 9.9) in addition to Fe and Mn and the excess of
metallic AI. After separation of the slag, the metallic part is crushed, washed with water to
remove the excess of CaC b and finally treated with mineral acids to dissolve the Fe phase,
while WC remains unattacked.
The WC consists of coarse, well-faceted crystals (40 mesh and below) and contains
approximately 0.2% Fe. The carbon content is stoichiometric and no nitrogen, free carbon,
or TJ-carbides are contained.
The main difference with respect to conventionally prepared WC powder is not only
the coarser crystal size, but that the carbides crystallize from the melt and are practically
free of internal subboundaries. The latter aspect improves the fracture strength of the
carbide [9.2].
A severe disadvantage of the process is the generation of heavy metal salts, mainly
FeCh, the disposal of which is an environmental problem. Another disadvantage is that no
grain size steering is possible, besides milling of the final WC. Further processing of
milled Menstruum WC affords special skill, because a milled 5-llm powder behaves
differently from a conventionally prepared grade.
Menstruum WC is used in hardfacing as well as in high-impact nonmachining
applications.

FIGURE 9.9. WC crystals in a solidified thennite mass; lOx. By courtesy of Kennametal Inc., USA.

333

TUNGSTEN IN HARDMETALS

The Rapid Carbo thermal Reduction (RCR) Process (The Dow Chemical Company)
[9.19,9.20]. In principle, the patented RCR process is another type of direct carburization.
The tungsten carbide powder is continuously synthesized by an extremely rapid carbothermal reduction of W0 3 in a graphite transport reactor. Calculated heating rates are in the
range of 100 million K per second by thermal radiation. The W0 3/C mixture during this
process is converted to a tungsten carbide precursor (WC I _ x ) within seconds.
There exist two basic methods of rapid heating:
Entrainment Method: The powder mixture is fed to a vertical tube furnace and falls
down by gravity in an inert gas atmosphere. The heating rate is between 10,000 and
100,000,000 K . S-I (providing small particle size). Residence times are on the order of
0.2 to 10 seconds (1800-2000 QC). A schematic sketch of a reactor is presented in Fig.
9.10.
Drop Method: The powder mixture is dropped into a preheated crucible (heating rates
between 100 and 10,000 K . S-I). Exposure time may vary between 5 min to 2 h
(1500 QC).
A combination of both methods offers the possibility to produce stoichiometric we.
The WC I _ x precursor is first prepared by the entrainment method, after which the
entrainment gas
cooled transport tube
solids feed

annular gap
insulating plug

<300 K

outlet

""""6U--- insulation
2200 K

'I~s~~~heating
~
element

inlet
<300 K

coolant out

expanded
cooling zone
coolant in

FIGURE 9.10. Continuous graphite transport reactor (entrainment process); patented by The Dow Chemical
Company, USA [9.19, 9.20].

334

CHAPTER 9

necessary amount of carbon is added by mixing, followed by the second step, generally
conducted in hydrogen atmosphere, for example by the drop method.
The method allows the production of submicron WC, but is preferable for average
grain sizes between 0.2 and 0.4 !lm (Fig. 9.11).

FIGURE 9.11. Ultrafine WC produced by the RCR process; 20,OOOx. By courtesy of The Dow Chemical
Company, USA.

The application of this procedure to preparing tungsten carbide is relatively new and
is presently performed on an industrial 500t/year scale.
The Chemical rizpor Reaction (CVR) Process (H. C. Starck) [9.21]. The process is
based on chemical vapor reaction between a metal-containing compound (for example
tungsten hexachloride) and a carbon-containing gas mixture (for example CH4 /H 2 ). The
principle has been known for many years (see also Section 3.5.4), but has only been
realized recently on a technical scale. The aim of the process is to produce a very fine
(3-500 nm), homogeneous powder with a narrow particle size distribution.
The separately heated reactants are fed into a horizontal hot-wall tube reactor under
defined flow conditions (Fig. 9.12). To fulfill the demand of particle homogeneity, each
individual grain must have the same evolution history (i.e., exactly the same nucleation
and growth condition). Therefore, nucleation and exposure time within the reactor must be
well defined in terms of chemical and thermal environment. Wall reactions are avoided by
shielding the wall with a narrow stream of argon, which is led into the tube through an
annular slit system. This also avoids any significant decrease in flow rate of the reacting
particle cloud in the wall near areas of the tube reactor.
A reverse flow filter operating at elevated temperature separates the carbide particles
from the gaseous components and minimizes the amount of gases adsorbed. This is further
decreased in the subsequent vacuum chamber. The powder is cooled and collected either
under vacuum or in inert gas. The process still runs on a pilot plant scale and is currently

335

TUNGSTEN IN HARD METALS

evaporation of the metal

component
separate heating ofthe
process gases
generation of a defined gas flow

, - - - - - - - - metal component

1.Irn=r!rn:::r=l------ non metal component

I-I-tfH+------ evaporater

* + + - - - - - gas preheater
,l-ooH----- annular gaps
it--+-t------ reactor

Ar
exhaust

powder separation
from the gas flow

cleaning of particle
surfaces from
adsorbed Hel

vacuum recipient
vacuum pump
Ar, Kr, NH 3 ,

j====F=F;;;;;;;;====:"'- exhaust recipient

- t - - - - - cooling recipient

powder cooling/
modification of particle surface
powder take out

+ - - - - - - - gas/vapor mixture
+ - - - - - - - vacuum/Ar

packing
container

FIGURE 9.12. The chemical vapor reaction process [9.21]. By courtesy ofH. C. Starck, Germany.

used to produce nanophase TiC and Ti(C,N). It is, however, also capable of synthesizing
nanophase We.
The Spray Conversion Process (Nanodyne Inc.) [9.22,9.23]. In comparison to all the
above-described methods, the process outlined here yields a final product which is already
an intimate mixture of WC and Co. The process is composed of three major stages:

1. Preparation of an aqueous solution of Wand Co in the desired W /Co ratio (with

respect to the final composition of the hardmetal).
2. Spray drying of the solution to obtain an amorphous precursor.
3. Thermochemical treatment of the precursor in a fluidized-bed reactor and
conversion into a nanostructured WC/Co powder.
A simplified flow sheet is given in Fig. 9.13.
Solutions of ammonium metatungstate (AMT) and soluble Co salts are mixed and
spray dried, yielding hollow spheres of an amorphous oxide precursor. The conversion of
the precursor to the nanostructured WC-Co powder is performed by carbothermal
reduction in a fluidized-bed reactor using gas mixtures, such as CO/C02 , CO/H2 , and
CH4 /H2 . The WC particles (grain size 20-50 nm) are agglomerated in hollow, porous
spheres (average size ~70 /lm). Hundreds of millions of nanocrystalline WC grains
embedded in a Co matrix are contained in one sphere. Any desired WC/Co ratio can be set
by solution mixing.

336

CHAPTER 9

addition of soluble
tungsten/cobalt
compounds in the
desired Co/W ratio

afterburner

filter

fluid bed
reactor

solution mix
tank

spray dryer
CO2 CO H2 CH.
reactant gases

product

FIGURE 9.13. Schematic diagram of the spray conversion process (Nanodyne Inc., USA) [9.22, 9.23].

Powders made by spray conversion are further processed similar to conventional

powders. Prior to compaction, the agglomerates are milled down in ball mills or attritors to
fragments of about 0.5 ~m.
9.2.1.5. we Powder Properties and Qualities. The important properties of tungsten
carbide powders today, regarded as relevant for their use in hardmetal production, are as
follows.
Average particle size. The average particle size (APS) of the powder is the main
parameter for we powder classification. It is obtained by Fisher Subsieve Sizer measurements and is commonly given both as "as-supplied" and "lab-milled," according to
ASTM standard procedures (ASTM B330, and 430; see also Chapter 5; Table 5.12).
Unfortunately, the term "particle size" is not well defined, and also depends on the
corresponding measurement technique. Nevertheless, there exists an ISO standard (ISO
3252), which defines the difference between particle, grain, and agglomerate. This
standard is sketched in Fig. 9.14. From this drawing, it becomes clear that the often
used term "grain size" for WC powders is misleading, and should be avoided when talking
about measurement values (FSSS, laser diffraction, etc.). Strictly speaking, the term
"grain" is only correct in case of single crystalline particles (which is not the case for most
WC grades).
For most WC powders (APS: 0.5-4 ~m), there also exists a close relationship
between the APS of the WC powder and the average grain size* (AGS) of the WC in
the sintered hardmetal structure (as long as milling conditions in grade powder preparation
and sintering conditions are kept constant). In this case, the APS of the WC powder is used
as the main steering parameter for the hardmetal structure. This is shown schematically in
Fig. 9.15.

* In this case the tenn "grain size" is correct, because it refers to single crystals.

337

TUNGSTEN IN HARD METALS

grain boundary

we grain

we particle

we agglomerate
FIGURE 9.14. Physical constitution of we powders (schematic).

Very fine we powders (APS <0.5 /lm) are commonly strongly agglomerated. These
agglomerates survive any deagglomeration treatment and will finally co-determine the
values as-obtained by the FSSS measurement. In this case, it seems more sensible to assess
the average particle size by image analysis obtained by high-resolution scanning electron
microscopy.
Particle size distribution. (For determination methods, see Table 5.12.)
The particle size distribution of the we is very similar to that of the corresponding W
powder. The latter is a consequence of the powder layer height during hydrogen reduction
of tungsten oxides and can be somewhat further influenced by the reduction conditions
(see Section 5.4.3).
A certain decrease in the coarse fraction can be achieved by heavy milling of the
powder either after carburization or during grade powder production.
we crystallite (grain) size. we particles larger than 3-4 /lm are commonly polycrystalline. In this case, the size of crystallites (grains; single crystal domains) will codetermine the behavior of the powder during liquid-phase sintering (Fig. 9.15). The
reason for this behavior is the penetration of liquid cobalt along the grain boundaries
during sintering, which leads (at least partially) to disintegration of the former larger
particles. For example, if the average crystallite size of a 5-/lm and 10-/lm average grain
size we is about the same, the respective hardmetal microstructures will also be about
the same.
The crystallite size of coarser-grained WC particles is determined by sectioning of the
Cu-embedded we powder. The size is mainly determined by the carburization temperature
(compare Figs. 9.16 and 9.17), but also by the purity of the raw materials (W,C). The
higher the carburization temperature, the larger the crystallites, mainly as a result of
coalescence processes which are favored by the higher temperature (1650 Qe ~ 1900 QC).
Trace impurities which segregate at the grain boundaries, such as alkali metals, inhibit or
retard grain boundary migration and thereby crystal growth. Therefore, a coarse we

338

CHAPTER 9

we powder
WPOwder
Particle size
submlcron
<111m

0goo 0
00 o~
o

oP

oo~<,
cP 000

ocPoocP

fine/medium
14"m

coarse
>4"m

000

0003~0

000

0 0

8.0~<:Po
o
008
0cP00~0

0000 08
00&0 000 0

we POwde..

Hardmelal
Grain size
with Inhibitor

0 0 0
00 g~eo
9 00
ooOQ
o r!P orPo
elZlo OOgO
&99 Go 0
o QO<0000

f>.f>.:i1,.Df>.f>. 66 tatP

#tP~~~f>.
no inhibitor

~~\l~

0Qeo

O~~rn

!~

000O<DO
- DD, D~

~D,~

0 0 IG) ul
O~
\l \J ~

The w e grain size (crystallite size) depends on the carburization hiSlory; in p:uticular on the carburizarion temperature

FIGURE 9.15. Relationship between particle size of W powder, particle/grain (crystallite) size of the derived
WC, and WC grain size in the sintered hardmetal (schematic).

FIGURE 9.16. Coarse WC derived from a l2l-lm W powder; carburization temperature l650C; powder
embedded in copper. Note the small crystallite size of the particles; during liquid-phase sintering the large
particles will be disintegrated into the small grains (Fig. 9.15). By courtesy of Wolfram Bergbau und
Hiittenges.m.b.H., Austria.

TUNGSTEN IN HARD METALS

339

FIGURE 9.17. Coarse WC derived from a 12-l.lm W powder; carburization temperature 1900C; powder
embedded in copper. The average crystallite size of the particles is much larger than that of the powder in Fig.
9.16; several particles have even become single crystals. By courtesy of Wolfram Bergbau und Hiittenges.m.b.H.,
Austria.

crystallite size is obtained in high-temperature carburized WC using raw materials of high

purity.
Morphology and microstructure. In recent years, it has become more and more
common to include microscopic investigations into routine testing of WC powders.
SEM imaging is especially useful to determine the "true" particle size and the degree
of agglomeration of ultrafine powders (Figs. 9.8 and 9.l1), or to reveal the polycrystalline
surface structure in case of coarse powder (Fig. 9.l8).
Also, light optical microscopy can be helpful to detect coarse particles in submicron
we (Fig. 9.19), or to investigate the internal structure of coarse we in order to determine
the frequency of particles which are not totally carburized (W 2C or W 2C + W in the
particle interior) (Fig. 9.20).
Apparent density (scott density) and press density. (For determination methods, see
Table 5.l2.)
Both properties are determined by the average particle size and size distribution of the
powder, as well as by its agglomeration and the particle shape. Agglomerates and their
strength are influenced by the carburization temperature. Powder milling, depending on its
intensity, leads to a more or less complete breakdown of the agglomerates, and increases
the values for both the densities, although during subsequent grade powder milling further
comminution may occur. The powder density values are closely related to the shrinkage
during sintering. Controlled shrinkage is a very important property in "near-net-shape"
sintering of hardmetal parts.
Carbon content. The carbon content of a WC powder is commonly stated as total
carbon and free carbon (uncombined carbon present as graphite or amorphous carbon).

340

CHAPTER 9

FIGURE 9.18. Coarse WC powder obtained by high-temperature carburization of a 25-llm tungsten powder. The
smooth surface of the particles is characteristic for the high-temperature treatment (1900 0C); nevertheless, grain
boundaries are still visible which indicates that the particles are polycrystalline. By courtesy of Wolfram Bergbau
und Hiittenges.m.b.H., Austria.

The difference between these values accounts for the combined carbon (chemically bound
as carbide).
The total carbon content should be theoretically 6.13 wt% carbon, which means that
all carbon is chemically combined as we. However, in practice, a small amount of
uncombined carbon is always present (some hundredths of a percent).

FIGURE 9.19. Coarse we particle within a fine powder matrix; powder embedded in copper; light optical image.
By courtesy of Wolfram Bergbau und Hiittenges.m.b.H., Austria.

TUNGSTEN IN HARD METALS

341

FIGURE 9.20. Very large carbide particle (> 100 /.lm) consisting of an outer, polycrystalline ring of WC and a
core of W 2C; light optical image. By courtesy of Wolfram Bergbau und Hiittenges.m.b.H., Austria.

The proper total carbon depends on further processing of the powder (compact) and is
dictated by the hardmetal producer, due to its processing experience (carburizing, neutral,
or decarburizing conditions).
The determination method is described in ISO 3907 and ISO 3908 and should result
in a standard deviation of no more than + 0.0 I wt%.
Content of grain growth inhibitors. Grain growth inhibitors like VC or Cr3C2 are
sometimes added prior to carburization to minimize particle/grain coarsening. This
procedure has become industrial practice for so-called ultrafine grades (sintered structures
with grain sizes .:s0.5 11m). Further additions are commomly made to the grade powder
milling process. Typical additions to the powders prior to carburization are in the range of
0.1 to 1 wt% VC resp. Cr3C2.
Trace impurities. Trace impurities play an important role in hardmetal manufacture,
since they can cause numerous types of defects (porosity, inclusions, origin of local grain
growth, etc.). They should be kept as low as possible in the powder, in particular when a
detrimental action has already been demonstrated (as, for example, in the case of Ca, Si,
AI, and S).
The impurity content of a WC powder is the result ofthree main factors: (1) the purity
content of the starting W powder; (2) the impurity content of the carbon black or gaphite,
and (3) the carburization temperature (the higher the temperature, the more the trace
impurities are evaporated during carburization). A typical analysis is shown in Table 9.1.
Although the above characterization in regard to physical and chemical properties
seems to be quite rigid and complete, in some cases unexpected and unsatisfactory results
may occur during the further hardmetal production. The reason for this behavior is based
on still insufficient specification, because the carbon black quality, carburization temperature, as well as type and duration of milling after carburization are not clearly defined, and
can be changed by the supplier or may differ between diverse suppliers.

342

CHAPTER 9

TABLE 9.1. Trace Impurities in we Powder (upper limits

in Ilg . g-It
Element

Concentration

Al
As
Ca
Cd
Cu

20

10

Li
Mg
Mn
Mo
Na
P
Pb
Sn
Si
Ti

2
2
2
30
20
20
2

10

20
2
2

10
10
10

Product data sheet, Wolfram Bergbau- und Hiittenges.m.b.H.

Conventional tungsten carbide powder. The particle size range of these qualities is
between 1 and 10 Jlm (the majority between 1 and 2.5 Jlm). The average particle size has
to be met usually within the limits of + 10% relative. Particle size distributions for a given
average particle size can be adjusted within certain limits already during tungsten powder
reduction (see Section 5.4.4).
In order to minimize testing operations for the hardmetal producer, it has become
common in recent years to prepare homogeneously blended lots of 5 t or even 10 t. In
blending, it is strictly forbidden to use smaller lots of strongly diverging properties to meet
the specified values.
Submicron tungsten carbide powder. In principle, every we powder having an
average particle size below 1 Jlm can be regarded as "submicron." However, it has become
common practice to further classify these powders into the following three groups:
Submicron we
Ultrafine we
Nanosized we

<1.0 to 0.5 Jlm,

<0.5 to 0.10 Jlm,
<100 nm.

Fine-grained we powder grades have gained significant importance during the last 10
years as demonstrated in Fig. 9.21. While submicron qualities still dominate the market,
there is growing interest in using even finer grades for specific applications (machining of
highly abrasive materials). Ultrafine grades are offered today by several producers, based
on conventional powder processing (Fig. 9.22) or alternative routes (Figs. 9.8 and 9.11).
Nanosized powder precursors are also available, but until now could not be transformed
into pure nanocrystalline composites by conventional liquid-phase sintering due to their
high interface energy, which causes rapid grain coarsening. Alternative consolidation
techniques are necessary in order to exploit their potential.

TUNGSTEN IN HARDMETALS

343

45
'0

c:

40

35

01

GI

'0

.E
GI

'"'"e

30
2S

20

011

15

i:'"

10

.5
CI

..
011

011

g,

5
0
1980

FIGURE 9.21. Demand in submicron

statistics.

we

1985

1990

1995

year

powder during the period 1980-1995 based on current industrial

Coarse tungsten carbide powder. Coarse WC powders are available in the particle
size range of >8 to 50 ).lm. Their crystallite sizes vary from <10 ).lm up to about 50 ).lm
and mainly depend on the carburization temperature (strictly speaking, the maximum
erystallite size is determined by the particle size).
9.2.2. Other Carbide Powders [9.7, 9.8]
The following carbide powders are used in hardmetal manufacture: WC-TiC, WCTiC-TaC(NbC), TaC, TaC(NbC), as well as M0 2C, Cr3C2, and VC (the latter are used in
small concentrations as WC grain growth inhibitors).

ULTRAFINE STANDARD
PRODUCTION WC

.. R&D POWDER"
MADE IN PRODUCTION UNITS

FIGURE 9.22. Ultrafine we powders produced by conventional (reduction/carburization) manufacture;

40,000 x. The size of the spheres emphasizes the degree of fineness of the development grade. By courtesy of
Wolfram Bergbau und Hiittenges.m.b.H., Austria.

344

CHAPTER 9

WC-TiC mixed crystal carbides (solid solutions). The raw materials are W powder
(1-2 11m), TiOl, and carbon black. The proper removal of oxygen is essential for obtaining
a high-quality (pore-free) hardmetal. The most common compositions are WC/TiC 70:30
and 50:50 (wt%). The carburization is carried out mainly in vacuum aggregates, more
rarely under hydrogen. Resistance-heated (carbon or graphite heating elements) or
induction-heated furnaces are in use. Carburization temperatures are up to 2100 0c. A
double carburization for proper carbon adjustment and for the total removal of residual
oxygen and nitrogen is preferable. The second step should be performed at a temperature
200 K higher than the first.
The powder properties are either directed by powder milling of coarser powders
(high-temperature carburized), or by steering the particle size through the oxide precursor
and subsequent soft carburization (lower temperature, but longer carburization times).
WC-TiC-TaC(NbC} mixed crystal carbides (solid solutions). The raw materials are
the same as for WC-TiC grades, with addition of Ta205. Any furnace equipment capable
of producing a good quality of WC-TiC is suitable. Total reduction and complete diffusion
can only be reached by high temperature, good vacuum, and long exposure time.
Carburization temperatures are about the same as for WC-TiC.
Nb, which is a natural companion of Ta, can be tolerated up to 30%, but the exact
content must be known for optimal carbon adjustment. Typical grades contain 33-50 wt%
WC.
TaC and TaC(NbC). The preparation is similar to that of the above-described mixed
crystal carbides. The expensive TaC 1% NbC) as well as the cheaper mixed Ta/Nb
carbide (Ta/Nb = 6: 1 to 1: 1) are applied.

