INDUSTRIAL AND ENGINEERING CHEMISTRY Hildebrand and Buehrer, J. Am. Chem. Soc., 42, 2213 (1920). Hill, C. L., U. S. Dept. Agr., Tech. Bull., to be published (contains bibliography). International Critical Tables, Vol. 111, p. 27, McGraw-Hill, 1928. Ibid., Vol. 111, p. 370. Keyes and Hildebrand, J . Am. Chem. Soc., 39, 2127 (1917). Knorr, Ber., 30, 912 (1897). Kraut, Rhoussopoulis, and Meyer, Ann., 212, 255 (1882). Marcus, J . Biol. Chem., 80, 9 (1928). McKelvy and Simpson, J. Am. Chem. sot., 44, 105 (1921). Mondain-Monval, Compt. rend., 183, 1104 (1926). Palmer, “Carotinoids and Related Pigments, the Chromolivoids,” Chemical Catalog, 1922. Parks and Huffman, IND.ENG.CHEM.,23, 1139 (1931). Rising and Hicks, J . Am. Chem. Soc., 48, 1929 (1926).
Vol. 26, No. 4
(38) Rothmund, 2.physik. Chem., 26, 433 (1898). (39) Schorger, U. S. Forest Service, Bull. 119 (1913) (contains bibliography). (40) Seyer and Todd, IND.EXG.CHEM.,23, 325 (1931). (41) Shepard and Henne, Ibid., 22, 356 (1930). (42) Shepard, Henne, and Midgley, J. Am. Chem. Soc., 53, 1948 (1931). (43) Smith and Norton, Ibid., 54, 3811 (1932). (44) Takahashi, Nakamiya, Kawakami, and Kitasto, Sci. Papers Inst. Phys. Chem. Research (Tokyo), 3, 81 (1925). (45) Thorpe, J . Chem. Soc., 35, 297 (1879); Am. Chem. J . , 1, 155 (1879). (46) Timmermans, J., J . chim. phys., 20, 491 (1924). (47) Walden, Ber., 40, 3215 (1907). (48) Wenaell, Phurm. J.,44, 97 (1872); Pharm. Rev., 22, 408 (1904).
RECEIVED October 27,
1933.
Phosphate Fertilizers by Calcination Process Action of Silica and Water Vapor on Phosphate Rock D. S. REYNOLDS,K. D. JACOB,
AND
L. F. RADER,JR., Bureau of Chemistry and Soils, Washington, D. C.
N SEVERAL publications facture of superphosphates (20). A study has been made of the action of silica Because of the deleterious effect from this bureau (13, 18, and water o,l phosphate rock at high tern201 S8)1 i t h a s been peratures in relafion to the colafilization of the animals of f l u o r(so, i n e 34), on the its presence h e a l t h in of shown (1) that fluorine is an important constituent of all types fluorine content Of the rock and the Of phosphate rock has prevented of phosphate rock produced in the insoluble phosphafe into forms readily availthe general use of this cheap s o u r c e o f c a l c i u m and phosthe United S t a t e s a n d m o s t able as plant food. HJater vapor and silica are both importantreagents in effecting the decomposiphorus in mineral mixtures for foreign Countries; (2) that for livestock feeding; removal of a given type Of phostion of phosphate rock at high temperafures; the fluorine f r o m p h o s p h a t e phate rock the fluorine content water capor is more important than is usually roughly proportional and a rock is of interest in this conto the phosphorus content; and combination of the fwo is much more effective nection ( 6 ) . (3) that the principal phosphatic than either alone. I n the presence of suflcient constituent Of phosphate rock is silica and water vapor, heating the 40-mesh T R E A T ~ ~ EOF N T PHOSPHATE calcium fluorphosphate, identirock f o r 30 to 60 minuies at 1400” C. results in ROCK BY CALCINATION cal in crystal structure with PROCESSES fluorapatite,calo~z(p04)6. D~~~ the volatilization of 95 io 100 per cent of the have also been obtained which jhorine and the coneersion of 85 to 95 per cent ?Tarious processes have been of the phosphorus into the citrate-soluble condiproposed for the manufacture of indicate that the presence of this tion; a sintered or semifusedproduct is obtained. available p h o s p h a t e s by socalcium fluorphosphate is called calcination methods (1, 8, largely responsible for