J. Chem. SOC., Faraa'ay Trans. 1, 1982, 78, 1269-1278 Triplet State Electron Spin Resonance Studies of Aryl Cations Part 5.-The Aryl Cation/Aryl Radical Product Ratio in Photolysis of Arenediazonium Salts at 77 K1 BY HANNA B. AMBROZ AND TERENCE J. KEMP* Department of Chemistry and Molecular Sciences, University of Warwick, Coventry CV4 7AL Received 18th June, 1981 An e.s.r. study has been made of the factors (solute concentration, matrix, irradiation wavelength and time, post-photolysis storage) affecting the relative yields of 3Ar+ and Ar. in the photolysis of an arenediazonium salt at 77 K. The principal findings are: (i) (in contrast with the situation prevailing in solution) photoheterolysis is the sole primary process; (ii) production of Ar - during irradiation comes from photoreduction of 3Ar+ ; (iii) there is a very powerful solute-concentration dependence of the photo- conversion of 3Ar+ into Ar.(which does not proceed at very low solute concentrations); (iv) conversion of 3Ar+ into Are proceeds slowly in the dark. Mechanistic aspects of these observations are discussed. We have shown previously that the photolysis at 77 K of arenediazonium salts suitably substituted with good n-donors, eqn (l), yields two principal reactive intermediates, namely the triplet state aryl cation, 3Ar+, and the aryl radical, Are, both of which can be trapped and identified by their e.s.r. spectra,2 which are especially characteristic for 3Ar+. In the case of other arenediaz- onium salts, the electron configuration of Ar+ is singlet and no e.s.r. signal is expected or observed for this species, and only the absorption of Are is apparent.The ratio [3Ar+]/[Ar *I, denoted T/R, which could reflect the relative importance of eqn (1) and (2) if the trapping efficiencies of the intermediates are identical, has been measured2 and varies from 4.0 (for the 3-methoxy-4-morpholino compound) to much smaller ratios than 1 .O for most other arenediazonium salts. These figures refer to our normal experimental procedure of photolysis using irradiation from an unfiltered 200 W high-pressure Hg arc for periods of several hours using microcrystalline samples. We also noted during previous experiments that T / R varied with the irradiation time. When this point was explored in more detail, it was found that the variation could be extreme, and that T / R was also influenced strongly by (i) the wavelength of radiation, (ii) the type of matrix, (iii) the concentration of diazo salt and (iv) storage of photolysed sample in the dark for some hours at 77 K.In view of the current interest 1269c h) 4 0 TABLE 1 .-DEPENDENCE OF T / R RATIOS UPON EXPERIMENTAL CONDITIONS time of photolysis/min storage U.V. light visible light time in matrices dark/h 4 10 30 60 120 4 10 20 30 60 120 microcrystalline 0 powder 24 48 72 LiCl/H,O glass 0 (high [Dz+l) 5 24 48 72 (low [ Dz+]) 24 48 72 LiCl/H,O glass 0 1.21 0.442 0.304 0.273 0.036 0.015 0.014 0.013 9.17 3.87 2.67 2.51 - 0.736 0.333 0.232 0.172 0.051a 0.025 0.020 0.01 8 2.93 1.38 1.16 0.95 - 0.469 0.235 0.164 0.119 0.1 34a 0.071 0.059 0.053 0.686a 0.300 0.229 0.198 - 0.1 372a 0.149 0.107 0.079 0.014a 0.0067 0.0067 0.0060 0.3Sa 0.142 0.116 0.103 - 0.214a 0.118 0.07 1 0.052 0.036a 0.023 0.019 0.017 0.63tIa 0.38 1 0.309 0.303 - 3.41 4.40 1.68 2.11 1.70 1.87 1.41 1.68 0.674 3.64 0.366 1.96 0.301 1.84 0.306 1.53 0.275 1.37 3.88 3.50 1.91 1.73 1.58 1.69 1.93 1.73 8.02 9.73 3.60 3.99 2.38 2.63 2.19 2.42 - 1.42 - 0.585 - 0.522 - 0.485 - 3.50 - 1.54 1.38 - 1.28 - - - 7.18 5.16 2.6 1 2.23 1.59 1.73 1.43 1.56 0.083 0.041 0.045 0.027 0.041 0.023 0.038 0.022 9.50 8.58 4.04 3.24 3.67 2.94 3.38 2.37 - - a Different instrumental settings (1.0 mT field modulation at 10, 20 or 26 dB microwave power).H.B. AMBROZ A N D T. J. KEMP 1271 in solid-state photo~hemistry,~ we now report these rather unusual phenomena in detail, taking as our example the behaviour of 2,4,5-trimethoxybenzenediazonium hexafluorophosphate, Dz+PF;, which was typical in giving a very small value for T / R on prolonged U.V.photolysis. EXPERIMENTAL Samples were of three kinds, namely (i) microcrystalline, (ii) solutions in LiCl/water/acetone (ca. 6.0 mol dmP3 LiCl in 50/50 v/v solvent mixture) at high [Dz+PF;] (ca. 0.5 mol dm-3) and (iii) solutions in water/acetone/LiCl (ca. 6.0 mol dm-3) at low [Dz+PF;] (ca. 0.1 mol dm-3). Samples were made up and handled as in ref. (2a). PHOTOLYSIS TIMES Photolysis times were varied from a few minutes (i.e. much shorter than we have used before) to 2 h: the e.s.r. signals achieved during short photolyses were naturally of much weaker intensity and were measured with particular care (using slow scans, large time constants, etc.).All photolyses and measurements were conducted at 77 K, and e.s.r. spectra were recorded immediately afterwards (and then again after several set times of storage in the dark). Irradiation was performed using two types of light source: in one case this was a 100 W high-pressure Hg arc (unfiltered) and in the second a 200 W tungsten bulb (ie. giving no light at 2 < 350 nm); radiation from these sources is referred to as ‘u.v.’ and ‘visible’ light, respectively. E.S.R. MEASUREMENTS [3Ar+] was measured from the intensity of the sharp Hmin feature, while [Ar.] was measured from that of the integrated resonance near the free-spin value. The majority of experiments were conducted under the following conditions : field modulation, 2.0 mT ; frequency modulation, 100 kHz; time constant, 0.1 s; microwave power, 10 dB.RESULTS AND DISCUSSION Table 1 summarises the results of varying (i) irradiation wavelength, (ii) photolysis time, jiii) matrix and (iv) storage time of the sample following irradiation. Several trends are apparent which we discuss below. Fig. 1-3 illustrate the dramatic affect of using visible (as opposed to u.v.) photolysis for a microcrystalline sample. While U.V. light gives [Are] > [3Ar+] and a T / R ratio which decreases monotonically with photolysis time, visible light produces a large predominance of 3Ar+ and a T / R ratio which initially increases with photolysis time and then maximises before decreasing. EFFECTS OF IRRADIATION WAVELENGTH A N D TIME I N DIFFERENT MATRICES In all matrices, the use of U.V.light results in higher [Are] and lower T / R ratios than obtained with visible light under comparable conditions, with a single exception; when [Dz+PF;] < 0.1 mol dmW3 in a LiCl glass, [Are] becomes extremely small, i.e. T / R exceeds 20 (table 2). MICROCRYSTALLINE MATRIX Under U.V. irradiation, the T / R ratio is maximal (near unity) at shortest irradiation times, but falls rapidly as photolysis proceeds to an extent requiring adjustment of instrumental settings and some consequent inaccuracy in the measurement of T/R (probably ca. 50% when T / R z 0.1). Visible light produced lower concentrations of paramagnetic species in comparable periods, but the tendency to much higher T / R1272 TRIPLET STATE E.S.R.STUDY OF ARYL CATIONS TRIPLET RADICAL FIG. 1.-E.s.r. spectra of 3Ar+ and Ar. under different irradiation conditions: (a) 30min visible light photolysis, (6) 2 h U.V. photolysis. Field sweep, 0.7 T [spectrum (a) is displaced to right for clarity]. time of photolysis/h FIG. 2.-Dependence of T / R ratios (%) upon times of photolysis and storage in dark (typical for U.V. irradiation of microcrystalline samples).H. B. AMBROZ AND T. J. KEMP 1273 time of photolysis/h FIG. 3.-Dependence of T/R ratios (%) upon times of photolysis and storage in dark (typical for oisible light photolysis of microcrystalline samples). ratios was unequivocal, as was the appearance of a maximum after 30 min of nearly 10. These observations contrast strongly with those found with glassy solutions.LiCl GLASS (HIGH [Dz+]) When [Dz] x 0.5 mol dm-3, U.V. irradiation gives T / R ratios ’< 0.1, which cannot be measured sufficiently accurately to permit many conclusions to be drawn about the effect of photolysis time apart from the obvious one that 3Ar+ production is insignificant under these conditions. However, visible light yielded T/R ratios of up to a maximum of ca. 3.5 after 10 min irradiation. LiCl GLASS (LOW [Dz+]) When [Dz+] x 0.1 mol dm-3, T / R ratios exceeding 9 are obtained with both U.V. light (after short photolysis, but decreasing continually on longer irradiation) and visible light (maximising at 9.5 after 60 min photolysis). The striking contrast with the more concentrated glassy solutions of [Dz+] under U.V.irradiation stimulated a more detailed examination of the dependence of the T / R ratio on [Dz+] summarised in table 2. This (and similar experiments not detailed) reveal a remarkably abrupt change in T / R between 0.08 and 0.12 mol dm-3 [Dz+] covering nearly two orders of magnitude (making T/R extremely sensitive to slight variation in [Dz+]). Such an extreme changeover suggests a microenvironmental effect ; thus Dz+PF; is almost insoluble in LiCl/H,O, requiring acetone co-solvent for solubilisation, and at high [Dz+] the solute ions are probably in clusters solvated by acetone (especially after freezing) and ready electron transfer from the counter-ion to 3Ar+ can occur following1274 TRIPLET STATE E.S.R. STUDY OF ARYL CATIONS TABLE 2.