Determination of small amounts of methyl chloride in air - Analytical

F. A. Patty, H. H. Schrenk, W. P. Yant. Ind. Eng. Chem. Anal. Ed. , 1932, 4 (3), pp 259–262. DOI: 10.1021/ac50079a006. Publication Date: July 1932. ...
0 downloads 0 Views 589KB Size
July 15, 1932

INDUSTRIAL AND E N G I N E E R I N G CHEMISTRY

hydrolysis of the reaction products will gradually diminish. The source of this hydrolysis is evidently not traceable to acetyl derivative of hydroxy acids, for the behavior of the castor oil belies such an assumption, nor is it brought about by the breaking down of mixed anhydrides. Rather it appears to be due to the presence of either unstable acetylated monoand diglycerides, or some acetylated oxidation products of the fatty acids. The observed instability of an acetylated mono-n-valerin under the above conditions (it was found that the acetyl number of the parent compound cannot be determined at all by the official method (2) and only approximately by the proposed one) supports the former view. A very important step in the determination of the acetyl number under the present mode of procedure ( 2 ) is that excessive washing of the reaction mixture be avoided. In fact, Lewkowitsch (9) cautioned against more than three applications of wash water. Yet, even when this precaution is observed, it is necessary to use at least 1.5 liters of wash water. No such situation arises, however, in carrying out this determination under the proposed procedure, for when conditions are such as to make a blank determination desirable (see description of method), the volume of water necessary to hydrolyze the anhydrides of the soluble fatty acids which have been formed is small; in fact it is not enough to cause a measurable hydrolysis of the major acetylated product. The presence of insoluble (higher) fatty acids does not introduce an error, for the action of acetic anhydride upon them is such as to form their corresponding anhydrides (9). These are very stable, even on continued boiling. Since equivalent quantities of acetic acid are fcjrmed in this reaction, none of its anhydride is actually consumed, a condition which would make a blank determination unnecessary in this case. By way of comparison of the two methods for determining acetyl numbers, there is included in this report a set of typical data (Table 11). It will be observed that there is a satisfactory agreement in results when the necessary correction for

259

interfering substances has been applied. Hydroxyl and acetyl numbers are practically the same in the bracket below twenty. OF OFFICIAL AND PROPOSED METHODS TABLE11. COMPARISON FOR DET~RMINATION OF ACETYLNUMBER

ACETYL NUMBER 05cia1 Proposed method method

MATERIALS Oils: Olive Cottonseed Sesame Tobacco seed (extracted) Tobacco seed (expressed) Peanut Rye germ Corn

5.0 5.9 3.3 9.0 4.8 5.4 20.2" 16.7b 15.4b 146.3

4.9 5.5 3.8 8.6 5.3 5.1 21.0a 17.7 18.1 146.4

HYDROXYL NUMBER 4.9 5.5 3.8 8.7 5.3 5 1 21.5'' 17.9~ 18.3d 164.5

Castor Waxes: Beeswax 24.2s 26.3 26.8 Carnauba 44.7 44.7 46.2 a Analyses by Albert W. Stout. b Values calculated from saponification number of washed original sample. They will be several points loa-er if carried out by official method. c Filtered before titration, correction for blank was applied. d Titrated in presence of acetylated product. No blank applied. B Some constituents apparently lost along with coloring matter which dissolved in acetic anhydride during the process of acetylation. Original product was bright yellow, final a pure white.

