26 Artalyst, January, 1971, Vol. 96, pp. 26-36 Some Observations on Oxidation - Reduction Indicators of the Benzidine, Naphthidine and Diar ylamine BY E. BISHOP AND MRS. L. G. HARTSHORN (Chemistry Department, University of Exeter, Stocker Road, Exeter, Devon) A spectrophotometric investigation has been made of representative compounds of the benzidine class of indicators in an attempt to resolve fundamental problems concerning their oxidation mechanism in sulphuric acid media. In all instances the oxidation to the coloured compound is a single-step two-electron process. There is no detectable evidence for the formation of an intermediate, either of a benzidine from arylamines or of a semiquinone from a benzidine or naphthidine. Reduction or spontaneous decay of oxidised arylamines stops a t the benzidine stage.All oxidised indicators are unstable both intrinsically and in the presence of excess of oxidant : unsubstituted benzidine and naphthidine are, additionally, photo- sensitive in the oxidised state. The spontaneous decomposition of the oxidised form regenerates the reduced form and is a disproportionation, probably of 2: 1 stoicheiometry, but a kinetic study shows that the rate-controlling step is a unimolecular precursive reaction, which may be the relaxation of the triplet state of the coloured dication diradical oxidised form. Stability decreases on sulphonation and with increasing temperature, and increases with increasing sulphuric acid concentration and with progressive substitution of the amino hydrogen atoms. FOR many reasons, such as sparing solubility, instability of oxidised forms, side and induced reactions, low exchange currents, photosensitisation and the need with certain oxidants of providing an inductor, potentiometric measurements on di- and triarylamines and alkyldiaryl- amines (hereafter collectively termed arylamines) , benzidine, naphthidine and their derivatives are attended by severe difficulties. Interpretation of such results as can be obtained is often uncertain and equivocal.In few instances is it known beyond doubt how many electrons are involved in the oxidation of the indicator, and little beyond the fact of its occurrence is known about the decomposition or further oxidation of the oxidised form of the indicator. None of the indicators is sufficiently stable in the oxidised form to pennit accurate deter- mination of its formal potential by direct potentiometric titration, and few allow approximate estimation of either formal potential or number of electrons by this method.Instead, recourse must be made to the method of conducting a potentiometric titration of, say, iron(I1) with dichromate or cerium(1V) in the presence of a substantial amount of indicator that involves passing through the end-point region with a 10 or 100-fold diluted titrant while recording potentials and visual observations of the indicator colour. Rapid titration to avoid error from decomposition of the indicator produces inaccurate potentials ; slow titration to achieve accurate potential measurements permits significant decomposition of the indicator. Conse- quently it is more usual to report “transition potentials,” i.e., the potential at which a discernible colour first appears (all indicators of these classes are colourless in the reduced form), and “transition ranges,” i.e., the potential range from the first appearance of colour to its full development. Potentials and ranges are both dependent on experimental conditions, and are valid when the indicator is used under the same conditions.But the ranges do not unequivocally define the number of electrons involved in the indicator reaction, either by the slope of the potentiometric curve and the magnitude of the range, or by the amount of oxidant consumed. Manual spectrophotometric examination of a decomposing species is obviously difficult, and only when the decomposition is slow can reasonably accurate values of wavelengths of maximum absorption (Amax.) and of molar absorptivities ( E ) be obtained.Mechanistic 0 SAC and the authors.BISHOP AND HARTSHORN 27 investigations have been few, and are empirical and largely speculative. It is much easier to rationalise the oxidation of benzidines on a basis of a one-electron mechanism than on a basis of a two-electron mechanism. The little unequivocal evidence available supports a two-electron mechanism ; the remainder favours a one-electron mechanism. Kolthoff and Sarverl postulated that arylamines were first oxidised by an irreversible two-electron bi- molecular process to the colourless diarylbenzidine, and that this benzidine was then reversibly oxidised by a two-electron unimolecular reaction to a diquinonediimine.The blue or violet oxidised form is unstable and is destroyed by an irreversible process. Further, frequent