9.2.3. Binder Metals [9.7, 9.8]

Cobalt powder. Powder grades with average particle size between I and 5 11m are
available for hardmetal manufacture. They are produced by hydrogen reduction of cobalt
oxide, oxalate, or carbonate. Ultrafine Co powders (particle sizes of 0.3-0.5 11m) are
prepared through the reduction of cobaltous hydroxide by a mixture of ethylene-glycol
and diethylene-glycol ("polyol" cobalt), or by hydrogen pressure precipitation in autoclaves out of cobaltous sulfate solutions (using nucleation catalysts).
Nickel and iron powder. Fine Ni and Fe powder is obtained by hydrogen reduction of
the oxide or decomposition of carbonyl compounds.

9.2.4. Other Materials [9.7, 9.8]

As solvents or milling liquid in grade powder milling, several organic compounds are
employed: alcohols, acetone, hexane, heptane, gasoline, and tetraline. Hexane, heptane,
and acetone are the most preferred today.
The most popular lubricant is paraffin wax, other than bees wax, camphor, and
polyethylene-glycol. Besides its lubricating properties, its purity is essential. It must be
removable by heating without any residue.
As plastifiers for extrusion, cellulose derivatives are in use.

345

TUNGSTEN IN HARD METALS

9.3. HARDMETAL PRODUCTION

The preparation of a hardmetal specimen consists of several steps and can be
performed along different paths. Basically, one may distinguish between three or four
main steps:
Preparation of powder grades; this stage includes the blending of all necessary
components, and is combined in most cases with a more or less rigid powder
milling. A big percentage of the grade powder is granulated.
Consolidation by compacting or extruding, in some cases followed by shaping.
Sintering.
Postsinter treatment.

A generalized flow sheet, outlining the different modes of preparation, is presented in

by Fig. 9.23.

9.3.1. Preparation of Powder Grades [9.1, 9.7, 9.8, 9.24]

Weighting. The necessary amounts of all constituents (WC, other carbides, grain
growth inhibitors, additional carbon black, and binder metal) and of all further necessary
temporary additions (organic solvent, additional organic liquids, lubricant) are weighted or
measured precisely and loaded into a suitable blending-milling device. The usual
equipment for that purpose consists of ball mills, vibratory mills, rod mills, or attritors
(mostly ball mills and attritors).
WC

Co

Ni

Ta(Nb)C

Plastlfler

Kneading

Wet milling

Wet screening

,
,

P..slnterlng

Sinterlng-

Extruding

Lubricant"

I I Granulation I I
I Screening I I'-----'----''C::==-.-==:::r----'
I
Gran+Ulate
Grade~wder
Spraydrylng

Plastlfl.!ct mass

Blending

I Vacuum drying I

Cr,C2

Weighting

VC

TiC

Drying
Hotp..sslng

1~:---;:::::::=:;t:=:=;-;==::::l..---,
I
I I

,
,
,

'Cold isostetic pressing'

Green

Preslnter
Forming
Sinterlng-

I
I

Die compacting

~ompact
I

Graan forming

Sinterlng-

I
I

Block p....

Graan forming

Sinterlng-

.
-

Only for apraydrylng and

granulatton
Sinterhip may be Inclucled

... Dewax Included

+
Sintered hard metal

FIGURE 9.23. Schematic flow chart for hardmetal production, showing the variety of production routes.

346

CHAPTER 9

Powder milling. The goals in powder milling can be quite different, and can vary
between a soft deagglomeration of particle agglomerates combined with mixing of the
components (mixing mill), and heavy milling of the powder, whereby significant particle
comminution will occur. Depending on the goal, different aggregates and milling
conditions are applied.
In general, a fundamental knowledge of the milling process is required. Mill
dimensions, rotational speed, ball diameter, loading of milling media, and hardmetal
components must exist in an optimized relationship. In this way the uptake of iron
(abrasion) from the mill walls and of cemented carbide from the milling media can be kept
at a minimum. The weight ratio of milling media to powder together with the milling time
is responsible for the amount of disintegration. The higher the ratio, the more pronounced
the milling (comminution).
Any particular mill and mill charge combination has a limiting particle size, beyond
which no further comminution will occur. For most industrial aggregates, this is in the
range of 0.5-1 11m [9.25]. Only in optimized mills is further comminution possible, finally
reaching a size limit of about 100-150 nm [9.24]. This limit is inherent to the WC (i.e., to
its mechanical properties). This kind of milling is used only by specialists.
Commonly, either ball mills or attritors are used. Ball mills that were in use at the very
beginning of hardmetal production, and which had a volume of only about 10 liters, have
disappeared, and modem aggregates possess volumes of up to 500 liters (charge weights
up to 700 kg). This change has also significantly improved their milling efficiency (milling
times of 4-6 days in a lO-liter mill was reduced to 1-3 days in a 300-liter mill). Basically,
the ball mill is more effective for the coarser fraction of the powder, which is trapped
between the balls or studs. This is why a wide particle size distribution can be improved by
ball milling. In addition, proper ball milling provides an excellent carbide/binder
distribution and minimizes the danger of cobalt "flake" formation (which results in
porosities during subsequent sintering).
About 40 years ago, a large proportion of the ball mills was substituted by attritors.
The advantage of the latter is a much shorter milling time (one-fifth or one-tenth of the
time needed in ball mills), which is the main reason why they have become increasingly
popular. This means hours of milling instead of days. However, this advantage is combined
with certain disadvantages. The attritor mill produces a Gaussian particle size distribution
which enhances the carbide grain growth. Furthermore, heterogeneities can be trapped in
"dead areas," which can be only partly overcome by the attritor design and by pumping
and recirculating the slurry from the bottom to the top of the attritor.
The choice of the correct mill and milling conditions depends on the hardmetal grade
produced, mainly on the WC grain size. Attritors are used today for most powder qualities
(1-12 11m). They have become increasingly popular because of the significantly shorter
milling times. Optimized ball mills are still preferred for submicron qualities, which are
very sensitive to all types of heterogeneities.
During milling, several processes occur simultaneously: break-up of agglomerates,
blending, and comminution. In general, the surface energy of the mix is increasing, which
enhances the sintering. Sometimes, the carbides are milled alone or with only a small part
of the Co ("split milling") to achieve a more intense milling in a shorter time, while the
rest of the binder is added only in the late part of the milling. Cobalt, by its large. powder
volume (low density compared to WC), decreases the probability for a WC particle to be

TUNGSTEN IN HARDMETALS

347

hit by a ball, thereby prolonging the milling time. In any case, a uniform distribution of the
binder is important in order to avoid increased porosities in the sintered body, which is
particularly important for ultrafine grades. For these grades, longer milling times are
therefore mandatory. The use of ultrafine Co powder grades has further contributed to
achieving better distributions.
Grade powder milling is always conducted in an organic liquid to prevent local
overheating and oxidation. The enormous milling energy, which is converted to heat, must
be removed from the mill by cooling (mills equipped with water jackets).
Balls or parts of the propeller arms (in the case of attritors) can fracture during
milling, leading to hardmetal fragments which are not further disintegrated. Moreover,
agglomerated cobalt powders can be forged to flat platelets (flakes), which resist any
further comminution. Therefore, the slurry after ball removal is screened (400 mesh) to
eliminate any heterogeneities.
Lubricants. Either before milling or afterward a lubricant (wax, polyethyleneglycole,
camphor, etc.) is added in amounts of 1-3 wt%. It dissolves in the solvent (acetone, hexan,
alcohole) and is distributed homogeneously. It remains in the grade powder and protects it
from oxidation after distillation of the solvent. The solvent is recycled to the next batch.
Moreover, the lubricant improves the strength of the green compact.
Granulation. A big percentage of graded powder is granulated by either spray drying
or mechanical granulation. Granulation is a necessary prerequisite for automated compacting combined with near-net shaping. Both granulation methods yield a free-flowing
powder consisting of spherical particles It guarantees the highly reproductive filling of
the compacting dies. Spray-dried granules are commonly hollow spheres, while granules
made by mechanical granulation are full. Therefore, the filling height during compacting is
higher for spray-dried powder (lower apparent density).
Spray drying is a combination of the distillative separation of the solvent and the
granulation process and is the most frequently used method for granulation of powder
grades on a bigger scale. In processing, the sludge is forwarded to a spray nozzle (pressure
5-10 bar). On leaving the nozzle it is finely divided into droplets (10-50 /lm) forming an
upward-directed jet. The jet forms a taper, roughly fitting the upper end of the chamber.
The forming of granules is a rather complicated mechanism. Hot gas (100-130 0c) heats
the incoming particles and the milling liquid is evaporated. Particles of different speed and
moisture grade hit each other, resulting in quite spherical particles of 100-400 /lm.
Particles big enough settle. The particle size distribution is quite close. To obtain the
desired size, the working conditions must be adjusted.
The granules collected at the bottom are evacuated into a gas-tight container. Critical
properties of the granules, besides their size, are also the apparent density and hardness.
The density is responsible for the filling reproducibility of the compacting die. If the
hardness is too low, granules may be smashed during handling, while too high a hardness
causes granules to partially survive during compacting, yielding an increased residual
porosity during sintering.
Approximately 5% of the powder remains ungranulated, is separated by the cyclone,
and can be used for purposes where granulation is not required. The solvent is almost
totally recovered and can be recycled to the milling stage.
The capacity of a spray drier is between 120 and 180 kg . h -\. The size of the spray
drier is about 5 m in height and 2 m in diameter.

348

CHAPTER 9

9.3.2. Powder Consolidation [9.1, 9.7, 9.8]

Direct compacting. Direct compacting is applied for "near-net shaping" of parts.
This means that the final hardmetal specimen after sintering already has the desired
dimensions and only limited surface treatment is necessary afterward. For that purpose,
semiautomatic or automatic mechanical or hydraulic presses are used. The pressure
application is only from one direction, resulting in a slightly anisotropic density distribution in the green compact.
Cemented carbide powders do not deform during compaction. Therefore, the
maximum green density is about 65% of the theoretical density (35-50% porosity).
Compacting pressures are 100-300 MPa, usually around 200 MPa.
The pressing dies (fabricated of hardmetal) are made to the shape of the desired end
product, which means that the die must be greater to allow for shrinkage occurring during
sintering.
The height of the green compact corresponds to 35-40% of the filling powder height
in the die. Shrinkage of the compact during sintering is between 15 and 20% in each
direction. Shrinkage calculations in Europe are based on the as-sintered size, while in the
USA it is based on the compacted or presintered size. Thus, a value of 25% in Europe
corresponds to about 20% in the USA. The percentage of shrinkage is generally smaller in
the pressing direction than perpendicular to it.
Nevertheless, today's advanced granulation and pressing technology have made it
possible to achieve very rigid dimensional tolerances of the sintered product. It is easy to
understand that a prerequisite is to keep the properties of the raw materials, as well as the
conditions during grade powder preparation and granulation, strictly reproducible.
Direct compacting is mainly used in producing mining tools, metal cutting inserts,
and parts for construction applications. Around two-thirds of hardmetal production is
compacted by this method.
Cold isostatic pressing. The powder is compacted in a tight rubber or plastic
container inserted in a liquid under pressure. Besides wear parts and metal-forming
tools, mainly bigger pieces are compacted by this method and in most cases need further
indirect shaping. The subsequent forming can be done either in the "green" state or after
presintering. All types of machining and grinding operations can be used. In forming, the
sintering shrinkage must also be considered. The dust as well as the scrap originating from
these operations is collected and recycled internally.
Extrusion. Round or profiled rods or tubes can be produced by extrusion. The
amount of lubricant must be increased and additional plastifiers added. The corresponding
recipes are kept secret.
Injection molding. Injection molding can be regarded as a combination of extrusion
and precision die compacting. A highly plastified mixture of the hardmetal powder and
lubricants is screwed under pressure into a precision die. The hardmetal-to-plastifier ratio
is roughly 90: 10 by weight or 50:50 by volume. The procedure is only economic in case of
complex shapes and a large number of parts. Dewaxing is carried out very slowly and
requires special care, as for extruded material.
9.3.3. Sintering
A common sintering cycle includes several stages, such as dewaxing, presintering,
sintering, and cooling. For pressed-to-shape parts, it is customary to combine dewaxing,

349

TUNGSTEN IN HARDMETALS

presintering, and sintering in one aggregate and heating cycle, thus minimizing energy
consumption and handling. It is a prerequisite that the furnace is equipped with a wax
condenser to remove the lubricant from the sintering chamber. Only for parts which must
be shaped before sintering a presintering stage is performed at 700-1000 C to give the
compact sufficient strength to permit forming. This is commonly done in special furnaces.
Sintering is conducted either in hydrogen (for straight grades and low TiC-containing
alloys) or, more commonly, in vacuum. For higher TiC- and TaC-alloyed grades a good
vacuum is mandatory, in order to avoid residual gas porosity. In general, there is a
tendency today to use vacuum sintering for all grades.
A typical sintering cycle is shown in Fig. 9.24. At first, the charge is slowly heated up
to 400--500 C to render proper dewaxing of the compacts. The temperature is then raised
gradually to the optimal sintering temperature of 1350-1600 C, which depends on the
hardmetal composition and the WC grain size. Several holds at certain temperatures are
common during heating to al10w outgassing of carbon monoxide (which forms as a result
of residual oxygen still present in the densifying compact), unless there is open porosity.
Isothermal liquid-phase sintering lasts about 1-1.5 hours.
Densification starts below the liquidus temperature during heating (nonisothermal
sintering). This solid-state shrinkage is especially pronounced in submicron and ultrafine
grades. It is due mainly to particle rearrangement processes, which occur as a result of the
spreading of cobalt along the WC surfaces, and subsequent contraction under the action of
capillary forces. Both processes are activated by the partial dissolution of WC in cobalt.
At about 1300-1340 DC (binary eutectic temperature WC-Co 1310 DC; ternary
eutectic temperature W-Co-C 1275/80 DC) partial melting occurs, and more WC is
dissolved until the eutectic concentration (54% Co and 46% WC) is reached [9.8]. A
further increase in temperature results in additional dissolution of WC and complete
melting of the binder phase. In this stage, rapid final densification occurs and the sintered
body is practically pore-free.
Microstructural changes. At 1400 DC, the constitution of the molten phase corresponds to about 50% Co and 50% WC [9.8]. The liquid phase preferably dissolves the
- shrinkage-

1500

- - - outgassing - - -

1000
dewaxing~

~
:J

ai
Q;

Co

{!!.

~format i on

500

of eutectic
liquid

non-isothermal solid - state sintering

O~~--r-'-~--r-~~--r-~---r-'-~~

Time (hI

10

15

FIGURE 9.24. Representative sintering cycle of hard metal production, indicating the different stages of sintering.

350

CHAPTER 9

surface layers of the WC particles and grains which have been activated during powder
milling, as well as bridges between agglomerated carbides. The morphology of the
particles changes from rounded to prismatic (formation of low-energy interfaces).
Furthermore, the melt disintegrates large polycrystalline carbide particles into individual
grains by penetrating into the large-angle grain boundaries. During isothermal sintering,
further adjustment of the microstructure occurs due to solution/reprecipitation and/or
coalescence processes, which lead to a more uniform distribution of the cobalt binder and
a decrease in carbide contiguity. This is an important aspect for obtaining optimal
mechanical properties of the sintered parts. Too high a sintering temperature will promote
unwanted grain coarsening.
On cooling, the dissolved WC is precipitated onto already present WC grains. Certain
amounts of Wand C remain in solid solution, their amount depending on the gross carbon
of the respective alloy, and the cooling rate. The W content in solid solution at room
temperature varies between 18-20 wt% W in substoichiometric alloys and 1-2 wt% Win
stoichiometric alloys.
The solutes stabilize the cubic modification of cobalt down to room temperature. Due
to the different coefficients of expansion of the carbide(s) and cobalt, the binder phase is
under tensile stresses and the carbides under compressive stress. Uneven macroscopic
stresses can lead to distortion of the sintered body.
Shrinkage during sintering corresponds to 17-2S% on a linear scale, or 4S-60% by
volume. The amount of shrinkage depends on WC powder particle size, the milling,
granulation, and compacting conditions, and the grade composition.
The carbon balance of a hardmetal is a very important parameter. The W:C ratio
should be close to 1: 1. If the carbon content is too low (strongly sub stoichiometric), 11phases of the M6C type will be formed and lead to embrittlement of the alloys. If the W:C
ratio is higher than stoichiometric, graphite is precipitated (C porosity), which also leads to
product deterioration. Thus, only a small gap exists in terms of the carbon in which only
WC and Co are present. This gap is smaller, the lower the cobalt content. Most hardmetals
are slightly substoichiometric to obtain an optimal property relationship and to avoid
precipitates. Tungsten in solid solution improves the strength of the binder and makes the
material less prone to plastic deformation during machining (solid-solution strengthening).
The carbon balance is severely affected by any traces of oxygen present in the grade
powder, but also by the sintering atmosphere. The latter can be both decarburizing as well
as carburizing. Being able to maintain the right carbon balance during sintering comes
from experience (empyry) and requires very strict reproducibility in all the foregoing steps.
It is most difficult in very low Co grades, where the gap between graphite or 11precipitation is the smallest.
Pressure sintering. Sometimes, hardmetals exhibit residual porosity (A and B type),
which can occur for several reasons. Since the 1970s pressure sintering (HIP-ing) has been
applied to remove this porosity. For that purpose, the sintered specimens are loaded in a
HIP-ing device and heated again under Ar or He pressure of SO-ISO MPa (temperatures
and commonly 2S-S0 C below the vacuum sintering temperature). Hence any residual
porosity can be removed. Exceptions are gas porosities, which either can occur due to
incomplete outgassing of the carbon monoxide, or due to the presence of impurities.
In recent years, this rather expensive two-stage treatment (vacuum sintering and
subsequent HIPping) was substituted by the sinterhip process, which was developed in the

TUNGSTEN IN HARD METALS

351

early 1980. Sinterhip represents low-pressure isostatic pressing up to approximately 7 MPa

combined with normal vacuum sintering. At the end of the common sinter heating cycle,
the pressure is applied as long as the binder is still liquid. Void-free specimens can be
produced.
Sintering aggregates. The majority of modern hardmetal sintering furnaces are
batch-type aggregates which are resistance-heated (graphite heating elements). Examples
are shown in Figs. 9.25 and 9.26. More rarely, semicontinuous push-type furnaces, which
operate under flowing hydrogen, or inductively heated aggregates equipped with locks are
used. All of them are electronically monitored (temperature and pressure control) and
equipped to work under H2 or controlled H2/CH4 mixtures (sometimes pulsating). Special
heat treatments are common practice today and allow for tailormade properties in different
parts of the hardmetal piece (hardmetals with a gradient structure). Characteristic examples
are shown in Figs. 9.27 and 9.32.
Hot pressing. Hot pressing combines the two steps of compacting and sintering into
one. The graphite mold is filled with hardmetal powder. The mold itself is the resistance
heating element. Pressure is applied throughout the heating period. Hot pressing includes
high labor costs and is generally only applied for special parts and/or alloys, which are not
easy to sinter to full density otherwise.

9.3.4. Postsinter Treatments [9.1,9.7,9.8]

Coating. Coating was the logical consequence in the historical ' development of
cutting tools. Braised tools in the past were reground several times when their cutting
edge was blunt, unless almost the whole tip was consumed. Later, with less specialized
labor, higher costs, . and pressure in time, the throwaway insert was generated. Held
mechanically, indexable inserts having several cutting edges or corners came on the

FIGURE 9.25. Three-chamber vacuum smtering furnace for hardmetal production (Ipsen). By courtesy of
WIDIA GmbH., Germany.

352

CHAPTER 9

FIGURE 9.26. Sinterhip furnace for pressure sintering ofhardmetals (max 60 bar). Charges of up to 1000 kg can
be sintered within a 24-hour cycle. By courtesy of ALD Vacuum Technologies GmbH., Germany.

market. When they are discarded, over 99% of their original mass remained. Discarding
does not mean the material is completely lost. Today, the tips are collected and sold as
scrap.
Consequently, it became reasonable to apply a hard and wear-resistant surface to a
less hard but much tougher substrate by CVD. It is easy to understand that for a throwaway
part the lifetime is of utmost economic interest, and by the invention of coatings it could be
prolonged by a factor of 2 to 5.
Coating became one of the most significant developments in the history of hardmetals
which started in the early 1960s and is still underway today.
In coated parts, the carbide plays a different role than in noncoated parts, because it is
no longer the active component. It is the coating which has to withstand the erosion from
the machined material. The carbide part in turn has to supply the best mechanical support
for the coating (rigidity, toughness, and thermal properties) and allow perfect bonding to
the coating (good adhesion) in order to resist spalling. When the coating is abraded, the
cutting performance decreases to the value of the noncoated carbide.
TiC layers were the first to be applied by CVD. Others, like TiN, TiCN, Ah03, Hft:,
HtN, ZrN, AION as well as combinations and permutations, followed. The newest
generation are multilayer coatings, which are deposited on binder-phase enriched
substrates (to increase toughness); see Fig. 9.27. Further examples are TiC, TiCN, TiN,
sequences, or an 8-layer undercoat composition topped by AINO. Also, 13-layer
compositions can be found. The thickness of single layers was 3-7 ~m and multilayer
coatings range from 8-12 ~m in total.
Coatings are made by CVD, MT (medium temperature)-CVD, PVD, and plasmaactivated CVD. The latter technique was recently successful in producing adherent
diamond layers. The keenest edges are now produced by PVD coating.
Coated hardmetal is the dominant tool material for indexable inserts and will remain
so in the foreseeable future. More than 80% of all turning inserts and about 70% of milling

TUNGSTEN IN HARDMETALS

353

FIGURE 9.27. Cobalt-enriched hardmetal substrate, obtained by gradient sintering with CVD multilayer coating.
Outer layer: TiCN/TiN; middle layer: A1 20 3 ; inner layer: TiC/TiCN. By courtesy of Kennametal Inc., USA.

inserts are coated today. Also, the proportion of PVD-coated drilling tools is steadily
increasing.
Other postsinter treatments. The following methods are in use: postsinter forming,
spark erosion, grinding, electrochemical machining, ultrasonic machining, conventional
machining, lapping, honing, and polishing.

