184 



THE ALUMNI JOURNAL. 



The assay is best made with a deci- 

 normal solution oi potassium hydrate in 

 absolute alcohol, but it may also be ac- 

 complished with an aqueous one (see 

 below). Absolute alcohol is preferable 

 for the reason that it is always kept in 

 glass bottles, and therefore does not con- 

 tain matters producing a very dark tint 

 with caustic alkali, and that the solution 

 will not turn materially darker on keep- 

 ing, as would be the case if ordinary 

 alcohol had been used. The nitro- 

 glycerin should be in alcoholic solution. 

 The assay is made in this way that a 

 weighed quantity of the nitroglycerin 

 solution is added to a measured volume 

 of the decinormal potassa solution, so 

 that the latter shall remain in excess, 

 and that this excess is then determined 

 by decinormal acid. The alcoholic char- 

 acter of the solution causes a separation 

 of most of the salts which are formed, 

 together with most of the brown color- 

 ing matter, and it is easy to titrate the 

 remaining solution with phenolphtalein 

 as indicator. 



Example- — Assuming that a solution 

 supposed to contain io ( /c of nitroglycerin 

 is to be assayed. Into an Erlenmeyer 

 flask introduce 20 Cc. of decinormal al- 

 coholic potassa, heat it moderately, and 

 then add to it, in several portions, 10 Gm. 

 of the nitroglycerin solution, finally rins- 

 ing the vessel which had contained the lat- 

 ter with a little absolute alcohol and add- 

 ing this to the mixture. Test the liquid 

 with litmus paper to ascertain whether 

 it is still alkaline. If it is not, this is a 

 sign that the amount of alkali was in- 

 sufficient to decompose all the nitro- 

 glycerin, and that some of the latter is 

 still present. In this case, add another 

 portion (10 or more Cc. , carefully meas- 

 ured fiom a burette) of the volumetric 

 alkali, and consider this in the final cal- 

 culation. Place the flask on a water- 

 bath and heat it until the contents begin 



to boil. Then stopper it and set it aside 

 to cool. Now pour off the clear pale- 

 colored solution from the colored crystal- 

 line crust adhering to the bottom of the 

 flask, wash the latter with alcohol, add 

 the washings to the other liquid, then a 

 little phenolphtalein solution and de- 

 termine the excess of alkali with deci- 

 normal acid. L,et it be assumed that 

 13.5 Cc. of the decinormal alkali had 

 been consumed in the decomposition of 

 the nitroglycerin, the amount of the lat- 

 ter would be 13.5x0.0755267—10.1961 

 Gm., that is 10. 196 per cent. 



The method has been tried by the 

 writer upon various preparations of nitro- 

 glycerin and has, so for, yielded reason- 

 ably uniform and satisfactory results. 



In many cases, except where the color 

 produced by the action of the alkali upon 

 the alcohol is too dark, the ordinary 

 aqueous volumetric alkali solution may 

 also be used. To insure complete sa- 

 ponification in this case, the measured 

 volume of the alkali should be heated to 

 about 150 F. and, after the nitroglycerin 

 solution is added, the whole should be 

 heated on a water-bath for about fifteen 

 minutes. There is no advantage in cool- 

 ing it before titration, since none of the 

 salts will separate. The liquid is simply 

 diluted with water until its tint has be- 

 come light enough to permit in it the 

 recognition of the color of the indicator. 

 Now add some phenolphtalein solution 

 and then enough decinormal acid to the 

 neutral point. As it is, however, more 

 difficult in such a liquid to recognize the 

 point when the tint due to the indicator 

 disappears, than that at which it re- 

 appears, it is preferable to titrate beyond 

 the neutral point by adding a known ex- 

 cess of volumetric acid, and then to titrate 

 back with alkali until the phenolphtalein 

 tint just appears again. 



Several samples of solution of nitro- 

 glycerin recently assayed by the writer 



