摘要:

Spin Correlation in the Geminate Recombination of Radical Ions in Hydrocarbons Part 2-Time Resolved Single-Photon Counting Study of the Magnetic Field Effect BY BRIAN BROCKLEHURST~ Radiation and Surface Chemistry Group, Atomic Energy Research Establishment, Harwell, Oxfordshire, OX1 1 ORA Received 7th December, 1976 Pulse shapes of scintillations produced by beta-particles have been measured by time-resolved single-photon counting. Magnetic fields and deuteration of the solute enhance the fluorescence of para-terphenyl and other solutes. This is due to electron-nuclear hyperfine interaction in the radical ions which interconverts the singlet and triplet wave-functions of the ion pairs. The extent of the effect gives a measure of the relative initial yields of singlet and triplet ion pairs.The magnetic field effects vary considerably among the solvents studied-n-hexane, iso-octane, cyclohexane, methyl- cyclohexane, mns-decalin, mixed decalins, squalane and benzene. The radiolysis of aromatic hydrocarbons, M, in hydrocarbon solvents, RH, has been used widely to study ion recombination' and excitation processe~.~-~ Most of the recombination is geminate, i.e., non-random: the electrons are thermalised at distances, b, from the positive ion, which are much less than the Onsager escape distance (-30 nm at room temperature).' The main excitation processes are believed to be: RH--+ RH+ + e', RH* (1) RH* + M --f RH + M* (2) RH++M-+RH+M+ (3) e- + M --f M- (4) RH+ + e- -+ RH* (5) RH+ + M- -+ RH + M* (6) M+ + e- 3 M* (7) M+ + M- -+ M* + M.(8) The relative importance of solute ion recombination (6), (7) and (8), and energy transfer from the solute, (2) following (1) and (5), is still a matter of debate. When RH is aromatic, the most probable value of b is small, 4.2 nm in benzene;6 recombina- tion in the solvent, (5), is very fast, RH* is stable, and (2) predominates. In alkanes, b is larger, y6 nm, so that charge trapping, (3) and (4), is more likely. However, some alkanes are known to fl~oresce,~ i.e., they have relatively long lived excited states: Beck and Thomass have argued that, in cyclohexane, singlet excited states of the solute are produced by (2), rather than (6), (7) and (8). In geminate recombination, in the simplest case, the two odd electrons in the radical ions, M+ + M-, etc., were initially paired together in the same molecule in a singlet state: if they retain their spin correlation, they can only form singlet M*.9 t Permanent address : Chemistry Department, The University, Sheffield, S3 7HF.BRIAN BROCKLEHURST 97 At times -10-50 ns this correlation will decrease because of the electron-nuclear hyperfine coupling in the ions.The rate of decrease can be reduced by applied magnetic fields : the limiting value of the triplet-singlet ratio is also field independent, so the field effect persists for long times (many ps). A detailed theoretical description of this effect is set out in Part 1." Experimentally it has been demonstrated by pulse radiolysis,ll in steady state fluorescence measurements,12 and by time-resolved single- photon c o ~ n t i n g .l ~ * ~ ~ For luminescence measurements, this last has better time resolution (-0.5 ns) than conventional pulse radiolysis (which is limited by the pulse length) and greater dynamic range, down to intensities in the pulse tail -loM4 of those in the peak. In this paper, a more detailed accoimt of results obtained by this method is presented. EXPERIMENTAL The use of time-resolved single photon counting for measuring scintillation pulse shapes was first described by Bollinger and Thomas.1si16 The method is now commonly used with nanosecond flashlamps to measure fluorescence lifetimes :17 the electronic circuits are essentially the same. Our equipment will be described fully elsewhere. In brief, klullard XP 2020 photomultipliers are used in both channels.These give excellent time resolution: when tested with a Cerenkov light source, the full width at half maximum was 680 ps. At short times, a serious problem arises from late pulses in the " stop " photomultiplier,'* which give a sharp peak of 1 % intensity of the main peak delayed by 15 ns. The resolution can be improved and the delayed peak removed by deconvolution but this has not yet been attempted. (Further deconvolution from the excited state decay will give the time- dependence of the rate of excitation.) Measurements at longer times are limited by the back- ground due in part to photomultiplier noise, in part to random coincidences with other pulses. The background was measured at -10-20 BS after the pulses and subtracted. Biphenyl, para-terphenyl and perylene (zone refined by Dr.J. N. Sherwood, Stratliclyde University), anthracene (B.D.H. blue fluorescent), [2H,4]para-terphenyl (Merck, Sharp and Dohme, 98 % deuterium) and AnalaR benzene were used as received. Alkanes were passed down activated silica gel columns until no further reduction in u.v.-absorption could be obtained. Tram-decalin was separated from the commercial mixture (-2 : 1 cis :trans) by distillation and vapour-phase chromatography. Samples were degassed on the vacuum line and sealed off. They were excited through a thin glass window by a 10 pCgOSr source. (N.B. the strong pulses obtained include Cerenkov radiation.) A series of permanent magnets were used to provide fields which were measured with a Bell magnetometer. Some measurements were made with the sample carefully screened with soft-iron and mu-metal, but no effect due to the earth's field could be detected. RESULTS The ideal solute for these experiments would have a high fluorescence efficiency q, a short lifetime z, and known hyperfine coupling constants, a, for both anion and cation.Para-terphenyl (p = 0.93 z = 0.95 ns)19 is good in most respects: its a values are relatively small,l0 so that the evolution of the field effect is slow and decon- volution is hardly necessary; however, it is rather insoluble. Earlier work16 with charge scavengers has shown that the scintillation pulse tail is due to ion recombina- tion : here tests with perfluoro-methyl-cyclohexane added to scavenge electrons showed a large reduction in intensity, increasing with time, as expected.Fig. 1 shows scintillation pulse shapes obtained for para-terphenyl in mixed decalins : for clarity, only every eighth point in the tail has been plotted. The back- ground (not subtracted) was -100 counts per channel. In agreement with theory,1° both forms show an enhancement of fluorescence by magnetic field in the tail of the98 SPIN CORRELATION I N GEMINATE-RECOMBINATION OF RADICAL IONS pulse (the peak intensity is not affected); the intensities at long times are the same within experimental error, but at shorter times (10-100 ns) the deutero-compound gives more fluorescence : the spin correlation decays more slowly as expected, because of the smaller hyperfine coupling with deuterium.1° Plots of the intensity ratio (R, field on/field off) have been published previo~sly:'~ for [lHI4] terphenyl the ratio rises to a maximum at about 25 ns, then falls slightly before levelling off.channel number 50 100 150 200 250 t l n s FIG. 1.-Scintillation pulse shapes for guru-terphenyl, 0.005 mol dm-3 in mixed cis- and trans- decalins: 0 x , C18H14; 0 +, C18D14; x +, zero field; 0 0, 0.16 Tesla. The same isotope effect has been found for terphenyl in benzene and in squalane (2,6,10,15,19,23-hexamethyltetracosane) : the squalane results are similar to those in decalin but R is much smaller in benzene, reaching a constant value of 1.09. (The earlier statement13 that R falls over the region 75-175 ns has now been shown to be incorrect.) [1H14]Terphenyl has been studied in a wider range of solvents over the same time-range as in fig.1; the time dependence of the field effect is the same in all cases, but the extent of the effect varies. The enhancement ratio is constant or nearly so (see below) between 75 and 175 ns: this plateau value, R,, averaged over this range is plotted against the magnetic induction in fig. 2. Error bars (&standard deviation calculated from the numbers of counts) are shown in one case; the others are comparable. There is probably no significant change with field above 0.1 T; pulse radiolysis measurements have shown that the effect is constant from 0.1 to 0.7 T.'O The decalin mixture gives similar results to trans-decalin at both con- centrations : methylcyclohexane is nearly the same as cyclohexane.BRIAN BROCKLEHURST 99 1. R P 1.1. 1. 1.1 A 8 A b O + $ * l! 4- I + i T P X W 0 + A B + I I 1 I 1 10 100 B / mT FIG. 2.-Magnetic enhancement ratio, R, in the plateau region (~75-175 11s after the pulse peak) against magnetic induction: 0, para-terphenyl in iso-octane; X , n-hexane; +, squalane; A, cyclohexane and El,., trans-decalin; all 0.001 mol dm-3 except my 0.005 mol dm-3. A few measurements have been made with other solutes - anthracene, perylene and biphenyl: all show similar behaviour to terphenyl, except that no maximum is observed in R, probably because of the longer fluorescence lifetimes, which reduce the time resolution. Some results are shown in fig. 3 : spacing between the points indicates the range over which channels were summed before calculating R. There is a marked solvent effect, as in the case of terphenyl and a concentration effect.Terphenyl concentration could only be varied in this range (up to -0.005 mol dm-3) in the decalins and in squalane : in both cases, the R against t curve shape hardly changes (unlike biphenyl) but R, increased with concentration. The field effect could still be detected at 3 ,us in squalane, in which there is a strong tail due to the high viscosity; iso-octane also gives a strong tail: it was found that in the dilute biphenyl solution (cJ: fig. 3) R rose to 1.40-1.45 at -400 ns and stayed constant to 800 ns. In other solvents, tails were weaker and measurements difficult because of the strong background : evidence was found for a further increase in R in cyclohexane and trans-decalin solutions beyond 200 ns, but it was not conclusive.Biphenyl was studied because of its greater solubility. DISCUSSION The agreement with experiment over the evolution of the field effect in time13 confirms the theoretical description ;lo in particular the observed isotope effect makes it possible to rule out other field effects, e.g., on fluorescence from triplet-triplet annihilation.*l However, the extent of the effect is smaller than the simplest theoretical prediction; in all the studies to date,11-13*20 the limiting value R, does not100 S P I N CORRELATION IN GEMINATE-RECOMBINATION OF RADICAL IONS 1.1 1.0- 50 0 * x 0O0 - 0 0 e + " + x + +** 0 .e+o *a* x x .+ I I I I I I I I channel number 150 I 0 0 . . . 0 200 ' 0 . . 0 50 100 I I I 0 . a 1.4- 0 0 0 R 1.3- 0 0 0 0 + + + 3.0 0 + 0 0 I 0 0 a 0 . ) I % X X t l n s FIG. 3.