9.4. HARDMETAL QUALITIES AND APPLICATIONS [9.1, 9.2, 9.4, 9.7, 9.8, 9.11,
9.26-9.28]
As mentioned earlier in the introduction to this chapter, the properties of WC-Co
based hardmetals can be varied widely and consequently their applicability is extremely
widespread. The properties are intimately connected with their microstructure (including
micro- and macroporosity) and surface conditions (grinding cracks and excessive roughness). These can be influenced by several raw material properties and processing
conditions:
The average particle size, size distribution, and internal structure of the WC
powder; its purity.
The nominal composition.
The production methods (especially powder milling and sintering).
Excess or deficiency of carbon (stoichiometric and substoichiometric alloys).
Postsinter treatments (HIPping, coating, grinding, polishing, etc.).
It is important to bear in mind that many of these items are interlinked.
The basic properties ofWC-Co and WC-(W,Ti,Ta,Nb)C-Co grades are summarized
in Table 9.2.

CHAPTER 9

354

TABLE 9.2. Basic Data for Different WC-Co and WC-(W,Ti,Ta,Nb)C--Co HardMetal Grades [9.1]

Gradea
(wt%)

WC-4Co
WC--6Co/S*
WC--6Co/M**
WC--6Co/C***
WC-25Co/M
WC--6Co-9 .5(Ti,Ta,Nb)C
WC-9Co-31 (Ti, Ta,Nb)C
a

Transverse
Mean thennal
Compressive
rupture
Young's
Fracture
expansion
Hardness
strength
modulus
toughness
coefficient
strength
(N. mm- 2) (N. mm- 2) (kN. mm- 2 ) (MPa . m- 1/ 2) (10- 6 . K- 1)
HV30
2000
1800
1580
1400
780
1700
1560

7100
6000
5400
5000
3100
5950
4500

2000
3000
2000
2500
2900
1750
1700

665
630
630
620
470
580
520

8.5
10.8
9.6
12.8
14.5
9.0
8.1

5.0
6.2
5.5
5.5
7.5
6.0
7.2

s... submicron; M ... fine/medium; c ...coarse.

Straight Grades. In regard to quantity, this group is by far the most important. Figure
9.28 details the impressingly broad applicability of the two-component system in which,
by varying the WC grain size in the sintered structure (0.4-7 /lm) and the Co content (330%), quite remarkable property changes can be obtained.
Applications can be divided into two main fields: nonmachining and machining.
Nonmachining. Thirteen percent of tungsten consumed by the hardmetal industry,
corresponding to 6% of world tungsten demand, are used in underground mining, road

FIGURE 9.28. Application areas of straight WC-Co hardmeta1s. By courtesy of Sandvik Hard Materials, Sweden.

TUNGSTEN IN HARDMETALS

355

construction and maintenance, rock drilling, oil and gas drilling, as well as tunnel boring
(Figs. 9.29-9.31). Consumption is dominated by coal mining.
Conventional coatings have failed to find a niche in these marketplaces, due to the
aggressive nature of the applications [9.26].
Grain sizes for these applications are commonly in the medium-to-coarse size range
(for certain applications up to 30 11m).
Hardmetals are also used as a support for polycrystalline diamond cutting tips, or as
matrix alloy for diamond grit, used in rock drilling and oil mining [9.2].
Further nonmachining applications include metalforming (drawing dies, rod mill
rolls, wire flattening rolls, slitter knives, extrusion punches, dies, etc.]; (see Figs. 9.32 and
9.33), structural components (plungers, boring bars, dies, pistons, hammers, etc.), and
fluid-handling components (seal rings, valve stems, and nozzles).
Wear resistance and toughness of parts can be further improved by controlled
redistribution of the binder phase ("dual property" carbide) [9.28], which is achieved
during sintering. The principles of such a gradient material are demonstrated in Fig. 9.34.
Machining. The grain size of WC in the machining industry is in the range
< I-211m. Straight grades are used for machining of cast irons, hardened steel, stainless
steels, nonferrous metals, nickel-based high-strength alloys, wood, plastics, composites,
etc. (see Figs. 9.35-9.37).

FIGURE 9.29. Examples of the broad variety of mining tools. By courtesy of Sandvik Rock Tools, Sweden.

356

CHAPTER 9

FIGURE 9.30. Modern button bit rock tools. The button bits exhibit a Co-depleted surface region, which is under
compressive stresses, and an adjacent region with enhanced cobalt content (higher toughness; DP-carbides)
[9.28]. By courtesy of Sandvik Rock Tools, Sweden.

FIGURE 9.31. Cutting head of a tunnel borer; WC drag bits for operation in soft material. Bore diameter, 3.5 m;
weight of machine, 180 t; cutting head speed, 12.5 rpm; crowding force, 6280 kN; cutting head drives, 4 x 240 kW. By
courtesy of Tarnrock Voest-Alpine Bergtechnik Ges.m.b.H., Austria.

TUNGSTEN IN HARDMETALS

357

FIGURE 9.32. Rod mill rolls for high-speed finishing blocks; WC-{15-20)wt% Co or, alternatively, CojNijCr
binder; WC grain size 3-4 ~m. By courtesy of Sandvik Hard Materials, Sweden.

Submicron alloys and, more recently, so-called ulfrafine grades play an important role
in this field. All these alloys have small additions of growth inhibitors (VC, Cr3C2, raC) to
restrict WC grain coarsening during liquid-phase sintering. VC is the most effective
inhibitor. A microstructure of an ultrafine grade is shown in Fig. 9.38.

FIGURE 9.33. Integrated CIC rolls (cast-in carbide) for intermediate stands of wire rod mills or finishing stands
in bar mills; up to 30 wt% binder; WC grain sizes: 3-5 ~m. By courtesy of Sandvik Hard Materials, Sweden.

358

CHAPTER 9

s..fIce u.. """ hlgll

wI_ 01 hInI pIIIM (We)

1I1d. low vaIImw 01 binder
_ rnIIIInce.
pIIIM
(Co) Iiii' IncnuId

gunnIII " - - ' IItqUI

FIGURE 9.34. Schematic presentation of a Dual Property [9.28] drawing mandrel, together with microstructures
of the surface/ intermediate/core zone. By courtesy of Sandvik Hard Materials, Sweden.

FIGURE 9.35. Dental endmill shown in different stages of manufacturing. Pressed part at the bottom left; assintered part in the middle; brazed onto a steel shaft (right); endmill together with the workpiece in the
background. By courtesy of United Hardmetal, Germany.

TUNGSTEN IN HARD METALS

359

FIGURE 9.36. Saw blades for wood cutting. By courtesy of Gebr. Leitz, Germany.

Plate-reinforced hardmetals have recently gained attention due to their improved

hardness/toughness combinations [9.30]. Their industrial use is near (Fig. 9.39).
WC-(W,Ti,Ta,Nb)-Co grades. These grades are the working horse of the cutting
materials in machining of steel, especially for long chipping alloys. They are used for all
kinds of machining operations, such as turning, milling, grooving, threading, drilling, etc.
Examples are shown in Figs. 9.40 and 9.41. The large variety of indexable inserts and
chip-breaker geometries can be assessed from Fig. 9.42. They demonstrate well today's
high standard in the machining industry.

FIGURE 9.37. Solid tungsten carbide tools for the machining of printed circuit boards. Drill rotation speed,
105 rpm; feed, 2-8 m . mi,n-I ; material, glass reinforced epoxy sheet. By courtesy of Kemmer Priizision GmbH.,
Germany.

360

CHAPTER 9

FIGURE 9.38. Microstructure of a straight WC-lO wt% Co alloy, etched with Murakami's reagent; IO,OOOx.
Such grades are used for the machining of printed circuit boards. Note that the "giant" grain in the figure is only
~lllm.

It would be far beyond the scope of the present book to go into more details. Larger
hardmetal producers prepare several thousand differently shaped hardmetal specimens
ranging in mass between less than 1 g (such as balls for ball-point pens, which are
produced to the tune of 5 billion parts a year) and some hundred kg (such as dies and
pistons in the synthetic diamond industry).
The quality palette of larger producers comprises up to 50-60 hardmetal grades and
8-10 different types of coatings.
Worldwide, there are around 30 big hardmetal companies with capacities between
250 and 3000 tons/year, besides a great number of smaller and very small companies
(5-10 tons/year).

FIGURE 9.39. SEM image of a plate-reinforced hardmetal; 4,OOO x [9.29]. By courtesy of Toshiba Tungaloy
Co, Ltd., Japan.

TUNGSTEN IN HARDMETALS

361

FIGURE 9.40. Hardmetal tools used in different machining operations: (a) turning; (b) slot milling; (c) threading;
(d) drilling. By courtesy oflscar Ltd., Israel.

FIGURE 9.41. Milling cutter for camshafts. The cutter is equipped with 66 inserts; milling cutters with diameters
>2 m are produced today; individual tools are equipped with up to 300 inserts. By courtesy of Walter AG,
Germany.

362

CHAPTER 9

FIGURE 9.42. Indexable inserts of different shape and chip-breaker styles. Note the small inserts with edge
lengths of only ~ 6 mm. By courtesy of Iscar Ltd., Israel.

REFERENCES FOR CHAPTER 9

9.1.
9.2.
9.3.
9.4.
9.5.
9.6.
9.7.
9.8.
9.9.
9.10.
9.11.
9.12.
9.13.
9.14.
9.15.

9.16.
9.17.
9.18.
9.19.
9.20.

H. Kolaska, Pulvermetallurgie der Hartmetalle, Fachverband fiir Pulvermetallurgie, Hagen, Germany

(1992).
A. T. Santhanam, in: The Chemistry of Transition Metal Carbides and Nitrides (S. T. Oyama, ed.),
pp. 28-52, Blackie Academic and Professional, London (1996).
S. SOderberg, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 49-56, ITIA, London (1997).
A. T. Santhanam, P. Tierney, and 1. L. Hunt, Metals Handbook, 10th edition, Vol. 2, pp. 950--977, ASMI,
Materials Park, Ohio (1990).
H. Kolaska and K. Dreyer, Meta1l45 (1991), 224-235.
R. W. Stevenson, Metals Handbook, 9th edition, Vol. 7, pp. 773-783, ASMI, Materials Park, Ohio (1984).
K. 1. A. Brookes, World Directory and Handbook ofHard Metals and Hard Materials, 5th ed., International
Carbide Data, Hertfordshire, UK (1992).
W. Schedler, Hartmetall for den Praktiker (Plansee TIZIT GmbH, ed.), VOl Verlag GmbH, Dusseldorf
(1988).
P. Ettmayer, Annu. Rev. Mater. Sci. 19 (1989), 145-164.
F. Klocke, K. Gerschwiler, and R. Fritsch, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 213-238,
ITIA, London (1997).
A. Egami, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 13-23, ITIA, London (1997).
B. F. Kieffer, unpublished results and personal experience of the authors.
M. Miyake and A. Hara, J Jpn. Soc. Powder Metall. 26 (1979), 16-21.
M. Miyake, A. Hara, and T. Sho, J Jpn. Soc. Powder Metall. 26 (1979), 90--95.
Y. Yamamoto, A. Matsumoto, S. Honkawa, S. Honkawa, and N. Shigaki, in: Proc. of 1993 Powder
Metallurgy World Congress (Y. Bando and K. Kosuge, eds.), pp. 785-788, The Japan Society of Powder
Metallurgy (1993).
Y. Yamamoto, A. MatsumOlo, and Y. Doi, in: Proc. 14th Plansee Seminar (G. Kneringer, P. Rodhammer,
and P. Wilhartitz, eds.), Vol. 2, pp. 596-608, Plansee AG, Reutte (1997).
P. M. McKenna, US Patent 3,379,503 (1968).
E. N. Smith, Met. Powder Rep. 35 (1980), 53-54.
A. W. Weimer, W. G. Moore, and R. P. Roach, US Patent 5,110,565 (1992).
S. D. Dunmead, W. G. Moore, A. W. Weimer, G. A. Eismann, and 1. P. Henley, US Patent 5,380,688 (1995).

TUNGSTEN IN HARD METALS

363

9.21. B. Gunther, T. Konig, and R. L. Meisel, in: Pulvermetallurgie in Wissenschafi und Praxis, Band II (H.
Kolaska, ed.), pp. 41-58, DGM Infonnationsges. m.b.H., Oberursel, Gennany (1995).
9.22. L. E. McCandlish, B. H. Kear, and S. 1. Bhatia, US Patent 5,352,269 (1994).
9.23. L. E. McCandlish and P. Seegopaul, Proc. European Can! on Advances in Hard Materials Production,
Stockholm, pp. 93-100, EPMA, Shrewsbury, UK (1996).
9.24. E. Rudy and B. Kieffer, private communication.
9.25. E. Lardner and S. Iggstrom, in: Proc. 10th In!. Plansee Seminar (H. Bildstein and H. Ortner, eds.),
pp. 547-579, Metallwerk Plansee, Austria (1981).
9.26. D. Kassel, G. Schaaf, and K. Dreyer, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 24-37, lIlA,
London (1997).
9.27. A. D. Landsperger and T. R. Massa, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 85-96, lIlA,
London (1997).
9.28. G. Gille and B. Szesny, in: Proc. 7th Int. Tungsten Symposium, Goslar, pp. 150-170, lIlA, London (1997).
9.29. Sandvik Hard Materials, company brochure H-9097 ENG, Sandvik DP Carbide.
9.30. M. Kobayashi, K. Kitamura, and S. Kinoshita, European Patent Application 0759 480 Al (1997).

10
Tungsten in Catalysis
10.1. METALLIC TUNGSTEN
10.1.1. Survey
Tungsten alloyed with nickel, cobalt, or rhodium in thin layers on alumina supports, also
sulfided, is used on an industrial scale as a catalyst in crude oil processing (hydrotreating,
hydrocracking, reforming, hydrodesulfurization, and hydrodenitrogenation), as well as in
Fischer-Tropsch synthesis (alcohol formation from CO + H2).

10.1.2. HydrodesulJurization and Hydrodenitrogenation [10.1,10.2]

HDS and HDN are processes for removing chemically bound sulfur and nitrogen
from various petroleum feedstocks. HDS and HDN became more important in recent years
since new oil drilling techniques made it possible to recover heavier crude oils. These oils
are higher in organic sulfur and nitrogen containing compounds such as thiols, organic
sulfides, and thiophenes, as well as amines, pyridine, and chinoline compounds. The sulfur
content of crude oils varies between 0.2 and 4%, while the nitrogen content ranges from
0.1 to 0.9%. The removal of sulfur and nitrogen by hydrogen, performed at 300-400C
and pressures up to 200 bar, can be described by the following overall equations:
CaHbS + CH2 .... H2S + CaHd
CaHbH + CH2 .... NH3 + CaHd
As catalyst, besides others, tungsten is used as nickel alloy in thin layers on alumina
supports. During the reaction, tungsten is converted, at least at the surface, to WS 2, which
also acts as a catalyst.
The HDS and HDN reactions occur during hydrotreating and hydroreforming of
crude oil, and the catalyst also promotes hydrogenation of unsaturated hydrocarbons and
hydrodeoxygenation.
There are several reasons for the reduction of sulfur and nitrogen concentrations in
petroleum feedstocks:
365

366

CHAPTER iO

To minimize the amount of sulfur and nitrogen oxides, which form during
combustion processes in order to reduce air pollution.
To prevent the poisoning of catalysts used in the subsequent steps of product fuel
preparation, such as reforming and cracking. Nitrogen compounds are a poison for
the zeolithe cracking catalyst and large sulfur amounts are poisonous for Pt/Re
reforming catalysts.
Last but not least, new environmental regulations require lower sulfur concentrations in fuels in general.
The necessity to improve HDS and HDN has recently boosted R&D efforts. The
search for better catalysts has also improved our understanding of the complicated
processes which occur at the catalyst surfaces.
In order to understand the importance of these technical processes, it must be stressed,
that approximately 33 million barrels per day (l barrel ~ 159 liter) are hydrotreated
worldwide.

10.2. OXIDES
10.2.1. Survey
Colloidal tungsten trioxide can be used for the photocatalytic reduction of organic
compounds.
W200S8 catalyzes the hydration, dehydration, hydroxylation, and epoxidation.
DENOx mixed oxide catalysts are used in power plants, waste incineration plants,
hydration, aldol condensation, ROMP, and dimethyl sulfoxide synthesis.
is particularly active in hydrogen oxidation in acidic media as a cathode catalyst
(x equals preferably 2.72-2.9) [10.3,10.4].

wax

10.2.2. DENOX SCR Catalysts [10.5-10.8]

The selective catalytic reduction (SCR) of NOx by NH3 using DENOX catalysts is the
best developed and most widely used method for treating flue gases of thermic power
plants in order to reduce the emission of toxic compounds. The efficiency, selectivity, and
economics of the method are favorable. It was developed in Japan.
Commercial catalysts are composed of: Ti0 2, 78%; W0 3, 9%; V20 S, 0-5%; Si02,
7.5%; AhO, 1.5%; CaO, 1.0%; S04, 0.5-2%.
Ti0 2 is in the anatas modification and acts as high specific surface support; V20S and
W0 3 are the catalytically active components; V20S is an oxidation catalyst; W0 3 improves
the selectivity of the catalyst by suppressing any S02 --+ S03 oxidation (by decreasing the
oxidizing properties of the catalyst system); W0 3 is further responsible for the stabilization
of the anatas structure and improvement of the activity of Ti0 2 (specific surface,
approximately 90 m2 g-I). The starting point in the production of the catalyst is a
precipitated Ti02 in anatas form suspended in water, which is doped with tungstate in
ammoniacal solution. The W0 3 is produced by precipitation, while Ti0 2 acts as crystallization nucleus and will be covered by a thin H2 W04 layer during which calcination

TUNGSTEN IN CATALYSIS

367

converts to W0 3 . In a second step, V20S and the other components are added. The
remaining components act as mechanical promotors.
The ceramic catalyst is a homogeneous mixture of all components. After mixing with
binders, plastifiers, and glass fiber the mass is extruded to honey-comb-structured
elements, which are dried and calcined (Fig. 10.1).
SCR catalysts are full catalysts, which means that the whole mass is catalytically
active.

FIGURE 10.1. DENOX catalyst. The catalyst elements exhibit a honey-comb-like structure (upper photo); view
into a reactor partly filled with catalyst modules (lower photo). By courtesy of Porzellanfabrik FrauenthaI GmbH
and Austrian Energy and Environment SGP Wagner-Biro GmbH, Austria.

368

CHAPTER 10

The ratio of NO/N0 2 in exhaust gases of power plants is usually about 95/5. The
nitrogen oxides, during the DENOX process, are reduced by NH3 at the catalyst surface
while the following reactions occur:
N0 2 + NO + 2NH3 -+ 2N2 + 2H 20
4NO + 4NH3 + O2 -+ 4N2 + 6H 20
2N02 + 4NH3 + O2 -+ 3N2 + 6H20
The temperature should be in the range between 150 and 450C. The side reaction
2S02 + O2 -+ 2S03, which is 100% when using TiOr V20S catalysts, is less than 1% in
case of TiOr V20 s-W03 systems.
DENOX-SCR catalysts have been used successfully for years to remove NOx from
power plant exhaust gases. New areas of application in recent times are removal of NOx
from exhaust gas of large-volume diesel engines used in ships or emergency powergenerating sets.
DENOX-SCR catalysts can drastically reduce the dioxine and furane content of
exhaust gases of waste incineration plants, far below the limits prescribed by law
(0.1 ng m -3). The dioxine concentration in the emission of a DENOX-equipped waste
incineration plant is 100 times lower than the concentration in nonincinerated waste.

10.3. HALIDES
WCl 6 and WCl4 0 are used in combination with organometallic compounds for
numerous organic syntheses (see Section 10.5).

10.4. CARBON COMPOUNDS

This section deals with carbide compounds and hexacarbonyl. Organotungsten
compounds are treated separately under in Section 10.5. For information about tungsten-containing biocatalysts (enzymes), see Section 14.2.

10.4.1. Survey
WC can act as catalyst for gas-phase oxidation/reduction reactions, alcane hydrogenolysis, alcane reforming, H2 oxidation in fuel cells, and H2 formation from water, and
also, for anodic oxidation of hydrogen and formic acid in acidic media and of hydrogen
and hydrazine in alkali media. It is only active after anodic polarization in the presence of
reducing agents. It is assumed that at the surface hydrogen tungsten bronzes are formed
[10.4].
W2C can be used as catalyst for the isomerization, aromatization, and hydrogenolysis
of alcanes.
Tungsten hexacarbonyl W(CO)6 catalyzes a wide range of organic reactions, such as
the metathesis of alkenes.