LITERATURE CITED (1) And& E., Compt. rend., 172, 984-6 (1921); Bull. SOC. chim., 141 29, 745-62 (1921). (2) Assoc. Official Agr. Chem., Methods, p. 326, 1930. (3) Benedikt, R., and Ulser, F., Monatsh., 8 , 41-8 (1887). 14) Cook, L., J . Am. Chem. Soc., 44, 392-4 (1922). (5) Jamieson, G . S., J. Assoc. Oflcial Agr. Chem., 8, 484-9 (1925). (6) Jamieson, G . S., Zbtd., 9,247-53 (1926). (7) Jamieson, G. S., Ibid., 10, 323-9 (1927). (8) Lewkowitsch, J., J . SOC.Chem. Znd., 16, 503-6 (1897). (9) Lewkowitsch, J., Analyst, 24, 319-30 (1899). RECEIVEDApril 23, 1932. Presented before the joint meeting of the Division of Agricultural and Food Chemlstry and the Division of Biological Chemistry at the 83rd Meeting of the American Chemical Society, New Orleans, La., March 28 to April 1. 1932.

Determination of Small Amounts of Methyl Chloride in Air F. A. PATTY, H. H. SCHRENK, AND W. P. YANT,U. S. Bureau of Mines, 4800 Forbes St., Pittsburgh, Pa. N CONDUCTIKG an investigation of the toxicity of small amounts of methyl chloride in air, it was necessary to check the computed concentrations by chemical analyses. Of the various methods (1, 4, 5 , 6) reported in the literature, none of them appeared suitable for the purpose a t hand. Attempts were made, therefore, to develop a method which would be satisfactory for the conditions outlined, and thus a procedure which proved satisfactory was devised. Allison ( 1 ) determined methyl chloride by absorption in glacial acetic acid and also by burning with an excess of oxygen in an Orsat gas apparatus. McKee (4) and Kicloux (6) also used a combustion method similar to that of Allison. Roka and Fuchs (6) heated methyl chloride with methanol and sodium iodide in a pressure flask to form methyl iodide, which they distilled into silver nitrate. It is obvious that none of these methods would be suitable for small amounts of methyl chloride in air-as, for example, 50 p. p. m. by volume (3). BUREAUOF MINES METHOD The procedure adopted is very similar to the Referees method for determining total sulfur in fuel gases. The air

containing the methyl chloride is mixed with natural gas and burned in a microburner. The halogen products formed combine with ammonia obtained from ammonium carbonate cubes placed around the burner, and also with ammonium hydroxide formed by the ammonia from the ammonium carbonate and the water in the products of combustion of the gas. The chlorides produced are collected and determined by the Volhard method. APPARATUS.Figure 1 shows the apparatus. The methyl chloride-air mixture is added as the primary air supply to burner a, shown in detail in Figure 2 and described later. Secondary air enters around the base of the burner. The products of combustion of the fuel gas, the methyl chloride and excess secondary air, are carried by convection up through the trumpet tube b and are impregnated with ammonia which emanates from the ammonium carbonate cubes piled around the burner (Figure 2). I n the presence of the water vapor some of the halogen reacts with the ammonia and is deposited on the walls of the upper part of the trumpet tube. The remainder enters a glass marble-filled absorption tower, c, where the surfaces are wet with ammonium hydroxide formed by the condensation of water vapor in the presence of am-

