V O L U M E 21, NO. 8, A U G U S T 1 9 4 9 reference blank containing 2.0 nil. of distilled water is treated similarly. At the same time one or more 2.0-ml. starch standards (100 micrograms suitable) receive the same treatment. The colorimeter is adjusted to zero with the reference blank and the samples and standards are then read. When a logarithmic scale colorimeter is used, concentrations are proportional t o scale readings. The pro ortionality constant is determined by the standards used. Coforimeters with a transmittance scale require B semilogarit,hniiccalibration curve. Cellulose. The solid sample is digested icoldj in 607, sulfuric acid for 15 to 30 minutes. The solution is made up to a measured volume with 60yosulfuric acid, so that a 0.5-ml. aliquot contains 10 to 200 micrograms of cellulose. If the solution contains a residue, it is filtered through a dry asbestos mat previously washed with 60y0sulfuric acid and then with water. A 0.5-ml. aliquot is then added to 2.0 ml. of water, and allowed to cool. Then 4.0 nil. of a 0.1% anthrone solution are added and the starch procedure above is employed. A reference blank is prepared from 0.5 ml. of 60% sulfuric acid, 2.0 nil. of water and 4.0 nil. of 0.1% anthrone reagent,. Standards are prepared froin 0.5-ml. aliquots of known aniounts of cellulose in 60% sulfuric acid. Mixtures of Cellulose and Starch. Mixtures of starch and cellulose (cotton) in the presence of each other are analyzed m follows:
953 The sample after boiling in water is filtered through a fine porosity sintered-glass filter or asbestos Gooch pad. The filtrate is analyzed for starch and the residue retained by the filter is digested in 60% sulfuric acid for 15 to 30 minutes. This solution i3 then filtered again if necessary and the filtrate is analyzed for cellulose as outlined above. 4CKYOI'LEDGMEST
The authors wish to acknon-ledge the technical assistance of Maria Frenkel. The work was done on a grant provided by the Nashua IIanufacturing Company and the Saco-Lowell Shops. LITERiTURE CITED
ISD. ENG.CHEM.,ANAL.ED.,18, 499 (19161. (2) Morris, D. L., Science, 107, 254-5 (1948). (3) Morse, E.E., IND.EXG.CHEM.,ANAL.ED.,19, 1012-13 (19471. (1) Dreywood, Roman,
RPCEIVED M a y 3, 1948.
Dichromate Reflux Method for Determination of Oxygen Consumed Effectiveness in Oxidation of Organic Compounds W. ALLAN RIOORE, ROBERT C. K R O S E R , AND C. C. RCCHHOFT U . S. Public Health Service, Cincinnati, Ohio Although the proposed method for the determination of oxJ-genconsumed has definite limitations, nevertheless it will be of value in estimating the strength of industrial wastes and sewage. Hydrocarbons and straight-chain acids are oxidized very slightly. The end products obtained in the oxidation of amino acids vary with the type of acid used. Branched-chain acids and alcohols as well as phenolic compounds are readily attacked. Sugars are quantitatively broken down to carbon dioxide and water. When 50% by volume of sulfuric acid is used in the reflux mixture, chlorides are quantitatively oxidized. The oxygen consumed values of industrial wastes containing high chloride concentrations can, therefore. be corrected for their chloride content.
