T H E PRECIPITATION OF PROTEINS I N PACKING HOUSE WASTES BY SUPER-CHLORINATION
w. J . FULLEN** With the present increasing agitation against stream pollution, civic organizations, municipalities, and industries are more than ever faced with the problem of satisfactorily treating industrial wastes. Unfortunately, the biological methods that have been so successful for domestic sewage are not always applicable. The high strength of these wastes frequently makes the cost of these methods entirely excessive, and the chemicals present often interfere with biological activity. Even when it is possible to mix the two the results are frequently unsatisfactory. I n facing this problem, sanitary engineers are therefore being forced to reconsider the almost discarded chemical precipitation method of sewage treatment. Geo. A. Hormel & Co., packers a t Austin, Minnesota, early arrived at the decision that biological methods were out of the question, and instituted a program of research to find the method of chemical precipitation most applicable to their waste. I n this connection, some of the underlying fundamentals have been investigated a t the Department of Bacteriology, University of Minnesota, and at their own laboratories in Austin. In order to relieve pollution in the Cedar river as much as possible, the Hormel company, several years ago, provided for preliminary treatment such as the removal of the paunch manure by suitable screening, and the removal of easily settleable solids by primary sedimentation. Waste from the stock yards was collected and used as fertilizer on neighboring farms, and domestic sewage coming from the plant was put directly into the city sewers. These modifications, however, did not materially relieve pollution in the river. The by-passing of all condenser and other clean water served to reduce the volume of the sewage in preparation for chemical treatment. With these provisions, the Cedar river received about three-fourths million gallons of packing house waste that contained approximately z 500 parts per million of volatile solids and that had a B.O.D. of about 1800. From 75 to 80% of these solids were in colloidal suspension and could not be removed by further sedimentation. The balance appeared to be in true solution. The object of the investigation was to find ways and means of removing all of the former and as much of the latter as possible. Various methods of chemical precipitation as a means of sewage clarification have been studied in the past, but none of them have been extensively applied in this country. In this connection there are various principles that BY H. 0.HALVORSON, A. R. CADE* AND
I 86
H. 0.HALVORSON, A. R. CADE AND W. J . FULLEN
may be employed. The proteins may be precipitated by adjusting the pH to the isoelectric point, such as in the Miles Acid Process,l or they may be precipitated by neutralizing the charge with the salts of various metals, such as iron or aluminum. It may be argued that this is not exactly a charge neutralization, but that the protein salts of the heavy metals are formed.. This is being investigated a t this time, but results are not as yet ready for publication. Proteins may also be precipitated by denaturing them with oxidizing agents or other coagulants. The Miles Acid Process was eliminated because good results can be obtained only in case the pH is very accurately controlled. This cannot be accomplished by manual methods in a plant that operates on a continuous flow, and since no fool-proof automatic methods have as yet been devised, it was felt that this process could not be relied upon for uniformly good results. Likewise, coagulation with aluminum or iron salts was rejected because it was felt that with manual control it would be difficult to avoid excessive ash content in the recovered product. Of the coagulants in the third class, chlorine appeared to be the most promising from an economic standpoint. That this chemical could be used for the precipitation of proteins was pointed out as early as 1840by Mulder,2 Thenard,3 Berzelius,&and DeVrij.5 I n 1897 Rideal and Stewart6 advocated the use of chlorine for the precipitation and quantitative determination of gelatin and'peptone in meat extracts. The authors stated that the precipitate so obtained flocced and filtered readily and was quantitatively weighable. According to them, the percentage of gelatin found by this method checked very closely with that found by any other then in use. They also noted that the proper drying of the precipitate was an important factor, since a t high temperatures the precipitate decomposed and became discolored. Their precipitation was accomplished by bubbling chlorine gas through the solution until coagulation was complete. It is apparent from this that they were using very large quantities of chlorine. In an article published in 1910Rideal again points out that the chlorine will completely precipitate all proteins and peptones, but that it does not throw down the amino acid or organic bases even though it does combine with them. There is no indication in the literature that Rideal ever attempted to make use of this in a practical way for the treatment of sewage, although he does mention in his publication of 1910that clarification observed when small amounts of chlorine were added to sewage might be due to this precipitation. This is in contradiction to his statement that when small quantities of chlorine are added, soluble compounds of the proteins are formed. ' U . S. Patent, 1,134,280. April 6, 1915. Jahresber., 19, 734 (1840); Jahresber. Chem., 44,489 (1848). MQm.d'Arcuei1, 2, 38. Quoted by Rideal. Jahresber., 19, 729 (1840). 5 Ann. Pharm., 61, 288 (1847). Analyst, 22, 228 (1897).
