FERTILIZER MANUFACTURE

E. C. HOUSTON, L. D. YATES1, and

HAUNSCHILD
Diammonium Phosphate Fertilizer R. L.
Tennessee Valley Authority,
‘from Wet-Process Phosphoric Acid Wilson Dam, Ala.

A process for making diammonium phosphate fertilizer from ammonia and wet-process
phosphoric acid was developed on a pilot plant scale. Acid containing about 24%
phosphorus pentoxide was ammoniated to a pH of about 5. The resultant precipitate,
which consisted principally of iron and aluminum phosphates, was separated from the
solution by filtration. The filtrate and additional ammonia were fed to a continuous vac-
uum crystallizer to produce diammonium phosphate. The crystals were centrifuged, mixed
Downloaded from pubs.acs.org by UNIV OF AGRI FAISALABAD on 05/16/19. For personal use only.

with filter cake, and dried to produce a granular fertilizer containing 18% nitrogen and
47% phosphorus pentoxide. Alternatively, the crystals and filter cake may be dried
separately to produce fertilizer materials of about 20-48-0 and 9-40-0 grades,
respectively. Potassium chloride and ammonium nitrate or concentrated superphosphate
may be mixed with the crystals and filter cake to produce a variety of grades of granular
fertilizers such as 18-18-18 or 9-27-27. A major portion of the work was concerned
with the development of satisfactory conditions for the precipitation step and for crystal-
lizing diammonium phosphate in the system ammonia-phosphoric acid-sulfuric acid-
wafer.
J. Agric. Food Chem. 1955.3:43-48.

and diammonium monium phosphate from electric furnace The capacity of the pilot plant was
Monoammonium
phosphates are attractive for fer- phosphoric acid. By crystallizing diam- about 1 ton per 24 hours. A flow dia-
tilizer use because of their favorable monium phosphate from acidic mother gram of the process is shown in Figure 1.
physical properties and high concen- liquor (mole ratio ammonium-phos-
tration of plant food, and because they phoric acid about 1.6; pH about 6.0), Raw Materials
provide an economical method for fixing they were able to eliminate any need for
ammonia in solid form. In the latter recovering ammonia from the crystallizer The wet-process acid used in the ex-
connection, diammonium phosphate is vapors. The process was carried out perimental work was made in a pilot
the more attractive because it fixes satisfactorily in an atmospheric saturator plant from uncalcined Florida phosphate
twice as much ammonia per unit of and in a vacuum crystallizer. Attempts rock. Typical compositions of the rock
phosphorus pentoxide. Greenhouse tests to substitute impure wet-process acid for and the acid made from it are as follows:
by Maclntire and coworkers (4) showed electric furnace acid were unsuccessful
diammonium phosphate to be an effec- in both types of equipment (<?). The Composition, Weight %
tive fertilizer and comparable to mono- impurities in the wet-process acid pre- p 2o5 so3 F Fe2 O,
ammonium phosphate as a source of cipitated in an unfilterable form and Rock 32.6 3.8 1 .4
nitrogen and phosphorus pentoxide. A caused eventual gelling of the solution. Acid 24.3 2.58 2.06 1.23
recent review of agronomic data (6) In the present work, innovations were
shows that monoammonium phosphate developed for separating and handling Arid
and diammonium phosphate are par- the impurities in wet-process acid to over- Ah Os CaO Insoluble H oO

