e
e XI. D. S..IXUDERS, Sac&
T h e B.O.D. of meat packing wastes can be reduced 60 to 889" by primary treatment followed by precipitation with a mixture of sulfuric acid and ferric sulfate at pH 4.8. A low analysis protein concentrate, usable as animal feed, can be recovered by coagulating the precipitates wTith open steam and drying. The product is a mixture of protein and fat.
T
HE liquid wastes from a modern meat packing plant may
contain from 500 to 2000 p.p.m. of 5-day biochemical oxygen demand (B.O.D.), resulting primarily from the protein fat, and partly digested forage in the waste. Floating fat is salvaged by flotation for industrial uses and fibrous niaterial may be removed by screening or sedimentation ( 2 ) . These operations may reduce the B.O.D. by 10 to 40%. The remaining organic matter consists of emulsified fats and nitrogenous material, chiefly protein. Chemical precipitation of the clarified waste offers the possibility of reclaiming these components in a usable form, while reducing the polluting strength of the waste. Halvorsen and associates ( 1 ) treated meat packing n-aste with chlorine and recovered a tankage containing 4570 or more of protein, \\-hile reducing the B.O.D. 80 to 85%. PRECIPITATION
The process here described depends on the precipitation of proteins and emulsified fats a t a pH of 4.0 to 5.2 and the coagulation of this precipitate with ferric or aluminum salts (3). The optimum pH is 4.8 for maximum removal of organic matter. The most practical reagents are salfuric acid and ferric sulfate because they can be applied in the same solution. Depending on the natural alkalinity of the plant water supply and the buffering effect of the organic matter in the waste, about 20 p.p.ni. cf ferric iron and 200 p.p.m. of sulfuric acid are required t o develop a pH of 4.8 and produce satisfactory precipitation. Figure 1 is a flow sheet of the process; it does not include preliminary screening or sedimentation of the waste. Anhydrous ferric sulfate is dissolved in water to make a 5%1, solution. Sulfuric acid, 60" Be., is added to this solution in the ratio of 3 pounds of 60" BB. acid per pound of ferric sulfate. The mixed
& Company, Chicago, III.
solution is further diluted 3 to 4 times with water. A Leeds & Northrup pH recorder-controller governs the addition of the solution to the waste in the air-agitated mixing chamber. One cubic foot of air per 20 gallons of waste provides adequate mixing. The calomel-antimony electrode assembly has proved entirely satisfactory with routine maintenance; this consists of \Tiping the antimony bar twice daily and renewal of the calomel electrode once T%-eekly. Detention time in the mixer averages 5 minutes. The waste next passes to the flocculator, with a retention time of 15 minutes, and a peripheral paddle speed of 1 foot per second. Final sedimentation is carried out in the usual mechanically cleaned and skimmed tank, with a 2-hour detention time. The precipitate is allowed to concentrate in the tank hopper for 12 to 24 hours prior to removal. 4 small quantity of fat is recovered in the scum trough of the final settling tank, as a result of acidification of the m-aste. PRODUCT RECOVERY
The precipitate can be processed into a low analysis protein concentrate (4). It is withdrawn from the settling tank and heated with open steam to 180' to 200 F. Coagulation occurs and the solids collect in a pasty layer a t the surface of the liquid. A limited fermentation in the settling tank hoppers has been found necessary t o accomplish this concentration, as the released gas carries the coagulated solids to the surface on heating. The vater which is separated is decanted to the plant influent for retreatment, and the concentrate, containing 8% or more solids, is further de>-ateredby vacuum filtration or by gravity draining to 75 to 85% moisture content. Drying is accomplished in a horizontal steam-jacketed cylindrical dryer with paddle agitator. The dried product contains 35 to 40% protein if primary sedimentation has been efficiently carried out, but under 30% protein if primary treatment is limited to fine screening. The ether-soluble content of the dried product is always high; it averages 25y0 or more because the emulsified fats are precipitated with the protein. This product is suitable for use as an ingredient of aninial feed, where a OW protein material is needed in forniulating animal protein feeds to a specific protein content. Its nutritive value has been determined as equal, on a protein basis, to other animal protein feeds. Hogs and cattle, on a fattening ration, showed normal gains when this product was used in an amount not to exceed 37, of the total feed. OXYGES DEMAND REMOVAL
I0
,-CHEMICAL
~
*pH
The removal of oxygen demand by primary sedimentation for 0.5 hour, followed by chemical precipitation according to the foregoing procedure is illustrated in Table I, which covers two 3-day periods of operation. The products of protein hydrolysis which are present in the tank water developed in the steam rendering of fats are
FEED TANK
CONTROLLER
'
--cVFiUlNT
SEDIMENTATION TANK
nitrogenous products of bacterial decomposition. The results shown crere obtained while treating fresh waste from a meat packing plant slaughtering all species, and process-
INDUSTRIAL AND ENGINEERING CHEMISTRY
1152
Biochemical Oxygen Demand, P. P. 1'1. Reduction. % Primary Chemical Primary Raw waste effluent effluent effluent Over-all 76.8 484 11.6 1832 2076
35
78.2
16
84.5
921
601
199
1649
1384
255
780
603
192
36
75
1160
840
130
27.6
88.5
E65
396
269
40
8 . 6
TABLE 11.
