June, 1921
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y REFERENCES
1-W. W. Scott, “Standard Methods of Chemical Analysis,” 2nd Ed., D. Van Nostrand Co., 148. 2-A. A. Blair, “The Chemical Analysis of Iron,” 8th Ed., Lippincott Co., 175. 3-C. M. Johnson, “Chemical Analysis of Special Steels,” 2nd Ed., John Wiley & Sons, Inc., 304. 4-Johnson, LOC.cit., 307; Blair, LOC.cit., 180. 5-Johnson, LOC.cit., 316; Blair, LOC.c i t . , 177. 6-J. W. Rothe, Milt.kgl. Tech. Versuchsanstalt zu Berlin, 1892, Part 111; Blair, LOC.cit., 177, 202. 7-It is important that the filter be not allowed to run dry, lest the hydroxide coagulate and retain traces of vanadium which would subsequently prevent a satisfactory deposition of cobalt and nickel. 8-Sodium peroxide was employed in two preliminary experiments for the oxidation and separation of chromium and vanadium from iron, cobalt, etc. A. A. Noyes, W. C. Bray and E. B. Spear [Tech. Quarterly, 21 (1908). 141, and also C. M. Johnson [Chem. Met. Eng., 20 (1919), 5881. The separation was complete (Expts. 1 and 2. Table 11), but the procedure was abandoned because it was necessary to repeat the separation twice and the precipitates were difficult to handle. 9--For the determination of chromium and vanadium see page 541.
543
IO-This precipitate represents a quantitative recovery of copper, and the percentage may, therefore, be determined by ignition to oxide or, preferably, by electrolysis in a small volume of solution. 11-If manganese is to be determined, the precipitate should be dissolved in 40 cc. of nitric acid (1 : 31, and reserved. 12-If a determination of manganese is desired dissolve any anode deposit in the solution described in preceding reference. 13-If a determination of manganese is desired add the precipitate to the solution reserved for manganese (two preceding references ) 14-The factors for NiO, COO and Cos04 a5e 0.786, 0.787, and 0.734, respectively. The use of the factor 0.75 on a precipitate weighing 2 mg. could, therefore, not occasion an error greater than 0.004 per cent on a 2%. sample. With large precipitates, ignition to metal in hydrogen must be carried out. Cf. Treadwell-Hall, “Analytical Chemistry,” Vol. 11, 4th Ed., page 139, John Wiley & Sons, Inc. 15-There is no difficulty a t all in precipitating traces of nickel in the presence of any amount of cobalt if this method is followed. The precipitate will contain cobalt, however, and must be purified as directed. 16-T~rs JOURNAL, 11 (ISIS), 632. 17-Johnson, LOC.cit., 8. 18-Unpublished method originating with Dr. I,. F. Witmer a t the Bureau of Standards. 19-D. L. Randall, A m . J. Sci., 141 2 4 , 313.
Improved Deniges Test for the Detection and Determination of Methanol in the Presence of Ethyl Alcoho1112 By Robert M. Chapin BrocxEMrc DIVISION,BUREAUO F ANIMALINDUSTRY,
u. s.
The examination of alcoholic products for methanol has been a problem of interest t o many chemists. If a certain few published papers are consulted the matter would appear t o be rather simple, a t least from the qualitative side. But a thorough survey of the voluminous literature, comprising a large number of methods with contradictory comments and conclusions, does not lead one t o undertake exacting work along this line with entire confidence. One of the most recent investigators, Gettler,3having reviewed fifty-eight existing tests, recommends subjecting the sample t o nine qualitative tests, sebuentially applied. I n passing i t may be noted t h a t his eighth test, a refractometric one, is essentially quantitative in nature, being based upon a numerical difference between physical constants, and is only secondarily of qualitative significance. Also his first seven tests are merely tests for formaldehyde, applied after treating the sample with a single oxidizing agent. If this oxidizing agent is capable of producing formaldehyde from any substance other t h a n methanol, all the seven tests must be subject t o a common source of error. PureIy qualitative findings, however, seldom afford solid ground for action in matters of commercial or legal importance. The question “How much?” is almost certain t o arise. It is a pertinent question here, inasmuch as several investigators4 have stated t h a t methanol is naturally produced in certain fermentations. If methanol, like fusel oil, is a normal constituent of alcoholic products, then the legitimacy of its presence in a n y case may be satisfactorily settled only b y quantitative examination. The analytical chemist needs, first, a simple but dependable qualitative test which shall possess semiquantitative value in t h a t Received February 16, 1921. of the Secretary of Agriculture. 8 J . Biol. Chem., 4!2 (1920), 311. 4 von Fellenberg, Milt. Lebensm. Hyg., 5 (1914), 172; Biochem. Z.,86 (1918). 45; Takahashi, J . Coll. Agr. I m g . Uniu. Tokyo, 5 (191.3, 301; J. A m . Chem. SOC.,89 (1917), 2721. 1
a Published by permission
DGPARTJIENT O F AGRICULTURE, WASHINGTON, D.
c.
it is able t o serve as a “limit test,” and, second, a quantitative method which shall enable him t o assert with positiveness very nearly the exact percentage present. The quantitative method must be subjected t o intensive study in order: (1) To develop its highest inherent precision. (2) To devise methods for the elimination of interfering substances. (3) In case elimination is impossible, to determine the size of the “blank” involved by the presence of each such substance.
