245 XXVI1.-A New Process for the Volumetric Estimation of Cyanides. By J. B. HANNAY, F.R.S.E., Assistant Lecturer on Chemistry, The Owens College, Manchester. THE processes a t present in use for the volumetric estimation of cyanides by solution of silver and iodine requiring, as they do, the absence of such frequently-occurring substances as ammonia and alkalis, and depending on solutions prone to change, are in so far faulty ; and as besides it is often desirable t o estimate the cyanide in an alkaline solution already containing silver cyanide, I have con- sidered that the description of a process (which I have used for some years), and which is interfered with by no commonly occurring bodies, may have some interest for this Society. The process is the converse of that for the estimation of mercury, of which I published an account (C'hem.Xoc. Jour., June, 1873, pol. xxvi, p. 565) some years ago, and which has just received such complete confirmation (Chenz. Soc. Jozw., December, 1877, vol. ii, p. 679) a t the hands of Professor Tuson and Mr. Neison; in fact, it was the publication of their paper that induced me to write the present note. The process for the estimation of cyanogen in compounds depends on the anomalous behaviour of mercuric cyanide with alkalis ; so that, if to an alkaline solution containing cyanogen a mercuric salt is added, no precipitlate takes place until all the cyanogen is combined with mercury ; that is to say, as long as the decomposition 2KCN + HgCl, = Hg(CNX+ 2KC1 goes on, all the mercury existing as cyanide is not affected by alkalis.The cyanide is dissolved in water, placed in a beaker on a black slab (or black velvet), rendered alkaline preferably with ammonia, and a standard solution of mercuric chloride is added in successive quantities, with frequent stirring, until a permanent bluish-white opalescence is produced. The end of the reaction is sharply marked, and half a drop of a centinormal solution is sufficient to produce a strong opalescence. To test the accuracy of the process, as well as to find if other sub- stances had any interfering action, a solution of potassic cyanide, con- taining 0.00G51 gram per C.C. (decinormal), was prepared, and the strength checked by silver estimation. A decinormal solution of mercuric chloride was also prepared, containing 0.0271 of the chloride per C.C.A portion of this solntion was reduced to half strength, that is, containing0.01355 per c.c., so that one C.C. of this was equal to one C.C. of the cyanide. It was found that 20 C.C. of the cyanide gave a246 HANNAP ON A NEW PROCESS, ETC, pcrceptible opalescence, when a small fraction over 20 C.C. of the mercury solution had been added ; the excess mas less than -03 of a C.C. in an average of five estimations. To find how little would really show the turbidity, I prepared a centinormal solution, and reducing it to half strength, 1 found that 20 C.C. of the decinormal cyanide required 200.1 C.C. of the mercury solution to producs a decided opalescence; but that was the bighest, other experiments requiring between -05 to :08 C.C.of excess. This indicates a quantity of .0000651 of the cyanide, or an error of 005 percent. A large number of experiments were made on the action of alkalilie sulphates, chlorides, and nitrates, but with the same result as Tuson and Neison have already shown. I also found that very large quantities of ammonium salts prevent the appearance of the opalescence when small quantities of cyanides have to be esti- mated; but as the above anthors have shown that the interference begins only when 15 times more ammonium salt is present than of mercury, i t will be seen that in ordinary chemical analysis such a state of affairs seldom arises. The kind of impurities in which I was more inter- ested, however, were those present in samples of commercial potas- sic cyanide, principally potassic cymate and thiocyanate, as caustic alkalis and alkaline carbonates are entirely without action.Solutions of the cyanate and thiocyanate were prepared, coutaining 9 5 gram per c.c., and the following experiments were tried. Cyanate : 20 C.C. of the cyanide were measured off, and 10 C.C. of the cyanate solution added ; and after rendering alkaline with ammonia, mercuric chloride (decinormal half strength) was added, when it was found that 8 single drop over 80 C.C. caused an opalescence. Two other experiments were tried, using 20 C.C. and 50 C.C. of the cyanate; the two took 20.05 and 20.08 to produce an opalescence. The fractions were done by com- pleting the process with centinormal solution: thus it will be wen that cyanic acid has no effect on the process.Three other experiments were done, using very dilute solutions, but still there was practically no effect. Thiocyanate: as before, 20 C.C. of cyanide and 10 C.C. of thiocyanate were rendered alkaline with ammonia and the mer- cury solution added. Two others with 30 C.C. and 50 C.C. of thiocyanate respectively took 20.6 and 21 C.C. of mercury solution. I suspected that the thiocyanate could not be quite pure, so I recrystallised some of what had appeared to be pure salt, and I found that it now had no effect; three quan- tities of 20 C.C. each, with 10.30 and 50 C.C. of thiocyanate, taking 20.01, 20.04, 20.07 for the production of the opalescence. It will thus be seen that thiocyanic acid has also no effect upon this process. As it is often desirable to estimate cyanides in a solution containing silver, as in electro-platers' baths, I added silver nitrate to three quan- tities of cyanide solution in the following proportions :- It took 20.2 C.C. of the mercury.TILDEN ON TERPIN AND TERPINOL. 247 1. To 20 c.c., a quantity not sufficient to cause a precipitate. 11. ,, just sufficient to cause a precipitate. 111. ,, a large excess. On adding a little ammonia to I and dissolving the precipit,ates I1 and I11 in ammonia, each took a very small fraction over 20 C.C. of mercuric chloride solution to produce an opalescence. This shows that silver salis do not interfere with the process. It will thus be seen that we have now a process by which cyanogen in combination can be estimated accurately, even in the most complex mixtures ; and as mercuric chloride can easily be obtained pure and keeps unaltered in solution, the process is one of great facility.