reports suggest that an intermediate product occurs in the oxidation of the benzidine, and this is described as a molecular complex of the benzidine and the diquinonediimine. For the unsubstituted parent compound, and ignoring protonation- dipheny lamine diphenylbenzidine diquinonediimine w+-e intermediate semiquinone I? destruction This theory has since been tacitly accepted as applying generally to all arylamines and benzidines and, by analogy, to naphthidines. However, the intermediate diphenylbenzidine has not been isolated and characterised, except by reduction of the purple oxidation product, nor have the intermediate semiquinone and the destruction product been examined.Further, the oxidation product formulated as a diquinonediimine could not be expected to have an intense blue colour. It has been reported, but in only one instance, that the original reduced form of the indicator is regenerated during the decomposition reaction,2 thus suggesting a disproportionation. Many questions, even one so fundamental as how many electrons are involved in the oxidations, remain to be answered, therefore, before the behaviour of these indicators can be rationalised. This brief introduction could be annotated by 200 or more references to the literature, and be extended to include a detailed argument of the possibilities, but this has been done in a recent monograph3 and need not be repeated here. Earlier work was carried out without the benefit of modern instrumentation, and it was thought that fast scanning spectrophotometry might be used to define the number of electrons involved in the oxidations and to give some information on the formation of intermediates and on the decomposition reactions.As it seemed likely that free radicals would be involved, electron spin resonance spectrometry might also yield some useful information. EXPERIMENTAL REAGENTS- Indicators-These were purified by crystallisation from water or dilute sulphuric acid, or by precipitation from concentrated sulphuric acid by dilution and cooling, under a carbon dioxide atmosphere.2 They were examined for impurities by appropriate thin-layer chromato- graphic methods.Commercial diphenylbenzidine was found to contain about mole per cent. of dichromate, otherwise the compounds appeared to be pure individual authentic species : commercial samples contained traces of oxidised species. Determinate stock solutions, usually M, were prepared in water or concentrated sulphuric acid and diluted as required. SuZPhuric acid-Aristar or AnalaR grades, free from traces of oxidants, were used concentrated, or after dilution with distilled water. Concentrations were determined by precise density measurements.28 BISHOP AND HARTSHORN : SOME OBSERVATIONS ON OXIDATION - REDUCTION [Analyst, Vol. 96 Cerium(1V) solutions-A stock 0.1 M solution of cerium(1V) in M sulphuric acid was prepared from AnalaR ammonium hexanitratocerate(1V) by the usual method: and stan- dardised against freshly prepared primary standard 0.05 M arsenic(II1) by using osmic acid as the catalyst and tris(1,lO-phenanthroline)iron(II) as the indicator.More dilute solutions were prepared by diluting the stock solution with 1.0 M sulphuric acid. Dichromate solutions-Stock 0.01667 M aqueous solutions were prepared determinately from AnalaR potassium dichromate and diluted with water as required. INSTRUMENTS- Spectral scanning was carried out over a pre-set wavelength range on a Unicam SP800B double-beam recording spectrophotometer, with the appropriate solvent in the blank beam. The repetitive scan mode of the instrument was used in preliminary kinetic exploration of decomposition rates, the time of passing the wavelength of maximum absorption being measured by stop-watch for each scan.PTFE-stoppered 10-mm Spectrosil cells were used. A Perkin-Elmer 137 UV double-beam recording spectrophotometer was also used in a similar fashion as a check. Didymium and holmium glass wavelength standards were used for calibration. Precise absorbance measurements at fixed wavelength were made with a Hilger H700 single-beam spectrophotometer, with blank and sample cells mounted in a constant-tem- perature housing fed from a thermostat tank maintained to within +0.05 "C of the required temperature. Molar absorptivities were thus determined and precise values of Amax. checked on this instrument. Electron spin resonance measurements were made by courtesy of the Physics Depart- ment on a Decca electron spin resonance spectrometer, operating in the Q band, but were makeshift as proper solution handling components were not available. METHODS DETERMINATION OF NUMBER OF ELECTRONS AND OF APPROXIMATE VALUES FOR Amax, AND E- A series of solutions in 100-ml calibrated flasks was prepared, each containing the same amount of diluted indicator solution, so