The DenigBsl test seems most promising for both qualitative and quantitative application. It consists in treating the alcoholic solution with potassium permanganate and acid, whereby methanol is oxidized t o formaldehyde. The latter is detected by Schiff’s reagent in the presence of sufficient sulfuric acid t o prevent development of color from acetaldehyde. There appears no evidence t h a t other proposed oxidizing agents, such as bichromate and acid or persulfates,2 are inherently superior t o permanganate and acid. The latter agent is preeminently simple and convenient, requiring no heat for its action and finally affording a colorless solution. No reagent effects a quantitative yield of formaldehyde. All require strict adherence t o a standard set of conditions under which i t is assumed t h a t a certain concentration of methanol originally present results in a certain concentration of formaldehyde a t the end. Likewise, for the demonstration of formaldehyde there appears t o be no reagent any more convenient or reliable t h a n Schiff’s reagent, prepared according t o Compt. rend., 150 (1910). 832. Preliminary experiments have indicated t h a t persulfates, especially in strongly acid solution, may produce a notable quantity of formaldehyde from pure ethyl alcohol. T h e possibility of such a reaction has been noted by previous observers in the application of several oxidizing agents. Bichromate and acid, in comparison with permanganate and acid, appears to afford a high yield of acetaldehyde from ethyl alcohol, but a low yield of formaldehyde from methanol. 1
2
544
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
the Elvove‘ formula. I t s chief comparative disadvantage is the slowness of development of the final color. Q U A N T I T A T I V E METHOD
The Deniges method has been used with more or less modification by a considerable number of investigators. Since in routine analyses following the procedure of Elvove the observed margin of precision seemed unnecessarily large, the whole process has been subjected t o close scrutiny with a view t o attaining greater precision. It was decided t h a t 0.04 cc. of total alcohol should be the standard quantity for each test, which, including the necessary acid, should be made t o a volume of 5 cc. The nature and proportion of t h e acid is of very great importance. The highest yield of formaldehyde results from slow action of permanganate in presence of low hydrion concentration; b u t practical considerations prohibit a n inordinately long reaction time, while the total acid must be kept up t o a safely high figure. The conditions finally chosen were the addition of 0.2 cc. of phosphoric acid (C. P., 85 per cent), previously diluted t o 1 cc. for accuracy in measurement, and a n oxidation period of 30 min., instead of t h e 0.2 cc. of concentrated sulfuric acid and oxidation period of 3 min. employed by Elvove. Next, after deciding t h a t the necessary permanganate should be added in a volume of 2 cc., i t remained merely t o find a concentration of the permanganate solution such t h a t either more or less t h a n 2 cc. of i t would give a lower yield of formaldehyde t h a n exactly 2 cc. The desired strength was found t o be 3 ‘per cent. I n a similar way the volumes of sulfuric acid and Schiff-Elvove reagent were tested. Directions for t h e method may be given as follows: Dilute the solution, previously purified as necessary, t o 1 per cent by volume of total alcohol (Sample Solution A). Of this, pipet 10 cc. into a 50-cc. volumetric flask, add 10 cc. of a 4 volume-per cent solution of pure ethyl alcohol, a n d make t o the mark with water (Sample Solution B). Of t h e latter, likewise, dilute 10 cc. plus 10 cc. of the 4 per cent ethyl alcohol t o 50 cc. (Sample Solution C). I n t o 50-cc., tall-form Nessler tubes pipet 4 cc. of the three sample solutions. Prepare standard methanol tubes containing, respectively, 1, 2, and 3 cc. of a 0.04 volume-per cent aqueous solution of pure methanol, plus 1 cc. of 4 per cent pure ethyl alcohol, plus sufficient water t o make 4 cc. After the tubes are properly arranged in a rack the following operations are put through in strict parallelism, remembering t h a t each reagent is t o be added t o all tubes before any are mixed: 1-Add
1 cc. of a 1 in 5 volume solution of phosphoric acid
(C.P., 85 per cent), and mix. 2-Add 2 cc. of 3 per cent potassium permanganate solution, mix, and let stand 30 min. 3-Add 1 cc. of 10 per cent oxalic acid solution, mix, and let stand till a clear brown (about 2 rnin.). 1 THISJOURNAL, 9 (1917). 295. Fuchsin (0.2 9.) is dissolved in 120 cc. hot water. After cooling t o room temperature there are added 2 g. of anhydrous sodium sulfite dissolved in 20 cc. water, followed by 2 cc. concentrated hydrochloric acid. The solution is diluted to 200 cc. and is allowed t o stand 1 hr. before use. If well stoppered in an amber bottle it may remain fit for use for several weeks. The Schiff-Elvove reagent appears decidedly superior to the original Schiff reagent, and should supersede the latter.
