Extraction of Potash from Polyhalite 111. Effect of Particle Size, Sodium Chloride Concentration, and Temperature upon Hot Extraction by a Multistage Process J. E. CONLEYAND F. FRAAS Nonmetallic Minerals Experiment Station, U. S. Bureau of Mines, New Brunswick, N. J.
T
HE problem of recover-
calcined Polish and Texas-Yew I n the recovery of potassium sulfate f r o m polyi n g p o t a s h f r o m the M e x i c o polyhalite. Owing to halite by extraction of the calcined mineral with mineral polyhalite the difficulty of exactly simulathot water, high concentrations and high recoveries (K2SO4.MgSO42CaOS4.2Hz0), ing a continuous countercurrent in the extract liquor are desirable. Since it occurring in large quantities in process on a small scale, it has was believed that the optimum results might be the salt beds of Texas and Xew been necessary to study the exMexico, has been under investitraction in several separate steps obtained by continuous countercurrent extracgation a t this station continuor stages. A multistage extraction, laboratory multistage experiments simulattion process does not exactly reously since 1928 (see Parts I and ing countercurrent operat ion have been perI1 of this series, 7). The experiproduce the conditions likely to formed. The effects of particle size, temperature ments described in the present be encountered on a commercial of extraction, and sodium chloride concentration report were performed to ascerscale, but it is believed that the tain w h e t h e r a c o n t i n u o u s data obtained furnish some inhave been investigated. sight into what may be expected countercurrent e x t r a c t i o n of The results indicate that material as coarse f r o m a c o n t i n u o u s countercalcined polyhalite by hot water as -10 mesh should yield high recoveries and current extraction. The effects would yield s a t i s f a c t o r y resatisfactory concentrations at 100" C. Lower of sodium chloride concentracoveries and concentrations of temperatures of extraction proved less favorable tion, particle size, and temperap o t a s s i u m sulfate in the top ture of extraction were studied. liquor. Previous investigation but the presence of a low concentration of sodium by Storch (6) had shown that chloride was definitely favorable. Sodium chlothe rate of solution of uncalcined EXPERIMESTAL PROCEDURE ride retards the formation of secondary solid polyhalite was too slow to be of The apparatus used in making phases in the top stage and thereby allows the value in a n industrial process, the mu1t i s t a g e tests consisted attainment of higher concentrations. Using andStorch and Clarke (8) had of a heavy s u c t i o n flask placed within a steam bath and provided demonstrated that calcination of 2 parts of sodium chloride per 100 parts of with a motor-driven stirrer. A the polyhalite greatly accelerwater in the extraction of -10 mesh calcined reflux condenser attached to the ated the rate of solution in hot side tube was provided with a a countercurrent procedure should polyhalite, water, allowing most of the poone-hole stopper and connections give a top liquor containing f r o m 11 to 11.5 parts t h r o u g h which pressure could tassium sulfate to be extracted be a p p l i e d for sample removal. of potassium sulfate per 100 parts of water with readily. The flask was fitted with a rubber Additional experiments on the a recoaery of more than 95 per cent. stopper w i t h a c e n t e r hole for extraction of polyhalite after calthe brass stirrer which oDerated within a close-fitting brass' sleeve. cination in a s m a l l laboratory rotary kiln were reported by Clarke, Davidson, and Storch ( 2 ) . -4bent glass tube for sample removal was inserted through anhole in this stopper. Evaporation losses around the stirrer The hot extractions were made with a ratio of two parts of other were minimized by means of a rubber tube which fitted tightly water to one of polyhalite for a definite period, usually one hour, around the stirrer and was attached to the brass sleeve where it after which the solid residue was separated from the hot solu- extended below the stopper. The experimental procedure in most of the countercurrent tion and was washed with cold water. By this procedure extests essentially as follows: The desired amount of leach tractions of 85 to 95 per cent of the potash were secured m-ith liquorwas was weighed out and placed in the flask which was then concentrations averaging from 9.