-Time-resolved effect of magnetic field (0.16 T/zero field) on biphenyl solutions in 0, x, cyclohexane; and 0,+ iso-octane; 0 0, 0.01 ; x +, 0.001 mol dm-3. exceed 1.40-1.45. If the ion pair is initially singlet, the singlet content of the wave- function will decay to one quarter at zero field or one half at high field:1° at times less than the spin relaxation time, TI, the singlet mixes only with the M = 0 com- ponent of the triplet because of the Zeeman interaction; R, should then be 2. Possible explanations for the difference are : (i) there are other, field-independent, excitation processes occurring-reaction (2) ; (ii) the limiting value of the singlet content is not reached at zero field, because there are degenerate energy levels between which no oscillation occurs :lo (iii) the singlet/triplet excitation ratio does not reflect the spin constant of the wave-functions because of Franck-Condon factors etc.;1° (iv) triplet ion pairs are formed initially as well as singlets: they decay into singlets so reducing the effect.Reaction (2) can be ruled out: the lifetimes of excited alkanes are very short, -2-3 ns,22 and the excitation process (5) is very fast: some alkanes (iso-octane) do not fluoresce at all.' Any interfering excitation process must decay at about the same rate as the ion recombination since R, is very nearly constant in many cases. (ii) and (iii) above would be expected to depend on the nature of the solute : measured variations with solute are not greater than the experimental errors.Therefore, (iv) is probably the main cause; since (ii) and (iii) may make a small contribution the following discussion is only semi-quantitative. If (ii) and (iii) are neglected, one can derive from eqn (21) and (22) of Part 1 that the initial triplet/singlet ratio, before any decay of correlation occurs, (GT/Gs),, = (2 - R,)(R, - 2/31; R, = 1.45 gives an initial ratio of 0.70. Sources of triplet ion pairs are (a) spin reversal of an energetic secondary electron which excites a solvent molecule to a triplet state (or similarly an excited positive ion transfers energy to formBRIAN BROCKLEHURST 101 a triplet state) and (b) cross-recombination in clusters containing more than one ion pair.23 For (b), combining the cluster size distribution of Ore and L a r ~ e n ~ ~ with the recent statistical calculations of Atkins and L a m b e ~ t ~ ~ gives (GT/Gs),, = o.335.It is assumed that all ion pairs are equally likely to be scavenged; in practice, scavenging is probably less efficient for large clusters. The extreme assumption that only one ion pair per cluster can be scavenged gives an initial ratio of 0.15. Clearly, process (b) above does not provide enough triplet in pairs to explain the results. It is difficult to estimate the extent of (a),23 but it must increase with decreas- ing energy of the solvent triplet. For alkanes, this is -6 eV;26*27 one does not expect large variations between alkanes ; such evidence as there is 27 places the trans-decalin triplet -0.14 eV below that of iso-octane.Accurate measurements of the limiting value of R in various alkanes would be very interesting. The triplet state of benzene lies much lower at 3.6 eV. The measured value of R, is 1.09, corresponding to a value of (GT/GJ0 = 2.15. The greater relative yield of triplets in p-xylene compared to bicyclohexyl has also been attributed to process (a).28 In the pulse radiolysis studies,20 it was found that the limiting enhancement was less in cyclohexane than in squalane; this unexpected result led to the present survey of the alkanes. They appear to fall into groups (fig. 2 and 3) : in cyclohexane, methyl- cyclohexane, trans-decalin and the decalin mixture the enhancement rises more rapidly at low fields but reaches a lower value at high fields (at times up to 200 ns).These alkanes also give much stronger scintillations (-2 to 4 times) than the second group of n-hexane, iso-octane and squalane. This grouping does not correlate with viscosity (squalane much greater than the others) or electron range and mobility (iso-octane is exceptional; b = 9.5 nm).6 It correlates with fluorescence efficiency to some extent (first group high but n-hexane $ i~o-octane)~ and also with the mobility of positive charge; this is diffusion-controlled in n-hexane, but much faster in cycl~hexane,~~ methyl-cy~lohexane~~ and trans-de~alin.~' Such " hole conduction " probably occurs when the ion retains the configuration of the neutral molecule, unlikely except in the symmetrical unstrained ring compounds. The behaviour of squalane, iso-octane and n-hexane (fig.2 and 3) suggests that two processes are involved: this is borne out by total fluorescence measurements20 on a number of solutes in squalane over a wide range of concentrations: at mol dm-3, the R against log B curves show negative curvature: as the concentration is reduced they can become S-shaped like those in fig. 2. Theoretical calculations of the distribution in time of reactions (6), (7) and (8)31 show that (7) dies out very quickly: the two processes could be (6) and (8), the former predominating at low concentrations. However, nothing appears to be known of the efficiency of (6); there are theoretical arguments against it leading to excitation of the solute32 and experimental evidence4 that it leads to radical formation: RH' -k M 3 R.+ MH.. (9) Further the magnetic field effect on the low concentration process is abnormal: it develops slowly in time (fig. 3) which implies a small effective a value, but it requires a large field for full development - large a. (Little is known of the magnetic properties of RH+.)IO An alternative is that the final step is (8) but reaction (3) is slow enough at low concentration that much of the spin evolution occurs in RH+ + M- pairs: at high concentrations, the pair is effectively M+ + M- for all of its lifetime. The S-shaped curves can be explained qualitatively if charge transfer occurs part way through the spin decay: transfer leads to loss of coherence, the second stage starts from the ion pair triplet/singlet ratio produced by the first stage. At high field or zero field, this makes no difference to the final singlet content of the spin wave-102 SPIN CORRELATION I N GEMINATE-RECOMBINATION OF RADICAL IONS function (& or t); at intermediate fields it is reduced.More theoretical work on this point is required. The behaviour of the cyclohexanes and decalins is difficult to explain at present. Reaction (3) is an order of magnitude faster than in the other alkanes : correspondingly, the shape of the field dependence does not change with concentration, both in the present time-resolved measurements and in total fluorescence measurements.20 The extent of the effect does change with concentration however (fig. 2 and 3). A number of measurements of excited state yields in solutions in alkanes have been made using pulse radiolysi~.~-~ Triplet to singlet ratios are usually -1-2, but recently Salmon et aZ.33 have obtained somewhat smaller values and found con- siderable variations in yields and ratios between cyclohexane, cyclopentane, cyclo- octane and mixed decalins.These measurements include free ion recombination, excitation by Cerenkov radiation and energy transfer (if any) as well as geminate ion recombination. Also, they are summed over the whole pulse: some conversion of singlet ion pairs into triplets occurs before recombination. The extent of this could be calculated from the decay curves by integration, but this has not yet been attempted. A rough estimate can be made from the steady state measurements: there is a 10% overall enhancement in cyclohexane solutions.20 At fairly short times where there is decay at zero field but not at high field, the correction factor (to " no-decay ") is just the enhancement ratio : this region is short but the intensity is high.At long times, when the limiting values are reached, the correction factor is ( 3 4 - 2), i.e., 2.35 if R, is 1.45. The observed overall enhancement of -10% must correspond to a conver- sion of not more than 20% of the initially singlet ion pairs into triplet. The value of (GT/GJ0 of 0.70 obtained above would correspond to a value of 1.08 after allowing for decay. (If indeed the correlation in RH+ + M- pairs decays to a much smaller extent, this would account for the fall in (GT/G,) with decreasing concentration observed in cyclopentane, cyclohexane and decalin: the reverse effect is found in cyclo- ~ c t a n e .) ~ ~ ~ ~ These calculations are very rough, but they confirm that a significant part of the triplet yield must come from initially triplet ion pairs: it does not all come from singlet pairs which lose their spin correlation. A puzzle remains in that Beck and Thomas' were unable to detect any triplets at very short times. The author wishes to thank Dr. W. G. Burns for his support and encouragement, Prof. A. Hummel, Dr. R. S. Dixon and Dr. F. P. Sargent for communicating results before publication, Dr. J. N. Sherwood for gifts of chemicals, the U.K. Atomic Energy Authority for a Research Associateship and the University of Sheffield for leave of absence. A. Hummel, in Advances in Radiation Chemistry, ed. M. Burton and J.L. Magee (Wiley, New York, 1975), vol. 4, p. 1. J. H. Baxendale and P. Wardman, Trans. Faraday SOC., 1971, 67,2997. J. K. Thomas, Int. J. Radiation Phys. Chenz., 1976, 8, 1. G. A. Salmon, Int. J. Radiation Phys. Chem., 1976, 8, 13. A. Singh, Radiation Res. Rev., 1974, 4, 1. W. F. Schmidt and A. 0. Allen, J. Chern. Phys., 1970,52, 2345. W. Rothman, F. Hirayama and S. Lipsky, J. Chem. Phys., 1973, 58, 1300. G. Beck and J. K. Thomas, J. Phys. Chem., 1972, 76, 3856. B. Brocklehurst, Nature, 1969, 221, 921. lo B. Brocklehurst, J.C.S. Faraday 11, 1976,72, 1869. l1 B. Brocklehurst, R. S. Dixon, E. M. Gardy, V. J. Lopata, M. J. Quinn, A. Singh and F. P. l2 R. S. Dixon, E. M. Gardy, V. J. Lopata and F. P. Sargent, Chem. Phys. Letters, 1975,30,463, l3 B. Brocklehurst, Chem.Phys. Letters, 1976, 44, 245. Sargent, Chem. Phys. Letters, 1974, 28, 361.BRIAN BROCKLEHURST 103 l4 J. Klein and R. Voltz, Phys. Rev. Letters, 1976, 36, 1214. ’’ L. M. Bollinger and G. E. Thomas, Reu. Sci. Iitstr., 1961, 32, 1044. l6 P. K. Ludwig, in Advances in Radiation Chemistry, ed. M. Burton and J. L. Magee (Wiley, l7 W. R. Ware in Creation and Detection of the Excited State, ed. A. Lamola (Marcel Dekker, New York, 1972), vol. 3, p. 1. New York, 1971), vol. LA., p. 213. C. Lewis, W. R. Ware, C . J. Doemeny and T. L. Nemzek, Rev. Sci. Znstr., 1973, 44, 107. I. B. Berlman, Handbook of Fluorescence Spectra of Aromatic Molecules (Academic Press, New York and London, 1965). ‘O R. S. Dixon and F. P. Sargent et al., unpublished work. ’’ P. W. Atkins, Chem. in Britain, 1976,12,214; P. W. Atkins and T. P. Lambert, Ann. Rep. Chem. ’’ W. R. Ware and R. L. Lyke, Chem. Phys. Letters, 1974, 24, 195. 23 J. L. Magee and J.-T. J. Huang, J. Phys. Chem., 1972, 76, 3801. 24 A. Ore and A. Larsen, Radiation Res., 1964, 21, 331. ” P. W. Atkins and T. P. Lambert, Mol. Phys., 1976, 32, 1151. 26 H. Brongersma and L. J. Oosterhoff, Chem. Phys. Letters, 1969,3, 437; K. Hiraoka and W. H. ’’ V. A. Smirnov, V. B. Nazarov, V. I. Gerko and M. V. Alfimov, Chem. Phys. Letters, 1975,34, ” L. Walter, F. Hirayama and S. Lipsky, Znt. J. Radiation Phys. Chem., 1976, 8, 237. ’’ E. Zador, J. M. Warman and A. Hummel, Chem. Phys. Letters, 1973,23,363; M. P. de Haas, J. M. Warman, P. P. Infelta and A. Hummel, Chem. Phys. Letters, 1975, 31, 382. 30 A. Hummel and J. M. Warman, personal communication. 31 S. J. Rzad, J. Phys. Chem., 1972,76, 3722. 32 B. Brocklehurst, Chem. Phys., 1973, 2, 6. 33 P. O’Neill, G. A. Salmon and (in part) R. May, Proc. Roy. SOC. A , 1975,347, 61. SOC., 1975,72A, 67. Hamill, J. Chem. Phys., 1973, 59, 5749. 500.