ticularly suited for use on alkaline soils come their interference with the process. Rock 1.8 47.3 6.0 0.5
and on acid soils when their residual The work resulted in the development of Acid 1.29 0.26 59.4
acidity is neutralized with limestone. a process that consists in ammoniating
Monoammonium phosphate fertilizer wet-process acid, filtering off the result- Precipitation
has been made commercially in Canada ant precipitate, and feeding the filtrate
since 1931 and in the United States since and additional ammonia to a continuous The reaction of wet-process acid with
1946 from ammonia and wet-process vacuum crystallizer. Diammonium ammonia causes the precipitation of most
phosphoric acid by a process described phosphate is crystallized from acidic of the iron and aluminum and part of
by Atwell (7). That process is unattrac- mother liquor, as in the process de- the fluorine. The precipitate contains
tive for production of diammonium phos- scribed by Thompson, and separated by complex orthophosphates of iron, alumi-
phate because of the high vapor pressure centrifuging. The damp crystals from num, and ammonia. Exploratory work
of ammonia over diammonium phosphate the centrifuge are mixed with the filter showed that precipitation of the impuri-
solutions, which can result in excessive cake, which contains a significant pro- ties under the conditions that exist in the
loss of ammonia at the temperatures portion of the phosphorus and some ni- crystallizer always resulted in materials
encountered. trogen, and then are dried in a rotary that could not be separated readily from
Thompson and coworkers (S) have dryer. The mixture of crystals and filter the mother liquor by settling, filtration,
described a process for making diam- cake granulates on drying, so that the par- or centrifuging. Therefore, it appeared
1
Present address, F. S. Royster Guano ticle size of the product can be controlled desirable to precipitate those materials
Co., Mulberry, Fla. without regard for the size of the crystals. in a separate operation before crystal-

VOL. 3, NO. 1, JANUARY 19 5 5 43

WET-PROCESS ACID acid concentration on the filtering rate
of slurries prepared by batch and con-
tinuous ammoniation of the acid at a
rate of 10 pounds of ammonia per hour
per cubic foot of liquor in vessel is shown
in Figure 2. The highest filtering rate
was obtained with slurry prepared from
acid containing about 23% phosphorus
pentoxide. Acid concentrations in the
range 26 to 30% phosphorus pentoxide
had a decided detrimental effect on fil-
tering rate, for both batch and con-
tinuously ammoniated slurries. It there-
fore would be desirable to avoid operat-
ing in that range. The usual concen-
tration of wet-process acid produced
commercially is around 27% phosphorus
pentoxide.
Temperature of Precipitation. In
mostof the laboratory and pilot plant
work the thermal conditions were such
that the heat of neutralization brought
the solution to the boiling point in the
course of ammoniation. Little precipi-
tation occurred during the early stages of
batch ammoniation, but eventually a
TO SCREENING, point was reached where heavy precipi-
tation occurred. Significant variations
BAGGING AND
in the filtering rates obtained with slurries
SHIPPING
in which heavy precipitation was con-
Figure 1. Flow diagram of process for making diammonium phosphate fertilizer trolled to take place at different tem-
from wet-process phosphoric acid peratures are shown in Table III. De-
creasing the temperature only slightly
lization of the diammonium phosphate. was used in either type of ammoniator. below boiling caused a significant de-
Both batch and continuous precipitation The only agitation was that resulting crease in the filtering rate.

were used in the laboratory tests of the from natural turbulence. Filtering tests Rate of Ammoniation. The effect of
effects of variables. Batch precipitation rate of batch ammoniation, expressed as
were made in replicate with a test-leaf
was used in the pilot plant tests. In vacuum filter. The data are shown in pounds of ammonia per hour per cubic
the pilot plant, gaseous ammonia was Table II. foot of acid, is shown in Figure 3. This
added through an open-end pipe to a With batch ammoniation the precipi- variable was not fully investigated in con-
batch of wet-process acid contained in a tation of iron and aluminum leveled off tinuous ammoniation, although a few
cylindrical, cone-bottom tank without at about 97% by the time pH 4.5 was points obtained for the latter do not de-
viate much from the curve in Figure 3.
agitator until the pH reached 5.0. The reached. Data on that point were not
conditions were selected on the basis of obtained for continuous ammoniation. Although increase in ammoniation rate
caused a decrease in filtering rate, favor-
laboratory tests of the effect of variables In the case of batch ammoniation there
discussed later. The batch was then fil- was a progressive decrease in filtering able filtering rates were obtained at all
tered in a rotary vacuum filter. Typical rate with increase in pH from 4.5 to 6.0, ammoniation rates below 15 pounds of
data are shown in Table I. whereas continuous ammoniation did not ammonia per hour per cubic foot. A
Effect of Degree of Ammoniation. follow this trend, but a maximum filtering rate of 5 to 10 pounds of ammonia was
tests were made rate was obtained at pH 5.5. used in most of the pilot plant work.
Variables Laboratory
in which wet-process phos- Acid Concentration. The effect of Choice of a specific rate would be deter-
phoric acid was ammoniated to various
pH values, using batch or continuous
ammoniation, to determine the effect of Table I. Pilot Plant Data on Precipitation and Filtration
degree of ammoniation on the degree of
Initial temperature of wet-process acid, 0 F. 150
precipitation and on the filtering rate. Size of batch, gal. 75
Batchwise ammoniation tests were made Ammonia rate, lb. NH3/(hr.)(cu. ft.) 5
in a beaker in which 3 liters of acid were Temperature at time of heavy precipitation, F.
0
220°
ammoniated by introducing gaseous am- Terminal pH 5.0
Filtering rate, gal. filtrate/(hr.)(sq."ft. submerged) 1006
monia through a glass tube until the de-
Wash water rate, lb./lb. slurry 0.18
sired pH was reached. The continuous Ratio of wet filter cake to filtrate, by weight 0.234
ammoniation tests were made in a 3-
liter wide-mouthed flask provided with a Composition of Materials, Weight %
side-outlet tube. Ammonia and wet- P 2O5 N H3 S03 F Fe20s A/2O3 CaO . O