-5 t o 10 grams of potassium put in the steam bath and allowed to come up to temperature sulfate per 100 grams of water. Occasionally, concentrations (95" to 96" C.). A small portion of this hot solution was then as high as 11 grams of potassium sulfate per 100 grams of water removed for analysis through the samplin tube. The flask removed from the bath, the desired weigtt of calcined polywere obtained. Further calcination experiments on a larger a-as halite was rapidly added to the flask, the flask was quickly rescale in a 6 x 132 inch (15.2 X 335.3 cm.) gas-fired rotary placed in the bath, and stirring was commenced. Samples of kiln have been described by Conley, Fraas, and David- the resulting extract liquor were then periodically removed by applying pressure to the top of the reflux condenser, forcing son ( 3 ) . a mixture of solid and liquid on to a filter. The filtered liquor All of the preceding work has shown the desirability of cal- was caught in a tared, glass-stoppered bottle and reserved for cining the polyhalite prior to the extraction with hot water. subsequent analysis. If the object of the particular test was to obtain time-concenThe extraction procedure has involved a moderate amount of data on the first step only, the solid residue was usually agitation and a single displacement wash of the filtered resi- tration If concentrations were being studied on the second due to remove the retained extract solution. The industrial discarded. or any subsequent step, after the final liquor sample had been application of this means of extraction would be a n intermit- taken in the first stage, the extract mixture was removed from the steam bath and rapidly filtered on a Buchner funnel with the tent or batch process. of suction. In the meantime the liquor to be used for the It was thought that continuous countercurrent extraction use next fraction was being heated in a flask provided with a reflux might yield more favorable results than had been obtained in condenser. This solution was sampled, then quickly poured the batch experiments. A large number of extraction tests upon the warm residue obtained from the filtration. A small have accordingly been made with 100 to 200 gram samples of portion, usually 50 cc., of the fresh hot extract solution was used 1002
September, 1933
I N D U S T R I ,4L A N D E N G I N E E R I N G C H E M I S T R Y
to wash the first residue, and this wash was combined with the previous extract. The residue with fresh extract liquor was then quickly replaced in the steam bath and extraction continued with periodic sampling as in the first step. The time required for removal of extracted mixture, filtration, and replacement generally consumed from 5 to 7 minutes. P R O P E R T I E S O F POLYHALITE U S E D
Owing to the scarcity of the supply of Texas-New hIexico polyhalite a t the time this work was being done, most of the experiments were made on imported Polish polyhalite with physical and chemical properties similar to the American mineral, The screen analyses of the samples of polyhalite used are given in Table I, and the conditions of calcination and composition after calcination in Table 11. As shown in Table 11, samples 3, 5, 6, and 9 were Polish polyhalite containing varying amounts of halite (XaCl). The principal other impurity present in any appreciable quantity was anhydrite (CaS04). Samples 10 and 15 were polyhalite obtained from deposits near Carlsbad, N. hlex., and represented selected material of high purity. The chemical analysis of sample 10 indicated a composition of 97.4 per cent polyhalite, 1.05 per cent halite, and only 1.?2 per cent of other impurities, principally ferric oxide, alumina, and silica. Sample 15 was of even higher purity. The chief physical difference between the Polish and the New Mexico polyhalite was that of color. hlost of the Polish material was very light, almost approaching white, TI hile the American mineral was a bright salmon pink, probably because of higher iron oxide content.
1003
to 9-15 per cent of potassium sulfate and from 5.77 to 6.25 per cent of magnesium sulfate. Owing to the fact that syngenite (K2SOd.CaS04.HzO)always appeared as a secondary product during the extraction in the top stage, the final extract liquors contained much more magnesium sulfate than that corresponding to a 1:l ratio. Data from a typical experiment have been plotted in Figure 1 to show the general type of time-concentration curves obtained from the extraction of polyhalite with solutions initially containing considerable magnesium sulfate. In this TIME, MINUTES[CURVES I, 2.3) C
I 0
l
!