TUNGSTEN IN CATALYSIS

369

10.4.2. Preparation of Carbide Catalysts

In heterogeneous catalysis, the catalytic properties of an element or a compound are
always closely linked to an active and large specific surface. The production of high
specific surface tungsten carbides requires special methods. Typical examples are listed
below.
Reaction of W or an oxide with a carbon-containing gas, preferably a gaseous
hydrocarbon. The most important prerequisite is a fine-grained precursor
[10.9,10.10], but also the hydrocarbon exhibits its own influence. In using
CH 4 /H 2 mixtures, H2 reduces the WO j and CH4 acts as carburizing agent. If
C2H6 or C2H4 is used, the hydrocarbon also takes part in the reduction. The
temperature in this case is about 150K lower compared to CH4 /H 2, which is of
special importance during the preparation of supported catalysts (suppression of
reactions with the support). Moreover, C2H6 or C2H4 originating WC exhibits
25 m2 g-l specific surface compared to 10m2 . g-l in case of CH4 /H 2 carburized.
Pyrolysis of organometallic compounds containing cyclopentadiene and carbonyl
groups to produce W2C [10.11].
Reaction of volatile tungsten chloride or oxychloride with gaseous hydrocarbons at
1400-2900 K [10.9].
Reaction of W(CO)6 with H2 [10.8].
Reaction of solid carbon with vaporized tungsten oxides [10.9].
Mixing of WCI4 , suspended in tetrahydrofurane, with Li triethylborohydride and
subsequent heat treatment at 723-773 K to produce W2C [10.9].
Reaction ofW films with a gaseous hydrocarbon or by reactive sputtering [10.12].
CO2 laser pyrolysis of C2H4 mixtures with W(CO)6 or WCI6 [10.13].
Reaction ofWCI 6 and guanidine hydrochloride to W2C free of uncombined carbon
[10.11].
Reaction ofWO j with NH j at 700-1000K to ~-W2N and subsequent carburization
with CH4 /H 2 at the same temperature to ~-W2C [10.14].
During reactions with carbon-containing compounds, free (unreacted) carbon may
form by cracking, which is undesirable because it changes the catalytic properties and
shows an increased specific surface not due to the carbide phase. C2 hydrocarbons produce
more excessive carbon than CH4 It can be removed in all cases by hydrogen treatment at
973 K. The carbide phase itself is not in any case pure and may consist of W2C/WC
mixtures.

10.4.3. WC [10.14, 10.15]

WC behaves similarly to platinum in many catalytic reactions (e.g., gas-phase
oxidation reactions). Unlike many metal and oxide catalysts, it is capable of operating
under a wider range of conditions in oxidative as well as in reductive regimes.
Kinetically, under oxidative conditions, due to an oxidic surface layer, WC acts
similar to an oxide in heterogeneous oxidation reactions (the redox mechanism consisting
of successive oxidation and reduction steps at the catalyst surface). Under reducing

370

CHAPTER 10

conditions, the surface of the WC becomes free of surface oxide and reveals its metal-like
nature. Consequently, both reactants being simultaneously activated at the catalyst surface
enable their reaction within the adsorbed layer (typical behavior of a metal catalyst).
The heat of oxygen adsorption corresponds to 362 kJ mol-I at 523 K and
356 kJ mol-I at 723 K. Compared to metallic W (812 kJ . mor- I) it is lower, because
the carbon atoms in the metal lattice reduce the adsorption affinity toward oxygen.
The oxidation of hydrogen (at excess of O2) is generally used to characterize the total
oxidation activity of a catalyst.
The specific activity is 16.7 mol m- 2 . S-I . 107 at 55 kPa H2, 22 kPa O2 at 623 K.
The specific catalytic activity in the H2 + O2 reaction at 22 kPa H2, 2.8 kPa02' and
623 K declines in the order
W
WC
W0 3

5.0
3.5
5.0

X
X

10- 4 mol m- 2 . S-I,

10-6 mol m- 2 . S-I,
10-8 mol m- 2 . S-I.

The reactivity of CO in the oxidation reaction to CO 2 is less compared to the H2 oxidation.

The specific catalytic activity for the oxidation ofNH3 at 20 kPa NH 3, 10 kPa O2 and
573 K is 0.71 mol m- 2 . s 10 7 (activation energy, 63 kJ mol-I).
WC can also be applied as catalyst for the oxidative coupling of methane.
WC can therefore be regarded as a promising candidate as heterogeneous catalyst
with high stability and activity for oxidation reactions, such as the conversion of natural
gas to higher hydrocarbons.
High surface (30-100 m2 . g-I) WC, free of carbon and oxygen, is very active in
neopentane hydrogenolysis. Chemisorbed oxygen inhibits the hydrogenolysis, but catalyzes isomerization reactions not only of neopentane but also, for example, of 3,3dimethylpentane. Furthermore, dehydrogenation as well as isomerization reactions of
methyl-cyclopentane occur on the WC catalyst, suggesting the presence of a bifunctional
surface that catalyzes dehydrogenation and carbenium ion reactions typically occurring on
reforming catalysts.

10.4.4. W2 C [10.11, 10.14]

High specific surface (100-150m 2 . g-I) and unreacted carbon-free W2C can be
prepared by reacting WC1 6 with guanidine hydrochloride. It shows excellent catalytic
properties in unsupported form as well as supported either on SiC-coated alumina or on
Zeolithe Y.
Unsupported and SiC / Ah03-supported W2C catalysts are multifunctional. n-Alcanes
are transformed by aromatization, isomerization, and hydrogenolysis. Aromatization can
be increased by the addition of O2 traces.
A comparative study using the unsupported W2C catalyst (122 m2 . g-I) and a
commercial alumina-supported Pt catalyst (0.3%; 180m2 . g-I), in the isomerization of
n-heptane (523 K), revealed different activities:
isomers (2-methylhexane, 3-methylhexane, dimethylpentanes, 2- and 3-methylpentanes): 81% for W2C and 48% for Pt,
aromatics (benzene and toluene): 3% for W2C and 32% for Pt,
hydrogenolysis products (C I to Cs hydrocarbons): comparable.

371

TUNGSTEN IN CATALYSIS

Zeolithe-Y-supported W2C shows metallic hydrogenation properties and can also be

used as catalyst in the thiophene hydrogenation.
A big advantage of W2C over the usual noble metal catalysts is the sulfur resistance.
Noble metal catalysts get poisoned (inactive) by the presence of S-containing compounds
in a short time. The sulfur resistance of W2C also allows its use as catalyst for the
hydrodesulfurization of thiophene.
High surface area ~-W2C, prepared from ~-W2N by subsequent carburization, shows
similar behavior as described for WC in reactions with neopentane, 3,3-dimethylpentane,
and methyl-cyclohexane.
10.4.5. Tungsten Carbide Catalysts in Fischer-Tropsch Synthesis [10.16]
The properties of tungsten carbide catalysts depend strongly on their surface
chemistry, which can be clearly demonstrated when using them in FT synthesis.
The starting product may act as well reducing (H2), as carburizing (CO), and as
decarburizing (H2). The carbon content of the surface, which can be changed by the above
reactions, is highly responsible for the catalytic action and may vary from W to WC. Also,
if supported tungsten carbide catalysts are used, the reaction between the catalyst and
support during pretreatment and use may have a significant influence on the catalytic
activity. Therefore, different supports lead to different catalytic properties.
By using an unsupported catalyst, a comparison of the carbides and the trioxide
revealed the following results:
WC(%)

Hydrocarbons
Alcohols
Carbon dioxide

79.7
18.4
2.1

86.5

13.5

68.0
20.4
11.5

W2C favors the CO dissociation and formation of hydrocarbons with excess of CH4
and CO 2 and no oxygenated products, due to its more pronounced metallic character as
WC. W2C is more sensitive to H2 reduction (the surface can be transferred to W).
Alcohols are only formed on we. The formation of alcohols is related to the surface
stoichiometry of WC (presence of C vacancies; understoichiometric WC). WC also
promotes higher chain length in both hydrocarbons and alcohols compared to W03
When using supported catalysts, the formation of alcohols is, on the one hand, related
to the surface WC stoichiometry. Only layers with carbon deficiency (vacancies) produce
alcohols while high carbon concentrations produce alcanes only. On the other hand, by
reaction with special supports, anionic vacancies are formed, stabilized by mixed oxides,
and associated with the carbon vacancies in the mixed carbide. The optimal support for
alcohol production is Ti0 2. The bifunctionality is due to carbidic and oxidic phases.
10.5. ORGANOTUNGSTEN COMPOUNDS: CATALYSTS IN ROMP AND ADMET
[10.17-10.20]
As already pointed out in Chapter 4, the number of synthesized metallo-organic
tungsten compounds is immense and the reason why so many compounds have been

CHAPTER 10

372

prepared and investigated can only be understood if the field of application are examined
more closely.
A large number of tungsten organometallic compounds plays an important role, or
can be regarded as prominent "Candidates, as catalyst in organic syntheses. Tungsten as well
as molybdenum and rhenium catalysts are closely associated with metathesis polymerization.
The term "alkene metathesis" describes a reaction mechanism shown schematically in
the following equation:
catalyst
~

The alkylidene groups of an alkene pair are exchanged with one another in the presence of
a transition metal catalyst. In recent times, not only reactions between two C=C double
bonds but also between C=C and C=O bonds have been investigated. The reaction
includes the cleavage of double C=C bonds and the formation of new ones. Homogeneous
as well as heterogeneous catalysts are effective.
There exist two groups of catalysts:
a transition metal compound combined with a main group metal alkyl co-catalyst,
a well-defined transition metal carbene catalyst.
One of the most important reaction mechanisms in organic polymer chemistry that
can be catalyzed by tungsten compounds is "Ring Open Metathesis Polymerization"ROMP. It can be expressed by the overall equation:

111
u

hl

catalyst
~

The chain propagation and addition-type nature of ROMP chemistry will be

elucidated by the scheme:
H

LnM=<

R'

LnM

--

ILJ
()

/""'R
(
~

/.::::::::,C/
'--"
\

R'

R'

r(R

~C7
I
H

Ln~R,
~

'-..J

repeat

Polymer

The first step is coordination of substrate to the metal center forming the metallacyclobutane intermediate. The subsequent opening of the metallacyclobutane ring produces the
first new carbene? which reacts with the next equivalent of the monomer, etc. The final
polymer molecule contains an original alkylidene group at one end and the catalyst

373

TUNGSTEN IN CATALYSIS

fragment at the other. The driving force for the reaction is the loss of ring monomer strain.
The structure of the catalyst allows one to direct the polymerization reaction, and
consequently the properties of the polymer. This is the reason for the very intensive
investigations in synthesis of catalyst compounds. Today, these efforts are mainly driven by
the necessity to find "economic synthesis reactions." This requires that the highest
percentage of input molecules are used to produce the desired final product comprising
minimization of raw material consumption and waste production.
Typical ROMP products are linear, thermoplastic polymers Polyoctenamer (from
cyclooctene), Polynorbonene (from norbonene), and the network (thermoset) polymer
Polydicyclopentadiene (from dicyclopentadiene) combining toughness, solvent resistance,
chemical inertness, and long-time durability.
The second important reaction is the "Acidic Diene Metathesis Polymerization"ADMET. The overall equation is as follows:

catalyst

The reaction is driven by a shift in equilibrium, caused by the removal of one of the
reaction products.
The metathesis polymerization of acyclic alkenes is an equilibrium step propagation
via, dimers, trimers, tetramers, etc. The scheme is:

R,~R~R'

!j

'R

~\R

LnM=CIm"

~\R~n ~R'

nR

,~R'

For the equilibrium shift, it is important to remove the monomer reaction product
continuously.
Typical products are Polyoctenamer (from 1,9-decadiene) and 1,4-Polybutadiene
(from 1,5-hexadiene).
The main groups of metalorganic tungsten compounds in use as catalysts in ROMP
and ADMET are:

Five-coordinate Walkylidene complexes, where L can be alkoxide or halide or S03CF3,

are highly effective olefine metathesis catalysts in the presence of Lewis acids. The vast
number of compounds synthesized can be understood by the variation and combination of
different ligands.

374

CHAPTER 10

Tungsta Cyclabutan Complexes

They are the reaction product of

RO,

~NAr

w..'?"

RO"""

~CH-t-Bu

+ CH2=CH 2 (2 eqn.)

NAr

RO-rO
II

OR

where
OCCCF3 h(CF 2CF2CF3 )
Well-defined alkylidene metallacyclobutane catalysts exist for the polymerization of
olefines (alkenes) and acetylenes(alkines).

10.6. TUNGSTATES
Tungsten bronzes MxWOy (Me can be Ti, V, Mn, Fe, Zn, and Ni;x can be 0-1 or 1-2,
and y can be 2--4) may act as anodic catalyst in fuel cells (high electrical conductivity and
corrosion resistance in acidic media) [lOA] .
Sodium tungstate bronzes with Au, Pt, or Pd additions exhibit high electrocatalytic
activity (also in acidic media) in reducing oxygen [10.4], and catalyze the carbon
monoxide oxidation.
Sodium salt of 12-tungstophosphoric acid acts as catalyst for isomerization, polymerization, nitrile synthesis, dehydrochlorination, dehydrogenation, ketone synthesis,
ring-closure synthesis, oil bodying, aromatization, and desulfurization [l0.4, 10.21].

10.7. CHALKOGENIDES [1004,10.8]

Tungsten disulfide WS 2 [10.8] is used for electrode activation in fuel cells. If
produced by hydrogen reduction of (NH4h WS4, it exhibits a BET surface 70 times
larger than that of commercially prepared WS 2 . It acts as catalyst for the reaction

at normal pressure with a yield of more than 99.9%.

WS 2 and WSe2 as colloidal coating are used in photocatalysis.

REFERENCES FOR CHAPTER 10

10.1.
10.2.

R. J. Angelici, "Hydrodesulfurization and Hydrodenitrogenation," in: Encyclopedia of Inorganic

Chemistry (R. Bruce King, ed.), Vol. 3, pp. 1433-1443, Wiley, Chichester, New York (1994).
Shi Yahua, Kang Xiaohong, and Li Ladong, Paper presented at the lIlA Annual General Meeting
Huntsville, AI, USA; transcript of papers: lIlA, London, UK (1994).

TUNGSTEN IN CATALYSIS
10.3.
10.4.
10.5.
10.6.
10.7.
10.8.
10.9.
10.10.
10.11.
10.12.

10.13.
10.14.
10.15.
10.16.
10.17.
10.18.

10.19.
10.20.
10.21.

375

B. Broude, US Patent 3,507,701 (1970).

A. T. Vas'ko, in: Encyclopedia of Electrochemistry of the Elements (A. 1. Bard, ed.), Vol. V, pp. 69-126,
Marcel Dekker, New York (1976).
T. Nakatsuju, Paper presented at the lIlA Annual General Meeting, Huntsville, AI, USA; transcript of
papers: lIlA, London, UK (1994).
Ceram Frauenthal Keramik AG, company brochure.
I. Binder-Begsteiger, E. Megla, M. Tomann-Rossos, and G. W Herzog, Chem. Ing. Tech. 62 (1990), 5664.
G. Winter, Paper presented at the ITIA Annual General Meeting, Huntsville, AI, USA; transcript of
papers: lIlA, London, UK (1994).
S. T. Oyama, in: The Chemistry of Transition Metal Carbides and Nitrides (S. T. Oyama, ed.), pp. 1-27,
Blackie Academic and Professional, (1996).
S. Decker, A. Lofberg, 1. M. Basin, and A. Frennet, Catal. Lett. 44 (1997), 229-239.
F. Sherif and W Vreugdenhil, in: The Chemistry of Transition Metal Carbides and Nitrides (S. T. Oyama,
ed.), pp. 414-425, Blackie Academic and Professional, London (1996).
F. Reniers, M. Detroye, S. Kacim, P. Kons, M. Maoujoud, E. Silberberg, M. Soussi, El Begrani, T.
Vandevelde, and C. Buess-Hermann, in: The Chemistry of Transition Metal Carbides and Nitrides (S. T.
Oyama, ed.), pp. 274-289, Blackie Academic and Professional, (1996).
R. Ochoa, X. X. Bi, A. M. Rao, and P. C. Eklund, in: The Chemistry of Transition Metal Carbides and
Nitrides (S. T. Oyama, ed.), pp. 489-499, Blackie Academic and Professional, (1996).
F. H. Ribeiro, R. A. Dalla Betta, M. Boudart, 1. Baumgartner, and E. Iglesia, 1 Catal. 130 (1991),86-105.
N. I. Ilchenko and Yu. I. Pyatnitsky, in: The Chemistry of Transition Metal Carbide and Nitrides (S. T.
Oyama, ed.), pp. 311-326, Blackie Academic and Professional, London (1996).
L. Leclercq, A. Almazouari, M. Dufour, and G. Leclercq, in: The Chemistry of Transition Metal Carbides
and Nitrides (S. T. Oyama, ed.), pp. 343-361, Blackie Academic and Professional, (1996).
G. G. Hlatky, "Oligomerization and Polymerization by Homogeneous Catalysis," in: Enclyclopedia of
Inorganic Chemistry (R. Bruce King, ed.), Vol. 5, pp. 2728-2735, Wiley, Chichester, New York (1994).
K. B. Wagener, 1. M. Boncella, and D. W Smith Jr., "Metathesis Polymerization Processes by
Homogeneous Catalysis," in: Encylopedia of Inorganic Chemistry (R. Bruce King, ed.), Vol. 4, pp.
2242-2251, Wiley, Chichester, New York (1994).
R. H. Grubbs and S. H. Pine, "Alkene Metathesis and Related Reactions," in: Comprehensive Organic
Synthesis (B. M. Trost, ed.), Vol. 5, pp. 1115-1127, Pergamon, New York (1991).
1. Feldman and R. R. Schrock, in: Progress in Inorganic Chemistry (S. 1. Lippard, ed.), pp. 1-69, Wiley,
Chichester (1991).
GTE Sylvania, Technical Information Bul/etin, CM-9003 (9/80).

11
Tungsten Scrap Recycling
11.1. INTRODUCTION AND GENERAL CONSIDERATIONS [11.1-11.4]
The recycling of scrap containing tungsten is by no means a modem procedure. On the
contrary, it has been established for many decades. Even before anyone thought about
recycling of glass, paper, and aluminum-as is common today in most industrialized
countries-a considerable amount of tungsten was being reclaimed from different types of
scrap. The reason can be easily explained: As outlined in Chapter 2, most tungsten ores
contain less than 1 wfllo W0 3 and very seldom higher concentrations. In Table 11.1,
different types of scrap containing tungsten and the corresponding tungsten contents are
listed. It is evident that the least valuable scrap, like low-grade grinding sludge, contains
about 15 times more tungsten than an average ore. Most scrap materials are even richer in
tungsten than ore concentrates.
It is the high intrinsic value in anything containing tungsten, even in smaller
concentrations, which makes tungsten scrap an interesting and worthy material for
TABLE 11.1. Typical Tungsten Scrap Materials
No.

2
3
4
5
6
7
8
9
\0
II

12
13
14

%w

Scrap type
High-purity W
Oxide dispersed W alloys
(Th02, zr02, Ce02. La203)
Hard metal pieces (also containing Co, Ta)
Heavy metal W alloys
Tungsten - copper
Pure tungsen powder
W grinding sludge
W cutting sludge
Hard metal powder
Hard metal grinding sludge
Heavy metal powder
Heavy metal turnings
W - Cu powder and green compacts
Floor sweepings (different sources)

377

Group

?: 99
96-98

Hard Scrap
60-97
92-94
60-90
98- ?: 99
30-60
70-80
60-95
15-60
92-97
92-97
50-90
40-60

(solid pieces)

Soft Scrap
(fine particles,
powder, dust,
timings, sludges)

378

CHAPTER 11

reuse. Due to the low abundance of tungsten, which requires the mining and ore dressing
of a huge ore tonnage to gain a small amount of tungsten, it was always a "valuable"
metal.
Besides this intrinsic value, the tungsten world market was decisive in increasing the
recycled percentage of the total tungsten production in earlier times. More recently,
environmental and resource-preserving aspects became necessary.
The preservation of tungsten ore resources prolongs the availability of primary
tungsten and decreases the damage to the environment caused by disposal of tailings. It
also saves capital resources.
In regard to the environment, recycling requires less energy and chemical consumption, thus producing less waste. Moreover, the dumping of tungsten scrap with lower
concentrations of tungsten like sludges-as exercised in the past-does not take place any
more in the industrialized countries.
Because of economic considerations, the tungsten in cemented carbide scrap is not
the only value of interest. The value of the other constituents, such as Co and Ta, are also
important. Both metals, at least at this time, are more expensive than tungsten.
Last but not least, the pricing policy of "virgin" W products and the price stability of
the recycled products influences tungsten recycling in general.
In discussing tungsten recycling, it might first be of interest to elucidate the tungsten
flow in general. The questions are: how much tungsten is supplied by primary (ore) and
secondary (scrap) sources and how big are the losses? Which losses are unavoidable and
which can be partially avoided by increased or improved recycling in the future? An
estimated balance was published in 1987 and is shown in Fig. 11.1. The figures certainly
fluctuate with time but should also be valid today within a certain measure. As one can see,
about 113 of the total tungsten demand is supplied by scrap. This proves that scrap
recycling is a very important factor in the supply of tungsten raw materials.
It must be mentioned here that any in-house reuse of valuable material, for example,
powder generated during green forming of cemented carbide parts and similar pure

Primary tungsten
ScheelltaiwolframHe
ore concentrate. (66%)

Secondary tungsten
Tungsten scrap
(34%)

Total tungsten demand

(100%)

I
Scrap originated
during processing

Final products
(90%)

(10%)

Loss

Scrap from used parts

(66%)

(24%)

Dissipation

Discard

Recycle & Dilution

(26%)

(30%)

(10%)

FIGURE 11.1. Tungsten flow chart for primary and secondary raw materials.