process phosphoric acid were introduced Phosphoric acid 24.3 2.58 2.06 1 .23 1.29 0.26 59.4
separately through two glass tubes located Slurry to filter 25.1 9.2 2.67 2.13 1.27 1.33 0.27 48.5
near the bottom of the flask. The Filtrate plus wash 21.3 8.3 2.55 1.03 Trace Trace Trace 57.8
Filter cake 20.0= 5.3 0.88 5.00 5.58 5.58 1.17 50.0
slurry left the flask continuously through °
the side outlet, and the first 9 liters were Boiling.
discarded before making filtering tests to
Eimco Corp. rotary vacuum
6
filter, 18-inch diameter by 12 inches; monofilament poly-
ethylene cloth; vacuum, 10 to 14 inches Hg; filtering temperature, 180° F.
ensure that steady-state conditions had c
About 50% of P2O5 was water-soluble.
been reached. No mechanical agitation

44 AGRICULTURAL AND FOOD CHEMISTRY

 encountered in its use. It is planned to
Table II. Effect of Degree of Ammoniation® test it in the pilot plant.
Filtering Ratec,
Gal. Fittrate/(Hr.)(Sq. Ft.)
Crystallization
Terminal Degree of Precipitation,%b Batch Continuous
pH Fe2Os A/2O3 F ammoniation ammoniation In the original process, in which elec-
3.9 91 .0 75.1 15.0
tric furnace acid was used, the mole ratio
4.5 97.3 96.0 54.0 150 115 of ammonia to phosphoric acid in the
5.0d 97.8 96.9 54.4 135 115 mother liquor was controlled by con-
5.5 98.1 91.6 65.5 120 185
trolling the pH. Thompson et al. (8)
6.0 97.5 95.8 71.4 90 170
present data on the relation of pH to
a
Acid concentration, 24% P2O5; ammoniation rate, about 10 lb. NH3/(hr.)(cu. ft.); mole ratio for saturated solutions in the
unagitated; boiling. system ammonia-phosphoric acid-water.
b
Data apply to batch ammoniation only. When wet-process acid is used, the sys-
c
Filter medium, polyfilament saran; filter vacuum, 15 inches Hg; test leaf; single
tem is altered by other materials which
cycle; replicate tests.
d
10% of P2O.Í precipitated at pH 5.0. are present in such acid. The most im-
portant of these is sulfate ion, practically
none of which is removed in the precipi-
Table III. Effect of Temperature of tation step. The crystallizer system
.The variables discussed were
Precipitation11 ra 0n
evaluated on the basis of therefore would be essentially the sys-
Temp, of Ammoniation Filtering test-leaf filtering tests using polyfilament tem ammonia-phosphoric acid-sulfuric
Precipi- Rate, Lb. Rateb, Cal. acid-water. The data available from
saran and single-cycle operation, without
tation, NH3/(Hr.) Filtrate/(Hr.) the literature on that system in the diam-
washing. That procedure was conven-
°
F. (Cu. Ft.) (Sq. Ft.)
ient and gave good reproducibility and monium phosphate range (3) were inade-
220= 10 175
215 10 116 good correlation. However, it was found quate for purposes of crystallizer control
170 10 19 that the polyfilament saran was subject and operation. Therefore, tests were
to blinding when used on the pilot plant run in the pilot plant crystallizer using
220= 5 226
170 5 31 rotary vacuum filter. Monofilament pure materials to determine the relation-
polyethylene was not subject to blinding ship between pH and the composition of
Acid concentration, 23% P2O5; batch
“
the mother liquor, the solid phases in
ammoniation; unagitated; terminal pH, and, therefore, was the preferred medium
5. in the pilot plant and also for design of a equilibrium with the solution, and the
6
Filter medium, polyfilament saran; ratio of ammonia to water in the vapors.
large plant. Other monofilament ma-
filter vacuum, 15 inches Hg; test leaf; terials may be equally satisfactory. Data
single cycle. obtained with monofilament polyethyl-
c
Boiling.
ene, as presented in Table I, are there-