I
I
1
I
I
!
I
I
I
I
I
I l l
I
I 2 3 4 5 6 7 8 9 1 0 1 1 1 2 B 1 4 1 9 2 0
TIME HOURS . (CURVE4)
FIGURE 1.
CHAXGES IN POTASSIUN SULF.4TE EXTR.4CTION CONCEKTRATIOK DURING TYPICAL OF CALCISED POLYHALITE
USEDIN MULTI- particular series of tests four separate steps were used. The TABLE I. SCREESAS~LYSESOF POLYHALITE STAGE EXTRACTIOS EXPERIMESTS top three steps were carried out in the steam bath a t tem(I11 p w cent by weight In fraction) peratures approaching 100' C., but the final step was conS m P m NUMBER--QCREEU ducted a t room temperature (20' to 25' C.) as it was originally bIZE 3 5 6 9 10 & 15 believed that a considerable portion of the potassium sulfate BEFORE C1LCIS4TIOK 2.7 7.30 10 ... ... ... in the solid a t the end of the hot extraction steps was present 41.80 0.10 ... 40.8 HzO) formed during in the form of pentasalt (K2SO4.bCaSO4 13.0 12.45 19.30 ++ 20 28 2i:io l9.'05 10.0 8.33 . . . the process of extraction. From the equilibrium data available iE 49.3 31.30 20.80 ... + 105 21.2 9.27 15.6 15.27 (4,6, 10, 11)it seemed evident that higher temperatures would +150 i4:h be less favorable for the decomposition of this double salt. 4.51 2.6 8.3 7.38 1.76 +250 4.2" 12.78 9.6c 8.39 22.04 +326 13.55 . . . 3.0b 5.09 Curve 1 in Figure 1 represents an extraction starting with -325 11,53 11.45 ... ... an 8.7 per cent potassium sulfate solution. To obtain curve S A M P L E NUXBER---5 6 9 10 2 , the top- or first-stage extraction was repeated but stopped AFTER C A L C I N A T I O N after 10 minutes a t which time a concentration represented by point a was reached. The solid residue obtained by filtration was then treated with a 5.85 per cent potassium sulfate solution and samples mere taken a t the desired intervals. To obtain curve 3 the same procedure was repeated, yielding a +ZOO screen. b -200 screen. -150 screen. point b on curve 1 and b' on curve 2 . The time-concentration data were obtained for curve 4 in an exactly analogous EXTRACTIOX O F CALCINED POLYHALITE COKTAINING Cox- manner after yielding concentrations in the various stages corresponding to points c, c', and c" on curves 1, 2, and 3, SIDERhBLE SODIUM CHLORIDE B Y hIE.4Xs O F SOLUTIOSS HIGHIN NIGNCSILRI SULFATEAKD SODIUM CHLORIDE respectively. The first four-stage tests were made with the relatively I n the earlier multistage extraction experiments the extract fine material of samples 3, 5, and 6 in Table I. Extractions liquors from previous batch extractions were used as the initial of 82.6 to 85.4 per cent were obtained, as compared with leaching solutions. In these extracts the molar ratio of po- values of 77.3 to 87.0 by the standard laboratory batch protassium sulfate to magnesium sulfate was nearly 1:1, with cedure ( 2 ) . The over-all extraction was not essentially difusually a slight preponderance of the latter. The actual ferent in experiments 5 and 6, as compared with 3, in spite of initial concentrations in the top stage varied from 7.83 the much higher content of fines in 5 and 6. 7 -
+
.
.
I
-
. I .
I-
C
TABLE11.