-cONCENTRATION DEPENDENCE OF T / R RATIOS AFTER 4 min U.V.PHOTOLYSIS OF Dz+PF; IN LiCl GLASS (I, 11, I11 correspond to three separate series of measurements) [Dz+]/mol dm-3 (approximate) I I1 I11 comments 0.5 0.049 0.075 0.036 below 10% (i.e. too low to be significant) 0.25 0.169 - - - 0.125 - 0.1 - 3.24 9.17 1.27 more than 1000% ( i e . too high to be 0.063 26.90 17.91 0.03 1 41 S O - 0.015 27.40 - measured accurately) photolysis. At lower [Dz+], the solute exists as individual ion-pairs, which undergo photolysis to the separated ions, viz. 3Ar+ - * - N, - - PF;. The situation with visible light irradiation is less extreme : taking the maximum value of T/R realised, this shows the matrix-dependence : LiCl/H,O LiCl/H,O microcrystalline ([Dz+] 0.5 mol d ~ n - ~ ) < ([Dz+] 0.1 mol dm-3) - - 3.6 9.5 9.7 In summary, at [Dz+] > 0.5 mol dm-3, T/R < 0.1 and 3Ar+ is produced in insignificant quantity, while at [Dz+] < 0.8 mol dm-3, T / R > 20 and Ar is barely formed.(Indeed, the figures for [Are] include absorption from a species centred at g = 2.0 with a total width of 13 mT and features indicating a radical-pair with separation ca. 6.6 A, suggestive of a species [Are - - - PF,].) EFFECT OF STORAGE TIME In all irradiated samples (irrespective of wavelength or time of irradiation, the matrix or diazo concentration), the ratio T / R decreases slowly on storage of the sample in the dark at 77 K. Averaging over all of the 31 individual samples studied, the value of T/R (taken as 1.0 at the initial measurements) changed as indicated in fig.4 for the three different types of sample studied. Clearly T / R declines rapidly over the first 24 h period of storage, but subsequently much more slowly. In one experiment (see table 1) with a sample in glassy LiCl of Dz+PF, (0.5 mol dm-3), we also measured T / R after only 5 h of storage, and found that much of the change in T/R occurred in this initial period. The universal decrease in T / R is to be associated with growth in [Ar-] on storage, coupled with decrease in [3Ar+]. We observed both of these changes many times, and fig. 5 is quite typical, although they were harder to quantify exactly than simple measurement of the ratio T / R . However, the following trends were discerned. (i) The largest relative growths in [Ar *] were found after visible-light irradiation for short periods, reaching nearly 3 times the original concentration achieved after photolysis.The percentage increase in [Are] was smaller for samples subjected to longer visible irradiation, especially for the glassy samples at low [Dz+PF;] when it was as little as 20%.H. B. AMBROZ AND T. J. KEMP 1275 0 24 48 72 96 storage time in dark/ FIG. 4.-T/R ratios after storage in dark in different matrices (initial T / R measured after photolysis taken as 1 .OO). Matrices: 0, microcrystalline powder; x , LiCl/water glass, ca. 0.1 mol dm-3 Dz+; A, LiCl/water glass, ca. 0.5 mol dm-3 Dz+. t -/- I I I I / / I J I 2 4 6 24 tstlh FIG. 5.-Storage-time profiles of [3Ar+] and [Arm] after photolysis: both [3Ar+] and 1Ar.J at storage time zero are taken as 1.0.T = 77 K. Sample: 0.08 M Dz+ in LiCl/H,O/acetone glass. (ii) Following U.V. photolysis, storage produced only a 30-40% increase in [Ar-1, and then only following short irradiation times; samples irradiated for 1-2 h produced only a 3-1074 increase in [Ar.] on storage. (It should be noted, of course, that if a sample has a very small T / R value on conclusion of photolysis, even complete conversion of 3Ar+ into Ar , the presumed sole mechanism for producing extra Ar , will increase [Ar = ] only marginally.) (iii) One special feature of the dilute solutions of Dz+PF; in glassy solution is that no conversion of 3Ar+ into Ar* proceeds in the dark. (iv) In some