fore more reliable for purposes of design
extrapolation than those presented else-
mined by costs of ammoniation tanks as where in this report.
compared with costs of filters. In the pilot plant the filter cake was
Agitation. In the course of the ex- washed on the filter by means of water
perimental work on precipitation it be- sprays. Although the filter cake is not
came evident that vigorous agitation was discarded in the process, but is included
detrimental to the filterability of the in the final product, it is desirable to re-
precipitate. To obtain a quantitative move as much of the water-soluble phos-
indication of the effect of agitation, a phorus pentoxide as can be readily re-
test was run in which 3 liters of wet- moved by washing, since it is econom-
process phosphoric acid containing 24% ically desirable to ammoniate it fur-
phosphorus pentoxide was ammoniated ther in the crystallizer.
batchwise to pH 5.0 at a rate of 5 pounds The filter cake had a tendency to crack
Figure 2. Effect of acid concentra-
of ammonia per hour per cubic foot. very soon after emerging from the slurry, tion on filtering rate of slurry
The ammoniation vessel was a glass and the cracking hampered washing.
battery jar 7 inches in diameter. The The use of a drag cloth improved wash- Those testswere run in the following
filtering rate of the resulting slurry, de- ing by 15 to 20%, as did the use of a manner. A mother liquor having a
termined on a test-leaf filter, was 160 smoothing roller to seal the cracks. The given pH was prepared in the crystal-
gallons of filtrate per hour per square residual moisture in the cake was around lizer by feeding pure phosphoric and
foot. A motor-driven propeller was 50% by weight, and little improvement sulfuric acids and ammonia and evapo-
then placed in the slurry and four 1- in dewatering could be obtained by either rating water until crystals were present in
inch baffles were installed in the jar. drag cloth or roller. appreciable amounts. The feed of acids
After agitation for 30 minutes, the filter- When mother liquor was recycled to and ammonia was then stopped, and
ing rate had dropped to 28 gallons per the precipitation step (see crystallization) water was fed to the crystallizer while
hour per square foot. Aging of the satisfactory precipitation and filtration being removed at the same rate by evap-
slurry for 30 minutes without agitation were maintained when the variables dis- oration. The vapor from the crystallizer
had no effect on filterability. cussed in the foregoing were held within was sampled to determine ratio of am-
As the natural turbulence of ammonia the recommended ranges. monia to water in the vapor at the
sparging and boiling was adequate for Laboratory tests of continuous am- given pH while water was being fed.
the reaction, mechanical agitation was moniation gave results that were com- Samples of mother liquor and crystals
avoided in the pilot plant ammoniator. parable to those obtained in batch, but were taken at the beginning and end of
The action of a centrifugal pump also continuous ammoniation was not tested the period of vapor sampling. The pH
was detrimental to the filterability of the in the pilot plant. It is believed, how- was then changed by adding ammonia,
slurry, so gravity flow was used rather ever, that continuous ammoniation would and the procedure w=as repeated.
than pumping to transfer the slurry from be the preferred method in most plants The results are shown in Table IV.
the ammoniator to the pilot plant filter. and that no serious problems would be Diammonium phosphate and ammo-

