Defecation of Refinery Sirups PHOSPHORIC ACID AND LIME AS PIPE LINE DEFECANrS GORDON G. HALVORSEN AND A. R. BOLLAERT The Dicalite Company, 120 Wall S t r e e t , New York, N . Y .
Pressure filtration of phosphoric acid-lime defecated refinery sirups using diatomaceous filter aids is now possible with pipe line defecation. Acid and lime are introduced in the pipe line between the pump and filter, and thus form the floc beyond the point where it would be damaged. Addition of phosphoric ‘acid first, followed by the lime slurry, shows best results on cycle lengths, clarity of filtrate, color, and lime salt removal. Previously pressure filtration of the phosphate floc has met with little success owing to the comminution of the floc by centrifugal pump action. Poor-clarity liquor and decreased flow rates have been indicated by refinery and laboratory tests where the floc passes through the pump. Selection of filter aid for percentage of defecant is discussed. Past and recent investigations on phosphate defecation have shown the desirability of using this treatment; however, in the modern refinery this has been accomplished only by considerable initial outlay of equipment. The method described requires little change in filter station layout.
T
HE introduction in 1914-15 of diatomaceous filter aids and pressure filters replaced the bag filter and opened a new era in the clarification of cane sugar sirups. Phosphoric acid-defecated sirups could not be successfully filtered with the filter aid available a t that time; consequently, this defecant was discontinued in favor of diatomaceous earth clarification. More uniform processing resulted, together with cleaner and more compact filter stations. I n recent years the trend has been toward the return of phosphate defecation (6). The increased load placed on the char house by difficultly refined raw sugars has caused the refiner to investigate new processes. Phosphate defecation offers a solution for these sugars, provided the liquors can be readily filtered. The interest of refiners in phosphate
defecation has been prompted by the scarcity and high cost of bone char plus the increasing prevalence of difficulty in refining raw sugars. Progress in the use of phosphate defecation has been madefor example, the use of Dicalite Speedex in combination with phosphoric acid and lime on pressure filters and the Williamson and Jacobs clarifying systems. The advantages and disadvantages have been shown by Brown and Bemis (1). Other methods which have been reported successful are filtration of the phosphate floc (4) at low pressures (5-7 pounds) and filtration with cord filters operated by vacuum (8). Many investigators have shown the advantages of the phosphoric acid-lime treatment in the removal of coloring matter and the elimination of gums, albumins, and other colloids. Spencer (6,7) states: “Experience hasshown that between 20 and 40 per cent of the total color of the melt liquor is removed by phosphoric acid, which means that the boneblack required for decolorization will be from 40-60 per cent less than with untreated liquors.” Phosphoric acid-lime treatment removes up to 40% coloring material and 25% of the ash constituents (1). Use of diatomaceous filter aids in conjunction with phosphoric acid and lime has been proposed ( 3 , 4 )and used with varying degrees of success. Laboratory tests using air pressure as the driving force give excellent results; however, high-speed centrifugal pumps have exceedingly deleterious effects upbn the floc. Poorclarity liquors and short cycles result from the breakdown of the floc by centrifugal pumps. The effectof centrifugal pump action has been demonstrated in several refinery tests and is shown in this article together with centrifugal pump tests made in the laboratory. The success of filtering phosphate-defecated liquors depends largely upon the avoidance of breakdown of the calcium phosphate floc through mechanical action. Recent laboratory tests have shown that it is possible to introduce the acid and lime be-
,
Compressed Air
e Graduated Cylinder for Acid or Lime Addition
Manometer for
High Speed Centrifugol Pump
Figure 1. LabaPatory Setup for Phosphate Floc Filtration Tests
385
386
INDUSTRIAL AND ENGINEERING CHEMISTRY
Standard Turbidity Block - 2 0 m.a. -
L-_ _ 1-
__100
95
Figure 2 .