S
INCE the inception of a biochemical oxygen demand test in 1870 by Frankland (9) and in 1884 by Dupr6 ( 7 ) for the determination of the strength of waste products of human or industrial origin, numerous attempts have been made to devise a chemical method that would give the same results in a much shorter time. Inasmuch as the metabolic activities of the flora and fauna in different samples of waste do not necessarily follow a constant rate, a chemical method would not necessarily correlate with the biochemical determination of oxygen demand. However, it is frequently desirable to know in the minimum time the approximate oxygen-absorbing poiver of a waste. -4 chemical method for determining oxygen consumed is most satisfactory for this purpose, although it nil1 give a value that is not comparable to B.O.D. on toxic wastes and will give higher results on stabilized biological treatment plant effluents, because 11 is impossible for a chemical method to differentiate betneen organic matter in biologically stable and unstable forms. Because a chemical method for determining oxygen consumed seems desirable as an additional criterion of pollution control, it is also necessary to study such procedures more thoroughly in order to understand, apply, and interpret the data obtained with them. Of the various oxidizing agents available, in general, only four have been used t o any extent for determ~ningthe ouygen-con-
suming power of sewage and industrial wastes-namely, potassium permanganate, potassium dichromate, ceric sulfate, and iodic acid. Potassium permanganate is still used in the recommended procedure ( 2 ) . Stamm (22) carried out the oxidation with pernianganate in an alkaline solution and prevented the reaction of >InOl-- + Mn02 by the addition of a barium salt which allom a better end point to be obtained. Benson and Hicks ( 4 ) ,in determining the pollution in. sea mater, found that application of the Zinimerman-Reinhardt procedure in titrating the excess permanganate gave more reproducible results. Haupt (IO) tried to correlate the permanganate oxygen consumed with the B.O.D. on wastes from paper pulp factories. He found, however, that the chemical method gave much higher results. due to the fact that cellulose is not readily attacked by either dissolved oxygen or bacteria. -4s xvith most chemical methods, variation in conditions affect the result obtained. Matubara (16) found that increased values could be obtained by increasing the boiling time, increasing the concentration of potassium permanganate used, saponifying the fats or oils, and neutralizing water-soluble fatty acids. Lovett ( 1 5 ) states that totally different values can be obtained depending on whether 0.125 S or 0.0125 4 potassium permanganate is used. Kashkin and Karasik (15) added an initial excess of potassium permanganate calculnted to be equivalent to 0.3 to 0.5 m g . of oxygen, and determined the final excess by titration with oxalic acid a t boiling temperature. Shutkovskaya (21) compared the discoloration of the potassium permanganate by the sample
ANALYTICAL CHEMISTRY
954
Table I. Oxidation of Organic Compounds by the Dichromate Keflux Me'thod Compound Glucose
CsHliOa
Lactose
CizHzzOii
Acetic acid
CHICOIH
Lactic acid
CHaCHOHCOtH
Citric acid
CHzCOzH
Tartaric acid Malic acid Furoic acid
Oxygen Consumed 33% HzSOd by Volume 50% HzSOt by Volume Replicates Average Replicates Average 1036 1033 ... ... 1035 1023 1032 1027 1038 1021 1019 1017 1049 52.8 53.8 15.0 17.5 54.8 20.0 443 441 549 520 491 438 760 752 698 675 744 651
Formula
kz:'"
Salicylic acid
106G
4.4
2.6
1066
98.3
94.9
1066
57.7
48.8
787
14.1
3.2
...
531
528 526
527
533
0.37
1.1
CH~COZH
657 662
660
680 685
683
716
7.8
4.6
701 691
697
1107 1101
1104
1285
45.9
14.1
251 241 1534 1631
246
1963 1941 1602 1575
1952
1967
87.6
0.76
1589
1622
6.4
2.0
~HOHCO~H CH=C-COzH 1 0 CH=CH/ CaHaCOzH CsHd-OH \COzH
(0)
Glycine
CHz(NHz)COzH
Alanine
CHaCH(NHz)COnH
Tyrosine
O H C ~ H I C H Z C H ( N H Z ) C O(ZpH ) CH
C HaH
j
~
(
~
~
z
)
~
~
Isopropyl alcohol
CHI
Ethyl alcohol
CHI' CzH6OH
1532
1620 1587
1602
1622
6.6
1.2
202 197 163 164 1435
200
632 627 408 400 1659 1624 1332 1314
630
640
68.2
1.5
404'
86.9
1642
1268a 359 1678
14.5
67.9 12.6 2.1
1323
1641
83.4
19.4
1398
87.4
14.3
78.4 3.2
14.9
2 183 ~ ~ 167 213 215 206 842 834
Glutamic acid
\CHOH
667 676 1737 1738
Catechol 5-Me-2-isopropyl phenol (thymol) p C u m y 1 phenol
1533 0
1637 1527
252 261
~ H C (X HH ~C) O ~ H
cH/ a-Amino isobutyric acid
50% HzSOd
536 525
m-Hydroxybenzoic acid
Valine
33% HIS04 3.1
E::: ~
Benzoic acid