that the absorbance at the wavelength of maximum absorption would be between 1.0 and 1.5 (2 x to 1 0 - 4 ~ after making up to volume), and the required amount of sulphuric acid to give the desired concentration after making up to volume.To these were then added diluted cerium(1V) or dichromate solution in amounts corresponding to 0, 06, 1.0, 1.5, 2.0, 3.0 and 4.0 equivalents with respect to the amount of indicator present.After addition of the oxidant, the solution was made up to volume with water, quickly mixed and immediately scanned through the appropriate wavelength range to include both ultraviolet and visible peaks. The series of spectra was recorded on the same chart. The spectra over the range from 200 to 850nm were carefully examined for evidence of intermediate or other products formed during the oxidation process and also with excess of oxidant. Approximate values for Amax. for reduced and oxidised forms of the indicator were noted for precise checking later, and approximate values of E were calculated from the absorbances at the wavelength of maximum absorption. A graph of absorbance against number of equivalents of oxidant added gave the number of electrons involved.Rate of decay of oxidised form-The rate of decay of the oxidised form was explored by repetitive scanning of the samples at timed intervals. The main examination was carried out on the samples treatedwith two equivalents of oxidant, but other ratios were also examined. During the decay process the spectra were again examined over the range 200 to 850nm for evidence of other products. More precise studies of decay were made at constant tem- perature and at a fixed wavelength of Amax, in the manual spectrophotometer. From the results thus obtained, rate plots for orders of reaction from zero to fourth order were made. Precise measurement of Amax. and E values-By using the Amax, values from spectral scans as a guide, and taking advantage of the small induction period usually allowed by a purified indicator, the region t 5 to 10 nm about the approximate Amax, value was quickly surveyed on the manual instrument with a solution containing two equivalents of oxidant.The molar absorptivity at the wavelength of maximum absorption was then calculated. A similar check was made on the reduced form of the indicator.January, 197 I] INDICATORS OF THE BENZIDINE, NAPHTHIDINE AND DIARYLAMINE TYPES 29 Determination of the amount of reduced form regenerated-A solution containing the stoicheiometric amount of oxidant was allowed to decay completely, after its absorbance at the wavelength of maximum absorption had been determined immediately after mixing. After decay the spectrum was again scanned and showed no more than a slight elevation over the base-line in the visible region.The decayed solution was then divided into aliquots, which were treated one after the other with successively larger increments of oxidant, the required amount of sulphuric acid was added and the solution made up to volume and im- mediately scanned. From a graph of absorbance against amount of oxidant added, the amount of the latter required fully to oxidise the regenerated reduced indicator was found. This was checked against the maximum absorbance reached which, from Beer’s law, also gave the amount of regenerated reduced form. As will be explained, the second method is the more reliable. RESULTS AND DISCUSSION The primary questions concerned the oxidation of benzidines of the form When X=Y =Z=hydrogen, the parent compound benzidine (4,4’-diamino-l ,l’-biphenyl) appears ; when X=phenyl and Y =Z= hydrogen, the putative intermediate diphenylbenzidine formed from diphenylamine appears; when X=Y =hydrogen, the compound is a nuclear- substituted benzidine, X or Y, or both, being other than hydrogen give N-substituted benzidines.Addition of a fused benzene ring on the side opposite to 2 gives the corresponding naphthidines (4,4‘-diamino-l, 1‘-binaphthyl) . As many compounds as could be readily obtained were examined, and gave a reasonable cross-section of the whole. These were X=Y =hydrogen and Z=hydrogen, benzidine ; Z=methyl, 3,3’-dimethylbenzidine (0-tolidine) ; Z=methoxyl, 3,3’-dimethoxybenzidine (0-dianisidine) ; Z=hydrogen, X= hydrogen and Y = phenyl, diphenylbenzidine ; and X =methyl and Y = phenyl, NN’-dimethyl-NN’-diphenyl- benzidinedisulphonic acid.Naphthidine (X =Y =Z= hydrogen) and 3,3‘-dimethylnaphthi- dine (X=Y =hydrogen and Z=methyl) provided examples of the naphthyl analogues. Of the presumptive precursor diarylamines, diphenylamine, diphenylamine-4-sulphonic acid, N-methyldiphenylamine-4-sulphonic acid and 2-carboxydiphenylamine (N-phenylanthranilic acid) were also examined; the last should give a benzidine with X=hydrogen, Y=phenyl and Z=carboxyl, but the location of the