ISSN:0301-7249    

DOI:10.1039/DC9776300096

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Studies of the Reactions of Hydrogen Atoms by Time-Resolved E.S.R. Spectroscopy BY RICHARD W. FESSENDEN AND NARESH C. VERMA Radiation Research Laboratory and Department of Chemistry, Carnegie-Mellon University, Pittsburgh, PA 1521 3, U.S.A. Radiation Laboratory, University of Notre Dame,? Notre Dame, IN 46556, U.S.A. AND Received 29th Nocenzber, 1976 Time-resolved e.s.r. spectroscopy has been used to follcw directly the reactions of H atoms produced by pulse radiolysis of acid solutions. Detailed analysis of the time profile of the e.s.r. signal was carried out by means of modified Bloch equations. The increased signal found when a scavenger for OH such as t-butyl alcohol is present is shown to be mainly the result of slower H atom decay by radical-radical reaction. The reaction H + OH does not appear to produce any signal polarization.The decay curves observed in the presence of solute are readily accounted for by the treatment, and good plots of pseudo first-order rate constant against solute concentration are obtained. The absolute rate constants for reaction with H atoms are for methanol 2.5 x lo6, for ethanol 2.1 x lo7, for isopropanol 6.8 X lo7, and for succinic acid 3.0 x lo6 dm3 mol-' s-I. These values are in good agreement with earlier chemical measurements. The electron spin resonance signals of H atoms are readily detectable in experiments involving continuous in situ radiolysis of aqueous solutions with a high energy electron beam.lP2 These signals show strong effects of chemically induced dynamic electron polarization (CIDEP) with the low-field line in emission and the high-field line correspondingly in enhanced absorption.As a consequence of this effect, the amplitude of the lines is dependent on the first power of the irradiating beam current rather than on the square root as would be appropriate if the signal were proportional to the steady-state radical concentration. This linear behaviour has been exploited in several studies of H atom reaction rate^.^-^ In this method, the H atom e.s.r. signal under constant irradiation conditions is used as an indicator as solute is added to the solution. The loss of signal as a result of reaction with solute then competes with the other processes which control the signal amplitude. Most of the systems studied obey simple first-order competition mathematics.The analysis presented previously2 assumed that the excess signal (CIDEP) arose because of a non- equilibrium distribution of radicals in the various spin states upon their formation. Subsequent time-resolved experiments 6 9 7 showed clearly that the CIDEP was the result of selective reaction of radicals in certain spin states rather than being an initial effect. The pulse experiments reported here were undertaken to investigate in more detail the behaviour of the H atom e.s.r. signals and to determine if the results could be analysed in detail to give reaction rates with solutes. Previous pulse experiments on H atoms published by Smaller et aL7 analysed the decay curves under the assumption that the signal amplitude represented concentra- tion directly. Such an approach may give a good estimate of radical lifetime t Address for correspondence.RICHARD W.FESSENDEN AND NARESH C. VERMA 105 particularly for the later stages of the decay, but this is not certain. In addition, the use of high frequency field modulation further complicates any more detailed analysis. In our earlier work,6 and in that reported here, no field modulation is used so a detailed analysis using the Bloch equations is possible. The basis of this approach was introduced in the earlier paper ti and greatly amplified in a more recent paper.* Pulse experiments performed with this improved apparatus will be reported for H atoms in the presence of several solutes. Both the experimental results and the mathematical treatment represent a considerable advance over the earlier paper.EXPERIMENTAL The experimental apparatus and its operation were as described in ref. (8). Radiolysis was with 0.5 ,us pulses of 2.8 MeV electrons. The pulse repetition rate was 100 s-l and the curves to be shown represent the accumulated response to several thousand pulses. The position of the centre of the e.s.r. line was determined very accurately by comparing the response at two magnetic field positions separated approximately by the line width. Thus all curves represent the response at the centre of the line and can be compared for absolute amplitude. The solution flowed at about 0.5 cm3 s-l. The dose per pulse was measured by observing the decay of SO? in basic (pH 12.4) solutions of &SO3 (10 m mol dn~-~) saturated with NzO.The rate constant for second order reaction given by Hayon et aL9 is 1.7 x lo9 dm3 mol- s-l for this ionic strength. At an average beam current of 0.14 ,!LA collected at the cell the first decay half life was 27 p s for a concentration of 2.2 x rnol d ~ n - ~ . This measurement was repeated on each day data were taken. Although we have found* some deviation of dose from strict proportionality to the collected current, this deviation will not greatly affect the results found here as a rather narrow range of currents is used. The water was prepared by distillation and pyrolysis of impurities as described previously.* All solutions were made acid (0.1 mol ~ l m - ~ ) with Baker Analysed HC104. Other solutes were also from Baker or were materials used previously.2 ANALYSIS OF THE E.S.R.CURVES Analysis of the time profile of the e.s.r. signals is made in terms of the modified Bloch equations as described previously. The equations used are ik, = AuMy - (TF1 + k/R)M, iky = -AcoM, - (Tc' + @R)My - mIMz (4 where M , and My are the components of the macroscopic sample magnetization in a coordinate system rotating at coo while M, is the normal z component, col = yH, with HI the microwave magnetic field and y the magnetogyric ratio, Aco = co - o,, the offset between the angular frequencies of observation (co) and exact resonance (coo), and TI and T2 are the conventional spin relaxation times. The relative radical concentration (R) is initially unity. The term (Mo/Tl) represents the equilibrium magnetization in the absence of reaction and the term containing P represents the production of magnetization by chemical reaction.This latter effect (CIDEP) arises because of differing reaction rates for H atoms in different spin states. These equations are solved by numerical integration from the initial condition M, = My = 0 and M, =f (i.e., Mo is taken as unity). The e.s.r. signal is proportional to -ivy. The radical concentration (R) at any time is obtained by integrating the chemical decay equation k = -kl(R) - k2(R)2 (B)106 STUDIES OF THE REACTIONS OF HYDROGEN ATOMS where kl is the pseudo first-order rate constant for reaction of H with solute and k , is the second-order rate constant for disappearance of H atoms by the reaction H + H. The chemical reaction acts to decrease both radical concentration and the corresponding magnetization thus explaining the presence of the rate k/R = -kl - k2(R) with the terms Ti1 and T,' in eqn (A).We note that PedersenlO does not have the loss of magnetization due to reaction appearing together with T F ~ in the equations for M , and My. The chemical eqn (B) includes second-order reaction of H atoms by H + H only. In the cases where a scavenger for OH is present, a scavenger radical S is formed and reaction of H -+ So is possible. Such a reaction will change the decay of H as well as produce magnetization. When a reactant for H is also present it is possible that, after most of the H reacts with the reactant, the major radical-radical reaction will be H plus the other radical. Results to be obtained suggest, however, that such reactions of H with solute radicals are probably much slower than H + H and so are not particularly important.The parameters necessary to define one curve are H,, TI, T2,f, P, and the two rate constants as well as a scaling factor for the amplitude. The various values were determined in different ways. The value of Hl was determined by an experiment with bromomaleic acid as described previously.* The radical -O,CCH=kCO, formed by reaction of e; has long relaxation times so that the e.s.r. response is oscillatory at high power. The value of HI at this power setting is adjusted to match the period of these oscillations and that at lower power obtained from a calibration of the attenuator. The value off controls the details of the initial rise of the curves.Previously it was found that -O,CCH=kCO, formed in basic solution from -0,CC-CCO; by reaction with both H and e& showed a value offof 0.4-0.5 which persisted at high concentration. Since the H atoms in acid solution are formed both directly and from e z a similar value should apply. The value is negative for the low- field line. This parameter does not significantly affect the decay in any case. The values of Tl and T2 also affect the rise time of the curve particularly at low power where the chemical reaction does not increase the rate of relaxation. We expect that Tl = T2 for H atoms but the large values (>lo ps) implied by the observed rise describe a much narrower line than is observed. (In this work a width of about 70 mG between points at half the height of the absorption line was found.) This observed width is produced by field inhomogeneity.To take this effect into account, sets of Bloch equations should be integrated for a number of offsets from the centre of the line and the signal amplitudes added with suitable weighting factors to give the observed width. With the computational speed available (Hewlett-Packard 9830A calculator) this approach is much too time consuming. As an alternative we have used Pedersen's suggestion1' that, at low power, the field inhomogeneity can be cor- rected for by using a short T,. Under these conditions, a value of TI = 20 p s fits the rise of the curves. For the low-field line of the H atom, the value of P must be sufficient to keep the line in emission throughout the decay as is observed.A value of P = -20 for the low-field seems to fit the data well. As long as the value is sufficiently large, the shape of the curve is not strongly dependent on it. The correct way to set this value is by comparison of the absolute amplitude of the H atom line with that of a radical such as SOT which shows no polarization. This calculation should take the field inhomogeneity into account directly, and we are not yet ready to do so. Nevertheless, preliminary calculations Normalized values are used so the initial value of (R) is unity. Consequently we have taken T2 = 1.5 ps. The value of P is more difficult to determine.RICHARD W. FESSENDEN AND NARESH C. VERMA 107 show that the value of P used is in the correct range.The parameters discussed to this point were kept fixed throughout the calculations. For the low field line they are Hl = 4 mG, TI = 20 ps, T2 = 1.5 ps, f = -0.5 and P = -20. The values of the first- and second-order rate constants and a scaling factor were determined to fit the specific curves. RESULTS AND DISCUSSION The low-field e.s.r. line of the H atom was studied exclusively. The time dependence of this line in 0.1 mol dm-3 HC104 both with and without 2 mmol dm-3 t-butyl alcohol is shown in fig. 1. Note that these curves represent emission but are 0 10 20 30 40 t i m e / p s FIG. 1.-Time profile of the low-field e.s.r. line of H atom in pulse irradiated 0.1 mol dm-3 HC104 both with (upper curve) and without (lower curve) 2 mmol dm-3 t-butyl alcohol.The upper curve was taken with a beam current of 0.06 PA, the lower with 0.07 PA. A power setting of -25 dB was used for both. Note that the upward direction represents emission. The calculated curves used the parameters described in the text with a second-order half life of 7.8 ps (upper) and 4 ,us (lower). drawn in a positive sense. The parameters in the presence of t-butyl alcohol (in addition to those kept fixed as described above) are (first) second-order half life = 7.8 pus and scaling factor = -4. To determine the reason for the lower signal without t-butyl alcohol as scavenger for OH radical, we attempted to fit that curve allowing only P and the half life to vary. The curve shown has parameters P = -18 and half life = 4 p s . (No correction was made for the small change in current between the two experiments.) The decrease in half life shows that H + OH is faster than H + dH2C(CH3)20H.In fact, with the rate constants of 2k = 1.5 x 10" dm3 mo1-1 s-l for H + H'l and 2.0 x 10" dm3 mol-1 s-l for H + OH12 and yields of 3.6 for H and 2.8 for OH, the half life should decrease by just a factor of 2 if H + eH2C(CH3)QH were much slower than H + H. We take this observation to show that fact. Alternatively, it is possible to use the observed half life and the concentration to determine the effective rate constant for second-order disappearance of H. The value so determined is 2.3 x lo1' dm3 mol-I s-l, which is higher than the published value. Taken with the result for no scavenger, it appears Calculated curves are also given.108 STUDIES O F THE REACTIONS OF HYDROGEN ATOMS that H + H is the main reaction.The similar value of P for the two experiments means that H + OH does not produce polarization. This conclusion was reached earlier in a less quantitative way. Now that a satisfactory fit to the experimental curves has been shown, it is possible to add reactants for H atoms, and to determine the rates. The reactants selected were methanol, ethanol, isspropanol and succinic acid. These compounds were selected on the basis of their behaviour in the previous study using steady-state irradiation.2 Of these four compounds, ethanol and isopropanol showed good agreement with a chemical study13 while the other two showed differences of about a factor of 1.5, each in opposite directions.Reactants with very high rate constants such as aromatics were not chosen because very small concentrations would have been necessary with problems of solute depletion. Typical data are shown in fig. 2 4 for methanol, succinic acid and ethanol. For methanol, the rate was sufficiently low that it was possible to use a concentration (2 mmol dm-3) where OH would be scavenged but H would not react significantly. The curve obtained under those conditions (the uppermost in fig. 2) agreed with that obtained under similar conditions 0 10 20 30 LO t i m e / ps F~G. 2.-Time profile of the low-field e.s.r. line of H atom in 0.1 mol dm-3 HClO, solutions containing 2 mmol dm-3 (uppermost) to 120 mmol dm-3 (lowest) methanol. All curves were taken at a beam current of 0.07 pA and at -25 dB power setting.Parameters for the calculated curves are as described in the text with a second-order half life of 7.5 ps. Shown in the upper right is a plot of pseudo first-order rate constant for each concentration of methanol. Each point on this plot corresponds to a curve shown except that no point is shown for 2 mmol dm-3 methanol. but with 2 mmol dm-3 t-butyl alcohol instead of methanol. Successive curves were analysed by allowing only the pseudo first-order rate to change (small adjustments of (10% in the scaling factor were allowed to take into account small changes in beam current). A plot of the pseudo first-order rate constants so obtained is also shown in the figure. Excellent linear behaviour is found. The data for succinic acid are shown in fig.3 and similar excellent behaviour is found. In each case the shortest half life is about 2 ,us. The data for ethanol are shown in fig. 4 and here a slight problem is encountered. The plot of pseudo first-order rate constant is linear but does not go exactly throughRICHARD W. FESSENDEN AND NARESH C . VERMA 109 I P 0 10 20 30 40 time/ ps FIG. 3.-Time profile of the low-field e.s.r. line of H atom in 0.1 mol dm-3 HClO, solutions containing various concentrations of succinic acid. The concentrations range from 5 mmol dm-3 (uppermost curve) to 80 mmol dm-3 (lowermost curve). Each curve corresponds to one point in the plot of pseudo first-order rate constant. zero. Because of the procedure used this curve logically must do so. Similar behaviour was found for isopropanol.It was found that the intercept could be eliminated by increasing the rate of the second-order decay from that obtained with 2 mmol dm-3 t-butyl alcohol but keeping it constant for all concentrations of a given t i me/ p s FIG. 4.-Curves similar to those in fig. 2 aiid 3 but for 1 mmol dn~-3 (uppermost curve) to 15 mmol dm-3 (lowest) ethanol. In this instance, the plot of pseudo first-order rate constant does not go exactly through the origin, apparently as a result of selection of the incorrect rate constant for second- order reaction of H + H.110 STUDIES OF THE REACTIONS OF HYDROGEN ATOMS compound. Because of the time consuming nature of this work the reference curve with t-butyl alcohol was run considerably earlier on a given day than the rest of the experiments and a change in focus of the electron beam may have changed the dose for a given measured average beam current.Because of the linear behaviour of the pseudo first-order rate constant with concentration we believe the data for different concentrations of reactant to be taken under the same conditions. The rate constants obtained in this work are tabulated in table 1 together with TABLE RA RATE CONSTANTS FOR REACTION WITH H ATOMS k x 10F6/dm3 mol-1 s-l reactant chemical" this work competition SS-ESRb pulse e.s.r.' ~~ methanol 2.5 2.9 1.6 2.4 ethanol 21 25 26 13 isopropanol 68 78 65 79 succinic acid 3.0 2.3 3.5 - Ref. (13). Steady-state e.s.r. method, ref. (2). Pulse e.s.r. measurement by Smaller et al.' The kinetic analysis in their paper was much simpler and less exact than that used here.values from the previous steady-state e.s.r. experiments,2 chemical competition and earlier pulse e.s.r. experiments by Smaller et aL7 For the three alcohols there is good agreement between the present measurements and the chemical com- petition measurements ; all methods agree well for isopropanol. With succinic acid the present result is somewhat above the chemical one but below that obtained by steady-state e.s.r. The value obtained by the steady-state e.s.r. method is low for methanol while that obtained for ethanol by Smaller et aL7 is low. Thus the deviations seem to be random. In general, the present results agree well with the chemical ones, however. No explanation for the low value for methanol in the steady-state experi- ments is evident from the present work.In particular, the behaviour of H atoms is the same whether methanol or t-butyl alcohol is used to scavenge OH. It should be noted that in the steady-state work the experiments with methanol and ethanol were repeated many times and so accurately represent the result of that experiment. No reason for the disagreement between that experiment and the present one can be suggested . Previously,2 the behaviour of the H atom e.s.r. signal in steady-state experiments was explained by an initial polarization upon radical formation. Now that both a correct understanding of the origin of CIDEP and an accurate description of the amplitude of the e.s.r. signal are available it should be possible to use this information to explain the behaviour observed.In principle, integration of a curve of e.s.r. signal as a function of time for a selected group of radicals under conditions appropriate to the steady-state should give the observed behaviour. However, one must solve for the steady-state radical concentrations to use this approach. Because of the problems involved in the mixed order kinetics it does not seem worthwhile to do this until more computational speed and capacity is available. The experimental observation of acceptable competition behaviour must continue to justify the steady- state e.s.r. approach. The present pulse method makes practical direct absolute determinations of H atom reaction rates. The accuracy of the method is affected somewhat by the extentRICHARD W . FESSENDEN A N D NARESH C . VERMA 111 to which the second-order decay of K atoms can be corrected for, but the agreement of the results with other methods indicates this correction to be successful. The close agreement with the competition experiments13 also serves to verify the basis used to convert those rates into absolute values. K. Eiben and R. W. Fessenden, J . Phys. Chem., 1971,75, 1186. P. Neta, R. W. Fessenden and R. H. Schuler, J . Phys. Chem., 1971, 75, 1654. P. Neta and R. H. Schuler, Radiation Res., 1971, 47, 612. ' P. Neta and R. H. Schuler, J. Phys. Chem., 1972,76, 2673. P. Neta and R. PI. Schuler, J. Amer. Chem. SOC., 1972, 95, 1056. N. C. Verma and R. W. Fessenden, J . Chem. Plzys., 1973,58,2501. ' B. Smaller, E. C. Avery and J. R. Remko, J. Phys. Chem., 1971,55, 2414. a N. C. Verma and R. W. Fessenden, J. Chem. Phys., 1976, 65, 2139. lo J. B. Pedersen, J. Chem. Phys., 1973, 59,2656. E. Hayon, A. Treinin and J. Wilf, J. Anier. Chem. SOC., 1972, 94, 47. P. Pagsberg, G. Christensen, J. Rabani, G. Nilsson, J. Fenger and S. 0. Nielson, J . Pliys. Chem., 1969, 73, 1029. M. Anbar, Farhataziz and A. B. Ross, Nntl. Stand. Ref. Data Ser. (Natl. Bur. Stand., vol. 51, 1975). l3 P. Neta, G. R. Holdren and R. H. Schuler, J. Phys. Chem., 1971 , 75, 1449.

