MARCH, 1968 THE ANALYST Vol. 93, No. I104 The So-called Phenylthiohydantoic Acid: Its Structure and Application to the Gravimetric Determination of Cobalt* BY A. BASHAR AND A. TOWNSHEND (Chemistry Department, The University, P.O. Box 363, Birmingham 15) The compound previously known as phenylthiohydantoic acid is shown to be carbaminothioglycollic acid anilide. The precipitate produced when this reagent is added to cobalt solutions is tris(tliioglycol1ic acid anilido) cobalt (111), formed by hydrolysis of the carbamino compound. This pre- cipitate is not formula-pure, however, and cannot be weighed directly. It must be ignited to tricobalt tetra-oxide, Co,O,, for the determination of cobalt. THE reagent known as phenylthiohydantoic acid was first recommended for the detection of cobalt by Pozzi-Escotl because of the red - brown precipitate produced when it reacted with aqueous cobalt solutions.Later, Willard and Hall2 investigated the gravimetric deter- mination of cobalt with this reagent. Precipitation was quantitative in ammoniacal solutions, and cobalt could be separated from numerous metals, including nickel, iron and manganese. Unfortunately, the cobalt precipitate was not quite sufficiently reproducible in its com- position, so it had to be ignited, thus making the conversion factor much less fauourable. Also, the precipitate was extremely bulky, and apparently only less than 25 mg of cobalt could be handled. It was thought that the application of precipitation from homogeneous solution to this system would retain the selectivity and completeness of the precipitation of cobalt, and at the same time reduce the bulk of the precipitate (as, for example, with nickel dimethyl- gly~ximate~,~), so that larger amounts of cobalt could be dealt with.It could also improve the stoicheiometry of the precipitate (as occurred with the cobalt l-nitroso-2-naphthol pre- cipitate6), so that the precipitate could be weighed after drying. Finally, any co-precipitation would be minimised.6 In 1877, Jaeger' reacted together monochloroacetic acid, ammonium thiocyanate and aniline and identified the product (which he was unable to hydrolyse) as phenylthiohydantoic acid, I. Later, Rizzoa and Beckurts and French9 reported that the compound hydrolysed readily to give thioglycollic acid anilide, 111, and thus concluded that Jaegers' compound was carbaminothioglycollic acid anilide, 11.These observations seem to have been completely overlooked by Willard and Hall,2 by Cuvelier,lo who also investigated the gravimetric method for cobalt, and by Duval and Duva1,ll who recorded a thermogravimetric curve for the precipitate. CeH,NHC( :NH)SCH&OOH C,H,NHCOCH,SCONH, C6H,NHCOCH$SCH,CONHC,H, f C,H,NHCOCH,SH + CO, + NH, I J. 11 Iv II1,HL We have confirmed the earlier obser~ations~~~ on the facile hydrolysis of Jaeger's com- pound, which is apparent even during its crystallisation, for the melting-points of successive * Part of this work was previously summarised in Chem. Commun., 1967, 901. 0 SAC and the authors. 125126 BASHAR AND TOWNSHEND : THE SO-CALLED PHENYLTHIOHYDANTOIC [,41zalyst, Vol.93 samples fall, and the NH band at 3380cm-l gradually disappears. Hydrolysis to thio- glycollic acid anilide was completed in 40 minutes. A single recrystallisation of authentic carbaminothioglycollic acid anilide from water gave a product shown by elemental analysis to contain 23 per cent. of 111; this increased to 65 per cent. after three “recrystallisations.” Continued refluxing produced dithioglycollic acid anilide, IV, as previously noted by R i z ~ o . ~ Attempts to form ethyl and 