carboxyl groups is indeterminate. NUMBER OF ELECTRONS- In the indicator reaction, Index + ne + Indrea the number of electrons, TZ, is in all instances 2. Graphs of absorbance of Indo, against equivalents of oxidant added are linear for all compounds examined up to 2 +_ 0.02 equiva- lents per mole of Indred, whether this be a benzidine, a naphthidine, an arylamine or an alkylarylamine : the absorbance remains constant when oxidant in excess of the two equiva- lents is added.SPECTRA- Immediate& after oxidatiofl-All spectra show a cut-off in the far ultraviolet region because of high sulphate concentrations and aromatic ring currents. Usually Ind,,d shows a peak in the mid or near ultraviolet region; the wavelength is longer for the benzidine than for the arylamine, exceptions being for Z=alkoxyl, when a double peak appears, and the naphthi- dines, when there is a double hump or a complex spectrum: none shows any other absorption above the vibrational frequencies. The Ind,, spectrum shows a single peak in the visible region, except for Z=alkoxyl, when the peak is again split, and naphthidine, which shows an additional weaker higher frequency absorption.Starting with pure Indred and scanning30 BISHOP AND HARTSHORN : SOME OBSERVATIONS ON OXIDATION - REDUCTION [Arzalyst, Vol. 96 solutions containing progressively larger increments of oxidant, the ultraviolet absorbance of Ind,,a decreases proportionately, usually to zero, and the visible absorbance of Ind,, appears and increases proportionately until two equivalents of oxidant have been added; thereafter there is no further change. At the same time the absorbance of solvated cerium(II1) appears in the ultraviolet /(Amax = 252 rnm, E = 905 1 mol-1 cm-l) and increases proportion- ately, reaching a maximum when two equivalents of cerium(1V) have been added.The changes in these absorbances are precisely linear with the number of equivalents of oxidant added, terminating at 2 f 0.02 equivalents. These peaks show no further change when more than two equivalents of cerium(1V) are added, but the absorbance of the latter (Amax, = 317 nm, E = 5 900 1 mol-l cm-l) appears and increases in proportion; there is no evidence whatever of a further oxidation product of the indicator. The series of Indred - Indo, spectra shows a characteristic isosbestic point pattern (Fig. 1). The shape of the Ind,, spectrum is the same 0.8 L 2 0-4 0.2 0 I I - 0) i 0.6 - t L I t ’ 250 275 300 325 I I 1 I 350 400 450 500 550 Wavelength/ nm Fig. 1. Illustration of the isosbestic pattern on oxidation.Solutions 4 x M in N-methyl- diphenylamine-4-sulphonic acid and 2.0 M in sulphuric acid treated with a, 0; b, 0.60; c, 1-21; d, 1-72; e, 2.0; and f, 2.2 equivalents of cerium(1V): e and f are coincident as that of Ind,,a, unless the latter is an arylamine, with merely a bathochromic shift and an increase in molar absorptivity, which are nearly constant for a given type of compound. This indicates that there is no basic change in structure but merely a decrease in mean energy of the electronic transitions. Spectral measurements are collected in Table I. TABLE I SPECTRAL CHARACTERISTICS OF SOME BENZIDINE AND NAPHTHIDINE DERIVATIVES AND SOME ARYLAMINES IN 2.0 M SULPHURIC ACID AT 25 “C Indred Indo, & - Spectral Compound Amax/nm 6/1 mol-1 cm-l Amar/nm ~ / 1 mol-l cm-I shift/kcal mol-l Cox/Cred Biphenyl 252 20000 (in ethanol) Benzidine 248 20 300 426 69 500 50-8 3.3 3,3’-Dimethyl- 248 18000 438 - 49 benzidine 250* 18000* 435* 65 000 48.6 3.6 12 800 454 31 000 49 2.4 8 100 510 23 000 41.5 2.8 31 200t 560 60 000 61.9 1.6 benzidine { E 3,3’-Dimethoxy- NN’-Diphenylbenzidine 253t N-Phenyl- N-Methyldiphenyl- 3205 24 OOOj 511 44 000 62 1.7 Diphen ylamine 220: Cut-off 565 45 000 - - anthranilic acid 256: - 524 30 000 - - amine-4-sulphonic 294: 14 OOO$ acid - 400 7 000 30 {E 11 000 527 18000 43 1.6 Naphthidine 3,3’-Dimethyl- naphthidine Complex spectrum 543 35 000 - - * In M sulphuric acid.t In concentrated sulphuric acid. $ Arylamine form. 5 In benzidine form; oxidised with two equivalents of oxidant and immediately reduced with zinc dust.January, 19711 INDICATORS OF THE BENZIDINE, NAPHTHIDINE AND DIARYLAMINE TYPES 31 When Indr,d is an arylarnine, a slight change in shape occurs for the Ind,, spectrum, but when Ind,, is reduced to the presumptive benzidine and the new Indred spectrum com- pared with that of the arylamine, the change is seen to be minor.Arylamine oxidation occurs in a single step: no evidence exists for the intermediate formation of the benzidine. Lest this be due to use of a very strong oxidant [cerium(IV)], dichromate, vanadate, iron(II1) and other oxidants were tested : either no oxidation occurred, or the single-step two-electron oxidation took place. If the arylamine is oxidised with two equivalents of oxidant and then