XCV.-Asymmetric Optically Active Sulphur Compounds. d-Methylethylthetine Platinichloride. By WILLIAM JACKSON POPE and STANLEY JOHN PEACHEY. SINCE the discovery of the sulphonium compounds by Oefele in 1864, attempts too numerous to be recapitulated here have been made to ascertain whether the sulphur atom is capable of combining with four separate univalent groups, or whether derivatives of apparently quadrivalent sulphur are merely molecular, as distinct from atomic, compounds; many attempts have also been made to determine the space arrangement of the four groups attached to the supposed quadrivalent atom. These attempts have, up to the present, yielded negative or ambiguous results (D. Strömholm, Om sulfin- och tetinföreningar, Upsala, 1899). Although our earlier attempts (Proc., 1900, 16, 12) to resolve methylethylthetine into optically active components were unsuccessful, we have since succeeded in obtaining an optically active substance, CH, CO2H,PtCl d-methylethylthetine platinichloride, 2CH Cl 2 CHg which owes its rotatory power to the presence of an asymmetric quadrivalent sulphur atom, and have thus proved that in compounds of the type SX4, the sulphur atom is truly quadrivalent and that the four groups attached to it are, as in the case of carbon, situated at the apices of a tetrahedron the interior of which is occupied by the sulphur atom. CH d-Methylethylthetine d-Camphorsulphonate, CH12O-O 10 15 On treating the methylethylthetine bromide of Carrara (Gazzetta, 1893, 23, i, 493) in aqueous solution with a molecular proportion of silver d-camphorsulphonate (Trans., 1899, 75, 1128), filtering off the silver bromide, and distilling the filtrate at about 40° under a low pressure in a current of air, a residue is obtained which soon solidifies to a crystalline mass in a vacuum desiccator. It is extremely soluble in water, alcohol, or acetone, but may be recrystallised by solution in a little absolute alcohol and addition of anhydrous ether, separation being assisted by strong cooling. Thus obtained, the salt forms colourless, odourless, microscopic prisms melting at 115-117°, and was at first thought by us (Proc., 1900, 16, 12) to be a pure substance because successive fractions had the same composition and practically the same melting point and rotatory power; the apparently pure salt gave the following results on analysis, &c., after drying at 100°: 0.1109 gave 0.2002 CO, and 0.0724 H2O. C-49.24; H=7.25. C15H26O6S2 requires C=49·18; H=7·10 per cent. 0.4296 gram made up to 25.1 c.c. with water, gave ap +0·49° in a 200 mm. tube. Whence [a] +14·3° and [M] +52·4° 0.3843 gram made up to 200 mm. tube. Whence [a] 25.2 c.c. with water, gave a +0.43° in a +14·1° and [M] +51·6°. Since we have previously shown (Trans., 1899, 75, 1086) that the d-camphorsulphonic ion has [M] +51.7° in aqueous solution, the conclusion that we were here dealing with a pure salt and that no resolution of the base had occurred seemed justified. In order, however, to obtain further evidence, the investigation of this and the salt next described was continued as stated in our preliminary note in February (loc. cit.), and the more so in that our work on tin published in May (Proc., 1900, 16, 116) seemed to bear upon the points at issue. During these intervening months, larger quantities of the salt have been prepared and subjected to prolonged fractional crystallisation, with the result that a resolution has been effected. On fractionally crystallising the salt some 40 or 50 times from a mixture of absolute alcohol and ether, a sparingly soluble fraction was obtained which melted at 118-120°, and gave the following results on analysis, &c., after drying at 100°: 0.1275 gave 0.2304 CO2 and 0.0839 H2O. C=49.27; H=7.31. CH26O6S requires C=49·18; H=7·10 per cent. 0.7224 gram made up to 25.1 c.c. with water, gave a + 1·07° in a 200 mm. tube. Whence [a] + 18.6° and [M] +68·0°. This