260

ANALYTICAL EDITION

Vol. 4, No. 3

monia. This absorbs additional chloride introduced into a stream A method for the determination of small halogen products of combustion. of air. amounts of methyl chloride in air is described. The excess condensate drips from APPARATUS AND PROCEDURE The apparatus and procedure are in many the tower into a beaker. The FOR M E A S U R I N GM E T H Y L respects similar to the Referees method for total effluent vapor from the tower is CHLORIDE. The methyl chlosulfur in fuel gases. Tests of the method using led through a Cottrell precipitaride was measured with a calitor which collects any halogenbrated gas m i c r o b u r e t which amounts of methyl chloride ranging from 12.36 bearing smoke or fog that esused dry mercury as the disto 30.74 mg. have shown the error to be less than capes the tower. The precipitaplacing liquid. The buret was 2 per cent of the amount present for the lower, s u r r o u n d e d b y a w a t e r tor is made by sealing a platinum and less than 1 per cent for the upper limits of wire concentrically in a glass jacket equipped with a therthe range used. I n developing the method, a tube, d, which has a side outlet a t mometer, the water being kept in each end. The outside of the c i r c u l a t i o n by bubbling air range of concentrations from 200 to 1600 p . p . m. tube is w r a p p e d with tin or it. A by-pass was arthrough or 0.02 to 0.16 per cent by volume was used. copper foil to within about 4 ranged so that the buret could The method has also been used for the determinainches of each end, and the foil is be flushed with methyl chloride, tion of dichlorodifluoromethane vapor in air, bound in place with friction tape. and the waste gas c o n d u c t e d and is suggested for the determination of other A wire wound around the foil from the room to prevent conwith one end protruding through t a m i n a t i o n of the atmosphere organic halide gases and vapors. the tape serves as one terminal which was used as the secondary and the platinum wire as the air supply to the burner. The other, both being connected to samde of methvl chloride in the secondary of aFord spark coil, e. The current for the spark the microburet was measured a t atmospheric pressure and the coil is obtained from a 110-volt 60-cycle lighting circuit through prevailing temperature, after which it was corrected to 0" C. a toy transformer, f,adjusted to give 6 volts. It is obvious and 760 mm. Hg. A minimum of stopcock grease was that other designs of electrical precipitators may be used, pro- used. viding they readily permit a washing out of the precipitate. CHEMICALS AND REAGENTS USED. The methyl chloride Figure 2 is a detailed sketch of the burner. It is con- used was obtained from the Roessler and Hasslacher Chemical structed by placing a glass T tube tightly over the burner Co., and according to specifications was 99.5 per cent pure. stem of an ordinary laboratory microburner, g, and sealing the The silver nitrate solution used was approximately 0.025 N primary air intake with wax, h. A piece of platinum gauze, i , and was standardized against standard hydrochloric acid. is fused onto the burner tip to stabilize the flame. The con- The hydrochloric acid was standardized using sodium carnection between the stem