379

TUNGSTEN SCRAP RECYCLING

materials, will not be considered. Large amounts pass this way, but their actual volume is
unknown.
The loss of over 60% is considerable. It can be divided into three different sources:
dissipation, discard, and recycling & dilution.
Dissipate losses are wear loss of carbide products, arc erosion of electrical contacts
and electrodes, oxidation losses, losses by chemical decomposition, etc. It can only be
influenced to a small extent by quality improvements of the final products. However,
because the quality level today is already quite high, no drastic change can be expected.
Loss by discard is mainly a matter of strict organization and today is not only a
question of the price of tungsten. Typical examples for discard are noncollectable carbide
products, burned out lamps and lighting fixtures, welding electrode stubs, electrical contact
disks, carbide hard-facing materials, etc.
Loss by discard has decreased remarkably since 1970. A study of the U.S. Bureau of
Mines revealed how drastically the discard of obsolete cemented carbide parts was reduced
between 1974 and 1991 (see Fig. 11.2). It can be seen that within this period 24%, which
corresponds to nearly 1/4 of the total tungsten consumed in the USA, could be recycled.
Under the assumption of a steady and linear increase to the year 2010, a 54% recycle can
be expected. The reasons for the drastic change are:
change in product mix (braised tool tips versus throwaway inserts),
better awareness of the value of tungsten (improvements in communication,
education, and organization),
better understanding of the byproduct value (Ta, Co),
improved recycling methods (direct-indirect).
Further improvements can be expected to minimize discard in the highly industrialized countries of North America, Europe, and Japan, and also in the developing countries.
Loss by recycling & dilution includes residues containing low tungsten concentrations as a result of recycling, and are not uneconomical to recover. For example, tungsten
as trace constituent in recycling of tool steel, carbides, etc. is lost into the steel melt.
This last loss category is not likely to be altered. Only extremely high tungsten prices
could affect the situation.
The above-mentioned USBM study, by an overall calculation, resulted in a total tungsten
loss of 61 % from 1955 -1991. By subtracting the amount lost by dilution, the figure changes

~...l-- 2010(90%)

90,----

eO -.'

._-.9!~~-"

75
60

1991 (58%)

45
30
15

:
1974 (17%)

0+---+---+-~4---1-

1974

1980

1986

1992

1998

.
2004

2010

year

FIGURE 11.2. Percentage of tungsten recycling via cemented carbide scrap (USA).

380

CHAPTER 11

to 52%. Regarding the USA to be an industrially advanced country, the figure of 66% loss,
worldwide, in the above flow chart (Fig. 11.1), compares very favorably.
Considering recycling today, it is not sufficient to just get the material back for some
useful purpose. Optimal recycling should always be done at the highest level of utilization,
which includes many factors, such as conversion costs with special emphasis on energy
costs, purity of the scrap material, as well as environmental reasons. For example, it is best
to re-enter the same product flow, which originally generated the scrap, using a minimum
of conversion steps. This is called direct recycling. Direct recycling is in any case the most
economic and ecological way to reuse scrap and should always be applied, if possible.
However, sometimes market conditions may dictate a less economic end use, such as when
one type of scrap is in oversupply, when the optimal recycling system is unavailable, or
when traders do not understand the market situation. Another case is when a fast
generation of cash is more important than finding an optimal recycling possibility.
Such circumstances still direct far too much scrap to chemical processing, which is
still the most important recycling system, although in most cases not the most economic
and ecological one.

11.2. TUNGSTEN RECYCLING METHODS (1l.l-11.4]

A variety of tungsten recycling methods were developed. They can be divided into four
groups: hydrometallurgy, melting metallurgy, direct recycling, and semidirect recycling.

11.2.1. Hydrometallurgy

In principle, most tungsten contammg scrap qualities can be processed by this

method, which converts tungsten to "virgin" APT. However, it should be kept in mind that
it is an expensive, high-energy-consuming and chemical-waste-producing procedure. In
regard to the economy, preferably more impure scrap qualities, which are not easily
converted by other methods, should pass this way. The scrap must first be oxidized by air,
chemicals, or electrical energy, in order to transfer tungsten to the hexavalent state, which
is soluble in alkaline solutions. From here on the entire procedure, as for ore concentrates,
is required.
Typical scrap materials for hydrometallurgical treatment are grinding sludges, floor
sweepings, contaminated powder qualities, tungsten - copper, and solid pieces, which can
be dissolved by electrolysis.
Hydrometallurgical scrap conversion has already been discussed in detail (Chapter 5)
and will not be further treated here.
In regard to tonnage of tungsten fed into this conversion type, it is the method mostly
used (even for scrap) for converting in a more economic and ecologically preferable
manner. One of the main reasons, besides those already mentioned in the introduction,
might be that APT is the general intermediate for most tungsten products, with the
exception of melting metallurgy. Furthermore, no additional contamination does occur, but
in contrast purity will be increased.

381

TUNGSTEN SCRAP RECYCLING

TABLE 11.2. Tungsten Scrap in Melting Metallurgy

Product

Most preferable scrap type

(according to No. in Table 11.1)

Superalloys
Stellites
Cast eutectic carbide
Menstruum WC
Tool steel
Ferrotungsten
Melting base

1,3,6,9
1,6
4,6,9, II, 12
1,2,4,7, II, 12
1,2,4,6,7, II, 12
1,2,3,4,6,7,8,9,10, II, 12

11.2.2. Melting Metallurgy

In Table 11.2, the products are listed into which tungsten can be recycled as well as
the preferable scrap materials. Superalloys, stellites, Menstruum WC, and cast eutectic
carbides require very pure scrap while, for steel melting metallurgy products, impure scrap
can also be used. The advantage of the latter is that all oxidic substances (suc~ as grinding
media, etc.) end up in the slag, and traces of foreign elements, which are soluble in the
metallic melt, are distributed homogeneously.
Any tungsten going in this direction cannot be recycled again by other methods, with
the exception of re-melting. It finally ends up in loss by recycling & dilution.

11.2.3. Direct Recycling

As direct recycling is understood, the as-supplied material is transformed to powder
of the same composition by either chemical or physical treatment, or a combination of
both. Prerequisites for direct recycling are:
The composition of the scrap must be the same as for the final product.
The scrap must be of high purity, not only in regard to foreign substances but also
in quality (e.g., straight WC-Co hardmetal scrap should not contain TiC- TaCbearing qualities, and the WC grain size should be within a small range).
The process must offer the possibility of converting the scrap to a powder of a
metallurgically acceptable form.
No contamination by foreign materials during processing.
Typical examples, which will be discussed in more detail below, are the zinc process, or
the Coldstream process for reclamation of cemented carbide scrap, and the oxidation and
subsequent reduction of tungsten heavy metal turnings or thoriated tungsten (see Table 11.3).
Direct recycling is combined with a minimum of energy consumption, chemical
waste, and lowest production costs. A general comparison of direct and indirect tungsten
recycling is presented in Fig. 11.3.
We are still far away from full utilization of the direct recycling potential, because
huge amounts of the corresponding materials are still chemically converted to APT. The
reasons for that situation are:

382

CHAPTER II

TABLE 11.3. Direct Recycling of Tungsten Scrap

Process
Zinc
Bloating/crushing}
Coldstream
Oxidation/reduction

Product

Scrap

Hardmetal powder

Sorted hardmetal pieces

Heavy metal powder

Thoriated tungsten powder

Heavy metal turnings

Thoriated tungsten pieces, turnings, powder

Bad organization by poor sorting, or no sorting at all, at places where scrap is

generated. Not only in regard to foreign matter, but also in regard to quality
(grade), purity of the scrap is one of the main demands for direct recycling.
Fear of heterogeneous impurity content.
Lack of knowledge of potential consumers and the different processing parameters
necessary for recycling the powder in comparison to the virgin equivalent.
Any efforts to increase the percentage of directly converted material will improve the
environment and the economy in general.
Zinc Process [11.1, 11.2, 11.5, 11.6]. In this process, cemented carbide scrap is
treated in molten zinc or zinc vapor, which reacts with the binder to form intermetallic
phases and a zinc-cobalt alloy. These reactions lead to a volume expansion of the binder
~nd bloat the scrap. After vacuum distillation of zinc, the material is friable and can be
readily disintegrated. The condensed Zn can be re-used. The reclaimed carbide/metal
sponge contains less than 50 ppm Zn.
A flow sheet of the process is given in Fig. 11.4. In the first stage, the cleaned and
sorted hardmetal scrap is contacted with molten Zn at 900-1050C in an Ar/N2
atmosphere for several hours, making sure that all smaller pieces are completely
penetrated. The second stage is the distillation of Zn in vacuum (0.06-0.13 mbar) at
1000 - 1050 C, which again takes several hours. The cooled-down material is crushed,
ball milled, and screened. The top screen is recycled in the next batch. The crushing and
grinding involves fracture, primarily in the Co phase and not in the carbide phase.
The chemical composition of the final product is almost identical to the original
material, with the following exceptions:
pick up of approximately 0.1 % Fe (during milling operations),
depletion of carbon by 0.12-0.15%.
Direct

Indirect
Scrap type

Any

Limited and sorted

Powders with same composition as scrap ..-. Final recycled product - . .

"Virgin" equivalent

Low

+-- Energy consumption --+

High

No or very little

......- Chemical consumption ---+

High

None

High

Very competitive

+--

--+

Higher than W-concentrate conversion

Waste by product

Conversion cost

FIGURE 11.3. Comparison of direct and indirect tungsten recycling methods.

383

TUNGSTEN SCRAP RECYCLING

Cleaned, sorted, and crushed harctnetal scrap

Reaction in molten zinc (900-1050 'C; A,/N,)

Vacuum distillation (1000-1050 'C)

Vacuum distillation

Ballmill, screen, blend, C-adjust

Powder (- 200 Mesh) for use

FIGURE 11.4. Flow chart of the zinc process for cemented carbide reclamation.

The basis of a successful use of the reclaimed powder involves some special skill and
special equipment, such as excellent analytical capability, cemented carbide testing facilities,
optimal milling equipment, and a good understanding of the shrink factor during sintering.
The energy consumption is approximately 4 kWh/kg, which compares favorably to
virgin WC produced via the chemical route (12 kWh/kg). Compared to the features of
indirect conversion, the cost of zinc reclamation is reduced by 20-30% for WC-Co
grades and by 30-35% for WC- TiC- Ta(Nb)C-Co grades.
Hardmetal scrap

Return to tungsten ha,dmetal process

(ready to press)

FIGURE 11.5. Flow chart of the Coldstream process.

384

CHAPTER II

Coldstream Process [11.7] (see the flow sheet Fig. 1l.5). This process is a mechanical comminution and employs air blasts to accelerate particles against a target with
sufficient energy to cause fracture (particle velocities near Mach 2). The air is cooled by
expansion from the nozzles, protecting the scrap from oxidation. After screening or air
classification, oversize material is recycled.
A method combining the two above-described processes is one in which zinc
infiltration is followed by Coldstream crushing with a final zinc distillation step.
Bloating/Crushing. The process is based upon the embrittling effect when heating
cemented carbides to about 1800C followed by rapid quenching. Embrittlement is the
consequence of WC grain growth, formation of brittle phases, and thermally induced
stresses.
Oxidation - Reduction of Heavy Metal Turnings. Heavy metal turnings can be easily
oxidized by air at about 800C. The layer height in this step is critical, because if it
exceeds a certain limit then local overheating (exothermic reaction) can cause melting. The
oxidized product is subsequently milled and screened and, finally, it is reduced in hydrogen
atmosphere at 900 -1000 C in the same type of push-type furnaces used for W-powder
production.
11.2.4. Semidirect Recycling

Semidirect reclamation can be applied to binary- or ternary-phase alloys. One

component is dissolved chemically, leaving the other phase(s) intact. By the dissolution,
the integrity of the structure is lowered and attrition can take place. Typical examples are
presented in Table 11.4.
In conclusion, it can be stated that the importance of tungsten scrap recycling will
continue to grow considerably in the future. This will mainly be driven by economic,
ecological, and natural-resource preserving considerations. The share of direct recycled
scrap will increase as soon as the organization of scrap collection and separation at the
places of generation has improved. Scrap should always be kept as pure as possible. This
must be understood not only in regard to contamination by foreign materials, but also in
regard to the grade purity. As an example, different grades of cemented carbide throwaway
inserts should be kept separated. The old philosophy-scrap is chunk-is not any longer
valid. On the contrary, scrap is a very valuable material and has to be treated accordingly.
Pure grade scrap will bring a much better price or yield and is easier to convert.
TABLE 11.4. Semidirect Tungsten Recycling
Principle

Scrap
Hardmetal pieces

Method

Selective dissolution of the Co binder

Attack by acid (H 3P04 )

Leach milling (mixed chemical and
mechanical attack)
Electrolysis (acidic solution)
WC or WC + mixed carbides can be recycled to hardmetal grade powder preparation.

Attack by acqueous FeCI 3 solution

Selective dissolution of the Ni - Fe
binder
Powder consists of approximately 40-l.Im tungsten spheres and can be used in plasma spraying.

Heavy metal pieces

TUNGSTEN SCRAP RECYCLING

385

A further prerequisite for an increase in direct recycling is the necessity for the
consumer to learn to deal with reclaimed materials, which require modified production
conditions.

REFERENCES FOR CHAPTER 11

11.1. B. F. Kieffer and E. F. Baroch, in: Extractive Metallurgy of Refractory Metals (H. Y. Sohn, O. N. Cralson,
and 1. Th. Smith, eds.), pp. 273 - 294, AIME, Materials Park, Ohio (1981).
11.2. B. F. Kieffer and E. Lassner, in: Proc. 4th Int. Tungsten Symp., Vancouver 1987, pp. 59-67, MPR
Publishing Services Ltd., Shrewsbury, UK (1988).
11.3. B. F. Kieffer and E. Lassner, Berg-und Huttenmaemmische Monatshefle 139 (1994), 340 - 345.
11.4. G. R. Smith, "Tungsten," American Metal Market (August 1993).
1l.5. P. Barnard, U.S. Patent No.3, 595484, July 27, 1971.
11.6. B. F. Kieffer, Int. J Refract. Met. Hard Mata 2 (1986), 65 - 68.
11.7. 1. Walraedt, in: Proc. 7th Int. Plansee Seminar, Vol. IV; No.2, pp. 1-12, Metallwerk Plansee, Reutte,
Austria (1971).
11.8. R. Zhao, in: Proc. 4th Int. Tungsten Symposium, Vancouver, pp. 50-58, MPR Publishing Services Ltd.,
Shrewsbury, UK (1988).

12
Ecology
12.1. INTRODUCTION
In general, the production of a metal, its use, and the manner in which it is discarded
influence the ecology more or less. These aspects grew in importance due to the steady
increase in environmental awareness during the last decades. Although this awareness is
more pronounced in the industrialized countries, its worldwide acceptance will be a
necessity in the near future.
In regard to tungsten it can be stated that mining, ore dressing, chemical conversion,
or metallurgy create more or less of the usual environmental problems. The use of the
metal or its alloys in some applications leads to a dissipative loss and distribution, which is
harmless and negligible. Discarding is not a big problem, because the recycling of
tungsten-bearing substances has been very well organized for quite some time (see
Chapter 11).
The impact on the world's ecology caused by the production oftungsten is relatively
small in comparison to that of metals like iron, aluminum, copper, etc. because of the small
tonnage. If we assume the following situation:

yearly production, 40,000 t W;

average content in the ore, 0.5% W0 3 ;
average concentration in the ore, concentrate 65% W0 3 ;
yield in ore dressing, 85%;
yield in chemical conversion, 95%;

then the following quantities must be treated worldwide: 12.5 x 106 t of ore have to be
mined, disintegrated, and concentrated to yield 82 x 103 t of ore concentrates and
12.4 x 106 t of tailings. The latter must be deposited.
During the chemical or melting metallurgical conversion of the concentrates, roughly
1/3 of the original weight remains as residue (slag, precipitate, and gangue). This
corresponds to approximately 26 x 10 3 t of material, which must be dumped. Residues
from chemical conversion (precipitates and gangue), which represent the majority, also
contain chemicals.
Assuming that 90% of total tungsten consumption is converted chemically, it can be
calculated that between 25,000 and 60,000 t of sodium hydroxide or sodium carbonate are
387

CHAPTER 12

388

necessary for the dissolution and about the same amount of sulfuric acid for the
neutralization. This means that between 35,000 and 84,000 t of dissolved sodium sulfate
are generated of which only a very small percentage is reused while the rest is disposed to
the environment.
In principle, these figures can be neglected for a worldwide consideration. However,
taking into account that today's supply originates from a small number of bigger mines
besides a large number of smaller ones (especially in China), and that only 10-20 chemical
converting plants exist, it is evident that local problems exist or may arise in the future.
Good examples of how ecological problems can be solved successfully without any
dramatic influence on the economy will be presented below.

12.2. MINING AND ORE DRESSING [12.1,12.2]

Mining activities are usually associated with the more or less severe destruction ofthe
environment. In open pit mining they are more visible than in underground operations,
because large surface areas must be removed. In both cases, the deposition of residues after
separation of the ore as tailings from flotation or gravity separation represent another
environmental problem. Also, ore dressing creates waste water, which in the case of
flotation contains small quantities of chemicals.
Naturally, it depends on where a mine is located: whether it is far away from or close
to civilian and highly populated areas. Moreover, different countries have diverging
environmental regulations.
An extreme in this regard is the Mittersill Scheelite Mine, located in the Alps of
central Europe, in Austria's province of Salzburg. The mine is one of the largest tungsten
deposits in the Western World and is an outstanding example of how ecology and economy
can be matched optimally. The open pit, the underground mine, and the ore dressing plant
are situated in a National Park which mainly serves tourism (Fig. 12.1). Therefore, it is
easy to understand that quite rigid and extensive environmental regulations have been
imposed by the Austrian nature conservation authorities and that these have necessarily
increased the costs of the total operation. Nevertheless, the Mittersill mine and ore dressing
plant is still ranked among the most economic tungsten ore concentrate producers in the
world, although the ore content ranges only around 0.4--0.7% W0 3 and the concentration
is only possible by flotation.
The high economic level of the mine will be proven by the following. In contrast to
nearly all other western mines, Mittersill was only closed for a short period from 1994 to
1995 despite extremely low world market prices. It should also be noted that the
outstanding economy of the Mittersill mine has a double reason: low operating costs,
and full integration with the conversion plant of the same company which is capable of
converting low-grade ore concentrates of approximately 30% W0 3 .
Mining started in 1976 as a 100% open pit mine, changing gradually to 100%
underground until 1980. The following important regulations had to be obeyed:
After closing the open pit operation the whole area had to be completely
recultivated so that it could be reused as pasture. This also included the destruction
of the access road built exclusively for the mine (Figs. 12.2 and 12.3).

ECOLOGY

389

FIGURE 12.1. Entrance to the Mittersill underground mine. By courtesy of Wolfram Bergbau und Hiittenges.
m.b.H., Austria.

In developing the underground mine, all facilities for the miners, warehouses, and
maintenance shops, as well as the first crushing steps, had to be incorporated inside
the underground mine (Fig. 12.4).
An underground connection (a tunnel equipped with conveyer belts) had to be
constructed in order not to disturb the tourist traffic by heavy trucks (Fig. 12.5).
Special attention had to be paid to the wastewater quality leaving the tailings pond.
Table 12.1 informs about the limits.
The first tailings pond was located near the ore dressing plant. After 6 years of
operation its capacity was exhausted and the only available area was 10 km away.
This required the construction of a pipeline and pumping stations. Meanwhile,
several sections of the new tailings pond have been completed and recultivated.

390

CHAPTER 12

FIGURE 12.2. Specially constructed road to give access to the open pit mining area. By courtesy of Wolfram
Bergbau und Hiittenges. m.b.H., Austria.

TABLE 12.1. Limits Pertaining to

Wastewater Leaving the Tailings Pond
(Mittersill Mine and Ore Dressing
Plant)
Temperature
pH
Undissolved substances
Settling substances
Extractable substances
Dissolved Fe
Dissolved Al
a

::s30C

6.5-8.5
::slOOml/1
::sO.3 ml/I
::slOmg/1
::s2mg/1
::s2mg/1

By courtesy of Wolfram Bergbau und Hiittenges.

m.b.H., Austria.