* Berselius
PRECIPITATION O F PROTEINS BY SUPER-CHLORINATION
787
A great deal of work has been reported in the literature? on the reaction of chlorine with proteins and protein products in connection with antiseptic studies. To the best of our knowledge, however, none of these references call attention to the protein-precipitating power of this element, and none of the investigators attempted to make any practical use of this property of chlorine. In fact, the reports in the literature would lead one to believe that the amounts of chlorine required are too excessive for any such purpose. Our data show that native proteins are precipitated by comparatively small amounts of chlorine, while modified proteins such as gelatin require larger quantities, and peptones are precipitated only when very high concentrations are used. Our work confirms that of Rideal and other early investigators in showing that amino acids are not precipitated, although they may be reacted upon and sometimes decomposed by the chlorine. Our data also show that precipitation can be effected w e n in solutions that contain mixtures of various proteins and their decomposition products, although in such cases sufficient chlorine must be added to satisfy in part the demands of all the compounds present. I n the following tables may be found the effects produced when the chlorine is added to pure solutions of various proteins and their decomposition products, as well as solutions containing mixtures of the two. The quantities of chlorine indicated in these data are approximately the minimum amounts required for the precipitation.
TABLE I Effect of Chlorine on Various Proteins and Protein Derivatives Substance
Nitrogen Grams per IO0 cc
Chlorine Grams per IO0 cc
Egg albumin Fresh blood Gelatin Gelatin Peptone Peptone Peptone Tryptophane Tryptophane Glycine
0.0096 0,0096 0.0096 0.0096 0.0096 0,0096 0.0096 0.0096 o ,0096 0.0096
0.0202
2.1
0.0240
2 . 5
0,0403
4.2
’
CI/S ratio
0.0500
5.1
0,0580
6.1 16.0
0 . I530 0.3500
-
36.4
-
0 . I200
12.5
0,1920
20.0
Remarks
Precipitation-filtrate clear Precipitation-filtrate clear 50precipitate-filtrate milky Partial precipitate-filtrate milky No precipitation KO precipitation KO precipitation-filtrate milky No precipitation-filtrate red Precipitation-filtrate dark red N o precipitation
Chattaway: Trans. Chem. Soc., 87, 145 (19oj);107, 181 (191 5); Dakin: Brit. Med. J. Aug. 28,Oct.2~,Nov.2~,Dec.4,19r j;(1)8j2(1916);Dakin,dohen,Daufresne,and Kenyon: Proc. Roy. SOC.,89 B, 232 (1916);Dakin and Dunham: “Handbook on Antiseptlcs,” (1917); Raper, Thompson and Cohen: J. Chem. SOC.,85,371 (1904);Rideal; J. Roy. Sanit. Inst., 31, 33 (1910).Rideal and Rideal: “Chemical Disinfection and Sterilization” (1921); Smith, Drennan, R e h e and Campbell: Brit. Med. J., (2)129(I I j)’Taylorand Austin: J. Exp. Med., 27, I+,j (1899); Tilley: J . A g r . R ~ . , 2 0 , 8 j(1920); ‘&lei and Chapin: J.Bact., 19, 295 (~930);onney and Greer: Am. J. Publ. Hlth., 18, 1259(1928); Tonney, Greer and Liebig: Am. J. Puhl. Hlth., 20, 503 (1930).
