Determination of Ethylene A. W. FRANCIS
AND
S. J. LUKASIEWICZ, Socony-Vacuum Oil Co., Inc., Paulsboro, N. J.
For the determination of ethylene, 22% mercuric sulfate in 22% sulfuric acid is more satisfactory than 72 to 95% sulfuric acid activated with silver sulfate. The proposed reagent does not attack paraffins; it absorbs ethylene irreversiblyr and it does not absorb carbon monoxide or hydrogen. Carbon dioxide and olefins heavier than ethylene should be preliminarily removed from the gas sample. The reagent may b e used also for determination of total olefins.
B
ROMINE solutions and fuming sulfuric acid are unsatisfactory for the determination of ethylene because they attack paraffins (4, 7 ) . Concentrated sulfuric acid activated with 1% vanadium pentoxide or 0.6% silver sulfate (4)is unsatisfactory for the same reason ( 3 ) . A solution of 3.5% silver sulfate in 72% sulfuric acid (3), although satisfactory in this respect, has been found to be unreliable because the absorption of ethylene in it is reversible and therefore not quantitative. I n fact, after a few determinations a sample of air or inert gas actually desorbs ethylene from the reagent and increases in volume. The dissolved ethylene is not gradually hydrated, because even after 6 months’ standing it can still be desorbed. Moreover, it was found that 87% sulfuric acid containing 9% silver sulfate absorbs carbon monoxide, as well as liberates ethylene after use. A reagent which satisfies the requirements of complete irreversibility and selectivity for olefin gases with respect to most other probable gases present is a solution of mercuric sulfate in 22% sulfuric acid. Other mercury salts likewise absorb olefins. The solubility of mercuric chloride in water is so low, 7%, that absoiption of ethylene is slow. Mercuric nitrate ( 2 ) and mercurous nitrate solutions, which require nitric acid to prevent hydrolysis, may become mildly explosive when combined with large amounts of olefins. They also leave acid gases in the sample. Mercuric acetate, which forms a 25% solution in water without hydrolysis, is satisfactory except for the theoretical objection that the aqueous vapor pressure from it is slightly higher than that of the confining liquid in the gas buret, concentrated sodium sulfate. Furthermore, a drop of either solution in the other gives a heavy precipitate. Mercury complexes with olefins are reviewed by Huge1 and Hibou ( 6 ) , and the structures of some of them are elucidated by Adams, Roman, and Sperry ( 1 ) . Winbladh (9) determined isobutene with a solution of mercuric oxide in sulfuric acid, but other olefins by bromination. It is not clear from the abstract how he distinguished between isobutene and the other olefins.
curic nitrate ( 2 ) . Carbon dioxide also is absorbed slowly by mercuric sulfate and so should be first removed with potassium hydroxide solution. Carbon monoxide and hydrogen are not absorbed by the solution. TESTS
Following are some tests with this mercuric sulfate solution in comparison with Eberl’s silver sulfate solution. Both solutions were made up fresh and used for these analyses alone.
A. A gas mixture containing 35% of ethylene and 65% of propane by volume was prepared using Matheson’s chemically pure ethylene and propane. Samples of 100 ml. were analyzed in a conventional Orsat apparatus using Francis autobubblers. A sample was passed through the reagent (about 30 seconds each way) until the absorption was complete or until a constant reading was attained: Sample Absorption in HgSOa, ml. Absorptionin AgpSOd, ml.
1 35.0 34.6
2
35.0 34.8
3 35.0 33.8
4 35.0 33.4
5 32:8
Ten passes were required for complete absorption in mercuric sulfate solution, and the results agreed. With Eberl’s solution fifteen passes were required, and the absorption became incomplete as the reagent was used repeatedly. The fifth sample showed a decrease of about 6% of the apparent olefin content; but when the sample was passed into the mercuric sulfate solution, 2.2 ml. additional were absorbed. B. A synthetic mixture free from air was made containing 40% ethylene and 60y0 n-butane using Matheson’s chemically pure gases, and analyzed with the same solutions as in A. From 100-ml. samples the mercuric sulfate dissolved 40.0 ml. but the silver sulfate solution only 37.6 ml. When the latter sample was passed into the mercuric sulfate, the remaining 2.4 ml, of ethylene were absorbed. C. DESORPTION. A sample of 80.0 ml. of Matheson’s chemically pure n-butane was passed into the silver sulfate solution used in A and B (six samples). After twelve passes the volume had increased to 83.2 ml. This sample was reduced again to 80.0 ml. by passing into the mercuric sulfate solution three times. A fresh sample of n-butane showed no change in volume on passing into the mercuric sulfate reagent several times. A pipetful of Eberl’s silver sulfate solution was saturated with 600 ml. of ethylene, and left standing. Occasionally a fresh sample of 90.0 ml. of air was bubbled into this solution (one pass), returned to the buret, and measured as follows: Time, days Volume, ml.