ISSN:0301-7249    

DOI:10.1039/DC9776300104

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Yields and Reactions of Hydrogen Ions on Radiolysis of Water and Aqueous Solutions BY ALEXEI K. PIKAEV, SERGEI A. KABAKCHI AND ALLA A. ZANSOKHOVA Institute of Physical Chemistry of the Academy of Sciences of the U.S.S.R., Moscow, U.S.S.R. Received 2nd December, 1976 Yields of hydrogen ions have been measured by pulse radiolysis with optical registration of short-lived species and by use of chromate ions and pH indicators as scavengers of H30 + . It was found that G(H,O+) for the bulk of the solution (after the completion of reactions in spurs) is - 3-3.3 ions/100 eV. Some reactions of H30+ ions were also studied by the pulse radiolysis method. The rate constants of their reactions with pH indicators-bromophenol blue, phenol red and bromothymol blue-were determined. They are 1.6 x lo'', 7.2 x 1O1O and 8.8 x 10" dm3 mol-' s-l respectively.The earlier conclusion that free hydrogen ions take no part in the formation of C1- during pulse radiolysis of neutral aqueous solution of alkaline metal chlorides was confirmed on the basis of results of the investigation of the influence of CrO2- additions and temperature on G(C1t). The hydrated electron (e;J, atom H, radical OH, H2, H202, ions H30+ and OH- are formed as primary products during the action of ionising radiation on water. In the case of e&, H, OH, H2 and H202 the radiation chemical yields, the properties and the kinetics of many reactions involving their participation have been studied in In this respect H30+ and OH- ions have been little investigated. For example, values of G(H,O+) for the bulk of the solution (after the completion of reactions in spurs) have been measured only by the authors of ref.(4)-(6). More- over, the results of these measurements are not concordant: 2.9,4 3.255 or 3.38 ions/ 100 eV.6 Besides, there is only fragmentary information on the dependence of G(H,O+) on the time after passage of the ionising particle. It has been found that G(H,O+) at times of about s is 3.93' or 4.1 ions/100 eV.7 Buxtons concluded that the yield of H30+ ions which decay in spurs via reaction with e& was 1.1 ions/ 100 eV. In the present work, G(H30+) on pulse radiolysis of aqueous solutions of potassium chromate and some pH indicators was determined. Earlier it had been shown 's9 that these compounds were effective scavengers of H30+ ions in irradiated water.In the case of neutral solutions of potassium chromate the decrease in optical density of the solution within the optical absorption band of CrOf- takes place during the irradiation. This is a result of the reaction between CrOf- and H3Q+. Under certain conditions the extent of this decrease is equal to G(H,O+). In the case of solutions of pH indicators, G(H,O+) may be evaluated from the formation of the acid form of the indicator in weakly alkaline media. Apart from this, in the present work some reactions of H30+ ions have been studied. In particular, special attention was paid to their possible role in the forma- tion of C1, during pulse radiolysis of neutral aqueous solutions of alkaline metal chlorides, this system being often used for checking different hypotheses on the primary processes of water radiolysis.Different mechanisms for the formation of C1, have been proposed.1°-14 They include the appearance of this radical-ion viaA. K . PIKAEV, S . A . KABAKCHI A N D A . A . ZANSOKHOVA 113 reaction with OH in acid spurs or with " hole " H20+, through intermediate formation of ClOH-, as a result of reaction of excited C1- with OH. In our preceding work14 it has been suggested on the basis of the observed dependence of G(C1T) on the nature of the cation (K+, Rb+ and Cs') that [H30+ . . . OH], which is a hydrogen ion and a hydroxyl radical connected by a hydrogen bond, takes part in the formation of c1~. To obtain additional data on the mechanism of the process under consideration we have studied the influence of CrOi- additions and temperature on G(C1;) during pulse radiolysis of neutral aqueous solutions of alkaline metal chloride.EXPERIMENTAL The pulsed electron radiation generated by linear accelerator U-12 was used. The duration of pulses was 2.3 ,us and the energy of the electrons was 5 MeV. The transient optical absorption was registered by means of a fast spectrophotometric apparatus described in ref. (15) and (16). The solutions were irradiated in quartz cylindrical cells. Twice-distilled water purified additionally by radiolysis and photolysis was used for preparing the solutions. Salts KCI, RbCl and K2Cr04 were " chemically pure " grade; they were recrystallized twice from bidistillate. Other reagents were " pure for analysis " or " chemically pure " grade and were used without special purification.The dosimetry was carried out by two methods : by the measurements of optical absorp- tion of (CNS), at 475 nm in 5 x mol drn" oxygenated aqueous solution of KCNS (G(0H) = 2.9; E~~~ = 7.3 x lo2 m2 m ~ l - ' , ~ ~ ) and by the measurement of optical absorption of e, at 720 nm in deaerated water G(e,) = 2.8; ~ 7 2 0 = 1.84 x lo3 m2 mo1-1).2 The yields of C1, at room temperature were found from the optical density of the solution at 340 nm ( E ~ ~ ~ = 1.25 x lo3 m2 mol-l).lo The decrease in CrO2- concentration as a result of the action of the pulse was determined from the decrease in optical absorption at 370 nm. In the case of CrOi- solutions containing C1-, a correction for C1; optical absorption at this wavelength was applied.To increase the precision of the measurement of G(-CrO$-) we determined the dependence of molar extinction coefficient E of CrOz- ions at 370 nm on temperature. The data obtained are shown in fig. 1. We also found the values of E for I 4 0 3.3- € E ; 3.2 - 0 N N I 3.1 - 3.01 I I I I I I 20 30 40 50 60 70 80 t e m p I "C FIG. 1 .-Dependence on temperature of molar extinction coefficient E of CrOi- at 370 nrn in neutral aqueous solution. the pH indicators used. For bromophenol blue, bromothymol blue and phenol red they are 2.7 x lo3, 1.4 x lo3 and 1.7 x lo3 m2 mol-1 at 435, 454 and 400 nm respectively. In the experiments at elevated temperatures the sealed cell with the solution under study was placed into a special metallic holder through which water from a thermostat was passed.The holder was equipped with a thermometer; the irradiation was carried out at fixed temperature.114 YIELDS A N D REACTIONS OF HYDROGEN IONS ON RADIOLYSIS The yield values given in the paper were determined from the dependence of optical density of the solution on dose. Each value is averaged from 10-1 5 separate measurements. The error in yield determination was within the limits &lo%. RESULTS A N D DISCUSSION MEASUREMENTS OF G(H30+) WITH CHROMATE SYSTEM As has been noted, G(H,O+) in irradiated water may be determined using neutral aqueous solutions of potassium ~hromate.~ Chromate ions react very rapidly with hydrogen ions: CrOi- + H30+ + HCr04 + H20. (1) According to ref. (18), kl = 7.4 x 10l2 dm3 mol-I s-l.Obviously in solutions containing oxygen (scavenger of e; and H) G(-CrO:-) is a measure of G(H30+). In the present work this method was used for the determination of G(H30+) at different concentrations of chromate, ix., at different times after passage of ionising particle. The measured values of G(-CrO$-) for chromate solutions of different con- centration saturated with air or oxygen are given in table 1. The pH values of the TABLE l.-VALUES OF G(-CrOi-) ON PULSE RADIOLYSIS OF NEUTRAL AQUEOUS SOLUTIONS OF POTASSIUM CHROMATE (ROOM TEMPERATURE, DOSE PER PULSE <2 x lo1' eV ~ m - ~ ) * conc. of CrOi-/mol dm-3 1.1 x 1 0 - 5 1 . 6 x 1 0 - 5 4.6 x 10-5 5.0 x 10-5 5.3 x 1 0 - 5 8.0 x 10-5 8.0 x 1 0 - 5 1.0 x 1 0 - 4 gas saturating solution oxygen air oxygen air air air oxygen air G(-CrOi -) 3 .3 3.2 3.7 4.0 4 . 1 5 4.8 4.05 4.7 G'(-CrOi -) 3.3 3 . 0 5 3.6 3.6 3.7 4.2 3 . 9 4.0 * G'(-CrOq-) means the value of G(-CrO;-) corrected for the reaction between e, and O2 (see text). solutions were 6.5-6.9. From this table it is seen that, in the case of 4 x - mol d n r 3 solutions, G(-CrOi-) depends slightly on oxygen concentration. Obviously it is due to the fact that in aerated solutions hydrated electrons react not only with oxygen but also in some measure with CrO$'. This effect leads to some increase in G(-CrOi-) for aerated solutions in the comparison with oxygenated solutions. After the correction for reaction e i + Cr0;- the yields (they are marked in table 1 as G'(-CrOi-)) for aerated and oxygenated solutions coincide within the limits of experimental error (see table 1).This correction is equal to G(eJ x k ( e i + CrO$-)[CrOi-]/{k(e;q + 02)[02] + k(eFQ + CrO",)[CrO",]>. In calcula- tions of the correction the literature values of rate constants of e; reactions with CrOi- and oxygen (1.9 x 10" and 2 x lolo dm3 rno1-I s - ' , ~ respectively) were used. It is necessary to discuss specially the question of the possible contribution of reaction H + 0, to the values of G(-CrO$-). I n this reaction H02 radicals are Obviously, G'(-Cr02,-) is equal to G(H30+).A. K . PIKAEV, S . A . KABAKCHI AND A . A . ZANSOKHOVA 115 formed. At the pH values used these radicals dissociate during the pulse with the formation of hydrogen ions (pK of this dissociation is 4.5;9 rate constant is 7 x lo5 s-').'' However, it should be noted that H atoms are formed in spurs mainly as a result of the reaction e i + H30+.According to ref. (8) this reaction is complete by s after irradiation. Chromate ions at concentrations of 5 x 10-5-10-4mol dm-3, because of the high rate of their reaction with H30+, react with them in spurs at times of the order of 10-8-10-9s. Therefore, under these conditions the forma- tion of H atoms is suppressed to a considerable degree, and the contribution of H30+ as a consequence of reaction H + O2 to G(-CrOi-) is comparatively small. At lower concentrations of chromate this contribution increases. The values of G(-CrOi-) at different temperatures were determined ; results obtained are shown in table 2. On the basis of the data of this table it is possible to TABLE 2.-vALUES OF G(-CrO: -) FOR 5 X 10 -' Ill01 dm - SOLUTION OF POTASSIUM CHROMATE CONTAINING OXYGEN AT DIFFERENT TEMPERATURES (DOSE PER PULSE (1-2) x lOI7 eV ~ r n - ~ ) temp./"C 20 30 40 50 60 70 80 90 G(-CrOi - ) 4.1 3.9 3.8 4.1 4.2 4.2 4.1 4.2 conclude that G(H,O+) does not depend on temperature within the range studied (from 20 to 9OOC).Our value of G(H,O+) at times of 10-8-10-9 s coincides with the value obtained in work5 by other methods. At chromate concentrations corresponding to times of 10-8-10-7 s the yields of H30+ determined on the basis of G(-CrOi-) are slightly higher than the yield of hydrogen ions which are formed on water radiolysis but escaped from the decay reaction in spurs. As already noted, this is due to the con- tribution of H30+ ions which are formed via the dissociation of H02.The maximum value of this contribution at above-mentioned times may attain the value of G(H). PULSE RADIOLYSIS OF AQUEOUS SOLUTIONS OF pH INDICATORS In ref. (9) it was observed that during pulse radiolysis of aqueous solutions of some pH indicators in the alkaline form, the optical absorption spectra of their acid forms appeared transiently. Obviously this is due to the reaction between the alkaline form of indicator and H30+ formed during water radiolysis. This phenomenon was used in the present work for the evaluation of the yields of H30+ ions and rate constants of their reactions with pH indicators. To study the kinetics of the reactions under consideration it is necessary to have a radiolytic system in which interaction of e i , OH, H with indicators and the reaction ea; + H30+ are suppressed.Such a system may be created, for example, by the introduction of oxygen (scavenger of e& and H) and methyl alcohol (scavenger of OH radicals) in slightly alkaline aqueous solution of the pH indicator. Radicals O F , H02 and CH20H formed in this system have comparatively low reactivity towards many organic c o i n p o ~ i n d s . ~ ~ ~ ~ ~116 YIELDS AND REACTIONS OF HYDROGEN IONS ON RADIOLYSIS In the system under consideration, only two H,O+ reactions occur during the pulse (In- and InH mean alkaline and acid forms of indicators respectively) : H30+ + In- -+ InH + H20 H30+ + OH- --+ 2 HzO. (2) (3) The competition of these reactions is described by eqn (4) : where G(H,O+) is the initial yield of H30+ and G(1nH) is the yield of the acid form of the indicator.The rate constant of reaction (3) is known; it is equal to 1.4 x lo1’ dm3 mo1-’ s-l.’’ The determination of the dependence of G(1nH) on pH value and In- con- centration gives an opportunity to find G(H30+) and the ratio k3/k2 (and con- sequently the value of k2). The rate constants of reactions of H30+ with bromophenol blue, bromothymol blue and phenol red were measured by the method described. In this determination, loe5 - 2 x mol dm-, aerated solutions of pH indicators containing 0.1 mol dm-, CH30H were used. Fig. 2-4 show optical absorption spectra of the acid and alkaline forms of indi- cators obtained by ordinary methods and the transient spectra of irradiated solutions for bromophenol blue, bromothymol blue and phenol red, respectively.As seen, curves for acids and the left hand parts of the transient curves coincide satisfactorily. FIG. 2.-Optical absorption spectra of aerated mol dm-3 aqueous solutions of bromophenol blue containing 0.1 mol dm-3 CH30H: @-spectrum of solution of acid form (pH 1.5); 0- spectrum of solution of alkaline form (pH 7.5); @-transient spectrum of solution with pH 8.5 immediately after pulse.A. K . PIKAEV, S . A . KABAKCHI AND A . A. ZANSOKHOVA 117 '0 0.51 9 -0.5- -1.0- FIG. 3.-Optical absorption spectra of aerated mol dm-3 aqueous solutions of bromothymol blue containing 0.1 mol dmP3 CH30H: @-spectrum of solution of acid form (pH 4); 0-spectrum of solution of alkaline form (pH 10); @-transient spectrum of solution, with pH 8.8, immediately after pulse.1.0 0. E 0 X : eI \ Q -0.5 - 1.0 FIG. 4.-Optical absorption spectra of aerated mol dm-3 aqueous solutions of phenol red containing 0.1 mol dm-3 CHJOH: .-spectrum of solution of acid form (pH 4); 0-spectrum of solution of alkaline form (PH 10); @-transient spectrum of solution, with pH 9, immediately after pulse.118 YIELDS AND REACTIONS O F HYDROGEN IONS ON RADIOLYSIS This means that the reaction between indicator and H30+ occurs during the radio1 y sis. For the determination of G(1nH) the dependences of optical density of the solution at 400, 435 and 454 nm respectively for phenol red, bromophenol blue and bromo- thymol blue immediately after the pulse on dose per pulse were measured at different pH values. These dependences are linear within the dose range of (0.3-2) :>< 10'' eV ~ m - ~ .From this it follows that the reverse transition of the acid form of the indicator to the alkaline form and the reaction e, + H30+ in a bulk of the solution do not occur during the pulse even at comparatively high pH values. Hence it is possible to believe that during the pulse H30+ ions take part only in reactions (2) and (3). For determination of G(H,O+) and k3/k2 eqn (4) was solved graphically. This solution is shown in fig. 5. The intercepts are l/G(H,O+) and the tangents of the 0 1 2 3 [OH-] / [In-] FIG. 5.-Dependence of reverse value of yield of the pH indicator acid form on ratio of concentrations of hydroxyl ions and pH indicator at radiolysis of their aerated solutions containing 0.1 mol dm-3 CH,OH: 0-bromophenol blue; 0-phenol red; a-bromothymol blue.slopes are l/G(H30+) x k,/k,. The values of G(H,O+) and k , measured by such a method are listed in table 3. The rate constants k2 were calculated on the basis of k3 = 1.4 x loll dm3 mol-1 s-l. One of these constants is known. According to ref. (23) the value of k2 for phenol red measured by the electric pulse method is 7.0 x lo1* dm3 mol-l s-l. As can be seen, the values of the constants determined by two different methods coincide satisfactorily. From table 3 it follows that the values of G(H,O+) are approximately the same for all the pH indicators investigated. These values are the sums of the values of the yields of H30+ which escape from the decay in spurs, and G(H).The necessity to TABLE 3.-vALUES OF k2 AND G(H3O'). MEASURED FOR PULSE RADIOLYSIS OF AQUEOUS SOLUTIONS OF PH INDICATORS pH indicator range of pH indicator range of OH- concentrations concentrations k2/dm3 G(H30 +) used/mol dmw3 used/inol dm-3 mol-1 s-l bromophenol blue 5 x - 2 x 3.2 x - 1.6 x 1O1O 3.5 phenol red 5 x 10-5 - 2 x 10-4 2.5 x 10-6 - 7.2 x 10'' 3.6 bromothymol blue 2.5 x loU5 - 1.25 X - 8.8 x lolo 3.85 1.2 x 10-4 1.5 x 10-4 10-4A. K. PIKAEV, S. A. KABAKCHI AND A. A. ZANSOKHOVA 119 take G(H) into account is due to the formation of M30+ via fast dissociation of radicals HOz which appear in the reaction H + 0,. Therefore, the yield of H30+ which escape from the decay in spurs is 2.95-3.3 ions/100 eV. INFLUENCE OF CrOi- ADDITIONS AND TEMPERATURE ON THE YIELDS OF C1, FOR PULSE RADIOLYSIS OF NEUTRAL AQUEOUS SOLUTIONS OF ALKALINE METAL CHLORIDES It was mentioned that chromate, being a very effective scavenger of H30+ ions even at concentrations as low as 5 x lo-' mol dmR3, reacts with them not only in the bulk of the solution but also partially in spurs.In this case, if according to ref. (10) it is necessary to have free hydrogen ions present for CI, formation, the introduction of CrOi- should decrease G(C1;) (under the same other conditions). If free hydrogen ions take no part in C1; formation, then the introduction of CrOf- should have no effect on G(C1,). Fig. 6 shows the dependence of C1- yield during the pulse radiolysis of neutral aqueous solutions of KCI, on its concentration both in the absence of CrOi- ions and in the presence of these ions.It was found that the introduction of CrOi- has no .A /* I 0.25 0.50 0.75 concentration of CI-/ mol dm-3 FIG. 6.-Dependence of G(C12) on KC1 concentration in aerated neutral solution: .-in absence of CrOi-; 0-in presence of 5 x mol dme3 CrOi-. effect on G(C1,) within the limits of experimental error. Furthermore, we studied the influence of C1- on G(-CrOi-), when we found that the introduction of Cl- leads to a decrease in G(-CrOf). However, the influence of KCl, RbCl and CsCl is not the same; KCl is the most effective, then RbCl and CsCl. The dependence of G(-CrOi-) on C1- concentration for all the chlorides is linear as shown by a plot of l/G(-CrOi-) against [Cl-] (see fig. 7). On the basis of the results obtained it is possible to come to two conclusions. First, free hydrogen ions take no part in C1, formation.Secondly, C1- ions react with precursors of hydrogen ions. Hence the results provide further confirmation of the hypothesis of ref. (14) on the formation of these radical-ions via reactions: [H,O+ . . . OH] + Cl- --f C1+ 2 H20 ( 5 ) [H,O+ . . . OH] --+ H30+ + OH c1 +- c1- -+ a,-. In this case120 YIELDS AND REACTIONS OF HYDROGEN IONS O N RADIOLYSIS 0.75 r 0 0.25 0.50 0.75 concentration of CI-/mol d ~ n - ~ FIG. 7.-Influence of KCl concentration in aerated neutral 5 x niol dm-3 solution of K2Cr04 if the stage limiting C1; formation is reaction (5) (i.e., G(C1;) = G(C1)). In eqn (8) Go denotes the yield of C1 precursors. The dependence of G(-CrO:-) on C1- concentration is described by eqn (9) on yield of CrOi- conversion.Eqn (8) and (9) indicate a route to obtain the connection between the decrease in G(-CrOi-) [i.e., AG(-CrOi-)] and G(C1;) : AG(-CrOi’) = G(C1;). (10) Experimental data on the dependence of G(C1r) and AG(-CrOi-) on Cl- concentra- tion are listed in table 4. It is seen that eqn (10) is not valid. G(C1;) is considerably less than AG(-CrOi-). The values of k,/k, calculated from the dependences of G(C1;) and G(-CrOi-) on C1- concentration for different cations are listed in table 5. As seen, k,/k5 depends on the nature of cation. This ratio determined from the concentration dependences of G(C1T) and G(-CrOi-) is also different. These two findings are explained if we take into account the following experimental data.We studied the TABLE 4.-DEPENDENCE OF G(C11) AND AG(-CrO:-) ON KCL CONCENTRATlON IN 5 X lo-’ mol dm-3 SOLUTION OF CrOi- conc. of KCl/mol dm-3 G(CG 1 AG(-CrOi-) 0 0.05 0.1 0.2 0.4 0.5 0.6 0 0.22 0.45 0.70 0.95 1.46 1.67 0 1 .oo 1.38 1.70 2.35 2.54 2.65A. K . PIKAEV, S. A . KABAKCHI AND A . A . ZANSOKHOVA 121 TABLE 5.-INFLUENCE OF NATURE OF CATION ON RATE CONSTANT RATIO ks/ks (k6/kS)/mOl dm-3 cation calculation on basis calculation on basis K+ 0.90 0.34 Rb+ 1 .so 0.43 cs + 2.60 0.79 of G(C1,) of G(-CrO$-) kinetics of formation and decay of C1, in neutral solution at different concentrations of KCl, and we found that at [Cl-] < 0.1 mol dm-3 the formation of C1; occurs during a time which is longer than the pulse duration.Moreover, the time necessary for the achievement of maximum concentration of C1; decreases with increase of chloride concentration, and at [Cl-] - 0.2 mol dmm3 it becomes equal to the pulse duration. The formation occurs as a process of first order. The observed rate constants of C1, formation in 0.01 and 0.1 in01 dm-3 solutions are equal to (2.3 & 0.5) x lo5 and (6.0 & 1.0) x lo5 s-' (both in the absence and the presence of CrOi-). This finding also supports the hypothesis that free hydrogen ions take no part in the C1; formation. The decay of CI; occurs as a process of second order (the observed rate constant 2kII is (1.5 & 0.3) x 10" dm3 mol-l s-l). This value coincides with the values measured in ref. (24) and (25). The rate constant does not depend on the chloride concentration or on the presence or the absence of CrOi-.From the results discussed it follows that, first, even during the pulse quite apart from the process of C1F formation the decay of these radical-ions takes place: c1, + c1,- --+ c1- + c1, (1 1) and, secondly, the stage which limits C1; formation is not reaction (5) but reaction (7). This explains completely the lack of validity of eqn (10) and the difference of rate constant ratios k6/k5 calculated from the formation of C1, and from the conversion of CrOi-. The reason is that during the study of the competition of reactions (5) and (6), using the chromate as reference compound, we measure G(-CrOi-) and k,/k, which are not changed by kinetic factors. It is because of this that reaction (12) CrOi- + H30+ --f HCr04 + H,O HCr07 + H,O -+ CrOi- + H30+ (12) (13) has a comparatively low rate. However, in the case of the study of the competition of reactions ( 5 ) and (6), using Cl, as reference compound, the occurrence of reaction (1 1) during the pulse gives a lowered value of G(C1;) (especially at low concentrations of Cl-), i.e., in this case G(CIF) is not equal to the yield of Cl atoms, the product of reaction (5).This leads to the failure of eqn (lo), obtained from the suggestion that G(C1;) = G(C1). This is also the reason why k6/k5 values calculated from C1; formation are higher than the values determined from CrOi- conversion. Hence, experimental results confirm the mechanism of C1; formation in the presence of CrOi- including reactions (5)-(7), (1 1) and (12).For the study of the dependence of G(C1;) on pulse radiolysis of neutral aqueous solutions of alkaline metal chlorides on temperature, it is necessary to know the temperature dependence of the molar extinction coefficient of this radical-ion. With this purpose in mind optical absorption spectra of ClF at different temperatures but at the same dose per pulse were registered. It was found that with increase of has a small characteristic time and the reverse reaction:122 YIELDS AND REACTIONS OF HYDROGEN IONS ON RADIOLYSIS temperature the intensity of the optical absorption band of Cl; is decreased, its width is increased and the maximum is shifted to longer wavelengths. The temperature coefficient of the shift is -10 cm-l K-l. The short-wave part of the C1, spectrum has a Lorentzian shape and the long-wave part of the spectrum has a Gaussian shape.For calculations of extinction coefficient eqn (14) was used : 26 f= 4.32 x lo-' (1.065 W$2 + 1.571 W&)E,,, (14) where f is oscillator strength, is extinction coefficient at E , , ~ , Wly2 and W1Y2 are Gaussian and Lorentzian parts of the band width at half-height. This value was calculated from the known value of E,,,, at room temperature. In the calculations of E,,, for other temperatures it was assumed thatfdid not depend on the temperature within the range of 20-90°C. The data obtained are shown in fig. 8. It was found that the temperature coefficient of E,,, was -6.1 m2 mo1-l K-'. The E,,, values obtained were used for the measurements of G(C1,) for tempera- tures from 20 to 90°C.These yields depended on the nature of cation but were constant within the temperature range investigated. For example, for 1 mol dm-3 It was found that f = 0.46 1.2 1- € 1.1 E = - c 0 - hl \ 1.0- 0.9 m - 0.8' 1 temp I "C 20 30 A0 50 60 70 80 90 FIG. 8.