INDUSTRIAL A N D E N G I N E E R I N G C H E M I S T R Y
1910
LITERATURE CITED (1) (2)
Baker, W. 0.. and Rdullen, J. W., unpublished (1943), of. (7). Buizov, B. V., German Patent 521,903 (Dec. 20, 1927); J . Applied Chon&.U.S.S.R., 6, 1074 (1933).
Cotten, E. W., and Reynolds, W.E., J . Am. Chem. Soc., in press. Cragg, L. H., J . Colloid Sci., 1. 261 (1946). Fryling, C. F., IND.ENG.CHEM., A N ~ LED., . 16, 1 (1944). (6) Fryling, C. F., U. S. Patent 2,313,233 (March 9, 1943). (7) Fuller, C. S., Bell Sustem Tech. J . , 25, 374 (1946). (8) Graebe, C., and Manu, W., Ber., 15, 1G53 (1882). (9) Hantesch, A., and Freese, H . , Ihid., 28, 3237 (1895). (10) Houston, R. J., Anal. Chem., 20, 49 (1948). (11) Kharasch, M. S., private communication. (12) Kolthoff, I. M., private communication. (3) (4) (5)
(13) (14)
Vol. 42, No. 9
Reynolds, W. B., and Cotteii, E. W., unpublished work. Schulze, W. A., and Crouch, W. W., INI).ENG.CHEM., Qo, (1948)
151
I
(15) Stadler, O., Ber. 17, 2075 (1884). (16) Walker, H. W., U. S. Patent 2,382,684 (Aug. 14, 1945). (17) Wolfe, W. D., Ibid., 2,235,626 (March 18, 1941).
RECEIVED August 22, 1949. This article is taken from portions of the theaea submitted by E. W. Cotten in partiai fulfillment of the requirements for the degrees of master and doctor of science. The work w m carried out during 1944 and 1945 under a contract between the Office of Rubber Reserve, Reconstruction Finance Corporation, and the University of Cincinnati, in connection with the government synthetic rubber program. The preparation of this manusoript has been delayed in part because of Dr. Gotten's untimely death on June 4, 1947.
Milk Proteins and Lactose from Dried Skim Milk SAM R. HOOVER AND ELSIE L. KOKES Eastern Regional Research Laboratory, Philadelphia 18, Pa. T h e feasibility of recovering protein and lactose from dried skim milk has been investigated. Analysis of the solubility relations resulted in dewelopment of a simple process which consists of a four-stage countercurrent leaching of the powdered solid with five times its weight of 0.25% sodium chloride at pH4.1. The solubleextract contains about 14% lactose, the whey salts and added salt, riboflavin, and other minor constituents; 93% of the protein nitrogen is recovered in the extracted solid, which (dried) contains 86% protein, 2% ash, and 0 to 3% lactose. The recovered protein consists of the casein and whey protein, heat-coagulated during the drying process, and so represents essentially the protein of the whole milk.
D
RIED skim milk is a stable pure commercial product which contains 35% protein and 50% lactose. The feasibility of
separating these two major constituents waa investigated. It is obvious that dried skim milk could be dissolved in water and the protein precipitated in a manner analogous to the commercial process of manufacturing casein and lactose from skim milk. At first glance the use of dried rather than liquid skim milk in such a process would seem unreasonable, but in certain circumstances it might be useful. For example, plant operation would not depend on the seasonal surplus of skim milk. Moreover, the dry solid could be dissolved in a solution more concentrated than skim milk, say 20%, and thus reduce by 50% the amount of nater to be evaporated in lactose crystallization. Such factors as these might offset the cost of drying and storing the powder. The experimental results show that the protein recovered from dry skim milk contains both the casein and whey protein. The latter is not recovered in the precipitation of casein from skim milk. Commercial stunples of spray-dried and roller-dried nonfat milk solids were used. Table I gives the composition of the two samples, together with the average composition reported by the American Dry Milk Institute ( 1 ) . The protein content, calculated from the Kjeldahl nitrogen determination with the factor 6.38, which is applied to milk products, is higher than the possible recoverable protein, for skim milk contains nonprotein nitrogen. Menefee el al. (6) have reported that casein accounts for 79.0% of the total nitrogen of milk; the combined globulin and albumin fraction accounts for 11.3%; proteose (nonheat coagulable) 4.2%; and nonprotein nitrogen (nonprecipitable by trichloroacetic acid) 5.6%. The
proteose and nonprotein nitrogen fractions total %S%, and these fractions might be considered the amount nonrecoverable by extraction procedures. Howcver, in this work the recovery of protein wag calculated on the assumption that all but the nonprotein nitrogen is potentially recoverhle-that is, that a yield of 94.4% of the original nitrogen equals 100% recovery. ANALYTICAL METIIODS
Nitrogen was determined essentially by the A.0..4.C. method ( 2 ) , as was lactose in solids or in high voncentratiorir. Dilute lactose solutions were analyzed by the microprocoedure of Stiles, Peterson, and Fred ( 7 ) . Ash ims determined by the ralcium acetate method, with an appropriate correction for the calcium oxide produced (8). Solubility of the protein in borax waq determined by the method ot %Her (8,9). EXPERIMENTAL
PRELIMINARY ExrErmiEtiTs. Iktraction \i ith aqueous alcohol was attempted, but separation of the protein from the lactose was not complete. The best product was obtained by triple extraction of the skim milk powder with 50% ethyl alcohol a t pH 5.2 (acetate buffer). The product contained 12.76% nitrogen (moisture-free basis) and 8.07% ash (calcium acetate). The high ash content together with the necessity of repeated extractions, which would require large volumes of alcohol and consequent recovery problems, caused us to investigate aqueous system. Rmults with alcohol-water extraction were in general agreement a i t h the more detailed study of Leviton ( 4 ) . TABLE I.