90 Percent G l o r i f y
-
85
80
Clarity Curve for Washed Itair Sugar 281
yond tlic pump uncl thus iicfocatc in tlir pipe line in.n.hich the unfiltered liquor AOTYS to thv presses. Tlic avnilability of scvcinl types of metering aiicl propoationiiig devices and also pH control syAterns make it possiblo for the rc4nc.r to install this mctliod of treatment n-iih lit tlc clinngc, i n picqs rooin layout. FILTILiTIOS STUDlE4
A stainleas steel ~,l:Ltc-and-Eraiiic filter press \van u x c l which had a filtering area of 77.8 kquaro inchcs. Lime or acid 11 in the line aftcr the pump, or tlic floc was puinpod tlir pump, dcpending upon the tests being run. Tlic: conlact time bctwccn chemical trcatmcnt and actual filtration \vas 7 ininiitk for all tests, since this time period is comparable to rcfincry operation. Earlier lalmratory tests had slion-n t h a t tlic color removal is practically iii.;t:int~ncoux. There was no o b x r v c d difference in color reinoval o\-w R wide rangc of contact. tinic (10 scconds to 2-10 minutcb). I n all ii:sis washcd raw sugai' liquor (00" Brix) m ~ used. s Prccoatk ol' 10 pounds per 100 square lect \\-ere applied at 1.0 po~iiidp c Squ:iro ~ inch prcwurc. lion tcmpernturc was iiiaiiLtaiiict1 a t 176" F. for all t filtcr rloths (Filter lletlia G ~ r p o i ~ : i ~ i oSno,. 07-1-01)), \~arlictIE m ; of starinh, wcre u c d for oacli test. I ~ i g u i ~1c is a. Iliagrain of tlici 1nl)oi:iiory setup u.sd for filtration t - i a i \ t vatc filtrations w r v niacic in all t w t s a t 100 ml. p ~ ininutv. r
Distilled W a t e r
0
10
ZQ
- 100rnicroamps.
30 4 0 50 60 70 80 90 Percent Bleach or Color Removal Figure 3 . ('olcrr Rrmoiiil C u r i e for % a u h c d Kow Siligtic 2111
LOO
Vol, 38, No. 4
INDUSTRIAL A N D ENGINEERING CHEMISTRY
April, 1946
centrifugal pump. The same t:ffects were apparent with 0.010/0Pz06and 0.3% . - SDeedex. TableIgives results of the tests. Since the refiner was unable t o use phosphate treatment as a result of the poor clarification and short cycles obtained when the. floe passed through the Rump, this study was made t o determine (a) filterability of floc when formed in the line between pump and filter, ( b ) effect of the order of chemical addition, and (c) choice of filter aid for percentage of P20sused. PHOSPHATE FLOC TESTS
FLOC THROUGH PUMP.Tests were first made to establish whether the laboratory e q u i p ment gave results comparable to those in the refinery. Laboratory tests (Table 11-A) on 0.5% Speedex and 0.02% P205 show the drop in clarity as the pressure increases. This is more pronounced when 0.01% P2Os is used with 0.3% Speedex. Tests were also made using Special Speedflow which has a higher clarifying capacity and a flow rate 40% of Speedex.
FORMATION OF FLOC IN LINE PUMPAND FILTER.
BETWEEN
387
TABLE 11. EFFECT OF PLACE AND ORDER OF ADDINQ CHEMICALS ON CYCLE LENGTH AND CLARITY OF WASHED RAWSUGAR 281
,..
Stan$ PsOS, 0.5% Speedex; Speedex I50 precoat o . O l ~ o,P*OS, 0.5% Special Speedflow, Special Spee'dflow precoat 175
Sugar 298 Cycle Flow, length, % of .kv.,% min. standard clarity
color removal,
70
100
97.9
Sone
325
100
07.7
None
78
99.3
40
235
72
99.2
52
91
99.4
28
270
83
99.4
39
April, 1946
389
INDUSTRIAL AND ENGINEERING CHEMISTRY ACKNOWLEDGMENT
The cooperation of the operating and technical staff of the Western Sugar Refinery on plant tests is appreciated. Many persons in the laboratory of The Dicalite Company aided in these tests; t o them thanks are extended for valuable assistance. LITERATURE CITED
(1) Brown, J. M., and Bemis, W. A,, IND.ENG.CHEW,34, 419-22
(1942). (2) Brown, C. A., and Zerban, F. W., "Physical and Chemical
Methods of Sugar Analysis", 3rd ed., p. 1075, New York, John Wiley & Sona, 1941. cHnM., 34, 398-402 (1942). (3) cummins, A , B., (4) Frankenhoff, C. A., Ibid., 34, 742 (1942). ( 5 ) Knowles, H. I., Ibid., 34, 422-4 (1942). S W G r Cane Tech., Baton Rouge, (6) Meade, h o c . 6th C O W T . Intern. 1938, 1032.
(7) Spencer, G. L., and Meade, G. P., Cane Sugar Handbook, 8th ed., p. 297, New York, John W h y & Sons, 1945. (8) Wright, Arthur, MD.ENG.CHEM.,34, 425-9 (1942). PRESENTED on t h e program of t h e Division of Sugar Chemistry a n d Technology of the 1945 Meeting-in-Print, AMERICAN CHEMICAL SOCILTY.