material hence seems to be d-methylethylthetine d-camphorsulphonate, in which the anion has the molecular rotatory power [M]D +16.3°. d-Methylethylthetine d-Bromocamphorsulphonate, SO CH; + H2O. On treating methylethylthetine bromide with silver d-bromocamphorsulphonate (Trans., 1899, 75, 1131) in place of the camphorsulphonate, but adopting otherwise the same procedure, an oily residue ultimately remains which readily solidifies in a vacuum desiccator. This, after several crystallisations from slightly moist acetone, is obtained as a mass of colourless, silky needles melting at 166-168°, and from analysis and rotatory power determinations appears to be a uniform product. 1.7103 lost 0.0571 at 100°. H20=3.34. 0.2455 gave 0.3478 CO2 and 0.1157 H2O. C-38-64; H=5.22. C16H250 BrS2, H2O requires H2O=3·88; C=38.87; H=5.83 per cent. 0.3533 gram made up to 250 c.c. with water, gave a +1.68° in a 200 mm. tube. Whence [a] +59.5° and [M] +275°. VOL. LXXVII. 4 D Since the molecular rotatory power of the d-bromocamphorsulphonic ion is about [M] +270 (Walden, Zeit. physikal. Chem., 1894, 15, 196), it seemed that no resolution of the thetine had taken place. Prolonged fractional crystallisation of larger preparations of the salt from moist acetone resulted, however, in the separation of a sparingly soluble fraction having a molecular rotatory power slightly different from the above, although its composition is the same; about 30 crystallisations were undertaken in obtaining this product. 1.8654 lost 0.0666 at 100°. H20=3·57; 1H20=3.88 per cent. 0.2385, dried at 100°, gave 0.3513 CO, and 0.1243 H2O. C-40-17; H=5.79. 2 0.2293, dried at 100°, gave 0·3372 CO2 and 0·1172 H2O. C=40·10; H=5.67. C15H2506BrS2 requires C = 40·44; H=5.61 per cent. 0.6172 gram air-dried salt, made up to 25.1 c.c. with water, gave ap +3.05° in a 200 mm. tube. Whence [a] +62.0° and [M] +287°. 0·6594 gram air-dried salt, made up to 251 c.c. with water, gave ap +3.29° in a 200 mm. tube. Whence [a] +62·7° and [M] + 290.5°. Since ammonium d-bromocamphorsulphonate has [M], +275° in aqueous solutions of about the same molecular concentration as is here used, the examination of the thetine bromocamphorsulphonate, as well as the camphorsulphonate, affords very strong evidence that both contain an optically active thetine of which the ion has about [M]D +15°. CH2 CO2H, PtCl d-Methylethylthetine Platinichloride, 2CH¡>S<CH On dissolving d-methylethylthetine d-camphorsulphonate of [M] +68° or d-bromocamphorsulphonate of [M] +287° to +290° in absolute alcohol, adding a little concentrated hydrochloric acid and then slowly stirring in an alcoholic solution of the requisite weight of platinic chloride, a yellow, crystalline platinichloride is deposited, which, after exhaustive washing with absolute alcohol, melts with profound decomposition at 177-180°. The salt is very soluble in water, but insoluble in absolute alcohol, and may be crystallised from dilute alcohol; its racemic isomeride is described by Carrara (loc. cit.) as melting at 167°. 0-2469 gave 0.1619 CO2 and 0.0802 H2O. C=17·89; H=3.61. 0.1104 0.1391 AgCl. Cl=31·19. C10H22S2O4PtCl requires C=17·73; H=3.25; Cl-31.42; Pt = 28.71 per cent. 0-2005 gram of salt from camphorsulphonate, made up to 15 c.c. with water, gave ap +0·14° in a 200 mm. tube. Whence [a] +4·6° and [M] +30-8°. 