1 97.9
3 97.6
8 96.0
17 96.0
33 96.3
101 96.3
200
94.2
The volume of desorbed ethylene is evidence that hydration by the acid is extremely slow, if it occurs a t all.
The recommended solution is made up by dissolving about 57 grams of mercuric sulfate in 200 grams of 22% sulfuric acid, or by dissolving 41 grams of mercuric oxide in 216 grams of 29% sulfuric acid, and filtering if necessary to remove any undissolved particles. A concentration of acid below 13% permits hydrolysis (yellow precipitate), and a concentration above 24% diminishes solubility (white precipitate). This is illustrated in Figure 1, which presents data recalculated from those of Hoitsema (6). A solution of any composition below the curve may be made and used. The recommended solution has a specific gravity of about 1.37. Since there is some sludge formation on continued use, it may be preferable to use the reagent in a pipet packed with beads or tubes rather than in a Francis autobubbler, whose holes might become plugged. The ultimate capacity of the reagent is calculated on an equimolar basis as 25 volumes or 5 liters in a 200-ml. pipetful. After 3 liters of pure ethylene had been absorbed, however, the rate of solution had become excessively slow. A 90-ml. sample of air passed into this solution still returned unchanged in volume, indicating no reversible absorption. Since the reagent also absorbs other olefins, it is necessary to remove them first when present with appropriate concentrations of sulfuric acid ( 7 ) . Butadiene is dissolved as in the case of mer-
PERCENTAGE S U L F U R I C ACID
Figure 1. Solubility of Mercuric Sulfate in Aqueous Sulfuric A c i d
703
INDUSTRIAL A N D E N G I N E E R I N G CHEMISTRY
704 Table
dample No.
1 2 3 4
CsHa 14.0 13.9 13.3 7.3
.
Mass Spectrograph Analysis, M o l e
CaHa 0.5 0.3 0.5 0.6
C4H8 0.3 0.3 0.1 0.0
’
Iso-CIHI~n-CdHe 85.2 0.0 85.5 0.0 86.0 0.1 91.5 0.0
7’
Total Olefin 14.3 14.2 13.4 7.9
HnSOi Anilysia, Mole Totap Olefin 14.2 14.3 13.3 8.1
D. EFFECT OF HYDROGEN AND CARBON MONOXIDE.Samples of pure hydrogen and carbon monoxide and duplicate samples of a mixture of 51% carbon monoxide and 49% ethylene were analyzed with the mercuric sulfate solution in a pipet packed with vertical tubes, leaving the sample in the pipet 30 seconds between pasces: Passes 0 1 2 3 4 5 6 I
10
Pure Hr
Pure CO
M1.
M1. 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Mixture (49% CBHI) Ml. M1. 100.0 100.0 82.0 84.3 71.4 73.6 63.9 65.6 56.2 58.8 52.3 53.6 51.1 51.1 50.9 50.9 50.9 50.9
E. COMMERCIAL SAMPLES.Twenty-nine samples of experi-
mental gases which varied in total olefin content from 2 to 30% were analyzed both by the mass spectrograph (8) and by ab-
Vol. 17, No. 11
sorption into the mercuric sulfate reagent. Results by the two methods differed by an average of 0.7%, which is the order of uncertainty of the former method. Analyses of four of these samples, which happened to be free of ethylene but contained higher olefins, are indicated in Table I. ACKNOWLEDGMENT
The authors are indebted to W. A. Stover of this laboratory for making these 29 analyses available to them. LITERATURE CITED
(1) Adams, R . , Roman, F. L., and Sperry, W. N., J. Am. Chem. SOC., 44, 1781 (1922). (2) Cuneo, J. F.,and Switaer, R. L., IND. ENQ.CHEM.,ANAL.ED., 15, 508 (1943). (3) Eberl, J. J., Ibid., 14,853 (1942). (4) Gooderham, W.J., J . SOC.Chem. Ind., 57,390T (1938). (5) Hoitsema, C.,2. phys. Chem., 17, 604 (1895). (6) Hugel, G., and Hibou, J., Chimie et Industrie, Speoial No. 296 (Feb., 1929). (7) Matuazak, M. P.,IND.ENQ.CHEM.,ANAL.ED.,10, 354 (1938). (8) Washburn, H. W., Wiley, H. F., and Rock, S. M., Ibid., 15, 541 (1943). (9) Winbladh, R., Ing. Vetenskaps. A k a d . Handl., No. 138, (1936). PBRBE~NTED before the Diviaion of Bnalytical Chemistry at the Spring MeetC H ~ M I C SOCIETY, AL June 13,1945, ing of the Philadelphia Section, AMERICAN and the Division of Petroleum Chemistry at the Meeting-imprint of the AMERICAN CHE~YICAL SOCIETY, September, 1945.