SDrsy-dri,ed Roller-dried American Dry Milk Institute ( I )
COMPOSITION O F SKIM AIILK POwUEn Protein .4ah (!X 'i6 . 3 8 ) , Lactose, (CaAc), Water,
%
%
36.4 34.4 36.9
53.5 52.9 51 . O
% 8.4
2.0
8.2
3.0
...
%
...
Fat,
%
... ...
0.9
PRECIPITATION OF PROTEIN FROM SOLUTIONS OF DRIED SKIM MILK.. In a typical experiment, 1500 grams of spray-dried milk powder were dissolved in 6 liters of water, precipitated a t pH 4.1 by stepwise addition of 5 iV hydrochloric acid, and filtered on a Biichner funnel with filter cloth. The solid was stirred with 4 liters of water, and the suspension was adjusted to pH 4.6 and
INDUSTRIAL AND ENGINEERING CHEMISTRY
September 1950
attainment of steady-state operating conditions could be measured readily by analysis of the lactose concentration of each solution. I S 0 YL. a25U NoOL A four-stage countercurrent extraction of a 10% solids suspension was set up in 250-ml.centrifuge bottles; 150ml. of 0.25% sodium chloride were used as the extracting solvent and 2.2ml. of 5 N hydrochloric acid were added to the end bottle, stage 1, just before the addition of 15 grams of skim milk powder. The resultant 14.7 X pH, after the milk powder was stirred in, was 4.1 * 0.05. It was not necessary to adjust the p H in the succeeding stages; the buffer capacity BOLID of the isoelectric protein in the presence of the SOLVENT added salt held the pH constant (4.1). The suspension was stirred well and then centrifuged Figure 1. Countercurrent Extraction of Dry Skim Milk to separate the supernatant liquid, which was passed on to the next solid. Successive extractions were carried out until the system reached a steady state, again filtered. It was washed twiee more with 3 liters of water as determined by the concentration of lactose in solution. Lacat p H 4.6 and dried. The yield was 36.8% on a weight basis. tose concentrations after five batches of solid had passed through The protein contained 14.53% nitrogen (moisture-free basis), the estraction were 6.03,0.98,0.125, and 0.005% for the four 9.00% water, and 1.92% ash (moisture-free basis), and accounted stages of the extraction. The total nitrogen content of the for 91% of the original protein. The lactose concentrations in sixth, seventh, and eighth protein preparations removed were the first three solutions were 10.1,3.9,and 1.5%; recovery was 13.93, 13.96,and 13.93% (moisture-free bwis); this confirmed 88% of the total lactose. that steady-state operation had been achieved. The region of minimum swelling and optimum filterabilitJi of These results indicated that the countercurrent extraction was casein precipitates corresponds closely with the isoelectric point. The isoelectric point depends on the salt content, as was shown applicable. Because the lactose concentration of the solution many years ago (6'). Because of the soluble salts, the isoelectric was not bo high as desired, extraction of a 20% initial solids suspoint of casein in whey is 4.1; in the absence of salt it is 4.6. pension was carried out. Figure 1 illustrates and Tables I1 and An important experimental observation is that casein precipiI11 summarize the results of this experiment. A 14.7% lactose tated at pH 4.1,when washed with distilled water, becomes successively more acid, swells, and final1 goes into solution. The solution was produced, and a 35% yield of whole milk protein authors found that 0.25% sodium c6oride in the wash water containing 13.5% nitrogen or 86% protein was obtained. The prevents this pH shift, and the entire process of precipitation repeated 'manipulations required for the successive extraction of and washing can be carried out a t a constant pH of 4.1. the small samples in this type model experiment are responsible A similar experiment on roller-dried powder gave a 36.0% yield of protein on a weight basis. The protein contained 14.6370 for the lower yield of protein, for nitrogen determinations on the nitrogen (moisture-free basis), 9.08% water, and 1.87% ash supernatant lactose solution showed that only 12% of the total (moisture-free basis), and accounted for 94% of the total protein. nitrogen and therefofe only 6.4% of the protein nitrogen was in the extract. The four-stage extraction did not completely reThe amount of protein recovered in these experiments was move the lactose. A larger number of stages could be used or greater than the casein content of the skim milk powder. It is continuous extraction carried out, if