VOL. 3, NO. 1, JANUARY 19 5 5

 Table IV. System Ammonia-Phosphoric Acid-Sulfuric Acid-Water at 140° F.
Composition of Solution Composition of
Mole ratio Vapor, Ratio pH of 0.1 M
Grams/100 Grams Saturated Solution nh3h3po4 NH3H2O Solution of
NHS h3po4 HiSOi (cor. for H2SO4) pH by Weight Solid Phase
13.91 42.05 10.20 1.43 5.5 0.0029 4.5
14.02 43.75 9.30 1.43 5.6 0.0026 5.4
13.44 31 .4 14.60 1.55 5.8 0.0076 7.7
13.60 30.82 14.63 1.60 6.0 0.008 7.7
13.10 28.01 14.51 1.66 6.15 0.012 7.7
12.95 30.81 10.01 1.78 6.42 0.032 7.8
12.63 28.86 9.76 1.85 6.62 0.03 7.8

Table V. Pilot Plant Data on Crystallizer Operation

0
caused no trouble in the pilot plant and
Temperature, F. 140 remained below 0.3% through incidental
pH of slurry 6.0
removal with the diammonium phos-
Crystals in suspension, volume % 35-40
Crystallization rate, lb./(hr.)(cu. ft. mother liquor) 17-18 phate crystals. At times, however, diffi-
Loss of NH3 in vapors, % of input 2.6 culty was encountered in centrifuging.
It was found that this difficulty could be
_Composition of Materials, Weight %
avoided by continuously withdrawing a
P2O3 N
NH3
Hs so3 F Fe2 03 a/2o3 H-O
Feedliquor" 27.1 10.5 3.2 1.3 Trace Trace 46.5 portion of the mother liquor and re-
Mother liquor 22.3 13.6 10.7 2.5 0.11 0.15 40.0 cycling it to the raw acid prior to
Crystals 46.1 24.0 5.6 2.3 Trace Trace 3.3 ammoniation. A recycle ratio of 1 part
“
Filtrate plus wash (Table I) preconcentrated by atmospheric evaporation.
of phosphorus pentoxide as mother
liquor to 10 parts of phosphorus pentox-
ide as acid was found to be adequate.
The data in Tables I and V include such
nium sulfate were found to be the solid crystallizer to supply the necessary sup- recycling.
phases in equilibrium with the solution plemental heat for evaporation. Feed
at pH 5.8 through 6.62 (Table IV). liquor, sometimes preconcentrated by
Monoammonium phosphate replaced Centrifuging and Drying
atmospheric evaporation to about 26%
diammonium phosphate as a solid phase phosphorus pentoxide, was introduced The pilot plant centrifuge was a batch-
at and below pH 5.6. The tests indi- at a constant rate, and gaseous ammonia type machine with a perforated basket 18
cated that, from the standpoint of com- was introduced at a manually controlled inches in diameter by 9 inches deep.
position of the crystals, operation should rate as required to maintain the pH at Stainless steel 25-mesh screen wire was
be at or above pH 5.8. With regard to 6.0. The temperature was held at 140° used to line the basket, which had a
the ammonia-water ratio in the vapors, F., which required an absolute pressure speed of 1800 r.p.m. The average batch
there was a fourfold increase in going of about 4.5 inches of mercury. Crystal was 50 pounds. It was not convenient
from pH 6.0 to pH 6.4. It was conven- slurry was withdrawn at 20-minute to obtain design data with the pilot
ient to operate the pilot plant crystallizer intervals by means of a vacuum tank and plant centrifuge because of lack of control
near pH 6.0, which provided an ade- was centrifuged. Production rates were of feeding and the time required for
quate margin to avoid monoammonium around 70 pounds of crystals per hour stopping the basket. Centrifuging tests
phosphate formation and prevent sig- (17 to 18 pounds per hour per cubic foot were run in a smaller centrifuge 4.75
nificant loss of ammonia in the vapors. of mother liquor). Typical data from inches in diameter. With a cake thick-
Operation at pH 6.0 resulted in a loss the pilot plant crystallizer are shown in ness of 3 s inch and a centrifugal force of
of about 2.5% of the total ammonia. If Table V. 850 X G, the moisture was brought to
desired, the feed liquor can be pre- Because some precipitate passed the between 2.5 and 3% in 15 seconds.
concentrated to around 26% phosphorus filter and a small proportion of the iron It was desired to mix the crystals with
pentoxide by atmospheric evaporation and aluminum was not precipitated in the filter cake obtained earlier in the
without increasing the over-all loss of the prior ammoniation step, there always process because the filter cake contained
ammonia. was some precipitate present in the around 18% of the phosphorus pentoxide
A point of difference between the sys- crystallizer. Generally, the precipitate introduced as phosphoric acid and also
tem shown in Table IV and the system
resulting from the use of wet-process acid Figure 3. Effect of ammoniation rate on filtering rate
is that the latter contains fluorine. The
presence of fluorine did not, however,
interfere with crystallizer operation.
Calcium fluoride was identified as inclu-
sions in the crystal product.
The pilot plant crystallizer, shown in
Figure 4, was a continuous vacuum-type
unit with internal agitation and without
any positive classifying action. It was a
cylindrical cone-bottomed vessel 18
inches in diameter and 6 feet in over-all
height. Vapors were drawn from a top-
entering duct though a cyclone-type
mist trap to a barometric condenser.
Vacuum was maintained with a two-
stage steam jet ejector. Steam coils
were provided on the inside of the