FIGURE 12.3. Recultivation of the open pit mine area. By courtesy of Wolfram Bergbau und Hiittenges. m.b.H., Austria.

t'I1

-<

Cl

o
r

392

CHAPTER 12

FIGURE 12.4. Underground installed crushers. By courtesy of Wolfram Bergbau und Hiittenges. m.b.H.,
Austria.

An as yet undissolved problem is that, in the case of whole flotation, a huge amount
of energy is necessary to finely disintegrate the ore, of which only about 1% is valuable. If
no other valuable constituents are present, the separated gangue material is worthless and
cannot be used at present for any other purpose. The only possible reuse is a partial backfilling in the caverns of the underground mine.

12.3. CHEMICAL CONVERSION [12.1,12.2]

A chemical plant is always confronted with environmental problems and APT plants
are no exception. We present as an example a plant operated by the same company as that
mentioned above-the conversion plant of Wolfram Bergbau and Hiittenges. m.b.H.
located in Bergla/Styria (Austria). The facility was constructed in the mid-1970s and
has been in operation since 1977. Although it is a relatively modem plant, a series of new
regulations have been introduced recently and are summarized in Table 12.2.
It can be seen that the discard of sodium sulfate is the biggest pollution problem
caused by that plant. As described earlier in Chapter 5, this is still a worldwide problem
especially in the case of scheelite digestion, requiring a large stoichiometric excess of
sodium carbonate. Improvements for the future are necessary. Some can be found in the
current literature and are being used on an industrial or pilot plant scale. Examples are

393

ECOLOGY

FIGURE 12.S. Underground conveyer belt connection to the ore dressing plant. By courtesy of Wolfram Bergbau
und Hiittenges. m.b.H., Austria.

TABLE 12.2. Emission and Wastewater limits (Wolframhiitte BergJa)

S02 emission (ore concentrate roasting)
NH3 emission (APT calcination)
NOx emission (steam bloc)

:s2.0g/Nm3
:SS.Omg/Nm3
:s200mg/Nm3

Wastewater
Temperature

pH
AlH
CoH
Cu 2+
NiH

NHt

F-

SO~-

Organic hydrocarbons
a

:s30C
:s6.5-8.S
:s3 mg/l
:sl mg/l
:sO.S mg/l
:s0.5 mg/l
:s30mg/1
:s20mg/1
:s2S0mg/i
:sIOmg/i

By courtesy of Wolfram Bergbau und Hiittenges. m.b.H., Austria.

394

CHAPTER 12

leach milling (perfonned in China) [12.3], which decreases the amount of sodium
hydroxide for dissolution (see Section 5.2.6), and electrodialysis [12.4] of tungstate
solutions to recycle sodium hydroxide and avoid chemical waste (see Section 5.2.4.7).
These examples are mainly driven by the economy, but at the same time serve to improve
the environment. (To some extent, both milling and electrodialysis increase electric power
consumption, which itself creates environmental pollution elsewhere.)
12.4. POWDER METALLURGY
Powder metallurgical production methods for high-tech materials are very clean
operations and do not generate any waste or emissions. Internal and external scrap
recycling is excellently organized worldwide.
For infonnation about hazardous substances in the production of tungsten alloys, such
as thoriated tungsten, cobalt, or nickel-containing cemented carbides, see Chapter 14.
12.5. ENVIRONMENTAL CONSIDERATIONS ABOUT THE SUBSTITUTION OF
HIGH SPEED STEEL BY HARDMETALS [12.5]
Not only will technical aspects direct the selection of tool materials in the future, but
also ecological considerations. In regard to cutting tools, one may say that the tungsten
content of high speed steels is an impurity in regard to steel scrap and is finally lost by
dilution. Any loss in tungsten is connected with increased mining activities and chemical
conversion creating environmental destruction, decrease in natural reserves, and pollution.
In contrast, hardmetal tools and their valuable constituents with tungsten in the first place
can be recycled to a high degree. A special advantage is the possibility of direct recycling,
which has special environmental suitability.

REFERENCES FOR CHAPTER 12

12.1. M. Spross, Erzmetall44 (1991), 438--442.
12.2. M. Spross, Proc. Leobner Bergmannstag, pp. 606-612 (1987).
12.3. K. Vadasdi, in: The Chemistry of Non-Sag Tungsten (L. Bartha, E. Lassner, W. D. Schubert, and B. Lux,
eds.), pp. 45-60, Elsevier Science, Oxford (1995).
12.4. H. Li, M., Liu, and Z. Si, Chinese Patent 85100350.8 (1985).
12.5. E. Egami, in: Proc. 7th Tungsten Symposium, Goslar, pp. 13-23, ITIA, London (1996).

13
Economy
13.1. INTRODUCTION [13.1-13.10]
Tungsten's economy was generally ruled by the supply/demand situation, which itself was
a consequence of the world's general economic situation. Booming industry was always
associated with enhanced tungsten consumption and mine production as well as with the
search for new deposits. Times of general recession always showed minimum of the
consumption combined with mine closures. Wars reflected situations of highest industrial
activity and always maximum of consumption and prices.
Supply and demand have not always been evident, especially during the cold war.
Moreover, in tungsten trading, not only have economic considerations determined the rule
but also political influences played an important role, due to the strategic importance of the
metal. The accumulation of government stockpiles or their release have sometimes acted
severely as a disturbance or help. For example, in the 1950s, when China's supply was
disrupted during the Korean war, U.S. mine production was expanded to feed the
government stockpile and also concentrates were bought from abroad. This program
ended in 1959, and sales of excessive concentrates from stock started in 1965. Sometimes,
the release of government stock has had a stabilizing effect on the market, by avoiding
shortages.
Mine or market stocks whose size is partially unknown (China, CIS, and Hong Kong)
sometimes create a distorted picture of the current situation. For example, huge market
stocks and stocks of downstream products have been one reason for the abrupt decrease in
consumption between 1992 and 1994, besides a general worldwide decrease in production.
The extremely low tungsten price forced the rigid reduction of all inventories. Furthermore, these stocks are the reason why quite divergent figures very often occur in published
statistics.
In addition to stocks, another uncertainty is tungsten supply by scrap. Trading of
scrap is only partly published, so the amounts can only be estimated. Consequently, the
consumption figures given subsequently reflect only the amount supplied by primary
tungsten and not quantities originating from any scrap. An estimate discussed in Chapter
11 is that, worldwide, about 1/3 tungsten consumption originates from reclaimed tungsten
(including all types of recycling). This means that in the MEC (market economy countries)
the percentage is much higher than in the rest of the world.
395

396

CHAPTER 13

Mine Stocks
Market Stocks
Government Stocks

Intermediate Stocks

FIGURE 13.1. Parameters of tungsten economy.

Other parameters influencing tungsten's economy are shown schematically in Fig.

13.1. The different types of stock may evidently create a confusing picture of the situation
at any given time. No direct relationship exists between mine production, supply, demand,
and consumption. The supply may originate from mine production, from scrap, or from
stocks. The demand must not be necessarily balanced with consumption, because
economic reasons may dictate a build up or decrease of intermediate stocks.

13.2. WORLD MINE PRODUCTION

In Fig. 13.2, the share of different countries and continents or political areas in
available tungsten ore deposits is shown. In former times, the percentage of mine
production showed a similar picture, but since the late 1970s the situation has changed.
Due to the drastic fall in the price of tungsten, almost all MEC mines were closed.
Figure 13.3 presents the price of tungsten for the period from 1960 to 1996 while the dates
of important mine closures are indicated. This diagram is also of special interest because in
some cases it allows a rough calculation of the break-even point of the mines. Mittersill
mine is excluded from these examples, because its excellent economy is based on full
integration from the mine to highly pure Wand WC powders, while Panasqueira mine's
special economy is due to the presence of Sn in the ore.
Within the CIS (community of independent states) in 1990, 10 mines were still
working but today only one remains. Also in China, which was always the main supplier of

Australia
4"10

China
44"10

USA

7"10

Canada

15"10

FIGURE 13.2. Share of different countries, continents, or political area in known minable tungsten deposits
[13.13].

ECONOMY

397

180
160
140
startup Mittersill

120

Mt Carbine
Sal au
La Parilla

:J

r-

"'

ffi
CJ)

:J

100
80
60
40
Pine Creek
Cantung
Strawberry

20

Mittersill
Panasqueira

O~~~~~~~~~~'+~~~Tr~

60

65

70

75

80

85

90

95
Jan Mar

96

FIGURE 13.3. Tungsten price (LBM quotation) versus time (1960-1996), including closure dates ofMEC mines
[13.7].

tungsten ore concentrates, there was a decrease in mine production due to the loss of
governmental subsidization since 1991 and increased economic consciousness. The annual
output was 32,200 t in 1990 but decreased to 26.600 t in 1996. The decrease is based on a
considerable reduction in the output of state-owned mines while locally operated mine
production remained more or less constant. Nevertheless, China has tremendously
increased its share in the worldwide supply of tungsten. In 1989 it was around 69%,
and in 1995 it climbed to 80%.
It is of interest that in the 1960s and 1970s about one-half of the tungsten supply
(49%) came from politically unstable countries (China included). In order to change this
unsatisfactory situation, enhanced exploration in the MEC lead to mine openings and the
share of the MEC shifted to 58% in 1980, while a further increase to 69% was expected in
1990. As explained above, the prediction went completely wrong and, in recent years the
tungsten world has depended almost totally on China's supply, although today China
cannot be regarded as a politically stable country. Recently, the CIS (also politically
unstable) has became an important supplier, too.
Figure 13.4 shows the total world mine production from 1973 on and, in addition, a
split is made in the supply by China, CIS, and others in more recent years.
From 1975-1980, a yearly increase in mine production of 8% could be observed.
From then on, consumption remained between 40,000 and 50,000t Wja until 1989, when
a sharp decrease began.

CHAPTER 13

398

It should be emphasized that the tremendous decrease in total production is not

necessarily connected only with an equal reduction in W consumption. Supplies from
different stocks of varying size have contributed toward demand.
This decrease in mine production since 1991 is based on several reasons. In the MEC,
the low price necessitated the closure or mothballing of nearly all mines. In China the
mines have lost their governmental subsidization since 1991 and, together with a growing
awareness of the necessity to think economically, many mines were closed. Russia's
consumption fell dramatically since 1991, but mine production proceeded at full capacity
until 1994 with governmental support, which stopped in 1994. Huge stocks were
accumulated during that period.
The end of the cold war had a strong worldwide impact on tungsten consumption, and
consequently on mine production (see Section 13.3 for more details).
Any figures must be understood partly as estimates, because reports from China and
the former Soviet Union have not always been precise or available. Nevertheless, the order
of magnitude is correct and, since 1990, these countries have been more open and
cooperative.
In the future the following picture seems likely. Chinese output will not exceed
approximately 25,000 t W ja even at higher price levels. Further closures of state-owned
mines are already planned. Estimates are about 4000 t W ja for the CIS and 6000 t W ja for
the MEC. This corresponds to a sum of 35,000 t W ja. Even under much more favorable
market conditions than today, the maximum world mine production would be around
40,000 t and this must be compared with a yearly consumption of 40,000 t, which will
certainly increase.
Nonetheless one should keep in mind that predictions of future developments in any
direction has always been made but have proved wrong in most cases, as shown by the
following examples. In the early 1980s an annual 4% growth in W demand was postulated.
However, between 1980 and 1990 there was no actual growth and later there was a
decrease. Average worldwide demand forecasts were 70,000 t W ja for 1990 and 80,000 t
65
60

rI

65
50

45

.......

40

~35

030
25

~-

..... '"'""'-

_Others

DCIS

20
15
10

_China

.~ -~ -

-Total

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
year

~ ~ ~ ~ ~ ~

~ ~

FIGURE 13.4. Total world mine production (1970-1997) [13.3,13.7,13.8).

ECONOMY

399

W /a for the year 2000, but demand has remained at around 40,000 t w. A bigger share of
MEC mines in total world production was anticipated, but actually China now supplies
80% of the world's demand.
These mistakes are mainly based on the fact that the world's political and economic
development is no longer predictable.

13.3. PRICE OF TUNGSTEN

Historically, it might be interesting to note that the price for 1 kg of pure tungsten
metal in 1860 was 1400 Deutsche Mark, between 1861 and 1884 it decreased to about 200
Deutsche Mark, and from 1884 only a technical grade existed on the market and ranged
between 4 and 12 Deutsche Mark per kg up to 1902 [13.11, 3.12]. Meanwhile, its technical
importance has grown enormously and the quality of the metal has improved extensively,
but the current price of DM 30 per kg of tungsten powder.is not so different from that of
1900; if we take into account inflation since 1900, it is quite cheaper!
In contrast to many other metals, no terminal market quotations exist for tungsten ore.
Sales are negotiated between either producers and consumers, or merchants and users.
The London Metal Bulletin (LMB) publishes twice a week a price for wolframite ore
concentrate. This published price reflects concluded transactions of the previous half week
reported by producers, consumers, and traders. The price quotation is presented as a range
between high and low, and is related to relatively clean standard-grade wolframite ore
concentrate (>65% W0 3 ). The content of "dangerous" impurities (P, As, S, Ca), as well as
lower W03 concentrations, decrease the price (for example, if the concentration of Ca
exceeds 1%). Most likely, Ca is contained as scheelite and will depress the digestion yield.
The LMB quotation is based on a relatively small percentage of world production
(about 10%). Internationally, tungsten trading prices are indicated in US $/MTU W0 3 for
wolframite ore concentrates. MTU means "Metric Ton Unit" and corresponds to 10 kg
W0 3 . In the USA, prices indicated in US $/STU are quite common. STU stands for
"Short Ton Unit", equaling 20lb W0 3 .
Price quotations for ferrotungsten are published by the London Metal Bulletin and by
Metalsweek N.Y., in US $/kg W. Prices for scheelite ore concentrates always orientate
around the wolframite price. Besides these price quotations, since 1989 quotations for APT
have also been published (in MTU for Europe and STU for USA). Different price
quotations for USA and Europe exist, because the U.S. market is somewhat stronger and
not as competitive.
The necessity for APT quotations arose due to a change in I1\W material trading. In
former times, the main traded tungsten raw materials were ore concentrates of wolframite
and scheelite. In recent times, a change to APT took place because China, the biggest
supplier of tungsten ore concentrates in the past, renewed its trading strategy and gave
preference to the sale of more advanced (further upgraded) tungsten products, like APT,
blue oxide, yellow oxide, and ferrotungsten.
In order to encourage the world in following the new policy, prices for the
intermediates have been about the same as for ore concentrates in regard to W content,
which means that it was cheaper to buy APT instead of buying ore concentrate and
converting it themselves. The same happened with ferrotungsten. The "answer" of the

400

CHAPTER 13

MEC has been to impose an import duty of 151 % on W content in the USA and an
antidumping duty of 37% in the ED.
The new policy converted China from the most important ore concentrate supplier to
almost the main supplier of highly pure tungsten intermediates. This is demonstrated
clearly in Fig. 13.5.
The diagram in Fig. 13.6 presents the yearly average LMB quotations since 1960 for
wolframite, ferrotungsten, and APT (USA and Europe). The fluctuations are fairly large.
With few exceptions, the APT price follows the fluctuations of the concentrate. In
principle, it must be said that wars or warlike confrontations of a bigger extent have
always been connected with price maxima. The strategic importance of tungsten caused a
higher demand, and as a consequence the price went up. Tungsten as cemented carbide is
necessary to manufacture weapons and military equipment like trucks, tanks, guns, etc.
Consequently, a lot of ammunition like kinetic penetrators, scatter grenades, etc. are made
of tungsten heavy metal alloys or hard metals.
Thousands of metric tons tungsten
20~------------------------------------------~

16

ores & concentrates

EZ:J APTloxide

12

8
4

1989

1990

1991

1992

1993

1994

1995

1996

FIGURE 13.5. Change in Chinese tungsten supply (1989-1996) [13.7].

100~------------------------------------------,

90

80
70

30
20

10

1984

....... WoIframII8

1986

__ APT (USA)

1988

-0-

APT (Europe)

1990

1992

--6- FeW

1984

(Rotterdam)

1986

year

FIGURE 13.6. Metal Bulletin Quotations for wolframite concentrates, APT (USA and Europe), and ferrotungsten (1985-1996) [13.8].

ECONOMY

401

Also, increased tungsten demand originating from peaceful industrial booms leads to
price maxima and, correspondingly, recessions lead to minima. The historical maximum
reached in 1977 was the consequence of increased consumption.
Until 1977, one can state that supply and demand dictated the price of tungsten.
However, from then on the price control mechanism was no longer effective. Although
annual tungsten consumption within the period from 1977 to 1989 increased further from
<40,000 t W to 52,000 t W, the price fell continuously. Several circumstances were
responsible for the new situation.
Up to 1977, China was the biggest and most important tungsten ore concentrate
supplier. In its trading policy, China was always price-conscious and all relevant activities
were carried out by state officials. The situation changed drastically in 1977 with the new
liberalization policy. Producers were allowed to sell at least a certain percentage of their
production on their own. This was the starting point of a historical price decline. In order
to conduct business, the offers became cheaper and cheaper while the traders were
pressuring in the same direction. The MEC was flooded by cheap Chinese ore concentrates. The price went further down, and with the closure of the MEC mines China's share
in world tungsten supply increased.
The aforementioned change of raw material supply, from ore concentrates to APT,
etc., starting in 1987, did not improve the price situation, but had a dangerous impact on
MEC converters, as well as APT and ferrotungsten plants. Some of them were forced to
close.
The collapse of the communist world together with the sudden fall in worldwide
tungsten consumption lead to the historical minimum of the price of tungsten in 1993.
Since then, there has been a slight recovery to levels like in the period between 1986-1992.
Here, too, any future prediction seems as worthless as the general opinion expressed
at the 2nd International Tungsten Symposium, held in 1982 at San Francisco, that the
LMB quotation will not fall below 100 US $/MTU.

13.4. SUPPLY AND DEMAND

As it can be seen in Table 13.1, that supply/demand statistics are worldwide more or
less balanced. Primary tungsten supply to the MEC is illustrated in Fig. 13.7 (by
suppliers). The share of China increased while the supply from the MEC decreased.
The 1994 imports from CIS were additional. However, what is not at all obvious is the
origin of the supply, either from stocks or from mine production. Therefore, another
statistic (comparison of mine production and demand) shows a much less balanced
situation (Fig. 13.8). The following short historical consideration should explain the
situation between 1981 and 1996.
From 1982-1991, a large build-up of market inventories took place, originated by a
steady oversupply from China. Between 1992 and 1994, these inventories were a major
source of supply. In 1994, consumption overtook supply temporarily and increased the
flow of stock materials from China and Russia. In 1995, supply overtook demand and a
build-up of inventories occurred again. In 1995, the inventories were reduced.
Figure 13.9 gives a picture of the worldwide primary tungsten production/consumption situation from 1989 to 1994 for the three blocks MEC, China, and CIS. Figure 13.10

CHAPTER 13

402

TABLE 13.1. Analysis of Supply and Demand (1990-1996) (W content in metric tons) [13.8]
1990
Supply
Imports of concentrates
9950
MEC imports of concentrates from China
5600
MEC imports of products from China
12000
Subtotal
27550
Imports from CIS
China's export to Eastern Europe and North Korea
Concentrates
4500
Products
China's domestic consumption
9900
41950
Total
Q

Demand
11000
Europe
8250
USA
4650
Japan
3250
Other Western countries
Subtotal
27150
China's concentrate import
9900
China's domestic consumption
China's export to Eastern Europe and North Korea 4500
41550
Total
a

1991

1992

1993

1994

1995

7300 5800 2550 3150 5350 5350

150
4850 1200
400
200
50
12850 9450 14950 19850 20300 16200
25000 16450 17900 23200 25800 21600
200
400 4300 7000 4650

200
900
500
800
50
10000 10500 11000 11000 11000 11000
35000 27350 29800 39400 44800 37300
7700 8650 8300 11750 13000 8200
11400 4250 5000 7450 8850 8500
4450 3750 3350 5150 6500 5150
2850 1700 1700 2750 3000 3000
26400 16350 18350 27100 31350 24850
400 1450 1400
10000 10500 11000 11000 11000 11000
200
500
900
800
50
36400 27050 29850 39400 44600 37300

Other than from China and CIS, including supplies from stocks and from MEC production.

Thousands of metric tons tungsten

40T-------------------------------------------~

35
30
25
20
15
10
5
O~~~~~~~~~~~~~~~~~~~~~~

81 82 83 84 85 86 87 88 89 90 91 92 93 94 95
Gill China

(imports from)

m C.I.S. (imports from)

Gill M.E.C.
_

1996

production

U.S. stockpile sale

Stock redn. (closures)

FIGURE 13.7. Primary tungsten supply [13.6].

403

ECONOMY
Thousands of metric tons tungsten

60~------------------------------------------~

50

40
30

20

------ mine production

........... tungsten demand

10

O+-'-~'-'--.-.-r-r-r-'-'-'~~-'-.-.~r-~

75

79

77

81

83

87

85

89

91

93

95

FIGURE 13.8. World tungsten mine production and demand (1975-1995) [13.7].

Thousands of metric tons tungsten

70~----------------------------------------.