immediately reduced with zinc dust, a scan shows that the ultraviolet peak is shifted to a longer wavelength, thus indicating the benzidine form.Subjecting this solution to the oxidation and scanning process shows that the oxidised form is identical in spectral charac- teristics with the oxidation product of the original arylamine, but the maximum absorbance is now reached at one equivalent of oxidant per mole of original arylamine, which corresponds to two equivalents per mole of the presumptive benzidine produced by the zinc dust reduction. The example shown in Fig. 1 is an alkylarylamine; reduction with zinc dust and repetition of the process gives a similar set of spectra with the ultraviolet peak shifted to 320 nm and a corresponding shift in the isosbestic point. As the sulphuric acid concentration of the medium is increased, a considerable batho- chromic shift occurs in Amax,, which is caused by the solvent effect, as shown in Table 11.Belcher has observed this shift with naphthidine derivative^.^ In sulphuric acid media there is no detectable evidence in any of the spectra for the formation of an intermediate, even transiently, either of the semiquinone type from a benzidine, or of a benzidine from an arylamine. In acetic acid some evidence exists that a different reaction may occur, and that this is specific to the presence of acetic acid or acetate ion. (Kolthoff's solutions con- tained acetic acid.l) For example, a double-humped peak in the region 300 to 420 nm was observed in the oxidation of diphenylbenzidine in a mixture of concentrated sulphuric acid, glacial acetic acid and water (2 + 78 + 20 v/v after correction for the addition of aqueous acid cerate) ; peculiarities have also been noted with o-dianisidine in acetate media.2 This, however, opens up a large field of subsidiary study, which will not be pursued at present.On decompositiort-On standing, all oxidised solutions, whether containing a deficiency, the stoicheiometric amount or an excess of oxidant, faded more or less rapidly, eventually becoming colourless and occasionally depositing a precipitate. In scanning the spectrum at intervals during this decay, as the absorbance of Ind,, decreased, the absorbance of Indred appeared and increased with time, again giving an isosbestic pattern as illustrated in Fig. 2. The absorbance of Indo, fell virtually to zero, but the absorbance of Indred increased to only about one half of its original value.Apart from N-phenylanthranilic acid, no other peaks appeared and there was no evidence for the formation of a semiquinone or any other compound, except that the base-line of the whole spectrum rose by a small but detectable and fairly uniform amount, such as may be expected from the formation of an insoluble but disperse phase. There was no change in shape or in the Amax values of either In&, or In&,d peaks during the decay process, except that an arylamine decayed to the benzidine giving the same Indred spectrum as that after zinc dust reduction of the freshly oxidised arylamine. Even in solutions containing one equivalent of oxidant, which might be expected to encourage 0.8 - a 0.2 - I I I I I I I O t 275 300 325 350 400 450 500 550 t Wavelength /nm 10 Fig.2. Illustration of the isosbestic pattern on spontaneous decomposition. A solution M in naphthidine and 2.0 M in sulphuric acid treated with 2.0 equivalents of cerium(1V) 4 x and allowed to decay. Interval between scans a to e, about 12 minutes32 BISHOP AND HARTSHORN : SOME OBSERVATIONS ON OXIDATION -REDUCTION [Analyst, Vol. 96 formation of the semiquinone, no spectral shift was observed, although decay was accelerated. Evidence for autocatalysis was found in many instances. The cerium(II1) and cerium(1V) peaks, when present, complicate interpretation of the spectra, but their resolution is possible from the law of additive absorbances.Unsubstituted benzidine and naphthidine oxidised forms decay rapidly in daylight, but in darkness, or in the cell housing of the spectrophotometer, the decay is drastically retarded, thus indicating that the process is photosensitive. Belcher, Lyle and Stephen5 have reported that 3,3'-dimethoxynaphthidine is photosensitive at the wavelength of maximum absorption and they were therefore unable to determine the Amax. or E value for this compound. The curious behaviour of N-phenylanthranilic acid is noteworthy. It is directly oxidised, without any intermediate benzidine formation, to the bluish red Ind,,. The Ind,, band at 524nm immediately begins to decay and a new band at 436 nm grows and the colour changes to green. This decay is rapid with a half-life of 10 minutes, and the set of spectra scanned at 3-minute intervals shows a good isosbestic point, smeared only slightly by the over-all slower decay of both species.Re-oxidation