INDUSTRIAL AND ENGINEERING CHEMISTRY
468
while t,he nitrogen contents increased to a t o 0.7 weight 70, maximum of about 4.0 weight Yc at, a retort, temperature of 1300' to 1400" F. The specific gravities of both gas oils increased, and the aniline points of the light gas oils decreased as the retort temperat,ure was raised, indicating t h a t t h e molecular st,ructure became progressively more dehydrogenated. The low aniline points, high specific gravities, and high nitrogen contents of these gas oils indicate t h a t they would be poor charge stocks for catalytic cracking, would not be suitable for use as Diesel fuels, and would not be high-quality burner fuels, but they have been hydrogenated to high yields of excellent, gasolines. Residuum. The yields and properties of the residua are given in Table VI. The changes in t8heproperties of the residua indicate that they also became progressively more dehydrogenated as the retort temperature was raised. The specific gravities and the carbon-hydrogen ratios increased as the retort temperature was raised. The physical properties of the residua also changed as the ret,ort temperature was changed. The residua from the oils produced a t 1000" and 1100' F., while solid a t room temperature, were not brittle and could not readily be crushed or ground a t room temperature. On the other hand, those from oils produced a t retort temperatures of 1200' F. and higher could be ground readily in a mortar t o a fine powder t h a t did not cake when stored in the laboratory for several weeks. The residua from oils produced a t ret,ort temperatures of 1200" F. or higher probably could be pulverized for use as powdered fuels. CONCLUSIONS
At 1000" F. an oil n-as obtained that was similar in chemical characteristics to, but somexhat more volatile than, oil from K-T-U type retorte. B s the retort temperature was raised, t h e products became more dehydrogenated, the volatility increased, the production of C4and lighter olefins a t first increased and then
Vol. 47, No. 3
decreased, and the oil became highly aromatic. Oils were obtained which have been refined to good yields of stable gasolines having low sulfur contents, and having octane ratings which exceed those of t h e present-day premium gasoline. In the system considered in this study the effects of the retort temperature appear to be most pronounced a t about 1200" t o 1300" F. In this temperature range the composition of the naphtha changed from olefinic to aromatic, and the maximum yields of naphtha, of CSplus Ca olefins, and of C3-L crude were obtained. ACKNOWLEDGMEKT
This project was part of the Synthetic Liquid Fuels Program of the Bureau of Mines and was performed a t the Petroleum and Oil-Shale Experiment Station under t h e general direction of H. P. Rue and H. M. Thorne. Special thanks are due various members of the personnel of the station for their valuable aseistance in carrying out this project. The work was done under a cooperative agreement between the University of Wyoming and the U. S.Department of the Interior, Bureau of 3Iines. LITERATURE CITED
(1) Ball, J. S.,Dinneen, G. U., Smith, J. R., Bailey. C. W.,arid Van l l e t e r , R., ISD. ENG.CHEM.,41,581 (1949). ( 2 ) Brantley, F. E., Cox, R. J., Sohns, H. W., Barnet, W.I., and I\Iurph), W. I. R.,Ibzd., 44,2641 (1952). ( 3 ) Hull, W. Q., Guthrie, Boyd, and Sipprelle, E. hI.,I b i d . , 43, 3
(19.51).
(4) Sohiis, H. W., Jukkola, E.E., Cox, R. J., Brantley, F. E., Collins, TI-. G., and Llurphy, W. I. R. I b i d . , 47,461 (1956). ( 5 ) Thorne, H. M.,LIurphy, W. I. R.. Ball, J. S.,Stanfield, K. E., and Horne, J. W., I b i d . , 43,20 (1951). RECEIVEDfor review May 7, 1954. ACCEPTED November 2, 1954. Presented before the Division of Gas and Fuel Chemistry a s part of the Symposium on Synthetic Liquid Fuels and Chemicals a t the 125th Meeting of the AAIERICAN CHEWCAL SOCIETY, Kansas City, Mo., March 1954.
Pure Amines for the Flotation of Si from Rougher Phosphate Concentrate THOMAS H. LENTZ, DAVID E. TERRY, AND HAROLD WITTCOFF Research Laboratories, General Mills, Znc., Minneapolis 13, M i n n .