% Deviation from Theoretical Theoretical 1066
164 1435 257 b 175
1240 1153
1197
211
827 820 855 1611 1599
834
838 672 1738
1667
1667
1111
1113
747 747 2024 1872 2225 2217 2648
980 C 218
1605
1600
47.6
747 1948
2087 695d 1891
66.7 3.3 8.1
63.7 7.5 3.0
2221
2346
24.6
5.3
2608
2596
57.1
0.46
0.31
2568
1115
o-Cresol
2174 2158
2164
2267 2277
2272
2518
14.0
9.7
m-Cresol
2139 2155
2147
2334 2324
2329
2518
14.7
7.4
50.0 5.1 526 923 488 923 976 923 449 2522 2656 7.2 1.3 2526 2360 2370 CioH70H 2518 2380 91.1 263 3072 99.7 272 10.0 8.8 Benzene CsHa 280 7.5 100 100 2531 0.0 ... Pyridine 0.0 CsHsN 76.1 3130 93.3 205 210 752 747 Toluene CsHaCHi 215 742 ... 1185 2.9 1150 1150 ... Cellulose (C8HIoOs)s 0 CHICH(NH,)CO,H 31/rOz +3coz 3HzO NHs. Theory is 1258 but reaction obviously does not follow this course. CHiCH(NHz)COzH 02 CHrCOzH N H J + , C o t . Theory is 359 p.p.m. However, 88 shown with CHICOZH, there is an oxygen consumed of 54 p.p.m.; hence observed value would be expected to be high. b CH, CHa 'CHCOtH. Theory is 273. This reaction evidently takes place. COa 4- NHa+ )CHCH(NHz)COzH 4- Oz CHa CH/ Odor of isobutyric acid noticeable.
2-4,6-Trinitrophenol 2-Naphthol
(NO~IC~HZOH
...
++ +
+
++
+
...
V O L U M E 21, NO. 8, A U G U S T 1 9 4 9 on heating and allowing to stand for 5 minutes to standard colored-glass plates calibrated in p.p.m. of oxygen. The amount of research carried out with the permanganate method shows that it is not entirely satisfactory for the determination of oxygen consumed values. For this reason, workers in this field have turned their attention to the use of other oxidizing agents. Klein (14) made a comparison of the permanganate, dichromate, and ceric sulfate methods of determining the strength of sewage. He found that ceric sulfate gave values two thirds that of the dichromate but two to three times that of the permanganate. Bezel (5) also found that ceric sulfate was superior to permanganate because of the greater stability of the reagent and of the titer. Adeney and Dawson (1)were among the first to use dichromate in the presence of sulfuric acid to determine the organic matter in water. They heated the mixture of 100" to 110' C. for 2 hours and titrated the excess dichromate with ferrous sulfate using an outside indicator. Popova (19) also used dichromate to determine the "oxidizability" of sewage and found it gave results of about 8G% of the B.O.D. Ostrovskaya (18) and Rhame (20) made use of the iodometric procedure for determining the excess of dichromate present. This method, howeve;, required rather careful manipulation and gave rather wide variations in the calculated B.O.D. and the standard B.O.D. test. Ingols ( 1 1 ) modified Rhame's procedure by refluxing the sample and the oxidizing mixture for GO minutes at about 145' C. He determined excess dichromate iodometrically. In 1938 Dzyadzio (8) used potassium iodate in a G 5 to 80% sulfuric acid solution as the oxidizing agent, heated the mixture at 200" C., and determined excess iodate iodometrically. He claimed that the error obtained in the oxidation of 14 organic compounds did not exceed 2 to 3% and that the method is superior to the dichromate method. Johnson, Tsuchiya, and Halvorson (12) also used the iodic acid method. I n their work, they refluxed the mixture if the sample was high in volatile acids. Another approach to the problem of determining the oxygen consumed by sewage and industrial wastes was tried by Mohlman and Edwards ( l 7 ) , who used chromic acid as the oxidizing agent and absorbed the liberated carbon dioxide in 0.1 N barium hydroxide. The excess barium hydroxide was titrated and the oxygen consumed calculated from the amount of barium hydroxide used. This method, as well as that of Burtle and Buswell ( 6 ) , requires very careful manipulation and rather complicated apparatus. In the latter case, the precipitated barium carbonate is filtered and weighed. In samples of sewage and industrial wastes there are present a wide variety of organic compounds. In order to evaluate the usefulness of any chemical method for determining the oxygen consumed, it is helpful to known just how efficient that method is in the oxidation of various organic compounds. The dichromate method proposed was, therefore, studied with about 30 organic compounds of various classes. APPARATUS AND PROCEDURE
The reflux apparatus used consisted of a 300-ml. roundbottomed flask with a 24/40 taper-joint neck connected with a Friedrich's reflux condenser. All samples were run in duplicate and a blank containing 50 ml. of distilled water was run simultaneously.