-Dependence on temperature of molar extinction coefficient E,,, of C11 radical-ions in neutral aqueous solution of KCl. 0 8 - 0 E \ IdN 8 - 0 .A- 0 c - 0 ._ c 0 1 2 3 10-17x dose /eV ~ r n - ~ e l - n 8 E " FIG. 9.-Dependence of ClZ concentration on dose per pulse for aerated neutral aqueous 1 mol dm-3 solution of KCI at different temperatures: 0-20"C; .-3O0C; 0 4 0 0 ~ ; c)--5ooc; e--(jO"C; 0-70"C; @-8O"C; @-9O"C,A . K . PIKAEV, S . A . KABAKCHI AND A . A . ZANSOKHOVA 123 solutions of KCl, RbCl and CsC1, the yields of Cl;, independent of temperature, are equal to 2.1, 1.3 and 1.0 ions/100 eV, respectively. It is also illustrated by fig.9 which shows the dependence of Cl, concentration on dose for aerated 1 mol dm-3 solution of KCl at temperatures 20-90°C. Within the limits of experimental error the data obtained at different temperatures are the same. Hence the ratio of a probability of [H30+ . . . OH] decay to a probability of charge migration leading to C1- oxidation [reaction (5)] does not in practice depend on temperature within the range from 20 to 90°C. Obviously this ratio must depend on the degree of structuring of the water. Evidently, the concentration of structured water within the temperature range of 20-90°C is changed but slightly.This con- clusion does not contradict literature data. For example, according to ref. (27), thermal vibrations disrupt the hydrogen bonds in water only at temperatures higher than 200°C. Spectroscopic investigations2' show that on increasing the temperature from 20 to 80°C the concentration of monomeric water is increased only by 6% (from 6 to 12%). A. K. Pikaev, Sol'vatirovannyi elektron v radiatsionnoi khimii (Izd-vo " Nauka ", Moskva, 1969). E. J. Hart and M. Anbar, The Hydrated Electron (Wiley-Interscience, N.Y., 1970). I. G. DraganiE and Z. D. DraganiE, The Radiation Chemistry of Water (Academic Press, N.Y., 1971). K. H. Schmidt and S. M. Ander, J. Phys. Chem., 1969, 73, 2846. B. Cercek and M. Kongshaug, J. Phys. Chem., 1969, 73, 2056.G. S. Barker, P. Fowles, D. S. Sammon and B. Stringer, Trans. Faraday SOC., 1970, 66, 1498. S. A. Kabakchi, A. A. Zansokhova and A. K. Pikaev, Khimiya vysokikh errergii, 1974, 8, 255. G. V. Buxton, Proc. Roy. SOC. A , 1972,328, 9. S. A. Kabakchi, L. I. Kudryashov and L. T. Bugaenko, Int. J. Radiation Phys. Chem., 1970,2, 187. lo M. Anbar and J. K. Thomas, J. Phys. Chem., 1964, 68, 3829. l1 T, P. Zhestkova and A. K. Pikaev, Khimiya vysokikh energii, 1975, 9, 165. l2 G. G. Jayson, B. J. Parsons and A. J. Swallow, J.C.S. Faraday I , 1973,69,1597. l3 H. Ogura and W. H. Hamill, J. Phys. Chem., 1973, 77, 2952. l4 S. A. Kabakchi, A. A. Zansokhova and A. K. Pikaev, Doklady Akad. Nauk. S.S.S.R., 1975, 221, 1107. l5 A. K. Pikaev, G. K. Sibirskaya, E. M. Shirshov, P. Ya. Glazunov and V. I. Spitsyn, Dokfady Akad. Nauk. S.S.S.R., 1967, 200, 383. l6 A. K. Pikaev, G. K. Sibirskaya, E. M. Shirshov, P. Ya. Glazunov and V. I. Spitsyn, Dokfady Akad. Nauk. S.S.S.R., 1974, 215, 645. l7 G. H. Baxendale, P. L. T. Bevan and D. A. Stott, Trans. Faraduy SOC., 1968,64, 2389. l8 I. BrdiEka, 2. Elektrochem., 1960, 64, 16. l 9 B. H. J. Bielski and J. M. Gebicki, A h . Radiation Chem., 1970, 2, 177. 'O Ya. Ilan and J. Rabani, Int. J. Radiation Phys. Cheni., 1976, 8, 609. 21 W. Foerst, Angew. Chem., 1963, 75, 489. 22 G. Ertl and H. Gerisher, 2. Elektrochem., 1961, 65, 629. 23 G. Ilgenfritz, Doctoral Dissertation (George August University, Gottingen, Germany, 1966); cited from L. P. Holmes, A. Silzars, D. L. Cole, L. D. Rich and E. M. Eyring, J. Phys. Chem., 1969,73, 737. 24 M. E. Hangmuir and E. Hayon, J. Phys. Chenz., 1967, 71, 3808. 25 T. P. Zhestkova and A. K. Pikaev, Zzvest. Akad. Nauk. S.S.S.R., ser. khim., 1974, 255. 26 I. G. Calvert and J. N. Pitts, Photochemistry (Wiley, N.Y., 1966), p. 172. 27 I. V. Matyash, Voda v kondensirovaniiykh sredakh (Izd-vo " Nauka ", Kiev, 1971), p. 5. 28 0. D. Bonner and G. B. Woolsey, J. Phys. Chem., 1968,72,899.

ISSN:0301-7249    

DOI:10.1039/DC9776300112

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Polarogram of the Free Hydrogen Atom some Simple Organic Radicals and of BY P. TOFFEL AND A. HENGLEIN Hahn-Meitner-Institut fur Kernforschung Berlin GmbH, Bereich Strahlen-chemie, D 1000 Berlin 39 Received 28th October, 1976 An apparatus is described which allows one to measure the polarographic currents of free H atoms produced radiation chemically in an aqueous solution under 50 atm of hydrogen. The electrode was a tiny mercury drop. The polarogram of the H-atom contains a weak anodic wave and two cathodic waves. The pH dependence of these waves is described and electrode processes are attributed to the waves. Methyl and ethyl radicals were also produced in solutions under high methane or ethane pressure as well as the 2-hydroxyethyl radical in an ethylene + NzO containing solution.All these radicals have separated anodic and cathodic waves. The anodic wave is shifted towards more negative potentials in alkaline solutions. In acid solution additional cathodic currents were obtained in the potential range between the two main waves. Various electrode reactions of these organic radicals are discussed. Hydrogen atoms in aqueous solutions may act as oxidizing as well as reducing agents. Ferrous ions and iodide anions, for example, are oxidized in acidic solutions, while Ag + is reduced. In their electrochemical behaviour, hydrogen atoms also exhibit oxidative or reductive capacities. Depending on the potential of the electrode, they are either oxidized to yield protons or reduced to yield molecular hydrogen and hydroxyl anions.These reactions are of interest in the understanding of cathodic hydrogen evolution, one of the most intensely studied processes in electrochemistry. In the present study, the methods of pulse radiolysis and polarography were combined to produce free hydrogen atoms in aqueous solutions and to study their reactions at a mercury drop electrode. The polarographic behaviour of many short- lived free radicals has been studied in this manner for the last few years.l In principle, the radicals are formed suddenly in the solution around the mercury drop by a 20 ns pulse of high energy electrons (12 MeV linear accelerator) and their polarographic currents are recorded as a function of time. After having carried out this experiment with different electrode potentials, one can const,ruct the polarogram of the radical, which shows the current at a given time after the pulse as a function of the electrode potential.Hydrogen atoms are produced radiation chemically via the following reactions of the initially formed hydrated electrons and hydroxyl radicals : in acid solutions : ea; + Hf --f H OH + H2 + H20 + H (1) (2) (3) (2) in alkaline solutions : e,; + N,O -% N2 + OH + OH- OH + H2 -+ H20 + H.P . TOFFEL AND A . HENGLEIN 125 In order to scavenge all OH radicals by dissolved hydrogen, it was necessary to operate under a pressure of 50 atm of hydrogen gas. The rate constant of reaction (2) is comparatively low (k = 4.5 x lo7 dm3 mol-I s-l), i.e., part of the hydroxyl radicals would have combined with each other or with H atoms, if the solution had been under normal hydrogen gas pressure.With alkaline solutions, it was necessary for nitrous oxide also to be present (at one atmosphere) to scavenge the hydrated electrons. An apparatus will be described below which allows one to carry out polarographic measurements under high gas pressures. Simple alkyl radicals such as CH3, CH3CH2 and CH2CH,0H can be produced by irradiating an aqueous solution of N20 and methane, ethane or ethylene respectively. Because of the low solubility of these gases and partly because of the relatively low rate constants for their reactions with hydroxyl radicals, high gas pressures also had to be applied to scavenge all OH radicals. The pressure vessel was therefore also used to study the polarographic behaviour of these organic radicals.Electrochemists have often used the method of electron photo-emission from an electrode to investigate the reactions of free radicals at electrode^.^*^ The ejected electrons solvate and react with solutes to form free radicals; for example, they react with protons to form hydrogen atoms [eqn (l)] or with dissolved methyl chloride to form methyl radicals. A photo-current is observed if the radical is reduced at the electrode. If it is oxidized, the electron is returned to the electrode, i.e., no photo- current is observed. In the case of the H atom, a photo-current that increased with negative potential according to i 0 9 4 cc 9 has been observed. At low negative poten- tials, deviations from this relationship were observed and attributed to an anodic process of the H atom. EXPERIMENTAL A simplified cross section through the apparatus is shown in fig. 1.The pressure vessel consisted of stainless steel; it had a 1 mm thick window to admit the electron beam. The solution was contained in a glass vessel at the lower end of the apparatus. It could be degassed by a hydrogen stream which was conducted into the solution via thin Teflon tubing. The mercury electrode, the gold counter electrode and the Ag/AgCl reference electrode were all mounted at the upper part of the pressure vessel. The mercury electrode consisted of the commercial Metrohm type E 410 micro feeder and the capillary (0.17 mm diameter). The micro feeder was mechanically connected to the upper part by a Teflon bush. The size of the mercury drop was controlled by measuring the capacity of its double layer.Pressurization and removal of air was achieved through three valves. Although the power supply of the electrode could be varied between +2 and -2V, the potentials which can be applied to a mercury electrode are limited to the range between +0.2 and - 1.7 V. The voltage drop at a working resistor of 20 $2 was recorded as a function of time using a differential amplifier type Tektronix 7A22 and a storage oscilloscope type 7613. The voltage between the mercury electrode and the Ag/AgCl reference electrode was measured with a digital voltmeter. Dosimetry was carried out using the Fricke dosimeter. The dosimeter solution was irradiated with 10 to 40 pulses and the concentration of ferric ions formed determined spectrophotometrically.The irradiated volume of the solution is about 15 % of the total liquid volume. Correction for this ratio was made. The dose per pulse in the irradiated volume was 10 krad. The accuracy of this determination is believed to be within a factor of two, the error being mainly due to the uncertainty in the value of the irradiated volume. In most of the solutions, sodium sulphate was used as supporting electrolyte at a con- centration of 0.3 mol dm-3. The solution and the pressure vessel were thoroughly flushed with hydrogen (or methane etc.), before the higher pressure was applied. Only those solu- tions were used which had a very low base current. Curve (a) in fig. 2 shows how the base126 POLAROGRAM OF THE FREE HYDROGEN ATOM ,actuator for drop formation pressure vessel upper part electric feedthrough mercury micro feeder pressure vesseI lower part Ag lAgCl electrode t------------1 10 cm FIG.1 .