F
LORIDA land pebble is t h e largest single source of phosphate and is responsible for nearly 75y0of t h e dollar value of phosphorus pentoside produced in this country. At present, there are three processes used in concentrating pebble phosphate: (1) washing and screening of 1-mm. material, (2) agglomerate tabling and Humphrey spiral concentration of - 1-mm. +28mesh matelial, and ( 3 ) froth flotation of the -28-mesh +ZOOmesh fraction. This paper deals with a phase of t h e last operation. I n 1942, Crago ( 2 ) developed a double flotation process involving anionic flotation of phosphate, subsequent deoiling of the concentrate, and cationic flotation of silica from t h e anionic float concentrate. Grades of plus SOTo bone phosphate of lime (BPL) and less than 2% silica n e l e available a t iecoveries of 90% of t h e phosphate. I n this operation, fatty acid soaps are used to float phosphate minerals n i t h a n aim to-ivard maximum recovery. The anionic phosphate concentrate is then treated Kith sulfuric acid t o remove f a t t y acids and flotation oils. After this treatment, amine flotation of silica from t h e anionic concentrate leaves
+
a cationic flotation residue of lon- hydrochloric acid-insolubles content and high bone phosphate of lime level. I n current commercial practice, tallow amine is widely used as t h e cationic reagent. This is a commercially available mixture composed largely of primary amines with chain lengths of 16 and 18 carbon atoms, Some of these amines are unsaturated, as tallow amines ordinarily have an iodine value in the range of 30 to 60. Economic considerations make tallow amines the most feasible products for cationic silica flotation. It was of interest, nonetheless, to explore the collecting properties of a series of pure amines and attempt to correlate chemical properties with performance in flotation. With this in mind, the folloning series of amines was investigated: Pure, saturated, primary amines with chain lengths of CIO, C12, CM,C I ~and , CI~. Pure octadecyl and octadecenylamines (saturated and monounsaturated, primary, C18 amines). Primary and secondary tallox3 amines. All reagents were converted t o acid salts, usually t h e acetate, t o
March 1955
Krrrl
0
469
INDUSTRIAL AND ENGINEERING CHEMISTRY
0.2
0.1
0.3
0'
0.4
I
Figure 2.
AMINE L E V E L , LE. PER TON
Figure 1. Effect of chain length on amine collector
I
I
0.1 0.2 0.3 AMINE L E V E L , L e . PER TON
0.4
Unsaturated CIS primary amines in silica flotation
Curve numbers show iodine value of amine collector. Figure i n parenthesis is ratio of octadecyl to octadecenylamine.
facilitate dispersion in water. I n another series of experiments, the effect of the salt-forming acid was studied. REAGENTS
The fatty amines were prepared in t h e laboratory. The Clo through Clo products were obtained by careful fractional distillation of coco amines through a Podbielniak column. Octadecylamine was obtained from crude stearylamine by distillation in the same manner. The octadecenyl (monounsaturated, C I ~ ) amine was obtained from distilled tallow amine by crystallization from absolute methanol a t -50" C. The filtrate was treated twice with methanol and finally warmed to volatilize the solvent. The concentrated h a t e was distilled for recovery of octadecenylamine. Primary amine content was 94.9'%, owing to a concentration of minor impurities from the alcohol crystallization. Table I summarizes t h e properties of the amines used. The amine number is defined as the milligrams of potassium hydroxide equivalent to 1 gram of a n amine sample. Salts of the amines were meoared bv treating the amine with an equivalent weight of acid. T o illustrate, 2.5 grams of Alamine &
.
-
26 (amine number, 192.5) required (2'5) (192'5) (60) = 0.51 gram (56108) of acetic acid (equivalent weight 60) for 100% neutralization. When convemions to salts of formic, glycolic, and hydrochloric acids were made, proper equivalent weights were substituted in the above equation. Concentrated acids were used in the neutralizations. Water solutions of 2.5y0 amine basis concentration were used, except for octadecglamine, where a l.Oyosolution mas prepared. ORE
Feed to the fatty acid flotation is typically 60 to 70% silica and 30 to 40% collophane. The associated silica is free and no grinding is required.
Fatty acid flotation of phosphate values results in a rougher concentrate of 65 t o 70% bone phosphate of lime and 20% silica. (Silica refers to components insoluble in hydrochloric acid.) Samples of these rougher concentrates were received from Florida and stored wet in lined drums. Separation was not a problem and grab sampling of the feed from t h e drum sufficed. Three different feed materials were used during the test program. Data in Tables I11 and I V were obtained from rougher concentrates assaying 69.2yo bone phosphate of lime. Another feed containing 74.670 bone phosphate of lime was investigated for the amine studies described in Table V. The studies on the effect of salt-forming acids, shown in Table VI, mere conducted on rougher concentrates assaying 72,4y0 bone phosphate of lime. FLOTATION PROCEDURE
Five hundred grams of the rougher phosphate concentrate ( T O O grams wet) was treated for 1 minute with 0.6 pound of sulfuric
Table I.