CA4LCIS.kTIOX
DATAFOR
Sample Source of polyhalite Size of polyhalite Temp. of calcination, C. Time of retention, min. Time of calcination above 300' C.,rnin. R a t e of calcination, Ib./hr. (kg./hr.) K S O I after calcination yo NaCl a f t e r calcination,'% Density after calcination, grams/cc. Extraction of 1 m - A T I O X l h
Original polyhalite incompletely extracted. Large amounts of syngenite. Anhydrite. Original polylialite incompletely extracted, and new crystals of polyhalite beginning t o form from s d u tion. Large amount of syngenite, about t h e same a s preceding sample. Anhydrite. 26.98 42.23 30.79 Polyhalite forming from solution. Syngenite still present in large amounts, but less t h a n in preceding sample. 28.4 43.6 28 .O Original polyhalite incompletely decomposed. Large amount of syngenite. Anhydrite. 52.4 27.9 19.7 Xew crystals ,>f polyhalite forming from solution. Syngenite corroded, smaller amount t h a n in precedina- samele. 68.8 17.2 14.0 Polyhalite forming from solution in large quantities. Syngenite corroded, smaller amount t h a n in preceding sample. Anhydrite less t h a n in preceding sample. a s anhydrite and a s undetermined form left a s residue during extraction of polyhalite. examination b y A. Gabriel, assistant chemist-pt!trographer, Nonmetallic Minerals Experiment Station
INDUSTRIAL AKD ENGINEERING CHEMISTRY
1006
TABLEVI. EXPT.
STEP
MCLTISTAGE EXTRACTION OF -10 MESH POLYHALITE WITH ADDEDSODIUM CHLORIDE BY SOLUTIOKS Low IN MAGNESIUM SULFATE AXD SODIUM CHLORIDE* NaCl
ADDED Grams
15B1
15B2
15B4
1
15B8
%
0 5 15 30
0 30 0 15 30 60 0 30 45 0 90 0 10 20 30 0 15 30 60 90 120 183 240 0
7.62 6.47 8.89 9.28 7.78 9.15 5.60
473.2
9.1
485
9.3
44s
9.15
2
0
420
8.6
3
0
440
9.0
1
0
8.5
0
427
419
415
KzSO4
.Kin.
9.0
8.5
C O l C E N T R A T I O l OF E X T R A C T
r
TIME
9.2
0
2
5B7
Grams 398
1
4
15B5
7.2
WATER ADDED Hz0 :&SO4
2
1
Vol. 25, Yo. 9
6.7
8.6
8.46
1
6.5
436
8.9
2
0
437
8.9
1
6.5
438
8.94
2
0
438
8.84
5 15 30 45 60 0 34 64 94 120 160 0 45 0 10 30 60 90 0
45
0 60 120 180 4 Calcined polyhalite contained 30.37% KzS04 and 0.13% NaC1. grama, respectively, were used. b Computed.
G . / l 0 0 g. H z 0 8.44 10.02 10.72 11.35 8.52 11.10 5.96
6.32 ... fi.71 7.31 , .30 8.01 7.39 8.02 8.90 9.10 l0:97 5.76 6.14 8.14 9.00 3.06 3.20 5.56 5.87b 6.786 6.36 6.68 7.22b 0 0 1.78 1.81b 1.88b 1.84 2.14 2.19b 2.36 2.42b 2.42 2.48b 2.61 2.68b 2.64 2.71b 8.50 9.40 8.80 10.34 9.19 11.06 9.36 11.38 11.44 9.41 9.52 11.66 1.60 1.63 7.25 ... 7.62 ... 7.77 7.75 8 ;3 4 b 7.65 8.25 9.09 9.72 11.85 2.00 2.05 5.86 7.12 7.82 ... 8.10 8.94 8.90 8.09 8.90 10.60 1.70 1.73 7.42 8.08 ... 9.02 8.12 Charge used was 161.5grams in all
...
.... ..