Vol. 13, No. 6
4-Add 1 cc. concentrated H&Oa (C. P.), mix, andletstanda few minutes for temperatures to become equal. 5-Add 5 cc. Schiff-Elvove reagent, mix well, and let stand till colors are sufficiently developed (0.5 to 2 hrs.).
Each 1 cc. of the 0.04 per cent methanol in t h e standard tubes is equivalent to volume percentages of methanol in tdtal alcohol contained in the sample as follows: S A M P ~SOLUTION E A B
.................... .................... C ....................
Per cent 1 5 25
For more precise results the determination is repeated on the appropriate sample solution with more closely set standards. The sharpest results are obtained with standard tubes containing not over 1 cc. of standard methanol. T o bring the sample into this range it is often best t o use only 2 cc. of a sample solution, adding thereto 0.5 cc. of the 4 per cent ethyl alcohol and sufficient water t o make 4 cc. Approximate readings may be made after 30 min., precise ones after 1 hr., but best under 2 hrs., for t h e colors fade later. T h e limit of detection is 0.2 cc. of t h e standard 0.04 per cent methanol. Tests on four “unknown” mixtures of methanol, ethyl alcohol, and water prepared by a n assistant indicated t h a t , including t h e necessary determination of total alcohol via specific gravity, the results need not be in error by more t h a n 1 part in 20. Q U A L I T A T I V E METHOD
A modification of Denig2s’ method is official a s a qualitative test in the U. S. Pharmacopeia IX. The U. S. P. test has been criticized as unreliable because a false reaction sometimes occurs. Ehmanl attributes the fault t o temperature and overcomes i t by running a blank with pure ethyl alcohol, adjusting t h e temperature until the blank remains colorless. I n t h e judgment of t h e present writer the difficulty is primarily due t o a n undesirably high concentration of total alcohol. Since the substitution of phosphoric acid for sulfuric acid considerably more t h a n doubles t h e yield of formaldehyde from a given amount of methanol, t h e concentration of the sample in the test here proposed need be only half t h a t employed in the U. s. P. test and still leave the proposed test more delicate t h a n t h e U. S. P. test a t its best. The proposed test has been run a t temperatures of 15’ and 35’ C. without experiencing difficulty with false reactions.2 It may be conducted a s follows: Dilute t h e liquid, purified as necessary, t o a content of 5 per cent by volume of total alcohol. T o 5 cc. add 0.3 cc. of phosphoric acid (C. P . , 8 5 per cent), mix, add 2 cc. of a 3 per cent solution of potassium permanganate, mix, and let stand until the permanganate is entirely decomposed (about 10 rnin.). Add 1 cc. of 10 per cent oxalic acid, mix, and let stand till a clear brown (about 2 min.). Add 1 cc. concentrated sulfuric acid, mix, add 5 cc. Schiff-Elvove reagent, immediately mix well, and observe the color after exactly 10 min. The solution may then possess A m . J . Phorm., 91 (1919), 594. I t may be that in the U. S. P. test the presence of sulfuric acid promotes oxidation of ethyl alcohol to formaldehyde at an elevated tempera1 2
ture.