Amine
Properties of Reagents Used
Amine Number Chain TheoActual Length retical (6) 10 367 12 303 14 265 16 232 18 208 18 210 Tallow ..
Iodine Value TheoActual retioal (1) .. 0.0 .. 0.0 0.0 0.0 0.0 2 4 91.7 94.7 .. 48 8
..
70Primary
Amine Theo- Aotua! retical (3) 100 94 5 100 99.8 100 99 3 100 99 4 100 99 4 100 94 9 .. 896
.
secondary Ditallow .. 105.0 .. 80.4 . a Registered trade name of General Mills' primary tallox amine. b Contained 81.4y0 seoondriry amine.
4 .Sa
INDUSTRIAL AND ENGINEERING CHEMISTRY
470
Vol. 47, No. 3
v)
W c
c U 2 w 0 2
0 0
E v)
w J m -I 3 v) 0
I -
0
*
'
2
0 w
P
a W
1 1
0
0 FORMATE X ACETATE A GLYCOLATE
1-1
0 HYDROCHLORIDE
0 2
01
AMINE L E V E L ,
I
acid per ton of feed to remove oils and fatty arid. (A4ctually, t h e feed had been deoiled prior to shipment from Florida. Thc sulfuric acid t,reatrnent was repcated to assure clean, fresh surfaces for flotation.) -Lfter this scrub, t h e feed n-as washed free of slimes and mineral acids. It was then made t o 507, solids and conditioned for 1 minute with 0.4 pound of sodium hydroxide per ton (flotation pH 7.8 t o 8,5), amine a t desired levels, arid 0.3 pound of pine oil per t,on. This conditioned pulp was transferred t o a BOO-gram Fagergren laboratory flotation machil!e, arid floated for 1 t o 2 minutes at, 25% solids. Silica ivas renioved as the froth product. Minneapolis t a p water was used throughout the test program. Typical hardness was 7 5 p.p.m. CaCOs or 4.4 grains per gxlloii, with a normal breakdown of calcium t o magnesium of 2 to 1. Wat.er temperature was normally 77" t o 80" F. Flotation concentrates were analyzed for per cent hydrochloric acid-insolubles according to standard methods, with t'he final hydrofluoric acid volatilization of silica eliminated. Per cent hone phosphate of lime was reported on both concentrates and tailings by a colorimetric determination of phosphorus on an acid-leached sample. This method n-as a modification of the well-known phosphovanadomolybdate complex nicthod ( 4 ) ) which permitted direct measurement of phosphorus on a colorimeter. DISCUSSION
f i i w
Figure 4,. Salts of Alamine 26 in silica flotatioii
I
i
The basic data demonstrating the effect of chain length of xinine are shown in Table I11 and from these data Figure 1 \vas prepared. As the chain length decreased from CIS to CI:, efficiency increased. T h e CIAand C,, amines were nearly identical in flotation, while the Clo amine n-as l c s efficient than C1a. This inversion of results with the CK amine can he explained on the following basis. As flotation collector niolccules are composed of
Table 11. lncreased liecorei.\- with Cleaned Tailings (0.4 pound of decylainine acctnte per ton) Flotation Product
Concentrate NiddlingQ Tailing (cleaner froth)
Wt.,
BPI,,
74.8
87.0 72 .i I3 0
c/c
6.2 2
yc
100.0
Curnula'/r H P L T3PL tive Assay, DistribuUnitr Yo BPL t,ion 63.1 4 5
2j
87.0 85 9
..
90.4' 13.2; " 3 4
72 1
Considered recovered by addition to concentrate, reslllting in final concentrate of 85.9yc BPL and 96.6% ~CCOT'CI'Y. '8
Table 111.
l c e t a t e of Dccylainine
con-cLusIoNs
Efficient amine collectors should produce a concentrate of low hydrochloric acid-insoluhles coritent with high recovery of phosphate. I n this investigation, amine levels as high as 0.3 and 0.4 pound per ton were used t o acquire complete data t o plot curves. These higher reagent levels lowered recovery, owing to inclusion of phosphate particles in t h e silica float. Increased recoveries were attained by cleaning the rougher froth (silica product) without additional reagents. T h e cleaner froth product was discarded as finished tailings, while t h e residue (or nonfloating product) could be added t o t h e primary concentrate without impairing its grade to any appreciable extent. Table I1 shows t h e significant increase in recovery. For this test 0.4 pound of decylamine acetate was used per ton of fecd.