...
in the original raw polyhalite. Tests 15B1 on the first stage and 15B2 on the first and second stages were exploratory. Tests 15B3 and 15B4 were made by the use of three successive hot extraction steps as in some of the earlier countercurrent exDeriments. but. as the data accumulated, it was noted that th'e concentration of the extracts from the second and third steps closely approximated each other, so that but two hot extraction steps were used in all later tests. Test 15B4 shows a fairly well-balanced example of the four-stage e x t r a c t i o n procedure, in which each stage yielded a solution close to the concentration of the starting solution for the preceding stage. The concentration of 11 grams of potassium sulfate per 100 grams of water in the extract liquor from the first
MULTISTAGEEXTRACMESH CALCINED POLYHALITE WITH SOLUTIONS PRACTICALLY FREE FROM SODIU>I CHLOOF
%
G./100 8 . Hi0
2 .06 5.48 6.82 7.36 1.00 6.70 0.50
2.2s 6.46 8.24 9.05 1.10 8.13 0.53
0 1.59 1.54 1.57
0 1.88 1.86 1.92
0 1.73 0
0 2.10 0
1:33 1.37 1.00
1:45 1.50 1.09
0:i4 0.13
6:30 0.50
7:59 0.53 1.51b 0.26
1:69
%
0'. 25
.. ..
..
.. .. ..
..
.. .. .. ..
1.00 4.73 6.50 7.06 7.16 7.13
.. .. ..
.. .. .. 1.00 6.48
0.23
.. ..
1:io 1.00 7.07 0
..
0
..
.. .. .. .. ..
..
.. .. .. .. .. .. .. .. .. ..
1.10 5.55 7.84 8.59 8.72 8.73 0
..
.... 0:47b 1.10 7.90 0.24
.. ..
1:44 1.10 8.41 0
..
...... .. ..
.. .. .. 0
1.38 1.37 1.37 1.39 1.61
.. .. .. .. .. ..
0 1.77 0
o:i5
0.14 0 2 :04
.. .. .. .. I
.
.. .. .. ,. .. .. ,.
.. ..
0 1.62 1.65 1.67 1.69 1.97
.. .. ..
..
.. .. 0 2.16 0
..
.. .. .. ..
1.'?5
2:is
.. .. ..
.. ..
.. ..
n'is 0137 1:68 1:51 tests excepting B1 and B2 in which 142.8 and 171.5
and satisfactory coordination of the second with the first step.
FIGURE 4.
TIONS
7
NaCl ...
Test 15B8 was made to determine if any advantage would accrue by a slightly longer time of extraction in the second step, No improvement was obtained by increasing the time from 90 to 180 minutes. All of these tests indicate that a high degree of extraction should be possible with a concentration in the top liquor of about 11 grams of potassium sulfate per 100 grams of water, and that material as coarse as - 10 mesh should be easily treated. Actual percentage recoveries have not been determined on this material, but, by comparison with results on the finer material, no difficulty would be anticipated in obtaining satisfactory results.
stage stands in the proper relation to the
ratio of water to potassium sulfate ratio of 9.15, to give a very high recovery, and the molar ratio of potassium sulfate to magnesium sulfate has the proper value of 1:l. Tests 15B5, 15B6, and 15B7 w e r e made to secure a properly balanced system by the use of two hot extraction steps. Test 15B7 shows reasonably good results, with the proper concentrations
MgSO4 G./100 0. H20
-10
RIDE
EXTRACTION OF CALCINED POLYHALITE OF Low SODIUM CHLORIDE CONTEXT BY SALT-FREE SOLUTIONS The presence of appreciable amounts of sodium chloride in the extract liquors might be a disadvantage in some cases. Accordingly, in order to determine the results which might be obtained in the absence of sodium chloride, a number of extraction experimentswasmadeupon calcined New Mexico polyhalite which had been mashed with water prior t o
September, 1933
I N D U S T R I A L A N D E N G I N E E R I N G C H E hZ I S T R Y