2-hydroxypropyl esters of Jaeger’s compound failed. A band at 1700 cm-l caused by the C=O stretch in a carboxylic acid is absent. This and the failure to undergo cyclisation to give phenylpseudothiohydantoin in the presence of mineral acidsl2 again point against structure I and favour 11.TABLE I SOME INFRARED FREQUENCIES, CM-I V-NH V-NH amide) amide) V-SH (amide) C6H, Amide (primary (secondary v-c=o Carbaminothioglycollic acid anilide 3380 3275 - 1660 1605 1540 Thioglycollic acid anilide . . . . - 3240 2630 1640 1600 1545 Cobalt precipitate with thioglycollic acid anilide . . . . .. - 3230 - 1640 1695 1650 Cobalt precipitate with carbamino- thioglycollic acid anilide . . .. - 3320 - 1635 1590 1560 Infrared spectra confirm these observations (Table I). Nuclear magnetic resonance spectra recorded at 60 Mc/s in dimethylsulphoxide (D,) gave singlets at -0.157 (lH), 7-37 (2H) and 6-67 (2H), all against tetramethylsilane, with a multiplet at 2.0 to 3-27 given by the phenyl group. As the resonance at -0-16 could have arisen either from a carboxylic acid or from a secondary amine group these results do not distinguish between I andII.However, a comparison of the resonances of the compound studied, its N-methyl derivative and thioglycollic acid anilide in acetone (DJ, which are summarised in Table IT, confirms that the compound is carbaminothioglycollic acid anilide, 11. TABLE I1 Shifts in p.p.m. against tetramethylsilane NUCLEAR MAGNETIC RESONANCE FREQUENCIES IN ACETONE (D,) Number of H atoms 1 5 2 3 2 1 N-methylanilide . . .. .. - 2.4 (m) 6.4 to 6-6 (s) 6.7 (s) 7.2 (s) - Thioglycollic acid anilide . . . . 0.6 (s) 2.1 to 2.8 (m) 6.6 (d) - -- 7.6 (t) Assignment . . .. .. . . NH c6H& CH2 CH, NH, SH Carbaminothioglycollic acid Carbaminothioglycollic acid anilide 0.5 (s) 2.3 to 2.6 (m) 6.3 (s) - 6.95 (s) - s = singlet, d = doublet, t = triplet, m = multiplet.POLAROGRAPHY- The alleged phenylthiohydantoic acid gave an anodic wave (E4 = 0-0 volt against an S.C.E.) at pH 4.0. Such a wave can be readily attributed to a thiocarbonic acid derivative, 11, but would not be expected for a thio-ether, I. At pH 6-8 the wave was unchanged after 1 hour, but at higher pH values hydrolysis produced a well developed anodic wave (E* = -0.5 volt against an S.C.E. at pH 7*2), identified as being due to thioglycollic acid anilide, 111. The gradual formation of this wave is shown in Fig. 1. By using this wave, the rate of hydrolysis was found to be instantaneous in 0-1 M sodium hydroxide or in a borate buffer (pH 9.2); in a phosphate buffer (pH 7-8), the plot of log (diffusion current) against time was linear for the first 11 minutes of the reaction and suggested a first order reaction.Hydrolysis at pH above 7 would be expected13 for the cleavage of the -S-CO- grouping of I1 but not for an S-substituted thiourea, such as I.March, 19681 ACIII : ITS STRUCTURE AND APPLICATION 127 Fig. 1. Polarographic anodic waves of carbaminothioglycollic acid anilide (2 x lo-* hf) in a phosphate buffer solution (pH 7.2) containing 2 per cerit. of ethanol. Measurements were made a t about 20" C after A, 0 minute; B, 11 minutes; C, 43 minutes; and D, 93 minutes (this is the wave of thioglycollic acid anilide). All curves go from t o - 1 to - 1.1 volts versus an S.C.E. Capillary characteristics : drop time is 3.7 seconds; mercury flow is 1.53 mg per second PRECIPITATION OF COBALT- When cobalt (11) was precipitated with carbaminothioglycollic acid anilide by using the procedure of Willard and the ligand involved in the precipitate was found to have only one nitrogen atom.As the precipitation occurred at pH 7 from boiling solution, it was reasonable to expect that this ligand was thioglycollic acid anilide, 111. Indeed, thioglycollic acid anilide gave precipitates with cobalt that were identical (colour, elemental analysis and infrared spectra) with those produced by Willard and Hall's method. In fact, thioglycollic acid anilide has been recommended as a reagent for the detection of cobalt because of the voluminous blood-red precipitate it gives with cobalt . 