the comNo increase in fhe citrate solubility of the 9,lo, 12, S*l 41)1which usually paratively low citrate solubility the fluorine in involve the heating of phosphate (17, 19) and fertilizer efficiency phosphorus is Obtained excess Of that combined as calciumfluorPhosPhate, rock with alkali salts (usually the (42) of the phosphorus in raw CaloF2(P04)6, and all that equicalent to one atom sulfates or carbonates of sodium phosphate rock, and for the slugof JEuorine in the fluorphosphate itself is tolatiland potassium), with or without gishness of the reaction between ized. From point, the percentage citrate added silica and carbonaceous phosphoric acid and phosphate material, in a nonreducing atrock (29). It may be concluded, Or less dimosphere up to about 1500” C. solubilitJ’ Of Ihe phosphorus is therefore, that the conversion of the phosphorus of phosphate r d y proportional to the percentage eolafilizaWith the exception of the work lion of the second atom of fluorine in the fluorof Guernsey and Yee (12) and rock into a condition that will Kuusk (24), very little inforpermit its ready utilization by phosphate. mation is available on the facplants is largely dependent upon tors involved in the production the extent to which the calcium of available phosphates by calcination processes. Inasmuch fluorphosphate is decomposed by the processing treatment. It is reasonable to suppose that any method which would as the products obtained by such processes have properties effect complete or nearly complete removal of the fluorine similar in many respects to those of basic slag, which is an from phosphate rock might also increase the availability of important phosphate fertilizer material in Europe, it seemed the phosphorus, provided that substantial quantities of desirable to make a thorough laboratory investigation of the calcium meta- and pyrophosphates, which have compara- behavior of phosphate rock towards various reagents a t high tively low plant-food values (S), are not formed in the proc- temperatures, in order to determine the factors influencing ess. Also, the treatment of phosphate rock by such methods the reaction and to determine, in particular, what relation should make it possible to effect a more complete recovery exists between the amount of fluorine removed from the rock I U. S. Patents through 1922 are listed by Guernsey and Yee (18). and utilization of the fluorine than is possible in the manu-
I
’”
April, 1934
I N D U S T K I A L A ?J D E N G I N E E R I N G C H E M I S T R Y TABLEI.
407
CHEMICAL COhlPOSITIOh' O F NATURAL PHOSPHATES (Results on air-dry basis except as indicated)
T Y P EOR SOURCEOF
PHOSPHATE
P?05
CaO
%
%
Florida land pebblee 31.62 47.69 35.81 Florida hard rock 50.82 34.06 lennessee brown r o c k P 48.08 30.83 Tennessee blue rock 45.17 50.55 35.63 Tennessee white rock Wyoming 46.06 30.08 Morocco 52.50 34.77 Fluorapatite 48.63 37.40 a Determined b y the Willard and JVinter method ( . i fI . b Total Si. C Total Fe. d Total S. e Ignited a t 400-500° C.
coz
H?0 AT 105' C.
F" %
Fe?OsC
%
%
%
3.87 3.87 3.75 3.85 3.83 3.48 4.20
1.63 0.69 2.27 3.40 0.92 0.87 0.12
3.66 2.18 1.73 2.58 3.04 4.12 4.08
1.37 0.56 0.82 6.61f 0.35 3.088 1.45
..
2.85h
SOsd
..
% 0:50
0:46 0.48 0.35 0.95 0.08
i Contained 5.19 per cent 50s as acid-insoluble sulfide. B Contained 1.30 per cent S O P as acid-insoluble sulfide. h
Contained lese than 0.2 per cent C1.
and the availability (solubility in neutral ammonium citrate solution) of the phosphate. As a part of this investigation a study was made of the action of silica and mater vapor on phosphate rock a t high temperatures and the results are prehented in this paper.