cases a small net decrease in [Ar J was found. This was typically of1276 TRIPLET STATE EAR. STUDY OF ARYL CATIONS the order of lo%, but the result seems genuine in that the decrease was reproducible and was large compared with signal drift.If there is some real concomitant decay of Are at 77 K, then this means that our estimates of the conversion 3Ar+ + Ar may be slightly lower than the true extent. To explore this point further, we investigated the photolysis of 2,4,6-trimethoxybenzenediazonium hexafluorophosphate, which we presume, from the established behaviour2b in aqueous solution at room temperatures, to give a (diamagnetic) singlet state Ar+ together with Are on photolysis (and, as before, no triplet Ar+ was observed). Storage of this material at 77 K following U.V. photolysis (10 min), both as microcrystalline samples and in glassy LiCl/H,O matrices, gave no enhancement of the Ar 0 signal, implying no conversion of singlet Ar+ into Are, and instead Ar= decayed ca. 10% over 72 h.We believe therefore that Are does decay on this time scale at 77 K. MECHANISMS FOR THE LIGHT AND DARK REACTIONS The most complete model takes the form: primary photochemical processes ArNiX- 14; 3Ar+ + N, + XL ArNzX-: Ar- +N,+X* secondary photochemical processes 3ArS $ ArX ( 5 ) Ar* 2 (Are)* 5 diamagnetic products (6) dark reactions 3Ar++ES +AR* +ES+* (7) (8) (9) ' 3Ar+ + X- + ArX Ar- + R - + diamagnetic products where ES is an electron source (the nature of which is discussed later) while ES+ is its electron-deficient counterpart. REACTIONS DURING PHOTOLYSIS The relative contributions of eqn (1)-(4) change as photolysis proceeds.Most significantly, and unexpectedly, no clear evidence for the primary photodissociation eqn (2) has been found for the arenediazonium salt examined; the ready production of Ar* during photolysis arises purely from the secondary reactions (3) and (4). The contributions of reactions (5) and (6) cannot be assessed accurately, but they are certainly significant and must influence the apparent rates of processes (I), (3) and (4)H. B. AMBROZ AND T. J. KEMP 1277 by providing competing routes for removal of 3Ar+. Under all the various experimental conditions utilised, i.e. different irradiation wavelengths, matrices and [Dz+], the photolysis features three principal stages: Stage I, i.e. the early stage when only process (1) is significant, Ar- is not detected and T/R assumes very high values.Stage 11, i.e. the middle stage when reactions (3) and/or (4) become significant, and Arm appears near g = 2.0. T / R falls steadily to reach 1.0, i.e. [3Ar+] = [Ar.]. Stage 111, i.e. the final stage. Now the secondary processes dominate, and a decrease in T / R to < 1.0 is observed. The exact point during photolysis at which a system reaches the various stages depends on such factors as the irradiation wavelength, the rigidity of the matrix, the energetics of the transformation 3Ar+ + Are and the (solid-state) oxidation potential of ES. The absence of any production of Are at the very lowest [Dz+] in LiCl/H,O/acetone glass even after prolonged irradiation suggests that a critical local concentration of 3Ar+ (and its counter-ion) is necessary for the 3Ar+ + Ar* conversion to proceed.(Cl-, despite its overwhelming preponderance, seems not to act as the major electron source: the function of PF; as the electron source is unambiguous for microcrystalline samples.) We associate the absence of 3Ar+ + Ar conversion at low [Dz+] with the separation of 3Ar+ and PF; by the N, molecule following photolysis of the ion-pair, Dz+PF;. At high [Dz+], the frozen solution contains clusters of the various solute ions, e.g. ArNzPF; 3Ar+ - - - N, - - - PF; hv -+ PF;+N,Ar and ready electron transfer between adjacent ions can cover. Stages 1-111 are more ‘compressed’ in time under U.V. irradiation: the rate of initial production of 3Ar+ is faster, as is the conversion (3) or (4). We expanded the sequence in time simply by reducing the incident U.V.light intensity by moving the sample either away from the source or nearer than its focus. Both procedures led to much higher T / R ratios, e.g. up to 1 .O for LiCl/H,O glass at high [Dz+], while the microcrystalline samples yielded T / R values of 5.5k0.2 (three determinations) for short (4 min) irradiations and 1.8 for a 10 