-Simplified cross section through the pressure vessel for electron irradiation and polarography. b UJ v -1.0 FIG. 2.-Base current as a function of electrode potential U (against Ag/AgCl). Solution: 0.3 mol dm-3 Na2S04, 50 atm H2. Curve a: fresh solution. Curve 6: after irradiation with 20 pulses. current changed with the applied potential. A solution which was irradiated with 20 pulses had an increased base current [curve (11) in fig. 21. The polarographic currents of the radicals were of the order of microamperes. The desired pH-values of the solutions were obtained by adding perchloric acid, sulphuric acid or sodium hydroxide.RESULTS H-ATOMS Polarographic current-time oscillograms at various electrode potentials are shown Positive signals correspond to cathodic currents, The current decreases after the pulse for two in fig. 3 for the hydrogen atom. negative signals to anodic ones.P . TOFFEL A N D A . HENGLEIN 127 reasons: (1) The electro-chemical consumption of H-atoms causes a decrease in their concentration at the surface of the electrode, i.e., a diffusion layer is built up with time. (2) The H-atoms disappear in the solution via reaction 2H -+ HZ. (4) The rate constant of this reaction is 2k = 2 x 1O1O dm3 11101-1 s-l. During the first 20-30 ps after the pulse, a disturbing signal due to electrons absorbed from the beam 0 -150 L d -- -500 aJ 0 CL Y -1000 mV mV mV mV pulse FIG.3 .-Oscillograms for the polarographic current of the H-atom (potentials against Ag/AgCl). appeared. Measurement of the polarographic current was therefore carried out at times longer than 30 ps. The initial concentration of H-atoms was 2 x mol dm-3. At 30 pus after the pulse, the bulk concentratioii had already dropped to 1.6 x It can be recognized that weak anodic currents are observed at potentials around zero volts and cathodic currents at more negative potentials. Polarograms of the H-atom at various pH-values are shown in fig. 4. They contain a weak anodic wave at potentials more positive than -0.18 V, a weak cathodic mol dm-3. 8 6 4 4 *-. 0 -2 -4 -6 % * --. 0 - 1.0 U l v FIG. 4.-Polarogram of the H-atom at pH values 0 = 1, A = 3, V = 6, 0 = 10.5 and at 30 ps after the pulse (potential against Ag/AgCl).128 POLAROGRAM OF THE FREE HYDROGEN ATOM wave between -0.18 and -0.9 V (pH > 4) and a stronger second wave for reduction of the H-atom at potentials more negative than -0.9 V.The second cathodic wave appears at more positive potentials in solutions of lower pH. Fig. 5 shows a plot of - 0.5 0 5 10 PH FIG. 5.-Half wave potential (against Ag/AgCl) of the second cathodic wave of the H-atom as a function of pH. the potential of the turning point of the second wave as a function of the pH. These observations are in agreement with the results of the electron photo emission ~tudies.~ The pK of the equilibrium H + H + + e ; is known to be 9.7. The polarogram of the hydrated electron, whose reduction potential is -2.9 V, should consist of a broad anodic wave over the whole potential range available at the mercury electrode.Fig. 6 shows polarograms for various H==H++ e& 4 ( D o -8 -6 5 0 -1.0 U l V FIG. 6.-Polarograms of alkaline solutions at pH values Q) = 10.5, (3 = 12, A = 13 (potential against Ag/AgCl). alkzline solutions. At pH = 10.5, where the hydrated electron is much more abundant than the H-atom in the equilibrium of eqn (9, the polarogram of the latter is still observed. At pH = 13, the expected polarograni of the hydrated electron appears. A superposition of the two polarograms is observed at pH = 12. The explanation for the apparent shift in the pK of the H-atom by about 2 units lies in the fact that the local concentrztion of OH- ions at the surface of the negatively charged electrode is lower than in the bulk of solution.The concentration of a z-fold charged species chalzgss with distance x from the electrode according to the relationshipP . TOFFEL AND A . HENGLEIN 129 where Co is the bulk concentration and ‘y, the potential. x may now be taken as the distance of closest approach of a solvated anion. It has been shown by Delahay4 that F- ions, which have a similar adsorption behaviour, experience a potential of only -0.13 V at this distance at a nominal potential of -1.0 V of the electrode. Applying this result to OH--ions, one calculates a pH which is lower by 2.3 units at the electrode than in the bulk of the solution. An additional reason for observing the polarogram of the hydrated electron at relatively high pH values is the slow establishment of the equilibrium of eqn (5).OH radicals are reduced at all potentials that can be applied to the mercury electrode. The cathodic currents are especially high beyond -0.8 V. Another product of the radiolysis of water is hydrogen peroxide; it is produced in small yield, however. It can also be reduced polarographically at potentials more negative than - 1.0 V. That the second cathodic wave of the polarograms of fig. 4 and 6 is caused by the H-atom and not by the oxidizing products of the radiolysis of water was demonstrated by using a polyethyleneoxide-coated mercury electrode. Polyethyleneoxide is strongly adsorbed at the electrode even at low polymer concentrations. The adsorbed macromolecules do not inhibit dissolved molecules from diffusing to the electrode surface and reacting there.However, OH radicals cannot reach the surface since they abstract H-atoms from the adsorbed macro- molecules to produce macroradicals which produce only very weak polarographic signal^.^ Fig. 7(a) shows the polarogram for an irradiated N,O-saturated solution in which both OH and Hz02 are formed. In the presence of polyethyleneoxide, no significant polarographic signal at strongly negative electrode potentials is observed. FIG. (10- 8 / a / 6 4 2 0 -2 Q 9 8 \ 6 4 2 /b/ - 2 -I - 6 t c t - - - - 0 -1.0 U I V 7.-(a) Polarograms of N,O saturated solutions containing no (A) or a small amount (0) mol dm-3) of polyethyleneoxide. ( b ) Polarograms of solutions under 50 atm of hydrogen in the absence (0) and presence (A) of polyethyleneoxide.The polarogram in the absence of the polymer is also shown for comparison. Ap- parently, the amount of hydrogen peroxide produced is too small to cause any measurable cathodic currents at the short times of our observation. It can finally be130 POLAROGRAM OF THE FREE HYDROGEN ATOM seen from fig. 7(b) that the polarogram of the H-atom, which is little reactive towards polyethyleneoxide, is not affected by this polymer. Additional proof for the existence of the second cathodic wave of the H-atoms comes from experiments with argon-saturated acid solution in which H and OH are formed in roughly equal concentrations. Superposition of the polarograms of these two species could, therefore, be expected.Fig. 8 shows that a polarogram is obtained - 1.0 U / v FIG. 8.-Polarograni of N20 (A) and argon (0) saturated solutions. whose shape resembles that of the H-atom. It is simply shifted upwards by the underlying cathodic current of the OH-radical. A comparison with the polarogram obtained for an N,O-saturated solution, in which mainly OH and H202 are formed, shows that the reduction signal at potentials more negative than - 1.0 V is much more pronounced in the presence of H. The anodic wave of the H-atom was stronger when sodium sulphate was sub- stituted by sodium perchlorate as supporting electrolyte. Furthermore, it was more pronounced at pH = 1 in sodium sulphate solution, where bisulphate ions were mainly present. F- and C1- ions also showed some promoting effect.All these anions are more strongly adsorbed at the mercury electrode than is the sulphate anion. A promoting effect has also been reported for the iodide anion.3 ORGANIC RADICALS The polarogram of the methyl radical which was produced via the reaction OH + CH4 -+ CH3 + HzO (7) is shown in fig. 9. It contains well-separated anodic and cathodic waves in neutral and alkaline solutions. The anodic wave shifts towards more negative potentials 6 2 4 2 .‘2 0 - 2 -4 -6 1 0 -1.0 u / v FIG. 9.-Polarogram of the methyl radical at various pH values A = 3, 0 = 4, 0 = 5 , A = 6, D = 11.5 (potential against Ag/AgCl).P. TOFFEL AND A . HENGLEIN 131 with increasing pH, while the cathodic wave is not changed. In acid solutions neither wave changes with pH.However, small cathodic currents are recorded here in the potential range between -0.2 and -0.9 V, which separates the two waves observed in neutral or alkaline solutions. Similar effects were observed for the ethyl radical. The reduction of the methyl radical has already been though its oxidation has not yet been recognized. The polarogram of the 2-hydroxyethyl radical produced via the reaction OH + C2H4 --f CH2CH20H (8) is shown in fig. 10(a). In- cluded in this figure is the well-known polarogram of the 1-hydroxyethyl radical which is obtained by the attack of OH on ethyl alcoh01.~ The broad anodic wave of this radical is in agreement with its strong reducing properties. The cathodic wave It contains well separated anodic and cathodic waves. 6 4 la) 0 -2 -L a > a 1 6 4 lbl 2 0 -2 0 -1.0 U I V FIG.l&(a) Polarogram of , 1-hydroxethyl (- CH2CH20H) and 0,2-hydroxyethyI (CH,dCHOH) radicals at pH = 6. (6) Polarogram of the 2-hydroxy-2,2-dimethyl ethyl (* CH2C(CH3)20H) radical at pH-values 0 = 4, = 11. of the 2-hydroxyethyl radical was not dependent on the pH in acid solutions. This fact seems remarkable since the 2-hydroxy-2,2-dimethylethyl radical which is formed in the attack of OH on t-butanol was shown to have a cathodic wave, the position of which was very sensitive to the pH of the ~ o l u t i o n . ~ Fig. lo@) shows the polarogram of this radical for comparison. DISCUSSION Various reactions which H and CH, may undergo at an electrode are compiled in table 1. In these reactions, either an electron is transferred without changes in the chemical bonds [reactions (9), (lo), (13), (14)] or the transfer is accompanied by the breakage and formation of bonds [reactions (1 l), (12), (15)-(17)].The standard redox potentials of these reactions are also given. They were calculated by using132 POLAROGRAM OF THE FREE HYDROGEN ATOM thermodynamic data for the stable particles H+ (aq), CH4 (g), H20 (aq), OH- (aq) and for H- (as) from tables. The standard free energy of H(g) was taken as 2.1 eV, that of CH3(g) as 1.7 eV and the energies of hydration of H, CH4, and CH3 as 0.2 eV (like helium). The energy of hydration of CH, was assumed to be equal to that of the relatively large H- particle, and electron affinities of 0.8 and 0.3 eV were used for H and CH3, respectively. The ionization energy of CH,(g) is 9.8 eV9; an upper limit of 4 eV for the hydration energy of CH+ was assumed.Furthermore, the C-0 bond strength in methanol was taken as 3.9 eV. The redox potentials would change somewhat if the reactions occurred via adsorbed radicals. The observed polarographic waves do not correspond to reversible processes at the electrode. Some of the products of reaction such as H-, CH; and CH,+ will react rapidly with the solvent before the back reaction can be established. However, the main reason for the irreversibility of the waves lies in the fact that the reactions have to be fast at the short times of our observation. They are only fast at sufficiently high overpotentials. Oxidation processes require much more positive, and reduction processes much more negative potentials than the redox potentials of the corre- sponding reactions.In all the reactions of table 1, major rearrangements of the TABLE l.