Vapor-Liquid Equilibria in Three Hydrogen-Paraffin Systems d
M. R. DEAN AND J. W. TOOISE T h e solubilities of hydrogen in isobutane were determined for temperatures from 100' to 250' F. and pressures from 500 to 3000 pounds per square inch. Solubilities in 2,2,4trimethylpentane were found for temperatures from 100" to 302.5" and in a mixture of isomeric dodecanes for 200' and 300" F. with pressures ranging from 500 to 5000 pounds per square inch. The compositions of the equilibrium vapor phases were also determined. The solubility of hydrogen increases with temperature and pressure but decreases as the solvent molecular weight inareases. The solubility of hydrogen follows Henry's law only in isobutane at 150' F. and lower. The hydrogen solubilities in. the two heavier hydrocarbons increase more rapidly with pressure at low pressures than at high pressures. Correlation with literature data shows that hydrogen is more soluble in paraffins than in aromatics of similar molecular weight. Vaporization equilibrium constants are computed from the data for both solvent and solute. The constants vary widely with pressure and to a lesser extent with temperature. The constant for hydrogen increases with an increase in solvent mQlecularweight.
T
HE solubility of hydrogen a t high pressures in several pure hydrocarbons was reported by Frolich (8)for 77" F. and up to 100 atmospheres, by Ipatieff (3) at higher temperatures for several petroleum fractions, and by Ipatieff (4) i n several pure aromatic hydrocarbons. More recently solubilities to 106 atmospheres in n-butane were reported by Nelson (7). I n none of these investigations was a comprehensive study of the composition of the equilibrium vapor phase made. Kelson, however, did present some data on the vapor phase. The purpose of this work was t o determine the solubility of hydrogen in a narrow-boiling mixture of isomeric dodecanes and in two relatively pure hydrocarbons-isobutane and 2,2,4-trimethylpentane-for a range of temperature and pressure. I n addition, the compositions of the equilibrium vapor phases were to be found, and a study made of the vaporization equilibrium constants of the components and the effect of temperature and pressure on these values. The hydrogen used was the commercial electrolytic waterpumped grade. The impurity, oxygen, was less than 0.2%. R a t e r vapor was removed by contacting with anhydrous calcium sulfate. The isobutane, with a tested purity of 99.5%,
Phillips P e t r o l e u m Company, Bartlesville, Okla. was obtained from the Phillips Petroleum Company, The major portion of the impurity was n-butane. The iso-octane, obtained from Rohm & Haas Company, was considered to be essentially pure 2,2,4-trimethylpentane. The physical properties were determined by the National Bureau of Standards: p ~ ; ~ i ~ g FSp. Gr.,
.
2,2,4-Trimethylpentane Isomerio dodecane mixt.
210.5; 350-2
d:" 0.691~a 0.756
Refractive Index, n v
Mol. Weight
1.3915a 1.4227C
114;22b 170
Determined by National Bureau of Standards. C Experimental.
b Calculated.
The mixture of isomeric dodecanes was specially prepared for this work by polymerizing isobutylene, hydrogenating the resulting product, and fractionating in a 1/2-inch i.d. column, 36 inches long and packed with l/a-inch glass helices. A cut boiling between 350" and 352' F. was collected for use in this work. The particular combination of polymerization and fractionation was believed t o have yielded a product which was a mixture of isomeric dodecanes. This final product will be referred t o as dodecanes. The determined physical properties of the dodecanes are also listed in the table. The molecular weight was obtained by a method utilizing the principle of the freezing-point lowering of benzene. Vapor pressures of the hydrocarbons are useful in the correlations of experimental data. The vapor pressures of isobutane and 2,2,4-trimethylpentane were determined up to their critical temperatures by Sage (8) and Smith (Q), rekpectively. Since comparable data for the dodecanes were lacking, the vapor pressure curve was determined t o 300" F. The apparatus used was based on the principle of balancing a column of mercury against the vapor pressure of the air-free hydrocarbon confined in a Utube over mercury. The technique was developed by measuring the vapor pressure of our iso-octane and comparing results with values by Smith. Bridgeman ( I ) described this type of apparatus in detail. The dodecanes vapor pressures are believed to be in error by no more than t 0 . 0 2 pound per square inch a t the higher temperatures and less a t the lower temperatures: Temp., O F. 32.0 77.0 173.7 200.0
Vapor Pressure Lb./Sq. In. Abs.' 0.04 0.11 0.67 1.04
Temp., ' F. 228.7 233.6 297.1 300.0
Vapor Pressure, Lb./Sq. In. Abs. 2.00 2.15 6.49 6.52