this amount of lactose in the apparent that the whey proteins were largely denatured in the product is undesirable. original drying process and are insoluble. Further confirmation of this result was obtained when the borax sohbility of the recovered protein was determined. It was 81% soluble, which corresponds fairly well with the proportion of casein to the total T a B L E 11. LACTOSE CONTENT O F SOLUTIONS IN COUNTERprotein of the skim milk. CURRENT EXTRACTION OF SKIMMILKPOWDER These experiments showed that it is feasible to recover whole (Thirty gram8 of air-dry powder in 150 ml. of 0 2 5 7 sodium chloride conmilk protein from skim milk powder by a process similar to the taining 4.4 ml. of 6 N hydroclilbric &id) commercial casein precipitation process. Moreover, the essenLactose Content, Grams/100 M1. Batch No. of tial conditions for developing a countercurrent extraction process Extracted Solid Stage 1 ,Stage 2 Stage 3 Stage 4 were established. 5 13.85 4.50 1.20 0.24 6 14.75 3.80 1.31 0.28 COUNTERCURRENT EXTRACTION. If the initial adjustment of 7 14.70 3.45 0.86 0.31 pH to 4.1 could be maintained by the use of dilute salt solution 8 14.80 3.00 0.16 9 14.70 3.15 0175 0.26 as the extraction solvent, the extraction of successive batches of skim milk powder could be carried out in a countercurrent system. The advantages of such a process lie primarily in recovery TABLE 111. PROTEIN RECOVERED FROM SKIMMILKPOWDER BY of the lactose fraction, for a more concentrated extract would be COUNTERCURRENT EXTRACTION produced, lactose would be completely recovered, and the waste (Thi?ty grams of air-dry powder in 150 ml. of 0.25% sodium chloride condisposal problem caused by the dilute wash waters would be taining 4.4 ml. of 5 N hydrochloric acid) eliminated. Protein Batch No. of Yield (N X 8.38). Ashb, Lactose, I n countercurrent extraction, the solid is passed successively Extracted Solids' Grams % % % % through solvent which is passing in the opposite direction. Thus 7 10.0 34.8 88.5 2.19 2.80 the solution which extracts the fresh solid already contains an 8 10.2 35.4 87.0 2.30 9 10.5 36.4 86.5 2.30 3: 37 appreciable amount of solubles, and the solid which has been extracted by successively weaker solutions is finally washed by ' All results on moisture-free basis. b Ash was determined by the calcium acetate method, protein contains fresh solvent just before it is removed from the system. Such approximstely 80% casein which has an ash content of 1.85% hecause of ita combined phosphorus. Therefore about 1.6% of the determined ash is countercurrent extraction can be a continuous process or it can cons2itutional; remaining ash content of about 0.8% may be called "true be done in separate stages. In the experimental system deash. scribed below, a batch-type extraction was used because the RAW
MATERIAL
f
1911
pH
AOJUSTMENT
SOLVENT
INDUSTRIAL A N D E N G I N E E R I N G CHEMISTRY
1912
TABLEIV. STEADY-STATE COUNTERCURRENT EXTRACTION OF 3GGRAM SAMPLES
OF
ROLLER-DRIED SKIM MILK POWDER
(A = phosphoric acid-phosphate solution: 150 ml. of 0.25% aodium dihydrogen phosphate 20 ml. of N phosphoric acid. B = sulfuric acid-sulfate aolution, 156 ml. of 0.25% sodium sulfate: 20 ml. of N sulfuric acid) Solution, % ’ Nitroaen, Extrac% of tion Lactose Nitrogen Original 0.13 12 A 10.7 11 0.12 B 10.5 0 ReRults on moisture-free basis.
Solid“, % ________ Protein ( N X 6 . 3 8 ) Ash 3.19 87.3 78.6 4.57
Lactose 0.8 5.3
Phosphate and sulfate were nest tested as anions in this process, for hydrochloric acid is more corrosive and the residual chloride ion content of the recovered protein and of the lactose solution might be undesirable in certain applications. Steadystate operation was obtained after five batches of solid had been extracted. Eight further extractions were run and the samples pooled for analysis. The results (Table IV) are in general agreement with those presented earlier. The lactose concentration was lower than in the previous experiments because a larger volume of extracting solvent wits used. The ash content of the solid from extraction B was higher and the protein content lower, in part caused by the insolubility of calcium sulfate. I n these experiments the recovery of protein was good. The loss of 12 and 11% of the total nitrogen in the extract indicates a loss of about 6% of protein nitrogen, or conversely a recovery of 93 to 94y0of the protein nitrogen if mechanical losses are disregarded.