0 4
0 3
L B P E 9 TON
nodccylamine
Teti.a+rylamine Hexadecylamine
Octadecylamine
a
Effect of Chain Length of Amines on Silica Flotation pound/ Short
Kt.
0.1 0.2 0.3 0.4
80.3 77.7 76.2 74.8
0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4
76.5 73.8 73.2 71.8 7 7 ,B 75.3 71.6 68.8 78.j 75.2 78.8 74.1
87.6 86.1 84.1 84.8 83.7 87 6 80.2 84.1 84.1 84.1
4.15 2.63 1.90 1.70 4.22 2.49 1.80 1.64 5.06 2.78 2.69 2.28
26.2 26.8 28.2 22.1 24.7 28.4 31.2 21.5 24.8 26.2 28.9
0.1 0 2 0.3 0.4
87.3 84.6 76.7 73 6
77 2 78.5 82 0 82.7
11.72 8.85 5.00 3 78
12.7 15.4 23.3 27.4
Ton
Phosphate Concentrate
YG
Yo
BPI. 78.5 82 0 88.1 87.0
84.1
85.5
HCI-
c7~
insoluhlea 6.85 4 17 3.17 2 66
Based on caloulatcd headb of 69.2% RPL.
silica c; Tailinp, BPL" Wt. YG Recoi-ers 19.7 01.1 22.3 92.0 91 8 23.8 25 2 94 I 23.5
$13 0 !)].A
92.7 89 4 $14 7 92.8 86.6
87 2 90 8 91 2 89.7 90 1 97 3 RR 1 $11 0 87.9
March 1955 Tahle IV. Ratio of Octadecyl to Octadecenyl 1oo:o
. a
INDUSTRIAL AND ENGINEERING CHEMISTRY Unsaturated C18 Primary Amines in Silica Flotation Iodine Value 2.4
70:30
28
47:53
50
23:77
71
0:lOO
91.7
pound/ Short Ton 0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4
_Phosphate Concentrate silica yoBpLa ?’& HC1- Tailing, Recov% Wt. ery Yo B P L insolubles Wt. Yo 87.377.2 11.72 12.7 97.4 8 4 . 8 78.5 8.85 15.2 96.2 76.7 82.0 5.00 23.3 90.8 73.6 82.7 3.78 26.4 87.9 83.9 80.2 8.41 16.1 97.4 90.6 78.2 8 0 . 2 4.93 21.8 7 4 . 6 82.0 3.90 25.4 88.3 69.2 84.8 2.67 30.8 84.8
0 1 0 2 0.3 0.1 0.2 0.3 0.4
8 0 0 7 6 5 71 8 81.9 79.9 74.6 68.2
8 2 7 8 2 7 85 6 83.4 84.1 86.3 83.4
631 388 2 78 7.10 5.28 3.43 2.58
200 235 28 2 18.1 20.1 25.4 31.8
956 913 88 8 98.8 97.1 93.0 82.1
0.1 0.2 0.3 0.4
83.482.0 79.3 84.8 76.0 80.2 72.1 84.1
7.64 4.67 3.32 2.99
16.6 20.7 24.0 27.9
98.8 97.1 88.1 87.5
47 1
Table VI. Effect of Salt-Forming Acid on Amine Flotation of Silica Salt of Technical Tallow Smine Formate
pound/ Short Ton 0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4
Phosphate Concentrate yo Yo HC1Wt. % BPL insolubles 88.6 78.5 5.68 85.4 83.4 3.51 84.4 80.2 3.28 81.8 86.2 2.16 Acetate 90.8 78.5 8.57 8 6 . 1 82.7 4.01 8 4 . 0 ‘82.0 3.14 82.2 84.8 2.05 Glycolate 89.9 77.2 7.24 85.8 7 8 . 5 3.52 84.4 82.7 3.11 80.6 82.0 2.06 8.08 Hydrochloride 0.1 90.7 76.6 4 .24 0.2 86.0 82.7 2.84 0.3 83.9 83.4 0.4 81.9 82.0 2.28 “ Based on calculated heads of 72.4% BPL.
silica Tailing, Wt. Yo 11.4 14.6 15.6 18.2 9.2 13.9 16.0 17.8 10.1 14.2 15.6 19.4 9.3 14.0 16.1 18.1
%
BPLa Recovery 96.0 98.6 93.7 97.3 98.5 98.3 95.0 96.2 95.8 93.1 96.4 91.2 95.8 98.2 96.7 92.8
Based on calculated heads of 69.2% BPI,.