1 4 9 1 5 In all instances, irrespective of whether cobalt(I1) or cobalt(II1) was initially present in the solution, the analytical results for the precipitate agreed most closely with a formula of COL,.Thus carbaminothioglycollic acid anilide is merely a source reagent that generates the actual precipitant, thioglycollic acid anilide, in solution. The procedure of Willard and Hall, therefore, is essentially one of the earliest examples of precipitation from homogeneous solution; no other example of this type of generating reaction has previously been reported. It was possible, therefore, that the original aims of this investigation could be accom- plished by a detailed study of the use of carbaminothioglycollic acid anilide for precipitating the cobalt - thioglycollic acid anilide complex from homogeneous solution.CONVENTIONAL PRECIPITATION OF COBALT WITH THIOGLYCOLLIC ACID ANILIDE- To evaluate the procedure that is developed for precipitation from homogeneous solution, it must be compared with both the original procedure of Willard and Hall2 and also the direct precipitation with thioglycollic acid anilide. Bersin14 found that the latter reagent was suitable for the detection of cobalt in the presence of iron(III), aluminium and chromium(II1). He assumed that the precipitate was CoL,.4H20, which changed on heating to Co,L,. This form was not suitable for weighing, however, and had to be converted into a form that was suitable. This is the only recorded investigation of thioglycollic acid anilide as a gravimetric reagent. However, Misra and Sircar15 have investigated other thioglycollic acid amides as precipitants for cobalt, and found that the precipitates can be weighed after drying at 125" to 130" C.Thionalide, the 2-naphthyl analogue of thioglycollic acid anilide, is the best known of the thioglycollic acid anilide derivatives.16 When cobalt(I1) (25 mg) was precipitated with thioglycollic acid anilide, according to the method used with other thioglycollic acid amides,15 the voluminous red - brown precipitates The complexes were formulated as CoL3.1.128 BASHAR AND TOWNSHEND : THE SO-CALLED PHENYLTHIOHYDANTOIC [Analyst, Vol. 93 had a wide range of weights after drying at 125" to 130" C; all of the weights were excessive for CoL, chelates. At least part of the excessive weight was caused by co-precipitated reagent, which could be seen as white particles.This could be reduced by decreasing the digestion time from 3 to 4 hours to 30 minutes. The application of thioglycollic acid anilide in the Willard and Hall procedure2 gave similar results, and both were similar to the results obtained by using the original Willard and Hall procedure. All of these results are sum- marised in Table 111. TABLE I11 PRECIPITATION OF COBALT (25 mg) WITH THIOGLYCOLLIC ACID ANILIDE Method Weight of precipitate, mg* Drying temperature Misra and Sircar16 .. .. .. .. .. 266-7 to 291.87 126"to 130' C Misra and Sircar16 .. .. .. .. .. 239.4 to 266.75 125" to 130" C Willard and with thioglycollic acid anilide . . 