CO;HCHmCHzCH(Pr"z)COzH
is 980.
+ 41/zOt +5COz + NHI + 3Hz0. + + +
955 One gram of the organic compound under study was weighed out, dissolved in distilled water, and diluted to 1 liter. In the case of phenolic compounds, it was necessary to add alkali to effect the solution of the compound. With compounds such as benzene, it was necessary to homogenize the mixture of water and the organic compound in a colloid mill and dilute the emulsion to 1liter. The size of sample taken for the oxidation was based on the theoretical amount of oxygen necessary for complete combustion to carbon dioxide and water. This amount of sample was diluted with distilled water to 50 ml. and placed in the round-bottomed flask. To the sample 25 ml. of 0.25 N potassium dichromate were added, followed by 75 ml. of 95% sulfuric acid. A few granules of pumice were added to prevent bumping and the flask was connected to the reflux condenser. The mixture was refluxed for 2 hours, cooled, transferred to a 500-ml. Erlenmeyer flask, and diluted to about 300 ml. The excess potassium dichromate was titrated with 0.25 N ferrous diammonium sulfate using o-phenanthroline ferrous complex as an indicator, The end point is sharp, changing from a gray-green to red. When the concentration of sulfuric acid was 50% by volume or less, no difficulty was encountered in determining the end point. However, if a higher acid concentration, or stronger dichromate was used, it was necessary to dilute the refluxed mixture three to four times with distilled water in order to reach the correct end point. The standard ferrous diammonium sulfate wadstandardized each day. The blank determination rarely exceeded 0.2 ml. of 0.25 N potassium dichromate. The temperature of refluxing, using the 50% by volume of sulfuric acid was 145' to 150" C. EFFECT OF CHLORIDES
I n 1932 Bach (3) found that the oxygen consumed value of raw sewage was increased from 258 to 291 p.p.m. when the sodium chloride content varied from 20 to 2000 p.p.m. I n Figure 1 the effect of chlorides on oxygen consumed values obtained with 0.25 N potassium dichromate is shown. When 50% by volume of sulfuric acid is ustd, quantitative oxidation of chlorides is obtained over the range from 250 to 20,000 p.p.m. However, when 3370 by volume of sulfuric acid is employed the results obtained are somewhat erratic, the amount of oxidation depending on the amount of 0.25 N potassium dichromate used. With 50.0 ml. of the dichromate, the results are not so erratic as with 25.0 ml. The chloride correction in the latter instance is subject to a larger error. OXIDATION OF ORGANIC COiMPOUNDS
In Table I, the results obtained with 32 organic compounds are given. These compounds represent several different types, such as sugars, aliphatic and aromatic acids, amino acids, alcohols, phenolic compounds, and hydrocarbons. No attempt was made to repurify any of these compounds and for this reason the oxidation values obtained may be slightly low. With the two sugars, glucose and lactose, the oxidation to carbon dioxide and water is about 97% complete. Cellulobe (represented by filter paper) is 100% oxidized under the experimental conditions used. As expected, the straight-chain acids are hardly attacked. When (as with lactic acid) an OH group is introduced into the straight chain, slightly better oxidation is obtained. Approximately 51% of this acid is broken down to carbon dioxide and water, using the 50% mid concentration. No difficulty was encountered in oxidizing the branched-chain or aromatic acids, as shown in Table I. Heterocyclic acids, such as furoic acid, were not so easily attacked under the conditions set up. Furoic acid was oxidized to only 85% of completion. Of the amino acids studied only glycine and tyrosine were quantitatively oxidized to carbon dioxide, water, and ammonia. If we assume, as shown in the following equation, that a mole of acetic acid is formed in the oxidation of alanine: Theory
C O , H C H ~ C H ~ C H ( N H ~ ) C O ~ Ho2 + C H ~ C H ~ C O ~ H2 c 0 2 + K H ~ . Theory is 218. d C9HbOH-b 0 2 CHaCOzH HzO. Theory is 694. Hdwever, 50% H2904 gave oxygen consumed of 54 p.p.m. for C H ~ C O I Hwhich wlll account f o r higher value obtained.