Colorimetric Determination of
DDT
Color Test for Related Compounds MILTON S. SCHECHTER, S. 8. SOLOWAY, ROBERT A. HAYES, AND H. L. HALLER U. S. Department of Agriculture, Beltsville, Md.
Bureau of Entomology and Plant Quarantine, Agricultural Research Administration,
A colorimetric method has been developed for the estimation of small amounts of DDT down to about 10 micrograms. The method involves intensive nitration and the production of colors by the nitrated products in benzene plus methanolic sodium methylate. This color reaction can also b e used as a test for degradation products of DDT and some compounds related to it.
THE
extraordinary development of the insecticide commonly known as D D T (1, 7) has made the need for a sensitive method of detection and determination rather urgent. h method which could detect small amounts of D D T would find application in such fields of study as spray-residue determinations, water analyses, and pharmacological investigations. hiuch of the analytical work on D D T has depended on chlorine determinations. Either the “labile” chlorine split out on dehydrochlorination by alcoholic alkali can be determined, as recommended by Neal et al. (21)and by Gunther (9),or else the total chlorine can be determined by some method such as the Parr bomb, Carius, or Umhoefer (26),or by a modification of the Winter method proposed by Hall et al. ( I d ) . The labile-chlorine method determines only 1chlorine atom per molecule of DDT, whereas the total-chlorine methods determine 5 chlorine atoms per molecule. If D D T completely decomposes to dehydrochlorinated DDT, the former method would yield no chlorine while the latter group would determine 4 chlorine atoms per molecule. If a total-chlorine method is used as the sole method of determination, no measure of decomposition of the DDT can be obtained. Both labile and total organic chlorine must be determined in order to prove the presence of DDT or to detect its decomposition. All these chlorine determinations run into difficulty when the amount of D D T is less than about 1
mg., and they lack specificity. Furthermore, there is no method based on chlorine determinations by which the amounts of p , p f D D T and o,p’-DDT present in mixtures can be estimated. The terms used in this paper to designate D D T and related compounds are as follows: The generic term “DDT”, originally abbreviated from dichlorodiphenyltrichloroethane,refers to the technical product, which ordinarily contains 70 to 77% of p,p‘D D T [l-trichloro-2,2-bis(p-chlorophenyl)ethane]and 15 to 25% of o,p’-DDT [l-trichloro-2-o-chlorophenyl-2-p-chlorophenyl)ethane]. One of the minor constituents is l,l-dichloro-2,2-bis(pchloropheny1)ethane which has been designated as p,p’-DDD ( 2 6 ) . Gunther (11) has pointed out his error concerning the term “p,p’-DDD” made in a previous article (IO). This compound has been named “1,l-dichloro-2,2-bis(p-chlorophenyl)ethane” in the present paper in conformity with the latest Chemical Abstracts nomenclature. The chemical composition of technical DDT is described by Gunther (10) and by Haller Bartlett, Drake, Newman, and others (13). Dehydroch1or)inated p,p’-DDT [ l,l-dichloro-2,2-bis(p-chlorophenyl)ethylene]is a decomposition product and bis(p-chloropheny1)acetic acid (8, 2 7 ) , which has been called p,p’-DDA, is a metabolite of p , p ’ DDT.
A search was made t o find a suitable color reaction for D D T which could be made the basis of a colorimetric analytical method. Some of the exploratory work done in this direction is outlined below : Tests for the trichloromethyl group using pyridine and alkali, resorcinol and alkali, or @naphthol and alkali, as described by Snell and Snell (23), were all negative. Boiling ethanolic silver nitrate gave no precipitate of silver chloride. Nitration of DDT, reduction, and diazotization, followed by coupling with a suitable compound, give a color (orange with &naphthol). Although this line of attack could probably be developed into a method for the analysis of DDT, it was not pursued further because it was