242.8 to 327.1 125" to 130" C Willard and Hall,z with carbaminothioglycollic acid 235.3, 268.5 Room temperature, anilide over silica gel * CoL, would weigh 236-5 mg.t Range of precipitate weights. $ Mean weight and number of determinations. 3 After only 30 minutes' digestion. 283-3 (3)$ 251.6 (3) 292.7 (3) PRECIPITATION FROM HOMOGENEOUS SOLUTION- None of the methods described above provided a satisfactory means for the gravimetric determination of cobalt by the direct weighing of the complex. Although Willard and Hall's original method meets some of the requirements of a procedure for precipitation from homo- geneous solution, the voluminous nature of the precipitate formed by reaction with cobalt (11) indicates an extremely rapid prccipitation process. Thus the hydrolysis, in which the thio- glycollic acid anilide is formed, is so fast as to preclude a slow, homogeneous generation of precipitant in the solution.It was possible to slow down this hydrolysis sufficiently by using a reaction solution containing more than 20 per cent. v/v of ethanol. In this instance, the precipitate produced is compact and dark red. Even then, however, the weights of the dried precipitates were still high and variable. Other methods of precipitation from homogeneous solution were attempted, viz., urea hydrolysisG and precipitation from mixed s01vents.l~ In these instances, there was no signifi- cant reduction in the bulk of the precipitates, and their weights were equally variable. TABLE IV IGNITION OF PRECIPITATES Cobalt taken = 25.05 mg Cobalt recovery r Weight of Calculated as Calculated as oxide found, mg Co,O,, mg CO804, m g 34-40 24.45 25-26 34.10 24-23 25-05 34-80 24-77 25.55 34.50 24.52 25-63 34-40? 24.45 25.26 34-60? 24-59 25.41 34-20t, $ 24.30 25.17 34-80?, $ 24.77 25-55 Mean ... . 24.57 25-35 Cobalt in filtrate, mg* 0-008 0.014 0.000 0.018 0.007 0.008 0.003 0.009 0.009 Cobalt in washings, mg* 0.140 0.094 0.034 0.078 0-082 0.120 0.026 0.041 0.077 * Determined by the nitroso-R-salt method,21 after destruction of organic matter. $ Precipitation in 20 per cent. ethanol. After only 2 hours' digestion on water-bath.March, 19681 ACID : ITS STRUCTURE AND APPLICATION 129 IGNITION OF THE PRECIPITATES- Direct weighing of the dried precipitates was, therefore, impossible, even with the new knowledge of the precipitant. One controversy remained, however.Willard and Hall2 found that if the precipitate is ignited to oxide at 850" C, better recoveries are obtained if the ignited substance is considered to be Co203, rather than Co,O,, although the results are high for both. Duval and Duvalll also identify the oxide as Co20, in their gravimetric studies. This is surprising, because Co203 is converted18 into Co,O, above 265°C. Repetition of the experiments of Willard and Hall, including experiments in which more than 20 per cent. of ethanol was used, gave the results shown in Table IV. They show that more accurate results are obtained if the oxide is considered to be C O , ~ , , although there is little to choose between either oxide if solubility losses are also considered. Identical results were obtained by using thioglycollic acid anilide as the precipitant. In all instances, precipitation of cobalt is quantitative, and most of the cobalt not recovered is lost during the washing of the precipitate.EXPERIMENTAL SYNTHESIS OF REAGENTS- Carbaminothioglycollic acid anilide-This was prepared by Jaeger's method7 for "phenyl- thiohydantoic acid." The product was recrystallised three times from ethanol to give white crystals, m.p. 146" to 150" C. Value given in literat~re,~ 148" to 152" C. Found: C, 51-5 per cent.; H, 5.0 per cent.; N, 13.45 per cent.; S, 15.2 per cent. C9Hl~,02S requires