min irradiation. To summarise, low T / R ratios are produced by the following conditions: (i) the use of a softer matrix, i.e. acetone/ H,O/LiCl (if [Dz+] is sufficiently high), (ii) U.V. light, (iii) higher light intensities and (iv) long photolysis times. A perplexing feature is the appearance of maxima in the dependence of T / R upon photolysis time using visible-light irradiation (irrespective of the matrix or diazo salt concentration), which cannot be understood solely in terms of eqn (1)-(4).These were found to be associated with a very strange build-up profile of [Are] with time; these profiles (in all matrices) display an initial build-up followed first by a well-defined, short decay region, and secondly by a slow further build-up. Individual growth profiles of [3Ar+] are not marked by any complications of this type, i.e. the maxima result solely from the kinetic behaviour of Are. One possible explanation involves two parallel reactions (3) or (4) involving ES sites of markedly different efficiency : the more efficient results in an initial fast build-up of Ar*, which subsequently disappears in a photostimulated decay, the less efficient gives a slower parallel build-up of Are and the sum of the two gives the ‘switchback’ profile. PF; - - - N, - - - 3Ar+1278 TRIPLET STATE E.S.R.STUDY OF ARYL CATIONS REACTION IN DARKNESS During post-photolysis storage in the dark, 3Ar+ slowly converts into Ar* (e.g. fig. 5) at rates (fig. 4) which are comparable for all samples irrespective of their varied photolysis conditions. While eqn (7) is clearly very important under appropriate reaction conditions, there are occasions when 3Ar+ disappears without a compensating increase in [Are], indicating the role of eqn (8) and (9): eqn (9) becomes more important at elevated temperatures but we detected slow decay in Ar* even at 77 K. Very few comparable studies exist, but in addition to 'normal' reactions of carbenes in matrices (to give dimers or cycloadd~cts),~* ti it has been reported recentlys that the quintet state formed by interaction of two benzoylphenylmethylene species decays on warming to 90 K to a new triplet state, regarded as the diradical remaining on combination of the two carbene species.Some comparisons should be made with the situation prevailing in solution thermolysis7 and photolysis8 of arenediazonium salts. Two principal types of product are obtained on either thermolysis7 or photolysis8 in MeOH, namely ArOMe (from Ar+) and ArH (from Are). The ratio ArOMe/ArH increases with (i) the presence of substituent electron (a and IC) (ii) the presence of 0,, which is expected to suppress radical-chain reactions,' (iii) the use of short-wavelength light (3 13 nm)* and (iv) the exclusion of added electron donors such as pyrene (which induce large yields of ArH).8 There is a contrast between the solution and solid-state photolyses in that photolysis at any wavelength promotes formation of 3Ar+ from solid samples +bile short-wavelength photolysis of methanol solution is required to produce ArOMe (longer wavelengths yielding ArH). We thank the S.E.R.C. for provision of the e.s.r. spectrometer, the gaussmeter and the cryogenics, and for support of H.B.A. Part 4. H. B. Ambroz and T. J. Kemp, J. Chem. SOC., Faraday Trans. 1, 1-982, 78, 725. * (a) H. B. Ambroz and T. J. Kemp, J. Chem. SOC., Perkin Trans. 2, 1979, 1420 and 1980, 768; (b) H. B. Ambroz and T. J. Kemp, Chem. SOC. Rev., 1979, 8, 353. ' (a) G. M. J. Schmidt et al., Solid State Photochemistry (Verlag Chemie, Weinheim, 1976); (b) J. Burdett and J. J. Turner, in Cryochemistry, ed. M. Moskovits and G. A. Ozin (Wiley, New York, 1976), chap. 1 1 ; (c) P. D. Fleischauer, in Concepts in Inorganic Photochemistry, ed. A. W. Adamson and P. D. Fleischauer (Wiley, New York, 1975), chap. 9; (d) G. M. Parkinson, M. J. Goringe, S. Ramdas, J. 0. Williams and J. M. Thomas, J . Chem. SOC., Chem. Commun., 1978, 134 and references cited therein. V. P. Senthilnathan and M. P. Platz, J. Am. Chem. SOC., 1980, 102, 7637. C-T. Lin and P. P. Gaspar, Tetrahedron Lett., 1980, 21, 3553. H. Murai, M. Torres and 0. P. Strausz, J. Am. Chem. SOC., 1980, 102, 5104. H. G. 0. Becker, G. Hoffmann and G. Israel, J. Prakt. Chem., 1977, 319, 1021. ' T. J. Broxton, J. F. Bunnett and C. H. Paik, Chem. Commun., 1970, 1363. (PAPER 1/994)