60
50

40
30

20
10

89
~

90

prodn. MEC
cons. MEC

91

93

92

prodn. China

EillI prodn. CIS

cons. China

94

cons. CIS

FIGURE 13.9. Primary tungsten production/consumption (worldwide) [13.6].

presents the primary tungsten supply/demand situation in the MEC and the cumulative
inventory change between 1981 and 1994.

13.5. CONSUMPTION AND USE

The main proprotion of tungsten consumption is closely related to industrial
processes. Sixty percent of W is consumed by the cemented carbides. These materials
serve as tools for construction work (metals, ceramics, wood, and plastic), mining tools

404

CHAPTER 13

Thousands of metric tons tungsten

35~------------------------~----------------,

30
25
20
15
10
5

81
-

82 83 84 85 86 87 88 89 90 91
supply

-0-

consumption

92 93 94

[ill] cum. inv. change

FIGURE 13.10. Primary tungsten supply/demand (MEC) [13.6].

(rock, oil, and water drilling), and as wear parts. All these applications are linked to
industrial activities. The same is more or less also valid for the rest of tungsten consumers
like alloys, pure tungsten, etc.
Historical tungsten consumption development is marked by the steady increase in
industrialization and a broadening of the application of tungsten and its products.
Fluctuations in tungsten consumption are due to booms or recessions. In average, a
growth of 4% annually was predicted in the early 1980s. As already pointed out,
predictions no longer fit in our world. Consequently, the situation is quite variable at
different places of the world.
In the MEC over the last 15 years, the average growth of W consumption was only 12%. On the other hand, in the PR China the net growth was 5% and in countries of the
former East Block consumption went down by 80%. The figures of these countries may
reflect a wrong picture, because it is unknown how much of the consumption was fed by
inventories.
Within a period of about 20 years, especially in the MEC, a reduction in W
consumption was due to:
Improvements in quality (longer lifetime of tools, coatings of hard metals, etc.).
Substitution of tungsten or tungsten-containing products by other materials
(cermets or ceramics instead of cemented carbides, depleted uranium instead of
heavy metal).
A contrary development took place in China: the replacement of high speed steel by
cemented carbides (the latter contains a higher percentage of W).

ECONOMY

405

TABLE 13.2. Tungsten Consumption [13.3,13.6,13.8]

103 t W
Year

Source I [13.3]

1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996

38.0
37.6
34.4
36.8
36.0
46.0
48.0
51.2
48.0
40.8
39.9
39.9

Source 2 [13.6]

46.7
41.9
39.7
47.3
51.3
41.6
36.4
27.1
29.9
39.5
43.5

60.0
58.0
57.0
46.0
41.5
44.5
44.6
39.5

Source 3 [13.8]

45.1
39.3
28.7

Consumptions in 1996 and 1997 range around 40,000 t W la. Experts count on an
annual growth of 4.4%, which can be split into 1.8% in MEC, 5.3% in China, and 19.9%
in the CIS.

Table 13.2 shows primary tungsten consumption since 1973 and is also partly based
on estimates. This is clearly reflected by the diverging figures.
The 1996 consumption by end use, for different areas, is presented in Table 13.3.

TABLE 13.3. Tungsten Consumption by End use-1996 (share in %)

[13.8, 13.10]

WC
Special steel, superalloys
Mill products
Others

Western Europe

Japan

USA

China

62
25
9
4

49
34
11
6

62
7
25
6

56
36
5
9

406

CHAPTER 13

13.6. THE "INTERNATIONAL TUNGSTEN INDUSTRY ASSOCIATION" (ITIA)

The ITIA was formerly inaugurated in Brussels in February 1988, and is registered as
an association with scientific aims under Belgian law.
The Association is a very useful instrument for the distribution and exchange of
valuable tungsten-related information between the different groups and members. Some 60
companies from 19 countries are members of the Association including mining companies,
processors, consumers, trading companies, and assayers. Any company involved in
tungsten may join, subject to fulfilling the requirements laid down by the ITIA charter.
One can get a good impression of how small the "tungsten family" really is by taking
into account that all major mines and convertors are included in the 60 members of the
Association, besides traders, etc.
The principal aims of the ITIA are defined as follows:
To promote the use of tungsten and tungsten products.
To provide a regular meeting forum.
To collate from different sources comprehensive statistics covering tungsten
production, processing, consumption, and end uses. To ensure confidentiality,
these statistics are submitted directly by members to an independent firm of
accountants, who prepare a summary. Safeguards are taken to make certain that no
individual company's data can be identified.
To prepare market reports for each ITIA meeting.
To collate and circulate on a regular basis information relating to the ITIA's
activities and the tungsten industry, and to arrange for publication of a periodic
bulletin.
To organize symposia and seminars.
To liaise and exchange information on a regular basis with other metal trade
associations.
To maintain contact with governments and professional organizations. For example, the ITIA is accredited to the UNCTAD (United Nations Committee on Trade
and Development) Committee on Tungsten.
The International Tungsten Symposia, organized by the Association every third year,
enable not only close personal contact of the members, but furthermore, by the presentation of technical and market-related papers the current situation of the tungsten industry
and market is highlighted. The 7th Symposium was held in Goslar, Germany in 1996. With
the exception of the 6th (Guangzhou, China), all papers as well as discussions have been
printed and are available as proceedings.

ACKNOWLEDGMENT
Without the kind and generous supply of information by the ITIA and its secretary,
Mr. Michael Maby, this chapter could not have been written in its present form.

ECONOMY

407

REFERENCES FOR CHAPTER 13

13.1.
13.2.
13.3.
13 .4.
13.5.
13.6.
13.7.
13.8.
13.9.
13.10.
13.11.
13.12.
13.13.

P. T. Stafford, in: Minerals. Facts and Problems 2 (1983), 881-894.

A. P. Newey, in: Proc. 3rd Int. Tungsten Symposium. Madrid, pp. 19-33, MPR Publishing Services Ltd.,
Shrewsbury, England (1985).
T. F. Anstett, D. I. Bleiwas, and R. J. Hurdelbrink, Bureau ofMines Information Circular, IC 9025 (1985).
R. Kohli, in: 6th Int. Tungsten Symposium. Guangzhou; transcript of papers: ITIA, London, UK (1993).
E. Doronkin, in: ITIA Annual General Meeting. Huntsville. AI. USA; transcript of papers: ITlA, London,
UK (1994).
R. M. Bunting, "Tungsten Market Studies," in: IT/A Annual General Meetings. Huntsville. AI. USA
(1994) and Stockholm (1995); transcript of papers: ITIA, London, UK.
M. Spross, Berg und Hiittenmiinn:sche Monatshefte 141 (1996),359-362.
M. Maby, "Market Reports," in: ITIA Annual General Meetings. Huntsville. AI. USA (1994), Stockholm
(1995), Goslar (1996), and Corning, USA (1997); transcript of papers: ITIA, London, UK.
ITIA Newsletters 1993-1997.
Peng Pugang, in: 10th IT/A Annual General Meeting. Corning. USA (1997).
A. Rossing, Geschichte der Metal/e, Verlag Leonhard Simion, Berlin, 1901.
B. Neumann, Die Metal/e, Verlag Wilhelm Knapp, Halle a.S. (1904).
"Tungsten Market, Update Analysis and Outlook," Roskill Information Services Ltd., p. 3, London
(1995).

14
Tungsten and Living Organisms
14.1. INTRODUCTION [14.1,14.2,14.3,14.4]
Metals play a very important role in the metabolism of living organisms. Ninety-six
elements form the earth's crust, of which 22 are essential for life. Within the essential
elements for humans we distinguish three overlapping groups (Table 14.1), the most
numerous consisting of metals which are found in only low concentrations in living
beings. By "low" we mean, for example, that the total amount of an element in the body of
an adult human ranges, depending on the element, between one gram and some
milligrams. The low concentration is the reason why some of the essential elements
have been discovered to be essential by deficiency reactions of patients receiving long-term
parenteral nutrition, and today we cannot be sure of knowing all the essential elements.
Until now, tungsten was not included in the group of essential elements like its near
relatives chromium and molybdenum for higher developed beings. However, we know due
to recent findings that it is essential for some simple living organisms (anaerobic bacteria),
which contain tungsten enzymes.
Like many other metals, tungsten and its compounds exhibit various influences on
living beings ranging from toxic, via indifferent, to beneficial. Generally speaking,
between poisonous and healthy or medicinal the threshold is often very close. Sometimes,
it is only the concentration which is decisive, or in other cases it is the valence state of the
element, such as chromium, which is among the essential elements in the trivalent state,
but is very toxic as hexavalent ion. In other cases, like barium, it is the solubility of a
compound which differentiates between toxic and indifferent; the insoluble barium sulfate
can be ingested in large quantities without any reaction while soluble salts are highly toxic.
TABLE 14.1. Essential Elements of the Human Body
Group I

Group II

Group III

Components of major structural

elements (amino acids, fatty acids,
nuc1eotides, Cas(OH)(P04h)

Responsible for maintaining ionic

equilibria, electrochemical
functions, and membrane potentials

Enzyme components
(metalloenzymes and metal ion
activated enzymes)

H, 0, C, N, P, S, Ca

Na, K, Mg, Ca, CI

Fe, Cu, Zn, Mn, Mo, Co, Cr, Se, V,

Sn, Ni

409

410

CHAPTER 14

The influence of tungsten on living organisms depends on three items:

The Kind of Living Organism. As noted earlier, for some it is essential and for others
indifferent. Poisonous effects may occur in one species (rats) and not in others (mice).
Allergic reactions of humans occur, but are rare.
The Type of Tungsten Compound. Metallic tungsten in powder form ingested by
humans did not cause adverse effects as a substitute for barium sulfate in radiological
examinations. Compared to most other heavy metals, tungsten and most of its compounds
exhibit low toxicity, although there exist some acute toxic compounds like WCI 6 and WF 6.
Effectively, it is not the tungsten but the immediate reaction with water or water vapor
which generates the gaseous compounds H2F2 or HCI, respectively. They are highly toxic
and very rapidly damage eyes and the respiratory tract.
Diluted tungstate solutions also may have beneficial influence on humans.
The combination of tungsten carbide and cobalt is responsible for the so-called
hardmetal disease.
The Concentration of Tungsten. As for most substances, there exist concentration
thresholds between toxicity, indifference, and positive influences.
The distribution of tungsten in the body was checked by experiments with animals
(mostly mice and rats). Roughly, one-third of tungsten inhaled and deposited as W0 3
aerosol and one-half of ingested water-soluble tungstate are absorbed. Most of it soon
leaves the body in the urine. The remaining amount is distributed among blood cells,
spleen, kidneys, and bones. Three months after digestion or 6 months after inhalation, it is
transferred almost completely to the bones (very low concentration). The biological halflife of tungsten in bone is 1100 d.

14.2. TUNGSTEN IN BACTERIA (TUNGSTEN ENZYMES) [14.3]

In the early 1970s, it was found that the growth of certain microorganisms are
stimulated by the addition of tungstate. Ten years later the first naturally occurring
tungsten-containing protein was isolated, and now six tungstoenzymes are known. The
number is likely to increase in the future.
Tungsten-containing enzymes until now, can be found exclusively in microorganisms
of the following types: methanogens, acetogens, and hyperthermophiles. All of them are
obligate anaerobes and are more or less thermophilic. Their enzymes have similar catalytic
properties. They catalyze redox reactions involving carboxylic -+ aldehyde conversions.
The essential role of molybdenum in a variety of fundamental biological processes
has been known for a long time. Although tungsten and molybdenum have very similar
atomic and ionic radii and comparable coordination chemistry, a substitution of Mo by W
always causes an inactivation of the enzymes. This elucidates the enormous selectivity of
biological, chemical reactions. In contrast, the substitution of W by Mo was possible in
some cases without any loss in activity, but not in all cases.
The acetogens Chlostridium thermoaceticum and Chlostridium formiaceticum grow
by sugar fermentation to produce acetate, but can also grow on CO 2 and H2. In this
metabolism, CO 2 is converted to acetate. In the first step, CO 2 reacts to the acetate methyl
group and this reaction is catalyzed by the tungstoenzyme formate dehydrogenase (FDH),

411

TUNGSTEN AND LIVING ORGANISMS

which contains W, Fe, Se, and S. Two kinds of FDH could be isolated: one containing Mo
and the other W The W species is named FDH2.
Another tungstoenzyme contained in acetogens is carboxylic acid reductase (CAR),
which catalyzes the reduction of carboxylic acids to the corresponding aldehyde:

In the methanogen Methanococcus vanelli growth, stimulation by tungstate was

observed when formate was used as energy source, but not when it was growing on CO 2
and H2 . In contrast, the methanogen Methanobacterium wolfei grows only on CO 2 and H2
The tungstoenzyme formyl methanofuran dehydrogenase (FMDH) catalyzes the first step
in the conversion of CO 2 to CH4 (reduction of CO 2 to formaldehyde), which will be bound
to the methanofurane cofactor:
CO 2 + H2

+ 2e- + MPR --+ MPR-CHO + 2H+

FMDH is a molybdoenzyme in most methanogens with the exception of the above, which
contains a FMDHI (Mo) and a FMDHII (W).
Hyperthermophiles generally grow at > 90C. Pyrococcus furiosus uses sugars as a
carbon and energy source. The tungstoenzyme aldehyde ferredoxin oxidoreductase (AOR)
catalyzes the oxidation of aldehydes:
RCOH + H20

+ Fdox --+ RC02H + Fdred + 2H+

Thermococcus litoralis gains energy from the fermentation of peptides. The tungstoenzyme formaldehyde ferredoxin oxidoreductase (FOR) catalyzes the formaldehyde
oxidation to formate:
HCOH + H20

+ 2Fdox --+ HC02H + 2Fdred + 2H+

Some properties of the six tungsten enzymes known so far are summarized in Table
14.2. It is seen that some are very complex, having molecular weights of 300,000 and
more, and contain high concentrations of iron (CAR and FDH2 from Chlostridium
TABLE 14.2. Properties of Tungsten Enzymes
g Atom. moI- I
Microorganisms

Enzyme

Mol. weight

Fe

Opt. Temp. ("C)

Thermocuccus litoralis
Pyrococcus foriosus
Chlostridum jormiaceticum
Chlostridum thermoaceticum

FOR
AOR
CAR
CAR
FDH2
FMDHII

280,000
35,000
134,000
300,000
340,000
130,000

3.3
1.0
1.4
3.4
2.0
0.4

16
7

>90
>90
40
?
?
70

Methanobacterium wolfei

II

82
36
2-5

412

CHAPTER 14

thermoaceticum). Of interest is that CAR from Chlostridium thermoaceticum and from

Chlostridium formiaceticum are not identical.
All tungstoenzymes are located in the cytoplasmic actions of the cells and are very
sensitive to inactivation by oxygen.
Moderately thermophylic organisms are not obligatorily dependent on tungsten,
because if tungsten is omitted and molybdenum is offered instead they grow further. It
seems not to be the case for hyperthermophiles, which require tungsten.
Tungstoenzymes play an important role in the metabolism of carbohydrates and
peptides in the hyperthermophiles. Hyperthermophily is regarded as an ancient phenotype.
It is assumed that the earliest life forms were hyperthermophilic and W-dependent, which
later evolved into mesophilic molybdenum-dependent species and now are the overwhelming majority. As to the reason why W was required and not Mo, there exist two
assumptions: W is present in volcanic areas where hyperthermophiles can be found (e.g.,
deep-sea hydrothermal vents); or alternatively, the better temperature stability of the
valence state in question of the W compound.

14.3. TUNGSTEN AND ANIMALS [14.4,14.5]

Investigations have been performed with mts, mice, and guinea pigs. About
investigations related to the Hardmetal Disease, see Section 14.4.2.
14.3.1. Rats
Chronic Toxicity. 5 ~g . g- I Was Na2S04 to drinking water caused slightly increased
growth development and significantly decreased life span.
150 ~g . g - I W to drinking water increased the incidence of N-nitroso-N-methylurea
induced mammary carcinoma.
Reproductive Toxicity. At a dose range without toxic effects to the maternal organism,
the following influences could be observed: Increase in embryo lethality; decrease in
ossification of bones; tungsten accumulation in the fetus; no significant retention in the
placenta.
Toxico kinetics. 100 ppm W to adict containing 30 ppm Mo resulted in a decrease in
the activity of sulfite oxidase and xanthine oxidase. Tungstate antagonizes the metabolic
action of molybdate in the xanthine hydrogenase, sulfite oxidase, and aldehyde oxidase.
Toxicological Data. Tungsten LD50 (i.p.), 5000mg.kg- 1; sodium tungstate (s.c.),
223-255 mg . kg - I; phosphotungstic acid LD50 (oml), 3300 mg kg - I; WC-8%Co LDLo
(intmtracheal), 75mg.kg- 1; WC-15%Co LDLo (intratracheal), 50mg.kg- 1
In oml uptake, the toxicity was highest for Na2 W04, intermediate for W03, and
lowest for (NH4h W04.
14.3.2. Mice
Chronic Toxicity. 5 ~g . g - I W as Na2 W04 to drinking water caused no abnormal
growth development and no decrease in life span.
Genotoxicity. 5 ~g . g - I W, as above, showed no tumoric effect.

TUNGSTEN AND LIVING ORGANISMS

413

Reproductive Toxicity. Ingestion of 0.1 ml of25 mmoll- I Na2 W04 on day 3 and 8 of
pregnancy had no effect on plantation and gave rise to increased frequency of resorption.
14.3.3. Guinea Pigs

Orally or injected, Na2 W04 solution caused anorexia, colic, in-coordination of

movement, trembling, and dyspnea.
Intratracheal administration of WC-Co gave rise to no tumors.
14.3.4. Rabbits

0.005 to 5 mg . kg-I sodium tungstate decreases the concentration of blood alkaline

phosphatase.

14.4. TUNGSTEN AND HUMANS [14.2]

In comparison to the majority of other heavy metals, tungsten and most of its
compounds possess very low toxicity, if at all. Intoxications occur rarely, almost
exclusively by occupational exposure. As treatment in acute poisoning, Dimercaprol
(British Antilewisite) may be useful.
14.4.1. Important Data
Average Tungsten Level in Blood is 5.83.5IlgW .1- 1 (I Ilg W .1- 1).
Average Tungsten Level in Bones is 0.25 ppb W.
Demand per Day is 1-15 Ilg.
Occupational Health. MAK/NIOSH (TLV-TWA): I mg m - 3 W or soluble W
compounds, 5 mg . m - 3 WC or other insoluble compounds.
Patch Test. In 853 tested workers, 2% showed irritant pustular reactions when
exposed to Na2W04
14.4.2. Hardmetal Disease [14.6-14.9)

The association between parenchymal lung disease and hardmetal exposure was first
observed in Germany and several later reports exist from other countries. An extensive
summary of the related literature in regard to clinical surveys, epidemiological studies,
clinical presentation, pathology, and hypotheses on pathogenesis was presented recently by
D. Lison and co-workers [14.6,14.7].
This parenchymal disease (fibrosing alveolitis) was only observed to occur among
hardmetal (cemented carbide) and diamond workers, the clinical presentation of which
may vary from subacute alveolitis to progressive fibrosis. Although at first WC was
thought to be responsible, and later Co, now it is evident that only the presence of both
hardmetal components induces hardmetal disease, characterized in the early stages by the
presence of bizarre, multinucleated giant cells in the lung interstitium, alveolar lumen, and
the broncho alveolar lavage. In contrast, parenchymal toxicity could never be observed

414

CHAPTER 14

when exposure was to Co alone or WC alone. Moreover, an increased risk of lung cancer
was found in hardmetal workers.
Extensive investigations comparing exposure to WC and to Co separately, and to
WC-Co mixtures or hardmetal powder, have been conducted recently [14.6-14.9]. The
following methods have been applied:
in vivo animal studies by intratracheal administration,
in vitro cytotoxicity investigations using macrophages
in vitro genotoxic activity determinations counting the DNA single strand breaks in
human peripheral lymphocytes.
All these findings very clearly demonstrate that only the combination of cobalt with
tungsten carbide is a necessary condition to induce severe alveolitis leading to fibrosis. The
pulmonary response produced by hardmetal dust is much more pronounced than that
caused by pure cobalt or cobalt compounds, while WC alone shows almost no effect.
Hardmetal disease is not a consequence of one of the hardmetal components but is a result
of interaction between Co and WC particles, producing toxic activated oxygen species,
presumably hydroxyl radicals.
This physicochemical interaction was investigated by arachidonic acid peroxidation,
ESR spectroscopy, and electrochemical measurements, and the following model is
proposed: [14.7].
Although Co is thermodynamically able to reduce oxygen, due to its surface
characteristics the reduction rate is extremely slow (resistance to oxidation). WC itself
is also unable to react with oxygen, but has high electrical conductivity and an active
surface. When both are in contact with each other, a surface electron transfer from Co to
WC takes place leading to the reduction of oxygen by the migrating electrons and
dissolution of an equivalent amount of oxidized Co as C02+. A schematic drawing is
presented in Fig. 14.1.
A possible but not yet fully proven carcinogenity ofhardmetal dust could result from
a synergism between the activated oxygen species (genotoxicity-damage to DNA) and
the already proven repair inhibition by C02+.
There do not exist any special values for permissible concentrations ofhardmetal dust
on the work place. Therefore, the MAK values for insoluble W and Co are valid
(5mgWC/m3 and O.5mgCo/m\ The ITIA (see Section 13.6) is encouraging inducing
enhanced activity in regard to acute toxicity tests to be conducted on tungsten compounds
including hardmetals (ITIA Newsletter, June 1997).