MATERIALS USED
The phosphate rocks used in this study were representative of the majority of the domestic commercial types of this material produced a t present. For comparison, a few experiments were also made with Tennessee white-rock phosphate, commercial phosphate rock from Morocco, and PREVIOVS IKVESTIGATIONS O N HIGH-TEMPERATURE VOLAfluorapatite from Quebec Province, Canada. Except a s TILIZATIOX OF FLUORINE FRO?J PHOSPHrlTE ROCK stated otherwise, the Florida land-pebble and Tennessee A few data on the volatilization of fluorine from phosphate brown-rock phosphates used in most of the experiments were rock a t elevated temperatures have been reported by Rey- ground to pass a 200-mesh sieve and were ignited a t 400" nolds, ROSS,and Jacob (SQ), Rozanov (44, 45, &), Illem- to 500" C. to remove the greater portion of the water. The rninger, Waggaman, and Whitney ( S I ), Britzke and Pestov ( 5 ) other phosphates were ground to pass a 100-mesh sieve and and Reynolds and Jacob ( S T ) . However, these data do not were uqed without preliminary ignition. Partial chemical relate directly to the preparation of available phosphates analyses of the samples are given in Table I. The silica used in most of the experiments was specially by calcination methods. As indicated in several United States and foreign patents purified quartz flour which completely passed a 200-mesh (22, SS, 40, is), the effect of removal of the fluorine on the sieve. The coarser sizes of silica used in the other experiavailability of phosphates produced by various processes of ments (Table IX) were prepared in the laboratory from pure calcining phosphate rock with alkali salts has heen recognized quartz crystals. in a general way for several years. N o s t of these patents indicate the importance of silica, and some of them the imAPPARATUS AND EXPERIMENTAL METHOD portance of water Vapor, in the reaction. Iluusk (24, 25) Figure 1 is a diagrammatic sketch of the electric furnace made an extensive study of the action of silica on phosphate arrangement used in the investigation. rock a t high temperatures but failed to stress the effect of water vapor and did not control this important factor; he The platinum-wound furnace, E, which was similar to the one showed that, to a certain extent, a relation existed between described by Madorsky ( W ) ,had an inner heating unit, A , comthe amount of fluorine removed from the rock and the solu- posed of No. 26 B and S gage platinum wire wound over a distance of 29.0 em. on a spiral-grooved alundum tube; the outer bility of the phosphorus in citric acid solution. heating unit, B, consisted of N o . 18 B and S gage nichrome wire According to Caldwell (6), substantially complete removal (No. 4) wound on an alundum tube. The space between the of the fluorine is obtained by heating presintered phosphate inner and outer heating units was packed with 120-mc!sh RR The rock a t 1400" to 1450" C. in the presence of an excess of alundum "specially prepared for carbon determinations. temperature of the outer unit was maintained at approximately oxygen. No mention is made, however, of any effect of 1000" C., and that of the nichrome-wound furnace, C, used for silica and water vapor on the reaction. preheating the air and water vapor, at approximately 900' C.
AIR
-
- . ~ ,.. . . , . . . , .
,
AND
TO SCRUBBIN6
WATER VAPO
TOW€/?