H. 0. HALVORSON, A. R. CADE AND W. J. FULLEN
I88
TABLE I1 Effect of Chlorine on Mixtures of Proteins and Their Decomposition Products Substances
B
A
Tryptophane 17
Jl
!, Glycine 11
Peptone 7,
Gelatin t,
,) 71
Nitrogen Grams per roo cc. A B
Chlorine Cl/N ratio Grams per I o 0 cc
0.0074 o ,0070
0.0000
0,0145
2.0
0.0005
0.0240
0.0059
0.0020
0.0037
0.0048
0.0476 0.0817
3.2 6 .o 9.6
0.0074 0.0037
0.0005
0.0215
2.8
0.0048
0.0560
6.6
0.0070
0,0005
0.0215
2.8
0.0037
0.0048 0.0024 o ,0048 0,0048 0.0024
0.0301
3.5
0.0055
0.0037 0.0120
0.0180
0.0173
2.2
0.0173
2.1
0.0522
3.2
0.0522
2.5
It is t o be observed that in the case of pure protein solutions, precipitation can be effected by comparatively small amounts of chlorine. The chlorine requirement is increased somewhat in the presence of peptones, but considerably more in the presence of amino acids, while gelatin increases the demand less than either of the former, Thus albumin and blood proteins are precipitated when the CI/N ratio is 2 . 5 or less, whereas a ratio of 6.0 or above is required when amino acids are present. Since the Cl/N ratio is calculated from the total nitrogen present, it is apparent that nitrogen compounds which are not precipitated will lower the efficiency of the process to an even greater extent than is indicated by the ratios given in the above table. Table I11 further emphasizes this fact by showing that the percentage removal decreases materially when amino acids or peptones are present.
TABLE I11 The Percentage Removal of Nitrogen by Chlorine Precipitation of Various Organic Nitrogen Mixtures Substance
A
Albumin Albumin Albumin Albumin Albumin
B
Concentration Gms N per roo cc A €3
0.0418
Gelatin Blood Tryptophane Peptone
0.0208
0,0272
0.0208
0.0105
0.0208
0.0099 0.0254
0.0208
Grams N in precipitate
Percentage removal
0.0417 0,0441 0.0303 o,ozzo 0.0209
99.7 91.8 96.8 j r .6 45.2
I n the case of mixtures of gelatin and protein, it appears that the gelatin is precipitated even though the Cl/N ratio is less than that ordinarily required to precipitate it alone. Thus we see that in Table I11 where a nitrogen
PRECIPITATION OF PROTEINS BY SUPER-CHLORINATION
189
removal of 91.8is obtained, a considerable portion of the nitrogen must have come from the gelatin. The flocculent precipitate formed by the native proteins apparently occludes the fine colloidal precipitate formed from the gelatin, so that a clear filtrate is produced in a mixture of this type, whereas in a pure solution of gelatin the fine precipitate will not settle out. To produce a clear filtrate with a mixture of native proteins and gelatin, it is necessary to stir the solution gently for about I O minutes following the addition of the chlorine. In coagulating proteins, definite ranges of chlorine concentration are required before any precipitate is formed. Small amounts of chlorine do not produce proportionate amounts of precipitate, but instead all the proteins precipitate when a definite range ifi reached. This is illustrated in the following table which shows the results obtained with different concentrations of egg albumin. TABLEIV The Effect of varying the Chlorine Concentration on Albumin Solutions Concn. egg albumin Concn. Chlorine Gms. N per roo cc. No. pptn. Pptn. Pptn. starting complete 0.074
0.0074 0.00148 0.00074
CI/N ratio Pptn. Pptn. starting complete
0.145 0.016
0.160
0.170
2.1
0.0165
0.0170
2.2
2.3
0.0029
0.0032
0.0034
2.1
2.3
0,0012
0.0017 0.0017
1.9
2.3
2.3
The above data indicate that definite proportion of chlorine to nitrogen is needed before precipitation occurs, regardless of the concentration of the latter. This would imply that the chlorine requirements are independent of the concentration of organic matter. This has been checked by determining the minimum amount of chlorine required to produce precipitation after a reaction period of 1 5 minutes. The data are given in Table V. TABLEV The Effect of Concentration of Protein on the Cl/N Ratio Concn. of Chlorine Concn. of Egg Albumin Gms. per 100 cc. nece8881y Gms. N in roo cc. for precipitation
Cl/N ratio
0.I o 0 0
0.2250
2.25
0.0500
0.I125
2.25
0.0250 0.0200
0.0580 0.0436
2.18
0.0100
0.0224
2.24
0.0050
0.0102
2.04
0.0025
0.0051
2.04
0.001I
0.0019
1.80
0.0001
N o visible precipitate
2.28
n. 0. HALVORSON,
190
A. R . CADE AND
w.