14.4.3. Beneficial Influences [14.2,14.10,14.11]

In contrast to certain toxic or hazardous effects of tungsten metal or its compounds

and alloys, very beneficial influences on human beings have also been described. In this
case, an aqueous sodium tungstate solution is applied to the body. The roots of this
knowledge-tungsten prolongs life or heals diseases-can be traced to the Middle and Far
East. The first publication was written by Mukherjje in 1936 [14.6] and today there exists
much relevant literature all written in Japanese. A compilation is given in the book
Tungsten sui Soda by K. Kase also in Japanese [14.7].

TUNGSTEN AND LIVING ORGANISMS

415

FIGURE 14.1. Mechanism ofCo-WC interactive toxicity. See: D. Lison, "Health and Safety in the Hardmetal
Industry," paper presented at the EPMA Meeting "Hard Materials-Production and Application," Grenoble,
France (1995).

The therapy comprises application as an oral 2% aqueous sodium tungstate solution,

or as a gargle, as eye droplets, as tincture to the skin, as inhalation vapor, or as
subcutaneous injection. Experiments on 500 people in Japan showed remarkably successful results for a wide range of illnesses. According to the corresponding literature, Table
14.3 lists the types of diseases which have been healed, or at least in which positive
influences could be noted. Moreover, it was found, that typical symptoms of aging, such as
partial loss of memory, etc. can be considerably retarded. In oral therapy, a daily dose of
25 ml of the above solution corresponding to 500 mg Na2 W04 , or 331 mg W, is
recommended.
Unfortunately, no further systematic clinical tests have yet been performed. Personal
experience over the years of one of the authors of the present book and of some friends
have been thoroughly positive and satisfying.
TABLE 14.3. Experience with Sodium Tungstate Therapy (500 persons) [14.7]
Very successful

Cararact
Duodenum ulcer
Halitosis

Hangover
Receding gum
Stomach ulcer

Successful

Adenomatous hyperplasia
Angina pectoris
Arrythmias
Asthma
Beginning near-sightedness
Bruises
Burning injuries
Cancer
Chronic nose bleeding
Diminishing eyesight
Diabetes
Fungus tinea pedis
Gout
Haemorrhoids
Hair loss

Heat rash
Herpes
High blood pressure
Lumbago
Mercury poisoning
Migraine headache
Obesity
Parkinson's disease
Polyps
Rheumatism
Selenium poisoning
Sinusitis
Sunburn
UV allergy

No influence

Circulatory system problems

Glaucoma
Retina detachment

416

CHAPTER 14

REFERENCES FOR CHAPTER 14

T. C. P. Stamp, "Mineral Metabolism," in: Nutrition in the Clinical Management of Disease (1. W T.
Dickerson and H. A. Lee, eds.), Edward Arnold, London (1988).
14.2. E. Lassner, in: 6th International Tungsten Symposium, Guangzhou; transcript of papers: lIlA, London,
UK (1993).
14.3. M. W W Adams, "Tungsten Proteins," in: Encylopedia ofInorganic Chemistry (R. Bruce King, ed.), Vol.
8, pp. 4284-4291, Wiley, Chichester, New York (1994).
14.4. E. Lassner, W D. Schubert, E. Liideritz, and H. U. Wolf, "Tungsten, Tungsten Alloys, and Tungsten
Compounds," in Ullmann's Encyclopedia of Industrial Chemistry, Vol. A27, pp. 229-266, VCH
Verlagsgesselschaft, Weinheim (1966).
14.5. G. D. Clayton, ed., Patty's Industrial Hygiene and Toxicology, CD-ROM, Wiley Interscience, New York
(1996).
14.6. D. Lison, Crit. Rev. Toxicol. 26 (1996), 585-616.
14.7. D. Lison, R. Lauwerys, M. Demedts, and B.Nemery, Eur. Respir. 1 9 (1996), 1024-1028.
14.8. D. Anard, M. Kirsch-voIders, A. Elhajouji, K. Belpaeme, and D. Lison, Carcinogenesis 18 (1997), 177184.
14.9. D. Lison, P. Carbonelle, L. Mollo, R. Lauwerys, and B. Fubini, Chem. Res. Toxicol. 8 (1995),600-606.
14.10. H. N. Mukherjee, Calcutta Med. 1 30 (1936), 452.
14.11. K. Kase, Tungsten sui Soda, Tokyo, ISBN-4-257-03314-2 (1991).
14.1.

Index
Abundance of tungsten, 64, 66-68
Acoustic properties, 42
ADMET catalysts, 373-374
AKS-tungsten: see Non-sag tungsten
Alloys, 255---281, 317-319
AKS-tungsten: see Non-sag tungsten
heavy alloys: see Heavy metals,
heavy metals, 27~276
Mo-W,256
Nb-W,256
non-sag tungsten, 259-266
oxide dispersion-strengthened, 266--268
precipitation hardening, 268-270
stellites, 318-319
substitutional alloys, 256--259
superalloys, 317-318
tungsten composites, 27~281
tungsten copper, 276--279, 287, 289
tungsten-fiber-reinforced composites, 278, 28~
281
tungsten silver, 276--279, 287, 289
W-Re, 22, 256--258, 269-270
W-Ta,258-259
W-Ti,259
Aluminothermic production (Ferro-W), 311312
Ammonium metatungstate, 156,214-215
Ammonium paratungstate: see APT
Ammonium tungsten bronzes, 164
Anodes in X-ray tubes, 297
Anodic processes, 125
Application of hardmetals, 83, 353 - 362
Application of tungsten
aviation, 304
chemical industry and metallurgy, 301-302
cutting, 293
electrical engineering, 286--287
electronics, 289-291

Application of tungsten (conI.)

high temperature technology, 291-293
laser technology, 304-305
leisure time and sports, 300-30 I
lighting, 283-286
mechanical and engine engineering, 298
military applications, 302-304
plasma spraying, 293
radiation, 297
space aviation, 302
spark erosion, 293-294
thermocouples, 36--37, 293
vapor deposition, 293-296
welding, 293
X-ray targets and shieldings, 297-298
APT
crystallization, 205 - 208
flow sheet, 188
properties, 157,214
thermal decomposition, 209-212
trace impurities, 210
for UHP application, 211
Arsenides of tungsten, 145
Atom spectrum, 6
Atomic mass, 1
Atomic radius, I
Atomic volume, I
Auger transitions, 5
Beneficiation of tungsten, 179-184
Bergla conversion plant, 392 - 393
Beta-tungsten (j3-W), 12-13, 95, 15~151
Binary tungsten oxides
nonstoichiometric, 151-152
stoichiometric, 146--151
Binding energies, II
Bloating/Crushing, 384
Boiling point, 30
417

418

Bonding
in tungsten, 9--10
of tungsten, 243
Borides of tungsten, 135, 138
Brazing of tungsten, 243
Brittleness, 20-22, 240
Bromides of tungsten, 171-172, 284-285
Bubbles, 262, 265
Calcium tungstate, 154
Carbide catalysts, 369--371
Carbides of tungsten
cemented: see Hardmetals
cast, 118
Menstruum, 118-119,331
powder
production, 324-336
. properties, 33~343
preparation
by CVD, 113-114,334-335
in an auxiliary melt, 118-119
in salt melts, 119
WC,I40-142
WC I _X' 140
WzC, 139--140
Carbon black properties, 115
Carbothermic reduction ofW03, 107-109,330-331,
332-334
Carbothermic production (Ferro-W), 310
Carbo-silicothermic production (Ferro-W), 311
Carburization, 114-119,324-331
direct, 330-331
high temperature, 329--330
Catalysis, 365-374
Cathode reactions, 124-125
Cathodes in x-ray tubes, 297
Cemented carbides: see Hardmetals
Chalkogenide catalysts, 374
Chemical uses, 148-149, 153, 155-160, 164, 16~
170,173,365-374
Chemical vapor
reaction, 334-335
transport, 96, 100, 101
Chemical vapor deposition: see CVD
Chinese APT plants, 201-205
Chlorides of tungsten, 169--170
Chlorination of ore concentrates, 110
Cleavage, 15
Coating (of hardmetals), 351-353
Cobalt powder, 344
Cohesive energy, 7
Coils, coiling, 264
Cold isostatic pressing, 232-233, 348
Coldstream process, 384
Cold work tool steels, 3 I 7
Compacting of tungsten powder, 231-233

INDEX
Consumption and use, 403-405
Coordination coqtpounds, 174
Corrosion of tungsten, 53, 125
Creep, 28-30
Critical density, 31
Critical pressure, 31
Critical temperature, 31
Crystallographic properties of tungsten, 11-13
Crystallographic shear structures, 151
CVD
of tungsten, 111-112,244-245
of tungsten carbide, 113-114,334-335
Deformation of tungsten, 22-27
Deformation-mechanism map, 28-29
Denox SCR catalyst, 36~368
Density
of tungsten, 13
of states, 9
Die pressing (W), 231-232
Digestion
of oxidized scrap, 193
of scheelite, 191-193
of wolframite, 192-193
Direct carburization, 109, 330-331
Direct compaction (hardmetal), 348
Direct recycling, 381-382
Discharge lamps: see Gas discharge lamps
Discoveries of tungsten and its ores, 77
Dislocations, movement of, 15, 22, 24
Doped tungsten powder, 229
chromium doping, 230
lithium doping, 229
NS/AKS doping, 230
sodium doping, 229
Ductile to brittle transition temperature, 20-22
Ductilizing effect, 22
Ecology, 387-394
Economy, 395-407
Effective nuclear charge, 2
Elastic properties of tungsten, 10-11, 18-19
Electrical resistivity, 34-35
Electrochemical treatment, 126
Electrochemistry of tungsten, 123-129
Electrochromic device, 128-129
Electrochromism, 12~129
Electrodialysis, 19~197
Electroforming, 126
Electrolytic dissolution, 126, 193-194
Electron affinity, I
Electron-beam zone melting, 244
Electron configuration, 2
Electron emission, 39--42
Electron emitters (W, W-ThO z' W-La Z0 3, porous W),
289

419

INDEX
Electronegativity, 2
Electronic structure and bonding, 7-11
Electropolishing, 127
Enthalpy
of fusion, 30
of sublimation, 30
Entropy of fusion, 30
Equilibrium lattice constants (calculated), II
Equilibrium partial pressures (W-H-O system), 92
Evolution of the universe, 61
Evaporation rate of tungsten, 30
Evolution of massive stars, 61
Extrusion (hardmetals), 348
Fabrication of tungsten, 239--243
Ferberite, 69--72
Ferrotungsten, 309 - 312
Fiber-reinforced composites, 278--281
Field emission, 42
Fischer Tropsch catalysts, 371
Flotation
of scheelite, 183
of wolframite, 184
Fluorescent lamps, 285
Fluorides of tungsten, 167, 168
Foreign elements (influence on reduction), 103--107
Fusion electrolysis, 126
Fusion reactor, 293
Gangue separation, 195
Gas-discharge lamps, 284-285, 287-288
Geology of tungsten, 65--69
Globular tungsten powder, 245
Grain boundary
diffusion, 33
weakness, 20-22
Gravity methods (beneficiation), 195
Hafnium carbide, precipitation hardening, 268--270
Halides of tungsten, 110-114
Halogen lamps, 284
Halogen cycle, 285
Hardening (tungsten steels), 313
Hard metals: see Hardmetals
Hardmetals, 321-362
applications, 353--362
coatings, 351-353
granulation, 347
hotpressing, 351
nanocrystalline precursor, 335-336
postsinter treatments, 351-353
powder
consolidation, 348
milling, 346
preparation of powder grades, 345
sintering, 348

Hardmetals (cont.)
straight grades, 312, 254
WC-(Ti, Ta, Nb)C-Co grades, 359
Hardness, 18--20
Heat sinks, 290-291
Heavy alloys: see Heavy metals
Heavy metals, 270-276
Heteropolytungstates, 157-159
Hexagonal tungsten bronzes, 164
High-density liquid, 156-157
High kinetic energy penetrators, 271, 257-276, 302303
High-performance switches, 287
High-purity tungsten, 247-250
High speed steels (HSS), 309, 314
High temperature carburization, 329
High temperature steels, 317
High tension separation (beneficiation), 184
Hot strength of tungsten, 27
Hot work tool steels, 317
Huebnerite, 69--72
Hydrodenitrogenation, 365
Hydrodesulfurization, 365
Hydrogen tungsten bronzes, 164
Hydrometallurgical scrap conversion, 189, 192-193,
380
Hydrometallurgy, 184-205
Immunity in aqueous systems, 124
Impregnated tungsten, 246
Impurities
influence on microstructure, 15
influence on strength, 28
Incandescent lamps, 284-286
Indirect sintering, 238--239
Infiltration (W-Cu, W-Ag), 276-277
Injection molding (hardmetals), 348
Intergrowth tungsten bronzes (pyrochlore), 165166
Intermetallic phases and compounds, 45, 135-138
International Tungsten Industry Association, 406
Interstitial impurities, 15,21-23,28
Iodides of tungsten, 172, 173
Ion exchange, 197,200
Ionic radii, I
Ionization energy, 2
Ionization potential, I
Iron powder, 344
Isopolytungstates, 121-122
Isotopes, 2
Kinetics of the W-H-O system, 91-103
Lattice defects, 14-15
Low temperature brittleness, 20
Lubricants in grade powder milling, 344, 347

INDEX

420
Macro process, 331-332
Magmatic hydrothennal model, 65, 68
Magnetic separation (beneficiation), 184
Magnetic susceptibility, 36
Magnetoelectric effects, 36
Magnetohydrodynamic power generation, 293
Mass absorption coefficient, 5
Melting base, 312
Melting metallurgy, 307-320
Melting metallurgy scrap, 381
Melting point, 30
Menstruum we, 118-119,331-332
Mercury xenon arc lamp, 286
Metal halide lamp, 288
Metallic bonding, 9
Metallothennic production (Ferro-W), 311
Metal working of
heavy metals, 272-273
non-sag tungsten, 262-265
P/M tungsten, 240-241
W-Th02,268
Meteorites, 63
Metric ton unit (MTU), 399
Microp1asticity,24
Microstructures of tungsten, 16--17,240-241
Minerals, 69--74
Mining districts, 75
Mittersill scheelite mine, 388 - 342
Mixed carbide powders, 343
Mixed ligand compounds, 173
Molybdenum separation, 195--196
Monotungstates, 153--155
Morphological changes (oxide reduction), 9797
Nickel powder, 344
Nitrides of tungsten, 143--144
Non-sag tungsten, 259--266
Nuclear fusion reactions, 62
Nuclear magnetic resonance characteristics, 2
Ore
deposits, 68-69
sorting, 181-182
Organometallic compounds, 175--176,371-374
Oxide dispersion strengthened alloys, 266--268
Oxides of tungsten, 90-103, 145--166
nonstoichiometric binary compounds, 151
nonstoichiometric ternary compounds, 160165
stoichiometric ternary compounds, 152-160
volatile oxide hydrate [W02(OH)2]' 92
W02,I50
W02.72 (W 1S0 49), 149
W02.9 (W200 SS )' 149, 152
W03' 146--149,212

Oxidation
of soft scrap, 190
of tungsten, 85, 95
Oxidation-reduction (heavy metal recovery), 384
Oxobromides of tungsten, 171-172
Oxochlorides of tungsten, 170-171
Oxofluorides of tungsten, 168-169
Oxoiodides of tungsten, 173
Oxygen adsorption on tungsten, 85
Oxygen diffusion in tungsten, 85
Oxygen penneation through tungsten, 85
Passivation of tungsten, 124
Perovskite structure, 161, 163
Peroxotungstates, 159--160
Phosphides of tungsten, 144
Photoelectric emission, 42
pH-sensitive tungsten electrodes, 126
Physical vapor deposition, 245
Plasticity, 22-27
Plasitifier for extrusion (hardmetals), 344
Plasma spraying, 244
Plutonic hydrothennal vein deposits, 70
Polycondensation tungstate, 120
Polytungstates, 155--157
Porous tungsten, 246--247, 289--290
Preconcentration (beneficiation), 180
Pretreatment of ore concentrates, 189--190
Primary tungsten sources, 187-188
Primary tungsten supply, 402
Proton absorption (P-process), 63
Pseudoalloys: see Tungsten copper (silver)
Pyrometry, 38
Radiation shields, 291, 297, 299
Rapid carbothennal reduction, 332-333
Rapid neutron capture (R-process), 63
Reactions
with acids and alkalis, 46
with carbon: see carburization
with ceramics, 55
with elements, 44
with glass melts, 55
with metals, 46
with metal compounds, 51
with nonmetals, 46--47
with nonmetal compounds, 50
with organic compounds, 56
with water and aqueous solutions, 53, 86
Recycling, 377- 87
Reduction mechanisms (oxide reduction), 95--96
Reduction of tungsten halides, I 10-114
Reduction of tungsten oxides
with carbon, 107-109
doped tungsten, 229--231
with hydrogen, 88-107,215--225

421

INDEX
Reduction of tungsten oxides (cont.)
push type furnace, 221
rotary furnace, 223
Reduction sequence (oxide reduction), 93-95
Reserves of tungsten, 76
ROMP catalysts, 371 -373
Scavenging circuit (beneficiation), 181
Scheelite, 7(}-73
Secondary tungsten sources, 189
Selenides of tungsten, 167
Self diffusion, 32
Semidirect recycling, 384
Shaping of tungsten, 242-243
Share in minable tungsten deposits, 396
Short ton unit (STU), 399
Silica precipitation, 195
Silicides of tungsten, 143
Single crystal preparation, 246
Sintering
activated sintering, 235
direct, 235-238
of doped tungsten, 235
of hardmetals ,348-351
of heavy metals, 271-272
indirect, 238-239
stages, 234
of tungsten, 233-239
of tungsten copper (silver) alloys, 277
Slip, slip systems, 15, 24
Slow neutron capture (S-process), 63
Sodium tungstate, 153
Sodium tungsten bronzes, 162-164
Solar system formation, 63
Solid-solution alloys: see Substitutional alloys
Solvent extraction, 197-200
Solvents in grade powder milling, 344
Spectral emissivity, 36-38
Spray conversion of WC, 335
Sputtering, 245
Standard potentials, 124
Steels: see Tungsten alloyed steels
Stellites, 3 I8, 319
Stres!HItrain curves of tungsten, 25-26
Strength, 22-28
theoretical, 27
of wires, 27
of whiskers ,27
Structural energy, 8
Structural properties, 11-15
Substitutional alloys, 256-259
Subvolcanic vein deposits, 70
Sulfides of tungsten, 165
Superalloys , 317-319
Superconductivity, 35
Supply and demand, 40 1-403

Surface diffusion, 15
Surface tension, IS
Telluride of tungsten, 167
Tempering of tungsten steels, 313, 316
Tetragonal tungsten bronzes, 163
Texture, 17,21
Thermal conductivity, 33
coefficient, 34
Thermal diffusivity, 33
Thermal expansion, 31
coefficient, 31
Thermionic emission, 40
Thermocouple electromotive forces, 37
Thermocouples, 36-37, 293
Thermodynamic functions
atom, 6
bulk,31
Thermodynamics ofW-H-O system, 89-91
Thermoelectric effect, 36
Thoriated tungsten, 40-42, 266-270, 288
Total energy, 7
Tungstate catalysts, 374
Tungsten alloyed steels, 312-317
Tungsten blue oxide, 212-214
Tungsten bronzes, 16(}-165
Tungsten carbide: see Carbides of tungsten,
334
Tungsten consumption by end use, 82
Tungsten heating elements, 291-292
Tungsten heavy metal alloys: see Heavy metals
Tungsten hexacarbonyl, 142
Tungsten losses, 378-379
Tungsten metal powder production conditions,
215-218
Tungsten metal powder properties, 218, 225229
Tungsten ore reserves, 76
Tungsten oxide hydrate (volatile) W02(OH)2' 91
Tungsten oxides: see Oxides of tungsten
Tungsten oxygen system, 145
Tungsten price, 397, 399-400
Tungsten rhenium, 22, 243, 256-258, 269-270
Tungsten scrap materials, 372-386
Tungsten steels: see Tungsten alloyed steels
Tungsten thin films, 290
Tungsten tree, 81
Tungsten world production (1910- 1996), 81
Tungstic acids, 152,214
Twinning, IS
Vapor pressure, 30
Volcanogenic-sedimentary deposits, 73
W contacts, 286
W-Ag contacts, 287

422
W-Cu contacts, 287
WC
catalysts, 370-371
properties, 140-142
WC-Ag contacts, 287
WC-TiC,344
WC-TiC-Ta(Nb)C,344
W 2C

catalysts, 370-371
properties, 139-140
Wp: see Beta-tungsten
Welding, 243
Wire drawing, 262-263
Whole flotation, 180-181

INDEX
Wolframite minerals, 69-71
Work function, 39-42, 266--268, 286
World mine production, 396--399
Wrought P/M tungsten, 240-243
X-ray absorption spectrum, 3
X-ray diffraction diagrams (a.-W, f3-W), 1213
X-ray emission spectrum, 3
X-ray tubes, 297
Xe-arc lamps, 286
Zinc process, 382-383
Zirconium tungstate, 155