C
FIGURE 1. VERTICALCROSSSECTION OF FURNACE ARRANGEMENT
INDUSTRIAL AND ENGINEERING CHEMISTRY
408
Vol. 26,No. 4
The reaction tube, D, of McDanel refractor porcelain, carried I n order to obtain data on the extent to which fluorine is the platinum-platinum rhodium thermocou re and the furnace volatilized from calcium fluoride at various temperatures, charge. In some of the experiments (Table f11) known amounts of water were supplied to the furnace by saturating dry air at experiments were carried out with a silica-free synthetic fixed temperatures. In the other experiments the water was calcium fluoride and with fluorspar (Bureau of Standards boiled at such a rate that approximately 1.0 gram er minute sample No. 79). The results (Table 11) show that water this was was carried into the furnace; as indicated in Table far in excess of the amount actually required for the reaction. vapor and silica are both important reagents in effecting the In all experiments the flow of air through the furnace was main- decomposition of calcium fluoride at high temperatures, tained at 120 cc. per minute (dry air measured at 25" C.). Vary- that water vapor is more important than silica, and that a ing the dry-air supply between the limits zero and 240 cc. per combination of the two is much more effective than either minute had no significant effect on the reaction. The platinum-platinum rhodium thermocouple used for meas- alone. uring the tem erature of the charge was calibrated against a TABLE11. EFFECTOF WATERVAPORAND SILICAON VOLATILIstandard cou ye of the Bureau of Standards; it was protected ZATION OF FLUORINE FROM CALCIUM FLUORIDE against the fpurnace atmos here b a gas-tight McDanel re(Samples heated for 30 minutes) fractory porcelain tube. &e axiar temperature gradient from FLUORINE VOLATILIZED the hottest portion of the furnace outward was approximately SYNTHETIC CALCIUM 2.88 Q. FLUORSPAR^ + 2 . m Q . 4" C. per cm. over a distance of 3.8 cm. The charge was heated FLUORIDE^ HEATED S I L I C A HEATSD I N : TEMP. IN WATER V A P O R b Water vaporb Dry aird in a platinum boat 7.6 cm. long, 1.3 cm. wide, and 0.9 cm. deep, placed so that the center of the boat coincided with the hottest c. % % % 800 .. portion of the furnace. Temperatures were measured at a point mn ... approximately 1.3 cm. from the hottest portion of the furnace 1000 S:9 towards the gas exit end. Tests indicated that the maximum 1100 16.7 1200 31.5 and minimum temperatures of the charge were usually within 1300 92.1 10" C. of the average temperatures measured in this way. Dur1400 .. 84.8 ing a run the temperature was usually maintained within *5O C. a Containing 99.8 per cent CaFn. of the desired average temperature. b Water supplied a t the rate of 1.0 gram in 120 cc. air per minute (dry In making the experiments the furnace was brought to a tem- air measured a t 25' (3.). of Standards standard sample No. 79 containing 94.9 per cent perature about 10" C. higher than that at which it was desired CaF:Bureau and 1.89 per cent SiOr. 1:l mole ratio of d a 0 to SiOs. to heat the charge; at the same time the desired flow of gas was d 120 cc. per minute; dried over HzSOb and measured a t 25' C. passed through the furnace. The boat was then introduced, and, after heating for the desired length of time (reckoned from about EFFECTOF WATERVAPORAND SILICAON PHOSPHATE ROCK the third minute after the introduction of the boat), it was withdrawn, cooled, and weighed, and the residue if fused or greatly As shown in Tables I11 and IV, the volatilization of fluorsintered was ground to pass a 200-mesh sieve, or, if only slightly ine from phosphate rock increased with the moisture content sintered, to pass a 100-mesh sieve. of the furnace atmosphere. The results obtained on straight The fluorine content of the original phosphates, and also Florida pebble phosphate (Table 111) indicate that the that of the residues containing less than 0.5 per cent fluorine, optimum effect of water vapor is obtained when the water was determined by the Willard and Winter method (51); is supplied at the rate of a t least 0.067 gram per minute per the fusion-acid extraction method (36) was used on residues 5 grams of phosphate rock; larger quantities of water caused containing more than 0.5 per cent