J. FULLEN
Within experimental error, it appears that the quantity of chlorine required is dependent only upon the amount of protein present and independent of its concentration. The slight decrease which occurs in the dilute solution may be due to experimental error which is difficult to avoid in those cases. The chlorine-nitrogen ratio of from 2.0 to 2.3, which is necessary for the precipitation of proteins from pure solutions, can be lowered considerably by an adjustment of the reaction. I n pure solutions the final reaction is brought to a point somewhere between pH 2.0 and 4.0,depending upon the concentration of proteins present. I n reacting with the protein, the major portion of the chlorine is converted to hydrochloric acid, which causes a lowering in the pH. If some of this acid is neutralized so that the final pH is about 4.0, precipitation can be effected with considerably less chlorine. That it does not appear to make any difference which alkali is used for this purpose is indicated in the table below. The same end result is obtained whether the alkali or chlorine is added first. The amount of alkali needed depends upon the quantity of chlorine used, and since that in turn is governed by the concentration of protein present, the alkali dosage must be varied in accordance with the strength of the solution treated. I n samples of waste from a packing establishment, we have found that there is usually present more than enough alkali in the form of carbonates of calcium and sodium. I n those cases, then, the pH must be adjusted to 4.0 by adding a small amount of mineral acid or by adding a little excess of chlorine. The condition can be partially alleviated by preventing clean waters, which usually contain carbonates, from being mixed with the sewage. Tables VI and VI1 show the minimum amounts of chlorine which give a clear supernatant liquor when the reaction is adjusted to the optimum.
TABLE VI The Effect of Various Alkalies on the Chlorine Demand in Protein Precipitation Concn. Protein Gms. N per cc.
Minimum gms. chlorine required to precipitate
CI/N ratio
Com arative minimum amounts of alfali required for adjustment N/13o CaO N/ro NaOH N/IO Na2C0,
0.0077
0.0090
1.1
20
0.0134
0.0157
1.1
35
2.5
2.7
0.0192
0.022s
1.1
SO
3.7
3.6
cc.
1.4 cc.
I
.4cc.
In Table V above it was shown that for a certain range the Cl/K ratio is independent of the concentration of the protein. The same condition is true if the reaction is adjusted as shown in Table VIII. Here, as in the unadjusted series, the Cl/N ratio appears to be independent of the amount of protein present, particularly in the more concentrated solutions. There appears to be a slight increase in the ratio in dilute solutions, although this may be due to experimental error. The percentage error will naturally be high in the solutions that contain 2 s p.p.m. or less of nitrogen.
PRECIPITATION OF PROTEINS BY SCPER-CHLORINATION
191
TABLE VI1 The Effect of Alkali on the Chlorine Precipitation of Various Proteins Nitrogen Grams per I O 0 cc
Protein
Albumin 11
Gelatin ,I
Alkali
Chlorine Grams per I O 0 cc
Cl/N ratio 2.1
0.0096 0.0096 0.0096 0.0096
None CaO None CaO
0.0210
0.0096 0.0096
None CaO
0.0096 0.0096
None CaO
0.0100
0.0096 0.0096
None CaO
0.0240
2.5
0.0125
1.3
0.0096 0.0096 0.0096 0.0096 0.0096
None NaOH CaO NaOH None
0.0096 0.0096
None CaO
0.0130
1.3
0.0420
4.2
0.0210
2.1
0.0260
2.6
0.0150
1.5
0.0310
3.2 I .o
so-so mixture
albumin and gelatin 7,
11
50-50 mixture
gelatin and blood 7,
11
50-jo mixture albumin and peptone 77
,7
so-so mixture blood and albumin I!
7,
t
7f
Blood
,,
0.0280
2.8
0 . 0 1 IO
1.1
0.0110
1.1
0.010s
1.1
0.0270
2.8
0.0480
5 .o
0.02.lO
2.7
W each albumin, blood and peptone 1,
J7
TABLE VI11 The Effect on Chlorine Consumption of varying the Protein Concentration with an Adjusted Reaction Concn. Egg Albumin Gms. N per zoo cc.