+
CHaCH(xHz)COzH
+
0 2
-
+
CH~COZH
3"
+ CO,
about ss% breakdown is Obtained and, as pointed Out above, the acetic acid formed would not undergo further oxidation. Valine, representing a branched-
ANALYTICAL CHEMISTRY
956
Table 11. Comparison of Dichromate a n d Iodic Acid a s Oxidizing Agents % Deviation from Theoretical Dichromate Iodic acid 13.0 3.1 14.0 2.6 94.9 77.1 5.8 1.1 0.76 3.3 1.5 0.0 2.1 8.1 14.9 0.13 100 so.9 91.1 14.8
Compound Glucose Lactose Acetic acid Tartaric acid Benzoic acid Glycine Tyrosine Glutamic acid Pyridine Benzene Toluene
88.7
76.1
Oxidation of Chlorides B y Potassium Dichromate 0 He-, 50% by v d 05cc.hw KeCreOT used) @ HtSOe,33%by vd. (5occ.N/4 KrCreOl used) 0 H i SO,, 3 3 % by vd. ( 2 5 ~N/4 ~ .KtCreOi used)
Ji
I n Table I1 a comparison of dichromate and iodic acid as oxidizing agents is shown. The oxidation with iodic acid was carried out by refluxing the compound with potassium iodate in a mixture of phosphoric and sulfuric acids. Iodic acid is not superior to dichromate in the oxidation of many of these compounds and the analytical work involved is more complicated. RECOXI>IE>DED PROCEDURE
Dilute an appropriate amount of sample to 50 ml. with distilled water in a 300-nil. round-bottomed flask with a taper-joint neck, add 25.0 ml. of 0.2500 S potassium dichromate and 75 ml. of concentrated sulfuric acid, and reflux for 2 hours. For the best quantitative results, use 50% by volume sulfuric acid. Cool, transfer the mixture to an Erlenmeyer flask, and titrate the eycess potassium dichromate R ith approximately 0.2600 S ferrous ammonium sulfate, using o-phenanthroline ferrous complex as an indicator. Reflux a blank a t the same time, using the’same amount of reagents and substituting 50 ml. of distilled water for the sample. The ferrous ammonium sulfate must be standardized daily. Calculation.
O.C., p.p.m. = ( a
- b)
X normality (standard Fe)
x
8000
Volume of sample where
O.C. a
b
= oxygenconsumed = nil. of Fe(SH&(S04)2 used for blank = ml. of Fe(SH4)u(SOa)2 used for sample CONCLUSION
.
SODIUM CHLORIDE- PRM.
chain amino acid, is apparently broken down with the 33% sulfuric acid in accordance with the follotving equation:
CH3 )CHCH (’153
+ 0,-+
(NHJCO~H
CHa
SH,
+ co?+ )CHCO?H CH3
as evidenced by the odor of isobutyric acid in the reaction flask. K i t h the 50% acid concentration, complete oxidation is not obtained, about 81% of the valine being broken down to carbon dioxide, water, and ammonia. Glutamic acid is broken d o w about S6’370 of the theoretical with 5oyOby volume of sulfuric acid. K i t h the 33y0 acid concentration, it is possible that the following reaction takes place:
.Is with most wet combustion methods employed, the method proposed has its limitations. However, the results obtained are reproducible and the method should be useful in determining the approximate strength of sewage and industrial wastes. Hydrocarbons as well as straight-chain acids and alcohols, are scarcely attacked. I n contrast to this, the aerobic bacteria are able to utilize and oxidize the latter two types of compounds as a food source. Branched-chain aliphatic acids and alcohols are, as a rule, readily oxidized by the proposed chemical method, and no difficulty is encountered in the oxidation of sugars. Phenolic compounds are also oxidized quantitatively. Chlorides are shown to be quantitatively oxidized at the higher acid concentration. Correction for the chloride content of industrial wastes may be made. K i t h the lower acid concentration, however, the amount of oxidation is dependent upon the volume of 0.25 N potasqium dichromate used. LITERATURE CITED
Adene?;, W.E.. and Damson, B. B., Sci. Proc. Roy. Dublzn Soc., 18, 199-202 (1926).