Table V. Effect of Secondary Amines o n Silica Flotation Tallow Amine Aoetate Collector 89.6Yo primary 2 . 3 % secondary 8l.l’?Z0primary 10.2% secondary
47 2’To primary 4 1 . 9 % secondary
Pound/ Short Ton 0.1 0.2 0.3 0.4 0.1 0.2 0 3 0 3 0 4 0.1 0.2 0.2 0.3 0.4 0 3
-phosphate 7% Wt. % RPL 88.7 81.1 85.4 83.4 8 3 . 8 86.3 79.2 82.0 88.3 82.0 84 7 8 2 . 7 84.0 85.6 8’39848 8 2 3 8 2 7 96.0 7 4 . 8 87.8 82.0 87.7 80.2 87.6 84.8 86.7 7 9 . 4 94 2 78 5
Silica % % H C T Tailing, BPL” Wt. % Recovery
insolubles 4.22 2.27 2.09 1.45 4.59 3.10 2.10 233 163 10.83 3.70 3.85 3.65 3.00 8 88
4 8% primary 81 4Yo secondary a Rased on calculated heads of 74.6v0 BPL.
11.3 14.6 16.2 20.8 11.7 15.3 16.0 161 177 4.0 12.2 12.3 12.4 13.3 5 8
96.4 95.7 96.7 86.9 97.1 93.8 96.3 953 913 96.3 96.5 94.3 99.3 92.3 99.2
polar (in this case amino) and nonpolar (hydrocarbon chain) portions, it is to be expected t h a t an optimum balance of the two functional groups would afford maximum metallurgical efficiency. Under the test conditions, optimum results were obtained with the C12 to Cla carbon chains. For longer ( C I S and CIS) or shorter (Clo) chain lengths, t h e polar-nonpolar balance was disturbed, resulting in less efficient collection. Comparisons of this sort can never be strictly accurate, as the efficiency of the collector is related to its ease of dispersibility as well as to t h e ratio of polar to nonpolar groups. The fatty amines of higher molecular weight naturally dispersed less readily than those of lower molecular weight. I n another study (Table IV and Figure 2), the effect of unsaturation was observed with pure CISamines. Figure 2 shows t h a t iodine values between 25 and 100 provided acceptable flotation results. Apparently dispersibility is very important in flotation with fatty amines. Unsaturation, a t least qualitatively, is val-
uable because it makes the amine more soluble and, hence, more readily dispersible. It is for this reason t h a t tallow amines, which are a mixture of ClS and CISamines with an iodine number range of 30 to 60, are of such great importance in t h e beneficiation of phosphate. I t appears from t h e data of Table V and Figure 3 t h a t secondary amines do not contribute materially to flotation, especially in the presence of a strong catior io collector. The efficiency of the amine collector is directly related t o the amount of primary amine present. Salts of tallow amine formed with formic, acetic, glycolic, and hydrochloric acids (Table VI) were equally efficient in flotation. This was observed from the cluster of curves in Figure 4, which indicated each reagent to be identical after i t was dispersed for flotation. It seems t h a t the acid used as an aid to dispersibility may be chosen simply on a n economic basis. The amine acetate curve for Figure 4 is slightly different from the corresponding curve (90% primary) in Figure 3. This difference is attributed to the use of different feedstocks. ACKNOWLEDGMENT
Particular thanks are. expressed to S. R. B. Cooke of the University of Minnesota and to Donald W. Scott of Hibbing, Minn., for valuable advice and suggestions in performing the flotations and studying t h e data obtained. LITERATURE CITED
(1) .4m. Oil Chemists’ SOC., “Official and Tentative Methods,” 2nd ed., No. CD 1-25 (1946). (2) Crago, A,, U. S. Patent 2,293,640(1942). (3) J a c k s o n , J. E., Anal. Chem., 25, 1764 (1953). (4) Snell, F. D., a n d Snell. C. T . , “Colorimetric Methods of Analysis,” 3rd ed., Vol. 11, Van Nostrand, New York, 1949. ( 5 ) T e r r y , D. E., Eilar, K. R., a n d M o e , 0. A,, Anal. Chem., 24, 313 (1952). RECEIVED for review August 6, 1954. AWJEPTDD November 1, 1954. Paper 166, Journal Series, General Mills Research Laboratories.