of the decayed solutions gave an identical Ind,, peak and the same Indred - Ind,, isosbestic behaviour. When the decay of a solution treated with excess of oxidant was followed by the scanning method, the decay was slower to begin with because of re-oxidation of the regenerated Indred by the excess of oxidant, but the cerium(1V) absorbance decayed more quickly than could be accounted for by this reaction: the cerium(1V) absorbance is therefore useless in assessing the extent of reaction. ELECTRON SPIN RESONANCE SPECTRA- Without the proper components for work with solutions the electron spin resonance spectra have only a qualitative value.Solutions of diphenylbenzidine with a range of sulphuric acid concentrations and amounts of oxidant were examined. Cerium(II1) and cerium( IV) gave no signals at the operating frequencies. A half-oxidised solution showed a signal that grew gradually to a maximum, and showed a Land4 splitting g factor of almost exactly 2, indicative of a free radical with a single unpaired electron, and then decayed at a rate similar to the rate of decay of absorbance at 560 nm previously observed spectrophoto- metrically. The band was unusually broad, probably on account of solvent effects, and may arise from an averaging effect, so that it cannot be regarded as evidence of a semiquinone, particularly in view of the absence of optical evidence.As the amount of oxidant was in- creased, finally to two equivalents, this band disappeared and was replaced by a weak complex signal that can be related to, but does not with any certainty identify, a diradical with two separated electrons of parallel spin. KINETICS OF DECOMPOSITION- The spectral scanning method gave a clue to the pattern of decomposition, there being some indication of a small induction period with pure indicators and strong evidence of autocatalysis. Precise kinetic measurements were made by following the change in absorbance at the wavelength of maximum absorption of Ind,, with time, at controlled temperature. The decay rate was found to increase with rising temperature, to decrease with increasing sulphuric acid concentration (as shown in Table 11), and to depend on the nature of the TABLE I1 SOLVENT SHIFT AND EFFECT OF SULPHURIC ACID CONCENTRATION ON DECAY OF OXIDISED DIPHENYLBENZIDINE AT 25 "C Sulphuric acid, per cent.. . 18 22 36 63 Amax. Indox/nm . . . . 560 565 570 585 Half-lifelminutes . . .. 188 200 445 1300 substituents X, Y and 2. The stabilising effect of high sulphuric acid concentrations is implicit in some earlier work,3 and may be caused by the decrease in water concentration, or by increased protonation of the various basic species. The decay of the oxidation products of the unsubstituted benzidine and naphthidine is greatly accelerated by exposure to daylight as has been noted. Sulphonation of the indicator de-stabilises the oxidised form, as has several times been observed b e f ~ r e , ~ , ~ and could well result from a shift in the charge densityJanuary, 19711 INDICATORS OF THE BENZIDINE, NAPHTHIDINE AND DIARYLAMINE TYPES 33 distribution.Partially oxidised solutions decay faster than stoicheiometrically oxidised solutions, and addition of Indred to a decaying solution accelerates the process markedly, as with diphenylbenzidine, giving support to the hypothesis that the autocatalysis is caused by the growth of regenerated Ind,ed in the solution. The kinetics of the decay process do not fit any reaction order, but come closest to first order; indeed, some show an almost perfect first-order plot. It is possible to fit the decay process for diphenylbenzidine to an equation of the form- - k, [Ind,,] + k, [product] d [Indod dt in 2 M sulphuric acid ( k , = 2.3 x Rather than present artificially fitted rate constants, the speed of the decay process is represented by the half-life in Table I11 under the conditions specified.As Indred in the benzidine form minute-l and k , = 8.2 x lW3 minute-l). TABLE XI1 HALF-LIVES OF IND,,~ IN THE ABSENCE OF EXCESS OF OXIDANT, AND IN THE INITIAL ABSENCE OF IN&&, IN 2 M SULPHURIC ACID AT 26 “C; AND THE RECOVERY OF INDred AFTER COMPLETE DECAY Compound Benzidine . . .. .. .. .. 3,3’-Dimethylbenzidine . . .. .. 3,3’-Dimethoxybenzidine .. .. NW-Diphenylbenzidine . . .. .. Diphenylamine . . .. .. .. N-Phenylanthranilic acid .. .. Naphthidine .. .. .. N-Methyldiphenylamine-4-sulphonic acid]) 3,3’-Dimethylnaphthidine . . .. Half-lifelminutes .. 500* ..264 .. 190 .. 188 .. 96 .. 6 to 105 . . 170 .. 