i
CH~CO~H The theoretical amount of oxygen required for this reaction is 218 p.p.ni. and the experimental amount obtained was 211 p.p.m. Of the two alcohols studied isopropyl alcohol is quantitatively oxidized, whereas the straight-chain ethanol is oxidized to acetic acid which is not furher attacked. No difficulty was encoun:ered in oxidizing the phenolic compounds studied, the amount of oxidation ranging from 90% for o-cresol to 99.57, for p-cumylphenol, Benzene was about 10% osidized and pyridine was not attacked a t all under the experimental conditions employed. The one substituted aromatic hydrocarbon used was toluene. It v.-ould normally be expected that this compound would be easily oxidized to benzoic acid, which was completely osidized as shown in Table I. However, on the basis of complete oxidation only 24% of t,he theoretical value was obtained. Seither inoreasing the acid concentration to 66% by volume nor increasing the potassium dichromate strength to 0.5 -V had any effect on the oxidation of toluene. Various catalysts were also employed, such as selenium, copper, iron, nickel, and platinum, but still only about 26y0 of the theoretical value was obtained.
Am. Pub. Health Assoc. and Am. Water Korks Assoc., “Standard Methods for the Examination of Water and Sewage,” 9th ed.,
p. 122, New York, 1946. Bach, H., 2. anal. Chem., 89, 439-41 (1932). Benson, H. K., and Hicks, J. F. G., Jr., ISD. ENG.CHEM.,ANAL. ED.,^, 30-1 (1931).
Bezel, L. I., J . Applied Chem. (U.S.S.R.), 18, 361-6 (1945). Burtle, Jerome, and Buswell, A. hl., Sewage Works J . , 224-38 (1937).
DuprB, “Stream Sanitation,” E. B. Phelps, p. 66, New York. John Wiley & Sons, 1944. Dzyadzio, A. M.,Vodosnabthenie i Sanit. Tekh., N o . 8-9, 1J7-25 (1938).
Frankland, Sir Edward, “Stream Sanitation,” by E. B. Phelps, p. 65, New York, John Wiley & Sons, 1944. Haupt, H., V o m Wasser, 10, 60-77 (1935). Ingols, Robert. and Murray, P. E., Water & Sewage W o ~ k s95, , 113-17 (1948).
Johnson, D. W., Tsuchiya, H. X I . , and Halvorson, H. O., Abstracts of 109th Meeting, Ax. CHEM.SOC.,p. 2 S, Atlantic City, 1946. Kashkin, M . L., and Karasik. R. M . , M e d . ezptl. (Ukraine), No. 4, 9-13 (1940).
Klefn, L. J., Proc. Inst. Sewage PUT$ (England), 1941, 174-91. Lovett, M . J., Ibid., 1940, 194-5 Matubara, Tamenaka, Mitt. Med. Akad. Kioto, 28, 563-78 (1940).
V O L U M E 21, NO. 8, A U G U S T 1 9 4 9 (17) Mohlman, F. W., and Edwards, G. P., ISD. ESG.CHEY.,ANAL. ED.,^, 119-23 (1931). (18) Ostrovskaya, R. E., Lab. Prakt. (r.S.,S'.R.), Sammelband, 84-6 (1939). (19) Popova, E. S.,Tmv. comm. recherches e'purat. euuz (~kfoacow), No. 12, 143 (1930). (20) Rhame, G. -4.. Water & Sewage Works, 94, 192-94 (1947).
957 (21) Shutkorskaya, L. A., H i g . isanit. (C.S.S.R.),10, No. 10/11,54-7 (1945). (22) Stnmm, Hellmuth, Angew. Chem., 47, 791-5 (1934).
RECEIVED June 30, 1948. Presented before the Division of Water, Sewage, and Sanitation Chemistry a t the 113th Meeting of the . k M E R I C A N CHEMICAL SOCIETY, Chicago, 111.