TREBTMENT O F
R a t e Applied 12 Hr. Each Day. Millions of Gallons per Acre
precipitation applied to a pilot plant trickling filter 6 feet deep by 4 feet in diameter. The pH of the applied waste varied from 4.8 to 5.0. The first four and the last four sets of figures each represent consecutive days. Type of Slaughter Hogs a t plant capacity Cattle a t plant capacity Cattle and hogs a t p!ant capacity Cattle and ho.3 at plant capacicy Cattle below capacity Cattle and h o g s below capacity
EFFLUENT ON TRICKLING FILTER
B.O.D. Applied per Acre-Foot/ Day, Lb.
B.o.D.* P.P.nI* Influent
Vol. 40, No. 6
Effluent
B.O.D. Removal,
73
CONCLUSION
-1 process has been described which will produce a major reduction in the strength of meat packing wastes, while reclaiming a usable product. Unfortunately, the value of the product pays only a small part of the maintenance, operating, and fixed expenses. Attempts t o separate the fat from the protein in the final product have not been successful, and neither has separate reclamation of the fat been accomplished prior to precipitation. If the fat and protein could be recovered separately, a much brighter economic picture might be presented, as the protein, freed of fat, would be m-orth more, and the fat would have a value exceeding that of the present product. LITERATURE CITED
(1) Halvorsen, H. O., Cade, A. R., and Fullen, W. J., Sewage W o r k s J . , 3, 4 8 8 (1931). (2) Mortenson, E. X,, Proc. of Second Industrial Waste Conference, P u r d u e Univ. Eng. Bull. 60, 28 ( 1 9 4 7 ) . (3) Banders, M. D., U. S,Patent 2,204,703 (June 18, 1 9 4 0 ) . (4) I b i d . , 2,277,718 (March 31, 1942).
The effluent from this type of chemical precipitation can be successfully treated on a standard type trickling filter. Table I1 gives the results of application of the effluent from chemical
RECEIVED April 12, 1947. Presented before the Division of Water, Sexape, and Sanitation Chemistry at the 111th 3Ieeting of the . h E R I c A X CHEMICAL SOCIETY, Atlantic City, S . J.
D. R. TORGESON Paci$c Experiment Station, C . S . Bureau of Mines, Berkeley, Calif.
T
HE nork reported in this paper m s prompted by requests from industry for data needed in conducting heat-balance calculations for the process of calcining trona t o soda ash. Such calculations have been hampered mainly by lack of a suitable heat of formation value for trona. From heat of solution measurements involving trona (NazCO3.NaHCO3.2HZ0),sodium carbonate, and sodium bicarbonate, the heat of formation of trona from its constituent compounds now has been determined and the heat of calcination evaluated. MATERIALS
Three samples of trona were employed: Sample A, natural trona, was selected massive crystals from Marston, Wyoming, furnished by Westvaco Chlorine Products Corporation through the courtesy of J. A. Robertson. The crystal surfaces were scraped clean before crushing to -60 mesh. Analysis shoRed this material to be highly pure. Titration with standard hydrochloric acid gave 99.94% of the theoretical value for trona; the ignition loss a t 290" C. was 99.87% of the theoretical value; no chloride was present; and the water-insoluble residue was only 0.05%. The latter consisted of hairlike stringers presumably of clay. Sample B, synthetic trona, was prepared as follows: To 1000 grams of hot water there were added slowly, while stirring, 150 grams of sodium carbonate, 60 grams of sodium bicarbonate, and 200 grams of sodium chloride. Heating and stirring were continued until the solution became clear; the final temperature was about 100' C. The solution then was cooled slowlv, over a
period of 30 hours, to an end temperature of 35' C.. to permit formation of trona crystals. The crystals mere removed, drained, washed lightly with 95Yc ethanol, and dried for several hours a t room temperature in a stream of tank carbon dioxide; drying was completed after crushing to -60 mesh. This process yielded material containing sodium chloride and sodium carbonate as impurities which resulted from incomplete drainage of mother liquor and incomplete washing. Attempts a t more thorough washing were not too successful because almost invariably some sodium carbonate was leached from the trona itself; this left an indefinite product. I t was consideied better, therefore, to permit the sodium chloride and carbonate to remain and make correction for them in the heat of solution results. Analysis of sample B indicated 97.72% trona, 1.75% sodium chloride, and 0.5391, sodium carbonate monohydrate. Sample C also was synthetic trona. I t was prepared like sample B except that freeing of the crystals from mother liquor impurities was less complete. Analysis gave 95.66% trona, 2.37% sodium chloride, and 1 . 9 7 5 sodium carbonate monohydrate. Reagent grade sodium bicarbonate was treated n-ith tank carbon dioxide for several hours at room temperature. Titration with standard hydrochloric acid showed virtually 100% purity. Sodium carbonate m s prepared from reagent grade sodium bicarbonate by heating a t 290' C. to constant weight. This material was employed in the heat of solution measurements and also served as the basis for standardizing the hydrochloric acid used in titrating the trona and sodium bicarbonate samples.