Concn. Chlorine necessary for precipitation Grams per zoo cc.
Cl/N ratio
0.rooo
0.IIiO
1.17
0 .os00
0.0560
I . I2
0 . 0 2 50
0.0285
1.14
0.0200
0.0218
I .09
0.0100
0.0109
I .09
0.00~0
0.oosi
I . 14
0.0025
0.0031
1.24
0.0010
0.0014
I .40
0.0001
KO visible precipitate
192
H. 0.HALVORSON, A. R . CADE AND W. J. FULLEN
Since the Cl/N ratio is constant in solutions containing IOO p.p.m. or inore of nitrogen, it is possible to predict the nature of the curve that would be obtained if this ratio were plotted against protein concentrations. This is illustrated in Fig. I . The solid lines are plotted from the experimental data given in Tables VI1 and VIII. The dotted lines are predicted. It is assumed that with concentrations above 0.1000the Cl/N ratio will remain constant. Since the amount of HCl generated decreases with a decrease in nitrogen content, the amount formed when the solution contains less than z j p.p.m. of nitrogen will be only slightly greater than that required t o adjust L.0
r 1
-5.0
‘t
I
-4.0
I
-3.0
I
-2.0
LCO C ~ P C ~ ~ IwC PPOTPV E
I
-1.0
1
-0.0
srmmr
FIG.I Effect of Chlorine Concentration on The Chlorine-Nitrogen Ratio
the reaction to its optimum. Since conditions are then more favorable for precipitation, less chlorine should be required. This accounts for the drop in the curve for the unadjusted reaction. This tendency will continue until the chlorine added is the exact amount for the pH adjustment. After that, additional chlorine must be used to acidify the solution, and the Cl/N ratio will naturally increase. As the reactions in the unadjusted solutions approach the optimum, less alkali will be Qeeded to neutralize the excess acid formed, until finally none a t all will be needed. At this point, then, no saving in chlorine will be effected by the addition of alkali, and the two curves will meet. Since the points in the above curve that show deviations from a straight line occur with dilute solutions in which experimental errors are large, the entire curve should be regarded as theoretical until verified by more rigid experimental data. The shaded area shown in Fig, I represents nitrogen concentrations of between IOO and 300 p.p.m., which strength can be easily obtained in packing plant wastes. In this range the Cl/N ratio is independent of strength. If the strength is reduced below 100 p.p.m. of nitrogen, the ratio for the adjusted series increases with increasing dilution. Since normal packing house sewage usually contains an excess of alkali, this is the curve which will normally be followed. In case there should be no alkali present, it will be desirable, from
PRECIPITATION OF PROTEINS BY SUPER-CHLORINATION
I93
the standpoint of economy, to add some. The adjusted curve shows the desirability of concentrating the sewage to a point where it will contain IOO p.p.m. or more of organic nitrogen. This is generally accomplished by bypassing the clean water coming from the plant. Since such waters contain bicarbonates of both sodium and calcium, their elimination has the additional advantage of reducing the excess alkalinity. In addition to the optimum that exists a t approximately pH 4.0,there appears to be another optimum a t pH 2 . 0 or less, particularly with chlorinated egg albumin. Thus we find that the Cl/N ratio can be decreased to 1.0if a comparatively large amount of HzS04is added. This range has not been
10
L
10
0
I 1.o
I 2 .a
m.cmwmrnnm o
I
3 .o
1 11.0
FIG.2 Percentage of Initial Chlorine available in Various Mixtures of Chlorine and Gelatin. (Data taken from N. C. Wright)
investigated with other proteins, since the large amount of acid necessary would counteract any saving of chlorine that could be effected. In addition, other undesirable conditions would be encountered. In the reaction between chlorine and protein, a portion of the chlorine reacts with the amino groups to form chloramines, a portion may replace hydrogen in other parts of the molecule, some may react into double bonds, and another portion, in the form of hypochlorous acid, may react in such a way that oxygen is introduced into the molecule instead of chlorine. In this connection, and in the light of the researches of Dakins and Wright: our data are of particular interest. Dakin has shown that the percentage of chlorine forming chloramines depends upon the relative amounts of protein and chlorine taking part in the reaction. This was investigated in more detail by Wright. He confirmed the ED&: Biochem. J., 11, 79 (1917); 10, 319 (1916); R O C . Roy. Soc.,89B, 232 (1916); Brit. Med. J., 1, 852 (1916). 0 Wright: Biochem. J., 20, 525 (1926) and Personal Communication.