0.3710 gram of salt from bromocamphorsulphonate, made up to 15 c.c. with water, gave ap +0·22° in a 200 mm. tube. Whence [a] +4.5° and [M] +30.2°. The molecular rotatory power [M], +30.5° of the platinichloride is, as would be expected, just about twice that of the thetine ion, [M]D +15°, deduced from its salts with optically active acids; in spite, however, of this close numerical agreement, we can as yet offer no guarantee that these optically active materials are free from their stereochemical isomerides. We have thus proved that the asymmetric quadrivalent sulphur atom acts as a centre of optical activity just as does the carbon atom. Since we have previously demonstrated that the same is true of the asymmetric nitrogen (Trans., 1899, 75, 1127) and tin (Proc., 1900, 16, 42 and 116) atoms, we are now justified in concluding that the quadrivalent elements, C, Si, Ti, Zr, Ce, Th, Ge, Sn, Pb, O, S, Se, Te, Cr, Mo, W, and U of groups IV and VI, and the quinquevalent elements N, P, V, Nb, As, Sb, and Bi of group V of the periodic table are in their compounds with four or five groups of atoms, surrounded by groups arranged in tridimensional space, and can thus act as centres of optical activity. During the present work we have had occasion to examine the conflicting records concerning isomerism possibly occurring in methyldiethylsulphonium compounds. Krüger (J. pr. Chem., 1876, [ii], 14, 205) and Nasini and Scala (Gazzetta, 1888, 18, 67) obtained two sulphonium iodides according as they treated diethyl sulphide with methyl iodide, or methyl ethyl sulphide with ethyl iodide. Klinger and Maassen (Annalen, 1888, 243, 193) consider that trimethyl- and triethyl-sulphonium iodides are produced during the latter reaction, and that the platinichlorides of these form a double salt which is naturally not identical with methyldiethylsulphonium platinichloride. We hope shortly to publish details confirming Klinger and Maassen's conclusions, having isolated the same platinichloride from the product obtained by heating methyl ethyl sulphide with ethyl iodide and water, as by crystallising together trimethyl- and triethyl-sulphonium platinichlorides. Any objection to the tetrahedral nature of asymmetric sulphur compounds which might be based on the work of Krüger or of Nasini and Scala can therefore be set aside. GOLDSMITHS' INSTITUTE, NEW CROSS, S. E. XCVI.-A New Method of Estimating Potassium. By RICHARD HALIBURTON ADIE and THOMAS BARLOW WOOD. ERDMANN (J. pr. Chem., 1866, 97, 385) and Sadtler (Amer. J. Sci., 1870, [ii], 49, 189), among other observers, have described double potassium and sodium cobaltinitrites, prepared by precipitating solutions of cobalt salts with strong solutions of potassium or sodium nitrite, and noted the very sparing solubility of potassium cobaltinitrite. This fact has led to the use of a solution of sodium cobaltinitrite (de Koningh's reagent) as a qualitative test for potassium, and the authors' experience of the delicacy of this test has induced them to investigate again the nature and composition of the precipitate obtained, with a view to the possibility of using the reaction for the quantitative estimation of potassium. Two possible methods of procedure suggested themselves: (1) to precipitate the potassium by a solution of sodium cobaltinitrite, and to weigh the precipitate; (2) to precipitate as before, but to estimate the amount by titrating the nitrite groups with potassium permanganate solution. For both of these methods, it is necessary to know that the precipitate has a constant composition. To determine this, the authors prepared and analysed several samples, and found them to be dipotassium sodium cobaltinitrite, K,NaCo(NO2)6,H2O. Preparation of the Reagent.-The sodium cobaltinitrite solution was prepared by dissolving 113 grams of cobalt acetate in 300 c.c. of water and 100 c.c. of acetic acid; 220 grams of sodium nitrite were also dissolved in 400 c.c. of water. The solutions were filtered and mixed, and the nitric oxide removed by evacuation. After standing 24 hours, the yellow precipitate, which had separated, was filtered off, and the solution made up to a litre. Analysis of the Precipitate.-The sodium cobaltinitrite solution, when added to an equal volume of a 1 per cent. solution of potassium chloride acidified with acetic acid, gives an immediate yellow crystalline precipitate, which settles in a few minutes. Several samples were made and analysed with the following results: Calculated for |