46 AGRICULTURAL AND FOOD CHEMISTRY

some nitrogen. It was found that the
filter cake could be liquefied readily by Table VI. Pilot Plant Data on Drying"
agitation, such as obtained with a high- Lb. filter cake/lb. crystals (wet basis) 0.6
speed propeller. That property simpli- Moisture in feed, %
0
20
fied the handling of the filter cake. In Temperature, F.
Inlet gas 550
the pilot plant, damp crystals from the Exit gas 200
centrifuge were fed by scoop to a small Exit material 180
tank equipped with a propeller agitator. Drying gas velocity, ft./sec. 7

Liquefied filter cake was metered into Drying gas rate, lb./lb. material (dry basis) 4.3
the tank through an orifice. The mix- Retention time, min. 20
Screen analysis6, %
ture overflowed through a pipe which + 4 mesh 0.5
led into a rotary dryer, 1 foot in diameter —
4 +35 mesh 81.4
by 8 feet long. Typical data from the
—
35 mesh 18.1
dryer operation are shown in Table VI. Composition of Product, Weight %
The mixture of crystals and filter cake in P2P3
the production ratio—about 2 pounds of Total Available W.S. N S03 F Fe203 A/2Os CaO H2Q
crystals to 1
pound of wet cake—formed 47.1 47.0 42.6 18.0 4.9 3.9 2.3 2.4 0.5 0.7
agglomerates in the feed end of the “
Natural gas, cocurrent; 1-foot diameter by 8 feet long; 12 flights extending to within
dryer, which resulted in a granular prod- 2 feet of feed end; speed, 25 r.p.m.; slope, 2%.
uct in the 6- to 20-mesh range and per- 6
Includes crushed oversize, amounting to 25%.
mitted the use of relatively high gas
velocities without loss of product. No
loss of ammonia was detected by analysis
of the material as long as the temperature of associated impurities caused some mental work, since less standardization
of the material remained below 200° F. agglomeration of the crystals during has occurred in the western phosphate
and the inlet gas temperatures did not drying, ‘ The analysis was about 20% industry and less is known concerning
exceed 550° F. nitrogen and 48% phosphorus pentoxide. rock composition. Laboratory tests with
a commercially produced wet-process
acid made from western rock indicated
Materials of Construction
that no special problems would be in-
All of the equipment in the pilot plant volved when such acid is used in the
was made from stainless steel, with the process.
exception of the dryer, which was made The phosphorus pentoxide in the
from aluminum. Corrosion tests run product is practically all citrate-soluble
in the laboratory showed, however, that and around 90%. of it is water-soluble, as
plain carbon steel should be suitable as a determined by official methods of analy-
material for plant construction for the sis.
crystallizer, centrifuge, and dryer. The Bag-storage tests were made according
filter should be constructed of austenitic
stainless steel (300 series) to permit
periodic washing of the cloth with phos-
phoric acid. The results of corrosion Table VII. Corrosion Rates in
tests using a slurry of mother liquor and Diammonium Phosphate Mother
Figure 4. Pilot plant vacuum crys- crystals are shown in Table VII. That Liquor"