fluorine. Later, excellent but little improvement in the volatilization of fluorine. results for fluorine in all the furnace residues were obtained Several experiments with various types of phosphate rock by slightly modifying (35) the original procedure of Willard showed that preignition of the phosphate a t 400" to 500" C. and Winter. The majority of the experiments were run in to remove moisture reduces the amount of fluorine voIaduplicate, the results agreeing within less than 5 per cent tilized when the rock is heated a t 1300" C. for 30 minutes in a dry atmosphere. of the total fluorine volatilized. All samples for the citrate-insoluble phosphorus deter- TABLE111. EFFECTOF WATERVAPORON VOLATILIZATION OF minations were ground to pass a 200-mesh sieve, and the FLUORINE FROM FLORIDA LAND-PEBBLE PHOSPHATE analyses were made by the recently adopted modification [5 grams of ignited (400-500° C.) 200-mesh phosphate heated for 30 minutes a t 1200' C.] (26) of the official neutral ammonium citrate method. ComAPPROX.QUANTITY APPROX. QUANTITY parative determinations showed ' the same percentages of OF HrO SUPPLIED OF Hz0 SUPPLIED TO FURNACE F VOLATO FURNACE F VOLAcitrate-insoluble phosphorus in 200-mesh and 100-mesh .4TM03PHEREa TILIZED ATMOEPHERE~ T I L I Z E D samples. Gram/min. % Gram/min. % b 12.4 0.030 38.0 All silica determinations were made by the Hoffman and 0.000~ 18.6 0.067 50.8 Lundell (14, 16) modification of the Berzelius method. 0 003 29.7 1.0 51.7
18,
7
O
. I
C
EFFECTOF KATER VAPORAND SILICAox CALCIUM FLUORIDE I n 1856 Fremy (11) reported that calcium fluoride is decomposed when i t is subjected to the action of water vapor a t a red heat, the reaction presumably proceeding according to the equation CaF2
+ H20 =
CaQ
+ 2HF
According to Treadwell and Hall (CS), calcium fluoride is more or less completely decomposed by heating it with silica in moist air: CaFl 4HF
+ H20 + Si02 = CaSiOJ + 2HF + Si02 = SiFa + 2H20
(2) (3)
Svendsen (48) proposed to utilize these reactions in the manufacture of fluorine and potassium compounds from mixtures of calcium fluoride and potassium-bearing silicates.
b C
0.011 33.2 In 120 c c . air per minute (dry air measured a t 25' C.). Air dried over concentrated sulfuric acid. dpproximate moiature content of air saturated a t 25' C.
OF TABLEIV. EFFECTOF WATERVAPOR ON VOLATILIZATION FLUORINE A~VD CITRATE SOLUBILITY OF PHOSPHORUS IN DIFFERENT TYPESOF UNIGNITED PHOSPHATE ROCKWITHOUT ADDITION OF SILICA
(5-gram samples of 100-mesh rock heated for 30 minutes a t 1300' C.) FLUORINE VOLATILIZED;CITRATES O L W - ~ Si02 Hs0 SUPPLIEDIN 120 BILITY OF PzOj CONcc. OF AIR PER iLI1N.b 1.0 g. HzO in A a FOLLOWS: TEST TYPEO R SOURCE OF 0.023 1.0 Original 120 co.,air g. rock per min. OF PHOSPK.ATE ROCK g.
Z;JE
%
%
%
%
Florjdalandpebble 8.89 28.6 34.4 64.7 Floridahardrock 4.98 25.3 28.8 53.6 23.3 30.0 58.9 Tennesseebrownrock 7.44 9.09 37.2 39.0 58.9 Tennessee blue rock Tennessee white rock 1.90 23.5 25.0 54.0 Wyoming 7.16 18.5 26.3 58.0 46.5 Morocco 0.85 8.3 a Per cent of total PsOa. b Dry air measured a t 25' C. 0 Dried over conceptrated H2SO4. d Approximate molsture content of air saturated a t 25'
..
% 11.0 7.8 5.5 7.2 10.7 2.6 13.0
C.
% 4.4 3.7 4.2 2.8 9.3 5.1 3.9
INDUSTRIAL A N D E N G I N E E R I N G CHEMISTRY
April, 1934
EFFECT OF
Although the results (Table IV) obtained on different types of rock containing different percentages of silica show no general relation between the silica contents and the percentages of fluorine volatilized, the low results on the Morocco phosphate are probably due, at least in part, to its very low silica content. I n general, the citrate solubility of the phosphorus in the products obtained under the conditions of these experiments was lower than that of the phosphorus in the original phosphates.
409
TEhlPER-4TcRE
Table VI and Figure 2 show the effect of temperature on the volatilization of fluorine and the citrate solubility of various phosphates heated alone and with added silica for 30 minutes in the presence and absence of water vapor. With all charges heated (lOOOo to 1400" C.) in a dry atmosphere, less than 32 per cent of the fluorine was volatilized and the citrate solubility of the phosphorus in the products was less than that of the phosphorus in the original phosphates.
