The determination of small amounts of tetranitromethane in air.

Tetranitromethane (T.N.M.) occurs in the production of T.N.T. as a byproduct of toluene nitration. It is a highly toxic substance, and its toxic properties have been recently investigated by Sievers, Rushing, Gay, and Monaco (1947). According to them, the maximum allowable concentration of T.N.M. is less than 5 p.p.m. Elkins (1950) recommends 0-5 p.p.m. or 4 mg./m.3 of air as the safe maximum allowable concentration. Unless we are prepared to take very large air samples, the analytical method for determining T.N.M. should enable us to determine with precision as little as 20 jig., that is, the amount of T.N.M. in a 5 litre air sample, if the concentration of the contaminant equals the maximum allowable concentration. As the volume of absorption liquid usually does not exceed 10 ml., a reliable analytical method should register with sufficient precision a concentration of 2 ,ug. of T.N.M. per millilitre of the solution. Baillie, Macbeth, and Maxwell (1920) describe a method for determining T.N.M. based on the reduction of T.N.M. by hydrazine. The nitrogen evolved during the reduction is determined volumetrically. However, this method does not seem sensitive enough for industrial practice, since only 0*2 ml. of nitrogen is produced per 35 mg. of T.N.M. A colour reaction between dioxane and T.N.M. is described by Reid and Hoffmann (1929). This reaction is used to indicate the presence of dioxane, and we tried to invert and adapt it for the determination of T.N.M., but without success, its sensitivity being insufficient. The method of Kranz and Stepanek (1937), based on the fact that iodine is liberated when potassium iodide is added to a solution of T.N.M., can be used only for determining larger quantities of T.N.M. (0-05-0O12%). Ehrenberg, Fisher, and Lofgren (1945) determined small quantities of T.N.M. by means of the Griessllosvay reaction, which, however, is not specific for T.N.M., and other contaminants such as nitrous gases would interfere if present. The polarographic method of Cruse and Haul (1949) would be satisfactory with regard to sensitivity (9-8 to 98 ,ug./ml.), but if m-dinitrobenzene or s-trinitrobenzene are present, a preliminary separation of T.N.M. by vacuum distillation becomes necessary. Thus only the method of Sievers and others seemed to satisfy the requirements both with regard to sensitivity and simplicity. However, there remained several points to be cleared up before the method of Sievers and his colleagues could be accepted with complete confidence. The method is based on a colour reaction between T.N.M. and benzidine in ethanol. One of the problems studied in this paper is the stability of the colour produced by mixing alcoholic solutions of T.N.M. and benzidine. Sievers states that " the colour is sufficiently stable, but not stable enough to permit the use of the same set of standards for comparison with samples made up at other times ". When preparing the calibration curve according to the procedure of Sievers and others we noticed that the optical density of a sample made twice in the same way could not be reproduced. This induced us to study in the first place the influence of light on the intensity of the colour reaction. Other problems we studied were the optimum amount of benzidine, the influence of the design of the apparatus for measuring the optical density, the influence of ageing of the T.N.M. solution on the intensity of the colour reaction, the influence of various other contaminants such as nitrobenzene, HNO3, and finally the efficiency of the sampling procedure.