tallizer part of the plant required to handle phos- Corrosion
phoric acid—namely, the acid storage Rafe,
It may be desirable in certain in- tanks and the ammoniator—should be Material Mils/Year
stances to produce crystals without ad- made from A.I.S.I. Type 316 or 317 A.I.S.I. Type 304 stainless steel 0,05
mixture of filter cake—for example, to stainless steel. Rubber-lined steel is an A.I.S.I. Type 430 stainless steel 0.02
alternative for construction of storage A.I.S.I. Type 410 steel 0.08
make fertilizers that are completely 0.80
tanks. A.I.S.I. Type 502 steel
water-soluble. In such cases the filter Mild steel 0.20
cake, produced as a by-product, should Duriron Nil
be useful as a phosphatic supplement in Everdur 1010 1.5
Discussion
the preparation of mixed fertilizers. <*
Laboratory tests, 1 month’s duration,
Drying tests made with filter cake gave Estimates indicate that the production nonaerated, 140“ F.
results that were comparable to those of diammonium phosphate from am-
obtained with mixtures of filter cake and monia and wet-process phosphoric acid
Table VIII. Requirements for
crystals in terms of temperatures and gas should be attractive economically. The
rates per pound of water removed. As Diammonium Phosphate Fertilizer
principal requirements of the process are from Ammonia and Wet-Process
in the case of the mixture, the material shown in Table VIII.
Acid
granulated in the dryer, and there was The composition of the phosphoric
negligible dust loading of the exit gases. acid will affect the proportion of filter Requirement
The dried cake contained about 9% cake in the product as well as the filter per Ton of
Item Product
nitrogen and 40% phosphorus pentoxide, requirements. The acid used in the ex-
about 90% of which was citrate-soluble perimental work was slightly less pure Phosphoric acid (100%
H3PO. basis) 0.64 ton
and 50% was water-soluble. It could be than most commercially produced acid Ammonia 0.22 ton
readily reduced to powder by crushing. made from Florida rock. Therefore, it Steam, for evaporation 1.00 ton
Drying tests made with unmixed crystals is probable that phosphoric acid pro- Cooling water, to baro-
metric condenser 12.70 tons
showed that the maximum allowable duced from the usual grades of Florida
Heat, to dryer 716,000 B.t.u.
inlet gas temperature was about 400° F. rock would be satisfactory in the process. Electric power, miscel-
as compared with about 550° F. allow- The use of acid made from western rock laneous 30 kw.-hr.
able for the mixture. The few per cent should be preceded by some experi-

VOL 3, NO. 1, JANUARY 1955 47

 development by T. R. Mitchell, Jr., and
J. G. Getsinger. T. P. Hignett was in
administrative charge and assisted mate-
rially in the planning and guidance of the
project. Julius Silverberg directed the
testing of the product and the economic
evaluation of the process. Acknowledg-
ment is also made of the assistance given
by F. P. Achorn, V. J. Jones, and F. G.
Heil, who participated in several phases
of the work. A major portion of the
analytical work was done by F. M.
Youngblood and T. C. Woodis, Jr.