calcination until it waz practically free from sodium chloride. KOsodium chloride was added to the extract solutions in this series. From previous batch experiments in the laboratory it had been generally observed that the use of higher ratios of water to potassium sulfate resulted in a better percentage extraction, usually a t a sacrifice in the concentration. A relatively low initial concentration of potassium sulfate and a relatively high ratio of water to potassium sulfate of approximately 10 were used in test 13R, the results for which are shown graphically in Figure 4. In this test 96 per cent of the potassium sulfate was extracted with a top concentration of 0.66 grams of potassium sulfate per 100 grams of water. One noticeable effect of the high ratio of water to potassium sulfate was the increased solution of the magnesium sulfate content of the charge in the top step. A number of additional multistage leaching tests was made upon a sample of -10 mesh material taken from lot 15, washed prior to calcination in order to reduce the salt content to 0.13 per cent sodium chloride, and calcined in the 6-inch (15.3-em.) rotary furnace (3) a t 484' C. The details of extraction have been summarized in Table VI1 and the results plotted in Figure 4. In the absence of sodium chloride but with fairly high ratios of water to potassium sulfate, concentrations of the order of 10 grams of potassium sulfate and 7 grams of magnesium sulfate per 100 grams of water, combined with satisfactory recoveries, seem possible for material as coarse as - 10 mesh. TABLE VII. MULTISTAGE EXTRACTION O F -10 MESH POLYHALITE WITH N O ADDITIOS OF SODIUM CHLORIDE W[TH SOLUTIOSS L O W I N h!fAGNESIUM SULFATE5
n20:
-CONCENTRATIONO F E X T R ~ C T KzSOd \IgSOa G./100 8. G /io0 p Grams Min. % H20 % HzO 15.11 1 388.6 9 5 0 7.01 7.72 2.09 3.00 5 7.91 9.12 5.46 6.30 15 8 . 3 1 9.86 7.21 8.54 8.66 10.32 7.51 8 . 9 5 30 45 8.91 60 8.62 90 8.72 15.43 1 407.9 9.9 0 7.00 7.63 1.14 1.24 30 8.94 10.52b 7.166 2 407 9.9 0 5.15 5.46 0:50 0 . 5 3 5 5.98 15 6.66 7.25 1142 1155 30 6.94 ... 45 7.31 8.01 1:41 1155 15.13 437 0 9.6 6.99 7.60 1.00 1.09 30 8.71 6.17 10.20 7.23 5.45 0 437 5.14 0.50 9.6 0.53 7.37 8.08b 45 .. 1.616 0 3.17 3.27b 10.0 .. 456 4.31 10 4.50b 20 5.14 5.42b 5.80 30 6.16b 45 6.35 6.78b 60 6.51 6.968 90 7.04 7.58b .. 120 7.16 7.71b ,. .. Calcined polyhalite contained 30.37% KzSOa and 0.13% KaC1; 135gram charges used in A1 and '12, 150 gram8 in 4 3 . b Computed.
EXPT.
WATER
STEP
A D D E D IIERC.4PTOBESZOTHIAZOLE, DIPHENYLGUAR'IDINE, during vulcanization. Although it some sulfur reaction p r o d u c t s to ANI) hfIXTURES OF THE T W O may be that Rhitby and Cambron's render zinc oxide soluble for actiFORMUL~ e x p l a n a t i o n has some b e a r i n g Pale crepe vation. 100 100 100 Stearic acid ... 3.0 ... on v u l c a n i z a t i o n , it obviously Three per cent of stearic acid was Mercaptobensorhiazole 0 5 1.0 ... Diphenylguanid ine 0 5 ... 1.0 added to each of the series of stocks 1 P a r t I, IND. ENG. CHEM..23, 1467 Zinc oxide 12 2 12.2 12.2 Sulfur described above. I n the first seijes (1931); Part 11, I b z d , 24, 565 (193?). 3.0 3.0 3.0 ECHKICAL interest in the two-accelerator or multipleaccelerator effect has received considerable impetus within the past several years owing to improvements in factory processing methods which permit the utilization of its advantages. However, the literature records very little of scientific interest on the phenomenon.