June, 1921
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
a pale greenish tint, b u t should show no distinct blue or violet color against a white background (less t h a n 0.2 per cent methanol in the total alcohol). I n carrying out the qualitative test i t is essential not t o be misled b y any colors developing in less t h a n 10 min. Concentrated sulfuric acid often becomes decidedly weak in t h e ordinary laboratory reagent bottle, and a transitory color from acetaldehyde may accordingly appear. This is also likely t o happen if the Schiff-Elvove reagent is not mixed with the solution immediately after addition. The color arising from acetaldehyde will have disappeared in 10 min. after mixing, but, needless t o say, i t is a safeguard against error t o run a blank along with t h e test. On longer standing, t h e test can naturally detect smaller proportions t h a n 0.2 per cent. PURIFICATION O F SAMPLES
The directions given for both quantitative and qualitative work specify t h a t t h e original material must be “purified as necessary.” I n general, the test must never be run directly on any material unless i t is positively known t o contain only water, alcohol, and other substances known t o be innocuous. Alcoholic preparations vary so widely t h a t no entirely general methods of purification may be given. The analyst can generally determine approximately the nature and amount of the nonalcoholic constituents, and must decide whether, in addition t o purification, it will be necessary t o run a blank on a synthetic mixture. CARBOHYDRATES A N D GLYCEROL-These substances, against which Salkowskil has given warning, are t o be separated b y distillation, a step which is also neces- sary t o permit determination of total alcohol via specific gravity. F O R M I C A N D A C E T I C Acms-These acids are stated by RosenthaleF t o yield color with Schiff’s reagent. They can be separated, if necessary, by distillation after neutralization, but the present writer did not find t h a t 10 per cent by volume of either acid added t o pure ethyl alcohol produced any color by the qualitative test. F O R M A L D E H Y D E , T E R P E N E S , ETc.-These impurities are removed b y von Fellenberg3 b y treatment with sodium hydroxide a n d silver nitrate, followed by distillation. PHEXOL-AS noted by S c ~ d d e r ,phenol ~ interferes with the test t o a degree dependent on its concentration. It may probably be adequately separated b y distillation after addition of a liberal excess of caustic alkali. FUSEL oIL-This has been stated5 t o afford a slight false reaction after oxidation. The present writer obtained one sample of “fusel oil,” and two of C. P. amyl alcohol (rectified fusel oil), one of the latter being a n “analyzed reagent,” all from different manufacturers. Each sample was made into a 10 volumeper cent solution in pure ethyl alcohol, and the qualita-
.
Z.Nahr.-Genussm.,
28 (1914),225. “Der Nachweis organischer Verbindungen,” 1914. 8 Biochem. Z.,86 (1918).45. 4 J . A m . Chem. Soc., 27 (1905),842. 6 von Fellenberg, Biochem. Z . , 86 (1918), 45; Salkowski, Z. Nahr.Genussm., 36 (1918),262. 1 2
545
tive test was applied. The heaviest color was given by the presumably purest sample, namely, the “analyzed reagent.” Upon making the qualitative test quantitative b y running i t in comparison with known mixtures of methanol and ethyl alcohol and letting stand a n hour or more, t h e color produced was found markedly fainter than the color produced from ethyl alcohol containing 0.08 per cent methanol. By the regular quantitative test the color was indistinguishable, being clearly less t h a n t h e equivalent of 0.1 per cent methanol. Hence the present writer has been unable t o demonstrate interference by fusel oil, provided t h a t it be not attempted t o strain the test beyond the limit recommended, namely, 0.2 per cent. ACETONE-This ingredient, constituting u p t o 10 per cent of the “total alcohol,’’ does not appear t o affect significantly qualitative or quantitative results. SUMMARY
The Deniges test has been modified t o increase sensitiveness and precision, a n d is recommended for practical work in t h e detection of, and especially in the quantitative determination of, methanol in t h e presence of ethyl alcohol, inasmuch as the possible normal presence of methanol in alcoholic products renders purely qualitative tests unsatisfactory. Though capable of greater refinement, the tests are adjusted t o a minimum limit of 0.2 per cent methanol in total alcohol. Procedures for t h e removal of certain interfering substances are outlined.
National Research Council Notes The Research Information Service of the National Research Council has recently compiled information about funds for scientific research. From this compilation it appears that there are hundreds of special funds, trusts, or foundations for the encouragement or support of research, in the mathematical, physical and biological sciences, and their applications in engineering, medicine, agriculture, and other useful arts. The income from these funds, which amounts annually to a t least fifty million dollars, is used principally for prizes, medals, research scholarships and fellowships, grants, and sustaining appropriations or endowments. So numerous have been the requests to the Research Council for information about sources of research funds, availability of support for specific projects, and mode of administration of particular trusts or foundations, that the Research Information Service has created a special file which i t is proposed to keep up to date in order to answer the questions of those interested in such funds. Furthermore, in order to give wider publicity to the immediately available information, the Council has issued a bulletin under the title “Funds available i n 1920 i n the United States of America for the encouragement of scientific research.” The Research Information Service is prepared also to assist investigators by locating scientific publications which are not generally or readily accessible. It will also, as is desired, have manuscripts, printed matter, or illustrations copied by photostat or typewriter. The cost of copying varies from ten to twenty-five cents per page. No charge is made for this service unless an advance estimate of cost has been submitted and approved by correspondent. Inquiries concerning the bulletin or for information about research funds or requests for assistance in securing copies of scientific articles should be addressed, National Research Council, Information Service, 1701 Massachusetts Avenue, Washington, D. C.