80 200 : . - 23’11 Indrd regenerated, per cent. sot 50 46: 60 50, 51, 46 1005 - 65 40 35, 40 * Measured in darkness. Decay accelerated in daylight to about 50 minutes. t Value very high. $ In M sulphuric acid. Q Initial fast decay to green form. 11 Presumptive benzidine derivative : Ind,,d oxidised with two equivalents of oxidant, im- 7 At 30 “C. Decay much faster in daylight, or under continuous illumination a t wavelength mediately reduced with zinc dust, then re-oxidised. of maximum absorption (about 2-5 minutes). is regenerated in the decay process, the latter must be of the nature of a disproportionation with a minimum molecularity of two, and therefore the rate-controlling step must be a preceding unimolecular reaction that could be accelerated by energy exchange with Indred.Substituents in the 2 position appear to protect the amino group, as Belcher, Lyle and Stephen ~uggest,~ but this is perhaps more a matter of preventing the photochemical reaction, with the exception noted,5 than of hindering further oxidation. It is notable that the half-life of oxidised diphenylamine is about half that of diphenylbenzidine. Substituents in the X or Y position stabilise the oxidised form, and maximum stability is reached when both X and Y are substituted. The instance of NNN’N‘-tetramethyl-3,3’-dimethylbenzidine (t et ramethyl-o-t olidine) is strongly relevant. RECOVERY OF INDred- Determination of the amount of Indred regenerated would define the stoicheiometry of the decomposition reaction.Re-oxidation of the regenerated Indred gives precisely the same spectrum and decay pattern as the original indicator. Measurement of the absorbance at the appropriate wavelength of the regenerated Indred and calculation from Beer’s law will give the amount of Indred, provided proper correction for other absorbances such as cerium(II1) and decomposition products is made. Measurement of the absorbance of Ind,, for a series of solutions to which successively larger amounts of cerium(1V) have been added and a plot of absorbance against equivalents of oxidant do not give reliable results, because cerium( IV) is consumed in other reactions, as noted before. However, addition of a slight excess of34 BISHOP AND HARTSHORN : SOME OBSERVATIONS ON OXIDATION - REDUCTION [A~zalyst, Vol. 96 oxidant [checked by the appearance of a small cerium(1V) peak], measurement of the absorb- ance of Indo, and application of Beer’s law give reasonably reliable results, provided the measurement is made quickly enough to avoid significant decay and yet allow sufficient time for the full development of colour. This latter time is no more than a few seconds except for benzidine, which takes up to 30 minutes.It is not possible to check the stoicheio- metry by adding the integrated number of equivalents of cerium(1V) (e.g., four for a 2 : 1 disproportionation) so that there is no residual Indred left after decay. Indrea is regenerated even in the presence of a large excess of oxidant, and this phenomenon persists until about twenty-two equivalents of oxidant have been added, so clearly oxidant is attacking a de- composition product by ring fission faster than Indred is regenerated.If the disproportionation is formulated as m Ind,, 3 D + % Indrea where D is the primary decomposition (disproportionation) product, then excess of oxidant is consumed by three processes: (a) re-oxidation of regenerated Indred to In&,, (b) direct oxidation of Ind,, to D, and (c) attack of D to give fission products, and of these reaction (c) is the major process. Results are included in Table I11 and, on balance, favour a 2 : 1 dis- proportionation (50 per cent. regeneration) , although other stoicheiometries such as 5 : 2 (40 per cent. regeneration) are not excluded, and there are notable exceptions. CONCLUSIONS In sulphuric acid media all indicators are oxidised directly to Ind,, by a single two- electron step.Reduction of oxidised arylamine stops at the benzidine stage. Oxidation of arylamines does not proceed by a benzidine tran~formation,~ and is not, therefore, an intra- molecular reaction but an intermolecular reaction. Starting from an arylamine there are five possible products depending on which carbon atoms provide the bridge, and if the arylamine is substituted in the nucleus the possible products are multiplied in number accord- ing to the rings in which the substituents finally appear: there are fifteen possible products, for instance, for N-phenylanthranilic acid, although it is probable that a single product will preponderate in a given instance.Presupposing that the product is a p,@’-benzidine and the substituent 2 appears in the benzidine nucleus, the main reaction can be formulated as- bentidine base Y NX xi&x - protonated benzidine -2H+-2e L - arylamine (Y = phenyl) where Ind,, is a triplet diradical dication. It should be emphasised that benzidine has not been isolated except after zinc dust reduction of In&, for the case when X = 2 = hydrogen and Y = phenyl. The identity in shape of the benzidine and Ind,, spectra and the isosbestic pattern suggest that no gross change occurs on oxidation, and