I94
H. 0. HALVORSON, A. R. CADE AND W. J. FULLEN
results of Dakin and was able to show graphically the amount of available chlorine present when varying quantities of proteins were added to fixed amounts of chlorine. From his data, we have been able to correlate the amount of available chlorine with the Cl/N ratio. This has been done for gelatin, and is illustrated in Fig. 2 . I n a private communication, Wright has presented data which show that the position and shape of the curve varies with the pH, so that the relationship shown in Fig. z should not be regarded as definite for all pH values. I t can be observed from this curve that when chlorine is added to a protein, a considerable amount can no longer be accounted for by thiosulphate titration. With a chlorine nitrogen ratio of from I to j, as shown by the shaded area in Fig. 2 , more than j o q of the chlorine is unavailable. If we assume that all of this has reacted as an oxidizing agent, it is possible, from determinations of available chlorine in the supernatant liquor and precipitate, to show the percentage that has reacted in this way. The data are presented in Table IX. It is to be observed from these data that from 48 to 64% of the chlorine becomes unavailable when no alkali is used for the adjustment of the pH. With the alkali present, the unavailable chlorine ranges from 7 2 to 92%. Attempts have been made to determine the exact amount of this which has reacted in the form of hypochlorous acid with the introduction of oxygen into the protein molecule. Suggestive data have been obtained, but further work is required before definite conclusions can be drawn. By titrating with standard alkali, the total acidity has been determined. If this is calculated on the basis of HC1, we have found that its chlorine equivalent checks, within experimental error, with the unavailable chlorine plus one-half of the available. Since, when chlorine reacts with the amino groups to form chloramines, onehalf of the chlorine goes to HC1, the results indicate that all of the unavailable chlorine has reacted as an oxidizing agent introducing oxygen into the protein molecule. Since the method used for determining the amount of HC1 formed is not very accurate, particularly in the presence of protein buffers, it cannot be stated definitely that nc chlorine is introduced into the protein molecule except that which is in the form of chloramines. It should, however, be safe to conclude that the majority of the unavailable chlorine has been converted to HCl. We have also observed that the modified proteins are precipitated a t pH 4 or less as readily when the available chlorine has been removed by thiosulfate as when the chloramines are left intact. This would lead one to believe that the chlorine which has reacted as an oxidizing agent is of primary importance in modifying the proteins for the precipitation. It is of particular interest to note that the amount of available chlorine in the supernatant liquor is very low when native proteins are the only nitrogenous compounds present, and particularly so when the reaction is adjusted. The amount increases in proportion to the concentration of nitrogenous material that is not precipitated. I n the practical application, the available chlorine in the effluent ranges from 2 5 to 50 parts per million.
PRECIPITATION OF PROTEINS BY SUPER-CHLORINATION
I95
TABLE IX Distribution of Available Chlorine in Protein Precipitation Composition
parts albumin, no alkali
Total N (Grams)
Available Chlorine Grams in ppt. eguent from IOO cc. solution
P. m in
Gms. C1. PercentaQe added C1. used in oxidation
.oo16 .0003"
,0033
.0126
61
,0052
.OISI
64
,0008
,006I
,0063
58
75 parts albumin parts peptone, no alkali
.0047
KO.ppt.