and the arm of the T tube is made bonate prepared by heating sodium bicarbonate. The potasgas-tight by means of a rubber tubing collar. The fuel gas sium sulfocyanate was standardized against the silver nitrate. enters a t j and the methyl chloride-laden air a t k. Secondary Ferric alum was used as an indicator. Calibrated burets were air enters between the trumpet tube wall 2 (shown in entirety used in making all titrations. DILUTIONAND BURNING OF SAMPLE. The methyl chloride in Figure 1) and a cork base, m, which supports ammonium carbonate cubes, n. which was confined in the microburet was forced into a The cork is covered stream of air by opening the stopcock a t the top of the buret with a sheet of as- and slowly displacing with mercury. The air was measured 110 volts 8. e. bestos, 0. This cork and the rate controlled by means of a dry meter, and the mixis smaller than the ture led through glass tubing to the microburner previously base of the trumpet described. All connections were glass-to-glass held in place tube in order to per- by rubber tubing. The volume of the samples varied from mit secondary air to approximately 5.5 to 13.5 cc. of methyl chloride. The air be drawn by convec- was admitted through the meter a t the rate of about 20 liters . was tion or the stack- per hour. The time of burning for the 1 3 . 5 ~samples effect of the trumpet approximately 1 hour, for the 10-cc. samples from 20 minutes t u b e a n d t o w e r to 1 hour and 20 minutes, and for the 5.5-cc. samples from 45 minutes to 1 hour and 30 minutes. This gave a range of above. The above m m e r concentrations varying from about 200 to 1600 p. p. m. by design, and in fact volume, or 0.020 to 0.16 per cent. Blank determinations much of the other were made on room air and gave an average titration of e q u i p m e n t d e - 0.05 cc. of silver nitrate, the equivalent of which was 0.07 scribed, was an as- mg. of methyl chloride. This blank titration was subtracted s e m b l y of p a r t s from the silver nitrate values in the determinations. DETERMINATION OF CHLORIDES.The chlorides were readily available in t h e l a b o r a t o r y . washed from the trumpet tube, Cottrell precipitator, and FIGURE1. APPARATUS FOR DETERMINOther designs and absorption tower until the washings gave a negative test for ING METHYLCHLORIDE IN AIR p a r t s w h i c h will chlorides. It was found that this was best accomplished in carry out the prin- the tower by quickly pouring 4 or 5, 50-cc. portions of water ciple of the method will be satisfactory. This also applies to over the beads. This procedure prevented channeling and the fuel, which may be any combustible gas not containing an produced a solid column of water which resulted in more effiappreciable amount of halogen compounds. A blank determi- cient scrubbing. The washings were added to the condensate nation using the fuel employed must, of course, be made. which collected and dropped from the tower during the combustion, and the chlorides were then determined by the EXPERIMENTS MADETO TESTMETHOD Volhard method. The silver chloride was removed by filtraThe accuracy of the method was tested by making a series tion through a Gooch crucible prior to the sulfocyanate titraof determinations of carefully measured amounts of methyl tion.