C, 51.4 per cent.; H, 4.8 per cent.; N, 13.3 per cent.; S, 15.25 per cent. Carbam.inothioglycollic acid N-methylannilide-This was prepared similarly, but by using N-methylaniline (214.3 g) in place of aniline (189 g).After refluxing the reaction mixture for 24 hours, the product was twice recrystallised from ethanol to give white crystals, m.p. 138" to 140" C. Value given in literat~re,~ 147" C. Found: C, 53.5 per cent.; H, 5-4 per cent.; N, 12.4 per cent.; S, 14.0 per cent. Cl,H12N202S requires C, 53.5 per cent.; H, 5.4 per cent.; N, 12.5 per cent.; S, 14.3 per cent. Thioglycollic acid anilide-This was prepared by reported methods9 9 1 5 and recrystallised from aqueous ethanol to give white crystals, m.p. 105" to 108" C. Value given in literat~re,~ 111" to 112°C. REACTIONS OF THE REAGENTS- Attempted esterijcation-Treatment of "phenylthiohydantoic acid" with ethanol in the presence of sulphuric acid and with propylene oxide, by an adaptation of the method of Haggis and Owen,l9 yielded only starting material. Attempted cyclisation-Only starting material was recovered after treatment of "phenyl- thiohydantoic acid" with 20 per cent.hydrochloric acid, according to Mouneyrat's procedure.20 HydroZysis-Pure thioglycollic acid anilide was obtained by refluxing an aqueous solution of carbaminothioglycollic acid anilide for 90 minutes. The product was white crystals, m.p. 104" to 108" C. Value given in literat~re,~ 111" to 112" C. Found: C, 57-7 per cent.; H, 5.45 per cent.; N, 9.0 per cent.; S, 19.3 per cent. C8H9NOS requires C, 57-5 per cent.; H, 5.4 per cent. ; N, 8.4 per cent. ; S, 19.2 per cent. Oxidation to dnithioglycollic acid anilide-Refluxing an aqueous solution of carbamino- thioglycollic acid anilide for 48 hours in a vessel open to the air gave a white crystalline product, m.p.156" to 158" C. Value given in literature,8 160" to 161" C. Found: C, 58.5 per cent.; H, 4.5 per cent.; N, 9-1 per cent.; S, 18.5 per cent. C,,H,,N20,S2 requires C, 57.8 per cent. ; H, 4.9 per cent. ; N, 8.4 per cent. ; S, 19.3 per cent. All melting-points were uncorrected. Thin-layer chromatography-Chromatograms were run on silica gel G, 250 p thick, acti- vated at 110" C for 1 hour. The solvent system was xylene - glacial acetic acid (60 + 40 v/v). The spots were located by iodine adsorption. The RF values obtained (relative to a solvent- front movement of 10.0 cm) were as follows : carbaminothioglycollic acid anilide, 0.34, 0.52 (faint) ; thioglycollic acid anilide, 0.49 ; carbaminothioglycollic acid anilide, after partial hydrolysis, 0.34, 0-53 ; pure hydrolysis product, 0.53 ; "phenylthiohydantoic acid" (British Drug Houses Ltd.), 0.33, 0.50 (faint), 0.57 ; and dithioglycollic acid anilide, 0.52.The RF values were not reproducible to better than +0*05. Polarography-The studies were carried out in a Kalousek-type cell, with a separated standard calomel electrode, and the polarograms recorded on an Evershed-Tinsley or a Radiometer PO, polarograph. Solutions were de-aerated by bubbling nitrogen through130 BASHAR AND TOWNSHEND them. All of the buffer solutions were prepared from analytical-reagent grade reagents. Freshly prepared 0.01 M solutions of carbaminothioglycollic acid (in 20 per cent. ethanol) and thioglycollic acid anilide (in water or ethanol) were used.Hydrolysis rates were measured by following the change in the diffusion current at -0.3 volt (against an S.C.E.). Infrared stdes-All spectra were recorded by using potassium bromide discs on a Perkin-Elmer, Model 23, spectrophotometer. Nuclear magnetic resonance-Spectra were recorded on a Perkin-Elmer R10 