,0126
.0028*
.0032
,0151
,0026
.0052
,0163
63 60 48
50 parts albumin 50 parts peptone,
,005
.oo16 .oo17 .oozg
,0163 ,0201
56 60 63
.OOI2
.0163
52
.oor6 .oo18
.0201
59
,0227
62 84 78 72
100
25
no alkali
I
,0048* '0047
,0150
parts albumin, 7 j parts peptone, no alkali
,0066" ,0066 .0069
50 parts albumin, 50 parts peptone,
,0006
.0002
,0050
.OOI2*
.oooj
,0075
with alkali
.0020
,0008
,0100
parts albumin, with alkali
,0002
25
IOO
,0002
*
.0002
Albumin with no alkali Albumin with alkali 50 parts albumin 50 parts peptone, no alkali
,0006
,0050
84
.OOIO
.0062
.OOI3
.0075
81 80
.OOI2*
,0130
. oooo*
.037s
.oo15
.0 1 7 5
62 92
.0124*
.0067
.0450
58
.002j*
. 00 I 7
.0200
79
50 parts albumin
parts peptone, with alkali
jo
* Combinations in which the minimum amount of chlorine was used to give complete precipitation. The above theoretical studies are of value in determining the limitations of the process. For efficient and economical chlorine precipitation, the waste must contain a comparatively large amount of native proteins, and must be treated before extensive septic action has taken place. From our experience with the waste from the plant of Geo. A. Hormel & Co., we feel that packing
E. 0. HALVORSON, A. R. CADE AND W.J. FULLEN
196
house sewage can be successfully treated by chlorine precipitation. We have found that from various samples of waste from the Hormel plant, 40 to 85% of the organic nitrogen can be removed. Samples containing relatively high percentages of blood gave the higher results, whereas those that had become septic or otherwise contained large amounts of peptone or gelatin gave lower yields. Composite samples from typical runs usually showed removals of from 60-89yo. The following table presents some representative results of actual plant operation. TABLE X Data on Plant Operation Substance
Raw Sewage Effluent (parte per million)
Organic nitrogen Ammonia nitrogen Total nitrogen Total solids Fixed solids Volatile solids Sodium chloride 10,591 pounds of
166.84
44.0 14.4 58.4
17.31
184.I 5 4971.00 2431 . O O 2 540.00
Percentage Reduction 73 .o
68.0
35.0 -
3235.0
2582.5 652.5
74.4
2422.00 2 2 9 2 .o sludge obtained from 660,000 gallons.
-
The data reported in the above table are representative of data obtained from carefully cornposited samples on days when the packing plant was operating a t normal capacity. The significant figures are those showing reduction in organic nitrogen and volatile solids, the former being 73% and the latter 74'35. In the various analyses that have been made, these figures show a variation of from 60 to 80%. When the plant is operating a t normal capacity about 5 tons of sludge are obtained daily. The above figure of 10,591 pounds is representative. About two tons of this are obtained in a primary clarifier without chlorine precipitation, while the other three tons are obtained in a secondary clarifier following the introduction of enough chlorine t o give a Cl/N ratio of from 1.3 to 1.5. The following is a representative analysis of the sludge obtained by chlorine precipitation after it has been dewatered and dried. Moisture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protein.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ammonia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fat. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ash., .................................
3.70%
48.90% 9.50% 22.72%
11.60%
The engineering features and practical results of the plant, having been published elsewhere,"JJlare not repeated here for the sake of brevity. AS has been pointed out in previous publications, the cost of installation is about onethird of the estimated cost of a biological plant. The saving in depreciation and 10
l1
Municipal Sanitation, April 1931. Sewage Work Journal, 3, No. 3, 488 July (1931).
PRECIPITATION OF PROTEINS BY SUPER-CHLORINATION
I97
interest on installation cost is more than enough to pay for chlorine a t present prices. Since other operating costs are low, even if the recovered sludge be considered of no value, the process compares favorably from an economic standpoint with any other method that has been devised for the treatment of packing plant wastes. The ultimate value and use for the sludge is problematic. On feeding the dry sludge to rats we found that it was toxic. Young rats died in four days when fed the dry sludge as the sole source of nitrogen. By extracting the sludge with ether or petroleum ether, the toxicity was reduced to a point where it was doubtful whether the effects produced were due to toxic materials or to the lack of certain essential amino acids. On feeding the extracted material to hogs along with corn no ill effects could be observed, and the animals showed a greater gain in weight than controls getting corn only. These preliminary results warrant further investigation, the results of which may show that the precipitated protein can be rendered fit for hog food. In case this cannot be done, it is felt that the sludge will have real value as nitrogen fertilizer. Experiments are now in progress to determine its value in this respect. *Department of Bacteriology and Immunology, University of Minnesota. **Gee. A . Home1 and Co.