A

lNDUSTRIAL AND ENGINEERING CHEMISTRY

July 15, 1932

261

TABLE I. DETERMINATIONS WITH KNOWN AMOUNTS OF METHYL may be taken by air displacement in containers of suitable CHLORIDE size and delivered to the apparatus by sweeping into the VOL. O F burner with a stream of air; or if the samples are relatively CHsCl VOL. AVER- AgNOs CHsCl (99.5% AT CHsCl OF AQE TITRA-RECOV- RECOVsmall they may be delivered to the apparatus by mercury 0 ° C . 7 6 0 ~ ~TAKEN .) AIR CONCN. TION ERED ERY ERROR displacement. Owing to solubility of methyl chloride in cc. MQ. Liters P.p . m. Cc. Ma. % % water, the use of water as a displacing medium should be 9B. 6 600 21.14 30.00 -0.4 13.43 30.12 22.5 21.87 31.03 +0.9 610 100.9 13.71 30.74 22.5 avoided, though if absolutely necessary, it may be used if care 99.7 21.53 30.56 -0.3 710 13.67 30.68 19.0 20.0 100.0 30.32 13.51 30.32 670 2 1 . 3 7 0.0 is taken to avoid operations that facilitate solution of the 100.6 7 . 0 1250 15.85 22.47 22.32 9.95 4-0.6 gas in the liquid. If the water is not agitated a great deal or 100.3 1600 6.3 16.00 22.68 22.60 10.07 4-0.3 99.4 410 -0.6 15.85 22.47 22.60 24.7 10.07 left in contact with the gas-laden air, i t is possible to use it 99.6 940 10.7 -0.4 15.79 22.39 22.48 10.02 12.70 101.5 12.57 370 13.5 4-1.5 5.60 9.02 without vitiation of the results for practical health investiga99.2 12.57 19.0 12.46 290 -0.8 5.60 8.81 tions. Experiments were made in which methyl chloride99.4 12.52 17.2 330 12.43 5.58 8.79 -0.6 101.3 200 12.54 12.38 2 7 . 5 8.87 4-1.3 5.52 air mixtures were passed from a mercury-filled buret into 19.0 12.43 100.5 4-0.5 12,36 290 8.79 5.51 distilled water contained in a Haldane type gas-absorption RESULTS.The results of the determination, given in pipet at room temperatures ( 2 ) . Table I1 gives the results of Table I, show that the accuracy of the method with amounts these experiments. The practical use of methyl chloride ranging from about 20 to 30 mg. is in the neighborhood of 0.5 per cent, being a t least within 1 per of water as a discent in all cases. When amounts as low as 12 mg. are deter- placing liquid was mined, the accuracy is nearer 1 per cent, being better than 2 also studied. Three determinations were per cent in all cases. The equivalent of 0.1 cc. of the silver nitrate solution used made using 19-liter was 0.14 mg. of-methyl chloride. A titration error of 0.1 cc. s a m p l e s of 600 would cause an error of approximately 1 per cent in samples p. p. m. m e t h y l weighing about 12 mg. Therefore, the per cent error of the chloride-air mixture 2 method depends primarily on the weight of methyl chloride obtained in and dein the sample taken rather than on the concentration. This livered from a 20fact should be borne in mind in any attempt to extend further l i t e r a s p i r a t i n g bottle by use of disthe range or accuracy of this method. PROCEDURE USEDFOR ROOMATMOSPHERES CONTAININGtilled water as the METHYLCHLORIDE.The above method has been used in displaced and disthis laboratory to check concentrations of methyl chloride placing m e d i u m . ranging from 50 to 600 p. p. m. by volume in air which were The results obtained used for toxicity experiments with animals. The atmosphere were 98, 95, and 96 per c e n t , respecFIGURE 2. DETAIL OF BURNER to be analyzed was taken from the animal exposure chamber tively, of the calcuwith a small rotary air pump and delivered through a dry meter to the burner. A by-pass between the blower and the lated amount of methyl chloride in the air. The surface of meter permitted the excess air from the blower to return to the water was disturbed very little, but no special precauthe chamber. The rate of burning and size of flame depended tions were used to prevent motion other than removing and upon the concentration of methyl chloride in the air, a slower admitting water under the water surface, The line to the burner was flushed by drawing a few liters rate and smaller flame being used for higher concentrations. of room air into the aspirating bottle and then forcing this air The above procedure gave results which were within 5 per cent of the values computed from flowmeter measurements of to the burner. I n judging the size of samples, the allowable per cent error the volume of methyl chloride and air used in the mixture that entered the gas chamber. The results were always lower than and the probable error of method must be considered. the computed values and, aside from flowmeter errors and USE OF METHOD FOR OTHER ORGANIC HALIDE GASES control, may be partially explained by the fact that the AND VAPORS Cottrell precipitator was not used in this particular set-up of The same apparatus has been used to determine dichloroapparatus. fluoramethane, and dichlorodifluoromethane in concentraTABLE11. ABSORPTION OF METHYL CHLORIDE (from a 1 per tions of 2 and 20 per cent, respectively. In dealing with these cent methyl chloride-air mixture) BY CONTACT WITH DISTILLEDrelatively high concentrations, the samples were taken over WATERIN GAS-ANALYSIS PIPET mercury and diluted with air prior to combustion. The ABSORPTION OF CHsC1, BLANKUSINGROOM results agreed well with other methods of analysis, and point No. OF PASEEB ACCUMULATIVE AIR A ~ C U M ~ L A T I V E to the applicability of this method to the determination of INTO PIPET Sample 1 Sample 2 Sample 3 ABSORPTION % % % % other organic halide gases and vapors in air. 1