Vol. 42, No. 9
would appear to be relatively simple to operate on a commercial scale, and the cost of chemicals would be low. The product, the whole protein of milk, would be unique. Such a protein is not now available on the market. The lactose extract might be used directly in fermentation or in food products, or it might be crystallized by the conventional proceM. The economic aspects of such a process depend on the complicated price structure of the dairy industry and on future trends, which are difficult to predict ( 8 ) . Nonfat milk solids are stable currently at the government support level of 11 cents a pound for the roller-dried and 12.25 cents a pound for the spray-dried product. Casein is 21 cents a pound; lactose is 17 cents a pound for the crude product and 26 cents a pound for the U.S.P. grade. The recovery of lactose by crystallization from casein whey ranges from 60 to 70y0 to over 90%; the authors have not been able to carry out experiments on the extracts to determine what yields could be expected in large scale operation. It is evident that the commercial application of this process would depend to a large extent on the price and market for the whole milk protein. LITERATURE CITED
American Dry Milk Institute, Inc., Chicago, Ill., Bull. 911 (revised, 1948). Association of Official Agricultural Chemists, Washington, D. C.. “Official and Tentative Methods of Analysis.” 5th ed., 1945.
Cook and Day, “The Dry Milk Industry,” Chicago, Ill., American Dry Milk Institute, 1947. Leviton, IND. ENG.CXEM.,41, 1351 (1949). Menefee, Overman, and Tracy, J. Dairy Sci., 24, 953 (1941). Sdrenson and Sladek, Compt. rend. trav. lab. Carlsberp, 17, No. 14 (1929).
DISCUSSION
Rerovery of whole milk protein from dried skim milk cam be carried out in a batch or countercurrent estraction process; the conditions for this have been worked out on a laboratory scale. However, pilot plant study of engineering aspects will be required before operating costs can be determined. The process
Stiles, Peterson, and Fred, J . Bact., 12, 427 (1926). Sutermeister and Browne, “Casein and Its Industrial Applications,” New York, Reinhold Publishing Corp., 1939. Zoller, IND.ENQ.CXEM.,12, 1171 (1920). RECEIVED November 25, 1949. Presented before the Division of Agricultural and Food Chemistry at the 116th Meeting, AXEMCANCHEMICAL Socrmr, -4tlantic City, N . J.
n-Butyl Oleate from n-Butvl Alcohol and Oleic Acid J
J
DONALD F. OTHMER AND SANJEEV ANANDA RAOl Polytechnic Institute of Brooklyn, Brooklyn, N . Y .
A
CCURATE kinetic data on the esterification reaction are needed for the design and successful operation of industrial esterification units. Requirements are: ( a ) order and ( b ) rate of the reaction; ( c ) an over-all mathematical expression of the variation of the rate with temperature, catalyst concentration, and proportions of reactants; ( d ) equilibrium constants; and (e) variation of density of the various components with temperature. Previous studies have been made to show the utility of such data in application to production of butyl acetate (7) and dibutyl phthalate ( 1 ) . A number of aliphatic nsters of oleic acid have been investigated, and it was desired to supply data for t,he commercially important, n-butyl oleate, which is prepared from oleic acid and n-butyl alcohol catalyzed by sulfuric acid. REACTION OF BUTYL ALCOHOL AND SULFURIC ACID
Earlier investigations show that the reaction of butyl alcohol and sulfuric acid gives increasing amounts of butylsulfuric acid I
Present address, The Oliio Oil Compnny. Robinson, Ill.
at elevated temperatures ( 1 , 7). Two per cent sulfuric acid in butyl alcohol was used over a reaction temperature of 0’ t o 115’ C., and had an apparent, milliequivalent value of 94,which has been used in this investigation for the catalyst concentration factor. MATERIALS
OLEICACID. Crude oleic acid as received from the Nopco Chemical Company had the following analysis: acid value, 190; iodine value, 90; saponification value, 200; unsaponified, 3%. The methyl ester was prepared using sulfuric acid (0.03 mole %) as catalyst. The ester was fractionally distilled at 10 mm.; the fraction boiling a t 190’ C. was collected and redistilled at 10 mm. of pressure four times. The methyl ester was hydrolyzed, and the oleic acid was removed by distillation a t 100 mm. of The oleic acid thus removed was purified four times g%%lation a t 100 mm. of remure. The urified oleic acid was titrated with standard alka!, and the mofxular weight determined to be 290.092. The molecular weight of a sample of C.P. oleic acid supplied by the Amend Drug Company was .determined to be 289.802,which checked with that of pure oleio acid
.