Literature Cited
(1) Atwell, James, Ind. Eng. Chem., 41
1318-24 (1949).
(2) Beeson, K. C., Ibid., 29, 705-8 (1937).
(3) Chernova, K. S., and Korzh, E. V.,
in “Solubilities of Organic and
Plant for making diammonium phosphate fertilizer from wet-process phosphoric Inorganic Compounds,” supple-
acid mental ed., p. 397, Van Nostrand.
New York, 1952.
(4) Maclntire, W. H., Winterberg, S.
to a method described previously (5). phosphate with common fertilizer mate- H., Clements, L. B., and Sterges.
The results showed that the granular rials has been discussed in an earlier A. J., Agron. J., 42, No. 9, 442-6
product made from mixtures of crystals paper from this laboratory (7). Al- (1950).
and filter cake remained free flowing after though the early work indicated that the (5) Miller, Philip, Lanaeus, G. A., Sae-
6 months in five-ply plain paper bags. compatibility with ammonium nitrate man, XV. C., and Dokken, . N.,
No loss of ammonia in long-term bag was limited to nitrogen-phosphorus Ind. Eng. Chem., 38, 709-18 (1946).
storage could be detected by chemical pentoxide ratios below 0.75, that point (6) Pierre, W. H., and Norman, A. G.,
“Soil and Fertilizer Phosphorus in
analysis, nor could the odor of ammonia was found to be erroneous. Beeson (2)
be detected over the material. The has discussed the reaction between di- Crop Nutrition,” p. 236, Academic
Press, New York, 1953.
product that was tested was 88% minus ammonium phosphate and limestone in
(7) Thompson, H. L., Miller, Philip,
6 plus 35 mesh and 12% minus 35 plus the presence of considerable moisture, Dole, F. H., and Kaplan, Abra-
60 mesh. The moisture was 0.7% on which results in loss of ammonia. It is ham, Ind. Eng. Chem., 41, 485-94
bagging. planned to make tests of the storage (1949).
The addition of other fertilizer mate- properties of several representative mix- (8) Thompson, H. L., Miller, Philip,
rials such as ammonium nitrate, potas- tures containing common fertilizer mate- Johnson, R. M., McCamy, I. XV.,
sium chloride, or concentrated super- rials and diammonium phosphate made and Hoffmeister, George, Ibid.,
42, 2176-82 (1950).
phosphate during the mixing of the di- from wet-process acid.
ammonium phosphate crystals and filter Received for review September 17, 1954. Ac-
cake offers attractive possibilities of cepted December 6, 1954. Presented before the-
Acknowledgment Division of Fertilizer and Soil Chemistry, Sym-
making granular high-analvsis fertilizers on Fertilizer Technology, at the 126th
posium
such as 18-18-18, 15-30-15, and 9-27-27. The authors acknowledge the sub- Meeting of the American Chemical So-
The compatibility of pure diammonium stantial contributions to the over-all ciety, New York, 1954.

POTASH DETERMINATION IN FERTIUZERS

C. W. GEHRKE, . E. AFFSPRUNG,
Flame Photometric Determination and E. L. WOOD

with Ion Exchange Separation Department of Agricultural

Chemistry, Missouri Agricultural
of Interfering Anions Experiment Station, Columbia, Mo.

FLAME PHOTOMETER offers a sodium, and phosphate. Several other magnesium, and others) and an anion
THE
rapid and inexpensive method for investigators have studied the flame exchange resin is used to remove inter-
the determination of the potassium con- photometric determination of potassium fering sulfate and phosphate anions.
tent of fertilizers and for this reason sev- and observed that various anions affect
eral studies have been made on the ap- the intensity of the potassium line (7, 2,
plicability of flame photometric methods. 6, 8). Reagents and Apparatus
Schall and Hagelberg (7) adapted the This paper describes a procedure used
official gravimetric procedure of the successfully in the Missouri Agricultural Ammonium carbonate, analytical re-
Association of Official Agricultural Chem- Experiment Station laboratories, in agent, 1.5A4 solution.
ists to flame determination, and re- which basic ammonium carbonate is Methyl red indicator, 0.2 gram of the
ported on the interference of magnesium, used to remove the cations (calcium, indicator per 100 ml. of 95% alcohol.

48 AGRICULTURAL AND FOOD CHEMISTRY