TABLEv. EFFECTOF VARIOUS Ah10UNTS OF ADDEDSILICA ON VOLATILIZATIONOF FLUORINE AND CITRATESOLUBILITY OF PHOSPHORUS IN FLORIDA LAND-PEBBLE PHOSPHATE IN PRESENCE OF WATERVAPOR
ID
IMixtures of 200-mesh materials heated for 30 minutes a t 1300' C. in water vapor supplied at the rate of 1.0 gram of water in 120 cc. of air per minute (dry air measured a t 25' C.) 1 COMPOSITION OF MIXTURE F CITRATESOLUPhosphate rock" Added Si02 Total Si02 VOLATILIZIGD BILITY OF PgOs Crams Grama Grams % % 4.3b 58.6 0 0.45 5.00 19.lC 0 0.45 5.00 l6:4 57.9 0.25 0.68 4.75 9.0 50.7 0.50 0.90 4.50 6.4 55.5 0.75 1.13 4.25 17.3 64.8 1.00 1.36 4.00 33.9 79.7 1.50 1.81 3.50 25.3C 1.50 1.81 3.50 5i;l 84.1 2.00 2.27 3.00 45.9 82.7 2.50 2.72 2.50 0 Ignited a t 400' to 500" C. b Citrate solubility of PxOs in original phosphate (ignited a t 400' t o 600° C.) 0 8.2 per,cent. C Charge heated in a dry atmosphere.
The data given in Table V indicate that increasing the total silica content of the charge within the approximate limits 9.0 to 22.6 per cent (0.45 to 1.13 grams) has no pronounced effect either on the volatilization of fluorine or on the citrate solubility of the phosphorus, when Florida pebble phosphate is heated for 30 minutes a t 1300" C. in the presence of an ample supply of water vapor. Increases in the volatilization of fluorine and the citrate solubility of the phosphorus were obtained, however, when the total silica was increased within the approximate limits 22.6 to 45.4 per cent (1.13 to 2.27 grams); larger quantities of silica apparently have a depressing effect. I n the absence of water vapor, increasing the silica content of the charge within the limits 9.0 to 36.2 per cent (0.45 to 1.81 grams) did not have an important effect on the reaction. The frequently noted deficiency of fluorine in fluorapatite (7) (see also the analysis given in Table 1) may have been caused in part by the action of silica and water vapor on the fluorapatite a t high temperatures.
Temperature
- 'C
FIGURE2.
EFFECTOF TEXPERATURE ON VOLATILIZATION OF FLUORINE AND CITRATE SOLUBILITY OF PHOSPHORUS IN FLORIDA PEBBLE PHOSPHATE I n the presence of water vapor, but without added silica, the volatilization of fluorine from Florida pebble phosphate increased with the temperature within the range 800" to 1400" C. but amounted to less than 70 per cent a t 1400" C. Heating the rock a t 800" to 1300" C. decreased the citrate solubility of the phosphorus, and even at 1400' C. the solubility did not exceed 23 per cent. The results of a few experiments (see also Table IX) indicate that neither the volatilization of the fluorine nor the citrate solubility of the
ox VOLATILIZATION OF FLUORINE AND CITRATE SOLUBILITY OF PHOSPHORUS IN VARIOUS TABLE VI. EFFECTOF TEMPERATURE PHOSPHATES (200-mesh silica used in all experiments) -FLORIDA LASD-PEBBLEPHOSPHATE^--. TESNESSEEBROTN- --MOROCCO 6.0 G . PHOSPHATE
TEMP.
volatilized
F
Citrate soly. of
PzOs
volatilieed
c.
%
%
%
1000 1100 1200 1300 1400
9.6 12.9 12.4 19.1 26.9
.. .. ..
4:6/
+
ROCKP H O S P H A T E b
Citrate soiy of
volatilized
F
Citrate soly. of PzOs
70
%
3.5 Q. PHOSPHATE. 1.5 G. A D D E D SILICA
F
PzOs %>
5.0 0. P H O J P H A T B
volatilized
F
Citrate soly. of PzOs
%
%
HE.