5 10 20 30 40 50 60 70 80 90

100 a

5a 15 26.5 45 50 “a a a a a

5a

a 20 40 46 57.5 65 75 75 750

a

a

0;04

i.5 19.5 35.0 45a 62.5 66.5 71.5 72.6 72.5 72.5

0.03 -0.01 -0;Ol

c

a a a

a a a a

Not determined.

The sampling procedure, as described, may be modified in accordance with other common procedures for sampling gases. For example, it is obvious that samples of atmospheres

ACKNOWLEDGMENT The development of the method described in this report was a necessary part of an investigation of the response of animals to methyl chloride in air which was conducted cooperatively by the Roessler and Hasslacher Chemical Company and the Bureau of Mines. The development was carried out at the Pittsburgh Experiment Station of the Bureau of Mines under the direction of its chief surgeon, R. R. Sayers. LITERATURE CITED (1) Allison, V. C., and Meighan, M. H., J. IND. ENQ.&EM., 94-6 (1911).

11,

ANALYTICAL EDITION

262

(2) Burrell, G. A., and Seibert, F. M., U. S. Bur. Mines, Bull. 197 (19261, (3) Martinek, M. J., and Marti, W. C., IND.ENG.CHEM.,Anal. Ed., 3, 408-10 (1931). (4) McKee, R. H., and Burke, s. p., IND.ENG. CHEM.,15, 578-9 (1923).

Vol. 4, No. 3

(5) Nidloux, M., and Scotti-Foglieni, L., Compt. rend. SOC. b f d , 98, 225-8 (1928). (6) Roka, K., and Fuchs, O., 2. anal. Chem., 71, 381-6 (1927). RBOEIVBD October 22, 1931. Published by permission of the Director, U. S. Bureau of Mines. (Not subject t o copyright.)

Some Titer Points of Mixed Fatty Acids 11. Mixtures of Pure Fatty Acids GEORGEW. JENNINGS,355 Marlborough Road, West Palm Beach, Fla.

T

HE titer points of fatty acids and of mixtures of fatty

acids is of such importance commercially, both in evaluating fats and in using them, that a better understanding of their behavior is desirable. Continuing the work reported previously (S), the titer points of mixtures of relatively-pure fatty acids were determined. It was expected that these figures would furnish indications, at least, of the causes underlying the seemingly erratic results obtained in several cases with commercial fats and oils. This was found to be true, but the behavior observed needs elucidation and consequently no explanation can be attempted until further data have been secured. The particular fatty acids predominating in the fats or oils to be mixed apparently determined the character of the resulting titer curve. The several other fatty acids usually present, although in much smaller amounts,

influence the curve somewhat, but to just how great an extent has not yet been determined. The four fatty acids selected for this work were lauric, myristic, palmitic, and oleic. The selection was determined principally by the fact that mixtures of the glycerides of these (together with that of stearic acid) in varying proportions form the majority of the commonly occurring nondrying oils and fats. The other acids, such as caproic, caprylic, capric, linoleic, etc., are present in much smaller quantity. The lauric, myristic, and palmitic acids were the specially purified grade of the Eastman Kodak Company. The oleic acid was purified in the laboratory, using a high-grade commercial red oil as the source, by the lead salt-ether method with subsequent fractionation under reduced pressure. That none of these acids was strictly chemically pure is recognized. It is a very difficult matter to effect the absolute separation of the fatty acids from one another (Z?), especially when acids close to each other in the series are under consideration, and for the present purpose strict chemical purity is not essential. The acids used were, however, of a relatively high degree of purity. The titer point is not a value of as high an order of accuracy as are some of the other physical constants. It varf certain limits according to the method employed mining it, the exact technic exerting considerable influence upon the result. The amount of fatty acid used, the method of stirring, the differential in temperature of fatty acid and surrounding bath, and the number of times the titer has been read on the same portion of fatty acid are a few only of the variables which must be standardized because no particular method can be considered the only correct one. Because of this factor of inherent variability in the titer itself, absolute purity of the fatty acids was not considered essential.

PROCEDURE The fatty acids were kept dry and, just before determining the titer point, were weighed out carefully in the desired proportions in a 50-cc. glass beaker; they were heated to about 115' C., thoroughly mixed, and, after cooling somewhat, were transferred to the titer tube. Determination of the titer from this point on was in accordance with the method of the Fat Analysis Committee of the AMERICAN CHEMICAL SOCIETY (1).

The results are reported in the form of curves which are plotted with the titer points as ordinates and the percentage composition of the mixed fatty acids as abscissas. OF LAURIC AND FIGURE1. TITERCURVEFOR MIXTURES MYRISTIC ACIDS OF LAURIC AND FIGURE 2. TITERCURVE FOR MIXTURES PALMITIC ACIDS FIGURE3. TITERC

DISCUSSION OF RESULTS As it was desired to learn in the f i s t place the effect of mixing two acids adjacent to each other in the same homologous series (considering only acids with an even numberof carbou atoms), lauric and myristic were selected. The curve resulting