spectrometer, operating at 60 Mc/s. PRECIPITATION OF COBALT- A stock solution of analytical-reagent grade cobalt nitrate containing 1 mg of cobalt per ml was used. The precipitation procedures used were those previously published, or simple modifications thereof, as described in the text. The precipitant (0-7g) was added as a solution in water or ethanol (30ml). The precipitates were filtered through No. 4 sintered-glass crucibles, washed with hot water, and then dried, either by heating at tempera- tures within the range 110” to 140” C or at room temperature in a vacuum desiccator, to constant weight.Precipitations of cobalt (111) were carried out after oxidising cobalt(I1) with hydrogen peroxide in a potassium hydrogen carbonate solution.2 Precipitates produced from cobalt(II1) solutions were more compact than when cobalt(I1) was initially present. Elemental analyses for various precipitates that appeared to be little contaminated were as follows- Cobalt(I1) precipitated by the original Willard and Hall procedure2: found C, 51.5 per Cobalt(I1) precipitated similarly, but with thioglycollic acid anilide: found C, 51-4 per Cobalt (111) precipitated similarly, with thioglycollic acid anilide : found C, 51-1 per Co(C,H,NOS), requires C, 51.7 per cent.; H, 4.3 per cent.; N, 7.5 per cent.; S, 17.2 per Ignition of precipitates was carried out after filtering on to Whatman No.40 filter-paper, in porcelain crucibles, at 850” C . The authors thank Professor R. Belcher for his advice and encouragement, Dr. P. Zuman for his advice on polarographic techniques and for his interest generally, Dr. E. F. Mooney for the nuclear magnetic resonance spectra and their interpretation, Dr. J. K. Brown for the infrared spectra, and the micro-analytical staff of this Department for numerous elemental analyses. cent.; H, 4.3 per cent.; N, 7.3 per cent.; S, 16-7 per cent. cent.; H, 4-4 per cent.; N, 6-5 per cent.; S, 16.2 per cent. cent.; H, 3.9 per cent.; N, 6.9 per cent.; S, 17.4 per cent. cent.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. 20. 21. REFERENCES Pozzi-Escot, E., Annls. Chim. Analyt. Chim. Afipl., 1905, 10, 147. Willard, H. H., and Hall, D., J. Amer. Chem. Soc., 1922, 44, 2219 and 2226. Salesin, E. D., and Gordon, L., Talanta, 1960, 5, 81. Bickerdike, E. L., and Willard, H. H., Analyt. Chem., 1952, 24, 1026. Heyn, A. H. A., and Brauner, P. A., Talanta, 1961, 7 , 281. Gordon, L., Salutsky, M. L., and Willard, H. H., “Precipitation from Homogeneous Solution,” Jaeger, J. H., J . firakt. Chem. 1877, 16, 17. Rizzo, N., Gazz. Chim. Ital., 1898, 28, 356. Beckurts, H., and French, G., J.firaht. Chem. 1902, 66, 172. Cuvelier, V., Natuurw. Tijdschr., 1929, 11, 131. Duval, R., and Duval, C., Analytica Chim. Acta, 1951, 5, 84. Reid, E. E., “Organic Chemistry of Bivalent Sulphur,” Volume 5, Chemical Publishing Co., New Fedorofiko, M., and Zuman, P., Colln Czech. Chem. Comm., 1964, 29, 2115. Bersin, T., 2. analyt. Chem., 1931, 85, 428. Misra, R. N., and Sircar, S. S. G., J . Indian Chem. SOC., 1955, 32, 127. Berg, R., and Roebling, W., Angew. Chem., 1935, 48, 430 and 597. Howick, L. C., and Jones, J. L., Talanta, 1961, 8, 446. Young, R. S., “The Analytical Chemistry of Cobalt,” Pergamon Press, Oxford, 1966. Haggis, G. A., and Owen, L. N., J. Ckem. SOL. 1950, 2250. Mouneyrat, A., Ber. dt. chem. Ges., 1900, 33, 2393. Sandell, E. B. , “Colorimetric Determination of Traces of Metals,” Third Edition, Interscience Publishers, New York and London, 1959, p. 415. Received Sfitember 21st, 1967 John Wiley & Sons Inc., New York; Chapman & Hall Ltd., London, 1959. York, 1963.