Extraction of Carotene from Green leaves V. H. BOOTH Dunn Sutritional Laboratory, Cambridge, England Hot light petroleum (85" C.) extracted nearly all the carotene from dried grass meal but only very little from fresh green leaves. Higher temperature (130' C.) increased the extraction from leaves but destroyed some carotene. A mixture of cold light petroleum, acetone,.and quinol easily extracted the carotene from leaves. Extraction of carotene by hot petroleum is hindered by large particle size and to a less extent by water. In support of these conclusions it has been shown that rehydration of grass
I
N A collaborative investigation, initiated by the Crop Driers Association in England, a simple method was evolved for the rapid estimation of total carotene in dried grass meal (12). (The term "dried grass" is used for convenience and covers other dried fodder crops including alfalfa or lucerne. Alfalfa meal was in fact used for many of the experiments.) The extraction part of this technique had been worked out by J. R. Edisbury. In this method the meal is heated with light petroleum for 1 hour on a water bath a t about 90" C. Light petroleum, boiling range 80" to looo, is used in a Kjeldahl flask, whose long neck acts as a condenser. The temperature attained by the extractant is about 85'. One extraction dissolves over 9776 of the carotene. The solution is cooled and decanted directly through a column of bone meal (9) which adsorbs unwanted pigments. Experience has shown that the new method must be carried out at about the temperature stated. I n this it is different from the method of Kernohan (7). It is known that 40" to 60" light petroleum is inefficient for extracting carotene from fresh leaves. It was therefore not surprising to find that carotene could not be extracted from fresh grass by 80" t o 100" petroleum. This paper is concerned with the cause of the different behavior of fresh grass and of dried grass meal towards hot petroleum. REAGENTS
Light petroleum, boiling range 40" to 60'. Light petroleum, nominal boiling range 80" to looo. High boiling petroleum means petroleum boiling above 130 '. Petroleum-acetone is made by mixing 40" to 60" light petroleum with an equal volume of acetone and adding l gram of quinol to each liter of the mixture. Petroleum-ethanol is the foaming mixture of Moore and Ely (100 ml. of 95% ethanol and 75 ml. petroleum ether, 11). The ethanol mas purified by distillation from zinc and potassium hydroxide.
meal reduced the extraction rate; extraction of whole dried leaves by hot petroleum was incomplete; and the finer the disintegration of the dried leaves the better the extraction. By grinding fresh leaves to a fine pulp, almost complete extraction with hot petroleum was achieved. But carotene was lost by oxidation, and although this could be countered by grinding with solid quinol, such a technique offers no advantage over that using cold petroleum-acetone.
was used for the determination of control values as well as for the determination of carotene that remained after one extraction of fresh or dried leaves with 80" to 100" petroleum. The method can be used for most types of fresh and dried vegetable tissue. It is particularly convenient for grass and other plants with small leaves. A gram of the material is weighed into a ver thick 50-ml. squat beaker. About 1 gram of quartz powder g o t sand) and 10 ml. of petroleum-acetone are added a t once and the tissua is ground with a flabbottomed glass pestle in the beaker. T h t beaker is much more convenient than a conventional mortar. Dry samples are damped with water before extraction. The supernatant solution is decanted (without filtering in most cases, but through a filter when dried grass meal is being extracted) into a separating funnel containing water. The residue is ground again and about six or seven extractions (taking 5 or 6 minutes altogether) are sufficient to remove all pigment. The acetone, etc., are removed in a simple automatic washing arrangement in which drops of water fall through the petroleum-acetone solution in the separating funnel to an overflow made from a bent glass tube. The igments remain in the light petroleum. The solution is purifed by adsorbing all the pigments on a chromatographic column of alumina-sodium sulfate and eluting the carotene with 2% acetone in petroleum. Because aluminum hydroxide is hygroscopic and alterations in its water content change its power of adsorption, each portion must be heated before usd ( 2 ) . Alumina was therefore stabilized by mixing it with an equal weig!t of anhydrous sodium sulfate and heating for 12 hours a t 150 . The mixture may be stored in bottleq and is ready for use a t all times.
METHODS
Proving the Control Method. I n work of this type the choice of a control method is very important. The method of Moore and Ely (11) using the Waring Blendor appears to be one of the most popular of published methods for extracting carotene from plant materials. On the other hand the beaker method (2) using petroleum-acetone-quinol is much neater and requires fewer vessels. In order to study their relative merits the two methods were compared.
Method with Light Petroleum. The method with 80" to 100' petroleum was used as described above, except that the pigments were purified on a mixture of alumina and sodium sulfate as described below. This adsorbent is faster and more specific than bone meal, and the elution of carotene is under visual control. Method with Petroleum-Acetone. The method with cold petroleum-acetone-quinol mixture ( 8 ) , modified in minor detail,
One operator practiced the blender method for several days, following the directions given by Peterson (13) until a routine for interlocking replicates was well developed. RIinor modifications were essential for quantitative transfer of the extract from blender to separator. ,411 apparatus being ready, a specimen of fresh grass was obtained. One operator weighed and extracted sextuplicates of about 3 grams each by the blender method. Simultaneously another operator weighed and extracted sextuplicates of 600 to 800 mg. each by the beaker method. The latter completed his weighing and extractions and had the