the bathochromic shift indicatesJanuary, 19711 INDICATORS OF THE BENZIDINE, NAPHTHIDINE AND DIARYLAMINE TYPES 35 an elevation of the mean ground-state energy of the molecular orbitals, so that electronic transitions require a lower energy; this is supported by the increase in molar absorptivity.Except for the curious instance of N-phenylanthranilic acid, Indred is invariably re- generated on decomposition (as the benzidine in the case of arylamines), together with a weakly coloured sparingly soluble decomposition product, D. Regeneration of Indred is clearly the reverse of the main indicator reaction. Neglecting the possible participation of hydrogen ion, this is- Ind,, + 2e + Ind,,d The necessary electrons must come from somewhere, and as Indo, will itself consume oxidant, it is the obvious source- The spontaneous decomposition of Ind,, is therefore a disproportionation, several examples of which are known in this field of chemistry.3 However, the minimum value of $’ for a disproportionation would be 2, and the decay does not show second-order kinetics.A monomolecular rate-controlling step must therefore precede the disproportionation, and it is not unlikely that this should be the relaxation of the triplet state, catalysed by energy exchange with Indred, to the xIndo, + y D + x e $’ Indo, -+ 4 D Y Indred hd4+J=(==J:x DQDI diquinonediimine; it could also be the deprotonation of DQDI to give the free base. The final reaction could then be formulated as for which the over-all stoicheiometry would be (a + b)/c. The product D has been neither isolated nor examined, and even if it had it would be difficult to be sure that it had not been changed in the process. Discussion of the nature of D is necessarily speculative, but there is obviously a connection between the substituents X, Y and 2 and the stability of In&,.X and Y are those which matter in the decay process, and it is worth recalling that stability increases with increasing sulphuric acid concentration, so that protonation may be influential. Moreover, biphenyl is itself unreactive so that the activity must reside in the nitrogen atoms. When neither X nor Y is a hydrogen atom, then the oxidation product, whether formulated as the diradical dication or the diquinonediimine, has no lone pairs available for protonation, and it is not possible to formulate a further oxidation stage without breaking bonds or ejecting substituents. That the tetra-N-substituted derivatives are stable in the oxidised form is well known,3y6 but a quantitative definition of stability is lacking.That the oxidised form of N-methyldiphenylamine-4-sulphonic acid does decompose and regenerate a benzidine may be the result of the de-stabilising effect of the sulphonic acid group but is still difficult to understand. When either X or Y is a hydrogen atom, then protonation is possible and the diquinonedi- imine does offer electrons accessible to oxidation. A 2 : 1 disproportionation can be formu- lated as- a Indo, + b DQDI -+ D + c Indred where the primary decomposition product, D, is the diradical dication of the DQDI. This unlikely looking species would be very reactive and may polymerise by nuclear attack. A 3 : 1 disproportionation is easier to formulate but is contrary to the 50 per cent. recovery Of Indred.36 BISHOP AND HARTSHORN When X=Y =hydrogen, the 2 : 1 disproportionation would again require formation of the diradical dication DQDI, but with hydrogen in place of Y, which opens up further possibilities. A 3 : 1 disproportionation would allow the formation of a hydrazine that is capable of being oxidised to an azo compound, which could isomerise to a dimeric diquinone- diimine or could be oxidised to an azoxy compound. The formation of a linear trimeric bis-hydrazine would give a 5 : 2 stoicheiometry and a 40 per cent. recovery of Indred. There is also the possibility of a phenazine type of polymerisation. It is not profitable to enumerate further all the possibilities because the matter remains speculative until accurate information is available on the nature of D. However, the stability of Ind,, increases as 2, Y and X are successively substituted and as the sulphuric acid concentration increases, and the recovery of Indred suggests that the oxidation of Ind,, to D is a two-electron process. REFERENCES 1. 2. 3. 4. 5. 6. Kolthoff, I. M., and Sarver, L. A., J . Amer. Chem. SOC., 1930, 52, 4179. Crawford, A. B., and Bishop, E., J . R. Tech. Coll. Glasg., 1950, 5, 62. Bishop, E., “Indicators,” Pergamon Press, Oxford, 1970, Chapter 8B. Smith, G. F., “Cerate Oxidimetry,” G. F. Smith Chemical Co., Columbus, Ohio, 1942. Belcher, R., Lyle, S. J., and Stephen, W. I., J . Chem. SOC., 1958, 4454. Jordanov, N., and Daiev, Ch., Talanta, 1963, 10, 163. Received April 16th, 1970 Accepted August 19tk, 1970