-STANDARD REDOX POTENTIALS FOR THE OXIDATION AND RE- DUCTION OF H-ATOMS AND METHYL RADICALS (e-: ELECTRON m ELECTRODE) reaction standard redox IV potential no. H (as) + e - + H- (as) H (as) -+ H+ (as) + e - H (as) + H20 + e - -+ H2 + OH- (as) H(aq) + H+ (as) + e - + H2 CH, (as) + e- + CH; (as) CH3 (as) + CHZ (as) + e - CH3 (as) + H20 + e- -+ CH4 + OH- (as) CH3 (as) + H+ (as) + e - + CH, CH3 (as) + OH- (as) + CH30H (as) + e- 0.05 1.5 2.3 - 0.45 1.5 1.4 2.2 - 2.1 - 2.3 solvation shells of the particles involved are required. In an actual electron transfer process, the product is not formed in its lowest energy state of solvation but with a great deal of excess energy, which has to be supplied by the overvoltage.Oxidation of the H-atom, for example, occurs only at potentials more positive than about zero volt (against hydrogen electrode), i.e., at overpotentials greater than 2.3 V. On the other hand, strong reduction is only observed at potentials more negative by 2-3 V than the redox potentials of reactions (1 1) and (12). Under these conditions, one can expect a relatively wide potential range in which H-atoms are oxidized and reduced simultaneously with comparable rates.1° It must, therefore, be suspected that the first cathodic wave of the H-atoms (fig. 4) contains a weak anodic component and, vice versa, that the anodic wave contains some cathodic contribution.At - 0.18 V (against Ag/AgCl) where the polarographic current is zero (pH > 4) the anodic and cathodic currents cancel each other. The dependence on pH of the two cathodic waves of the H-atom is the only means for attributing chemical processes to them. Since the first wave is independent of pH, we attribute process (11) to it, and since the second wave still appears at low H+ concentration (fig. 4) we believe that reaction (9) is responsible. The shift of theP . TOFFEL AND A . HENGLEIN 133 second cathodic wave at lower pH-values is explained by the occurrence of reaction (12). It does not seem possible to draw any further conclusions from the observed polarograms, such as about the activation or barrier-less occurrence of such processes. The currents in the plateaus of the anodic and the first cathodic wave are rather low.This may be understood in terms of reactions (10) and (1 1) occurring via adsorbed H atoms with the result that the overall rate is determined by that of the adsorption. In the case of the methyl radical, the cathodic wave can hardly be attributed to reaction (16) because of the independence of pH. We would rather attribute this wave to reaction (13) as has already been done by Schiffrin.6 However, the weak cathodic currents in acidic solutions and at less negative potentials (fig. 9) indicate that a second process of reduction becomes possible in the presence of protons. This process is believed to be reaction (16). The anodic wave of CH3 is attributed to reaction (14). The shift of the wave to more negative potentials in alkaline solution is attributed to reaction (17). The waves of the 2-hydroxy ethyl radical are also explained by simple electron transfer from or to the electrode. The product of reduction should be ethanol. The reduction of the 2-hydroxy-2,2-dimethylethyl radical, however, is attributed to the formation of isobutylene and OH- because of the strong catalysing action of protons : CH3 CH3 H I 1 +/ I \ I CH3 CH2-C-OH + H+ + CH2-C-0 -L CH2=C(CH3)2 + H20. (18) CH H Isobutylene has indeed been found as a product of the reduction of the 2-hydroxy- 2,2-dimethylethyl radical by Ni+ .I1 A. Henglein, in Electroanalytical Chem., ed. E. J. Bard (Marcel Dekker, N.Y., 1976), vol. 9, p. 163. Yu. V. Pleskov, Z. A. Rotenberg, V. V. Eletsky and V. I. Lakomov, Disc. Faraday SOC., 1973, 56, 52. P. Delahay, Double Layer and Electrode Kinetics (Interscience, New York, 1966). R. M. Sellers, K. M. Bansal, E. Janata and A. Henglein, Ber. Bunsenges. phys. Chem., 1974, 78, 1085. D. J. Schiffrin, Disc. Faraduy SOC., 1973,56,75. * G. C. Barker, Ber. Bunsenges. phys. Chem., 1971, 75,728. ’ M. Gratzel, A. Henglein, J. Lilie and M. Schemer, Ber. Bunsenges. phys. Chem., 1972, 76, 67. * M. Pourbaix, Atlas of Electrochemical Equilibria (Pergamon Press, Oxford, 1966). J. L. Franklin, J. G. Dillard, H. M. Rosenstock, J. T. Herron, K. Draxl and F. H. Field, National Bureau of Standards, Report NSRDS-NBS, 26, 1969. M. Kelm, J. Lilie, A. Henglein and E. Janata, J. Phys. Chem., 1974, 78, 882. lo A. Henglein, Ber. Bunsenges. phys. Chem., 1974,78, 1078.

ISSN:0301-7249    

DOI:10.1039/DC9776300124

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

“Molecular” Products in the Radiolysis of Vinyl Monomers BY A. CHAPIRO, A. M. JENDRYCHOWSKA-BONAMOUR AND G. LELIEVRE Laboratoire de Chimie Macromol6culaire sous Rayonnement, C.N.R.S., 92190 Meudon, France Received 10th December, 1976 Radiolytic yields for gas production in styrene were found to be: G(H2) = 0.018, G(C2H2) = 0.0053. These values, which are significantly lower than those measured in benzene which has an electronic structure similar to that of styrene, are assumed to correspond to “ molecular ” products. In mixtures with cyclohexane, styrene produces a much stronger protection on G(H2) values than benzene, An analysis of the data leads to an estimate of the molecular product yields in benzene: G(HJ = 0.0195 and G(C2H2) = 0.0075. Preliminary data indicate that in methyl methacrylate G(tota1 gas) = 0.51 and G(CH4) = 0.05.Styrene also “ protects ” benzene with respect to G(H2) and G(C2H2). The most obvious chemical change occurring when a vinyl monomer is subjected to ionizing radiation is its polymerization. Numerous papers deal with radiation- initiated polymerization and the pertinent literature is extensively Experimental evidence conclusively demonstrates that under the usual irradiation conditions the polymerization proceeds via a conventional free radical mechanism. In this process the radicals generated in the medium by the radiolysis of the monomer M and eventually the solvent S add to the double bonds of the monomer thereby initiating growing polymer chains : M ---+ 2 Re (1) S---+2R (2) R*+CH2=CHX+R-CH2-C* (3) H X ........................................................................................................................................................ R MA + M -+ R MA+1.(5) It thus appears that the initiating radicals R become chemically attached to the polymeric chain end and are thereby eliminated from the reaction medium. It was further shown5 that primary radicals R- may escape scavenging by the monomer [reaction (3)] and combine either with themselves or with a growing chain RM;: R* + R* + R2 R + RM; -+ P,. In the most general case reactions (3), (6) and (7) compete, and normal competition rules apply to the system. If the concentration of monomer is high enough and the dose-rate is not too high, reactions (6) and (7) become negligible and in such cases (6) (7)A .CHAPIRO, A . M . JENDRYCHOWSKA-BONAMOUR AND G . LELIEVRE 135 all free radicah are captured by the monomer in reaction (3).5 Any product formed in such a system other than the polymer is expected to have a non-radical origin. Very few papers deal with the formation of non-polymeric products in irradiated monomer. Gaseous products were detected in the radiolysis of a number of olefins. Several papers deal with the radiolysis of ethylene and other gaseous olefins but irradiations were usually performed under non-polymerizing conditions and these studies will not be considered here. Hexadec-1-ene6 and various dienes 7*8 were irradiated in the condensed state. Hydrogen and trace amounts of hydrocarbon gases were found among the reaction products together with the polymer.Crystalline acenaphthylene polymerizes under irradiation, but also generates fair amounts of cyclic dimer 9*10 the precursor of which is a triplet state of the monomer. No cyclic dimers were found in the radiolysis of muconic or sorbic acids or their esters in spite of the very high yield of the cyclodimerization in the photolysis of the same compounds.ll The present study is devoted to the analysis of gaseous products arising in the radiolysis of conventional vinyl monomers either in bulk or in solution. Most experiments were carried out with styrene. Preliminary results are reported for the radiolysis of methyl methacrylate. EXPERIMENTAL Pyrex vessels 16 mm in diameter containing -5 g of monomer were carefully degassed and sealed under high vacuuni.They were subjected to cobalt-60 gamma-rays at 20°C and dose-rates of 68 and 1700 rad min-'. After irradiation the vessels were attached to a vacuum line mounted with a McLeod gauge and the gases were analysed. * The sample was first kept in liquid nitrogen and the non-condensible gases were pumped out and led through a liquid nitrogen trap into the gauge. Their volume was measured; they were then collected in a gas coil and injected into an Aerograph " Autoprep 706 " chromatograph and analysed. The sample was thereafter warmed to - 78°C in a dry ice bath and the same procedure was repeated in order to analyse the condensible gases. Each monomer or binary mixture was subjected to at least 3 or 4 different doses and it was checked that the gas yield was a linear function of dose. The G(gas) value was derived from the slope of the gas yield, against dose relationship, RESULTS 1.GAS YIELDS FROM CYCLOHEXANE A N D BENZENE In order to check the experimental procedure we determined values of G(H,) in cyclohexane, and G(H,) and G(C,H,) in benzene. The gas evolution was found to be linear with dose in the range studied, i.e., from 10 000 to 30 000 rad in cyclohexane and from 1 to 3 Mrad in benzene. The results are shown in table 1 which also includes the corresponding G values reported in the literature. TABLE l . - G ~ s YIELDS (MOLECULES PER 100 eV) IN THE RADIOLYSIS OF CYCLOHEXANE AND BENZENE cyclohexane present work literature benzene 0.039 l4 G(H2) 0.033 0.038 l5 5 .O 5.612,13 GW2) W 2 H J 0.020 0.019 l4.lS * The design of the gas analysis equipment was suggested by Dr.J. Belloni.136 “MOLECULAR” PRODUCTS IN THE RADIOLYSIS OF VINYL MONOMERS It appears that our G(H2) values are 10 to 20% lower than the values from the literature, whereas the values of G(C,H,) are in good agreement. 2. GAS YIELDS FROM STYRENE The gases evolved from irradiated styrene contain chiefly hydrogen and acetylene. No attempts were made to characterize the traces of other products. The yields were found to be linear with dose in the range of 0.5 to 3 Mrad for H2 and 2 to 4.5 Mrad for acetylene. For such doses significant amounts of polymer are formed but at the higher dose-rate used (1700 rad min-l) the molecular weights are fairly low: 4 0 O0Ol6 and the increase in viscosity of the system does not hinder the degassing.The G values were found to be: G(H2) = 0.018 and G(C,H2) = 0.0053. These values are significantly lower than the yields in benzene, a compound which has a very similar chemical structure. 3. GAS YIELDS FROM BINARY MIXTURES A. MIXTURES OF CYCLOHEXANE WITH STYRENE The G(H2) values measured in mixtures of cyclohexane and styrene are plotted in fig. 1 against the electron fraction of cyclohexane. It can be seen that G(H,) drops very steeply when small amounts of styrene are added to the saturated alkane. It is of interest to compare these results with the data from the literature con- cerning mixtures of cyclohexane and benzene.17-19 The broken curve B on the same figure shows the early results of Manion and Burt0n.l’ In this system the reduction of G(H2) in the presence of benzene is generally interpreted in terms of energy or charge transfer processes, though other mechanisms were also postulated.20*21 FIG.I .-Hydrogen (A)-mixtures I I I I I 3 - I I rs Y 2 - 1 - 0 0.25 0.5 0.75 1 electron f r a c t i o n yields in binary mixtures as a function of electron fraction of cyclohexane. with styrene; and (BFmixtures with benzene [based on data of ref. (131.A . CHAPIRO, A . M . JENDRYCHOWSKA-BONAMOUR A N D G . LELIEVRE 137 Whatever the actual mechanism, fig. 1 clearly shows that styrene has a more pro- nounced “ protective ” action than benzene which can be accounted for by scavenging of radical precursors. From the data plotted in fig. 1 it appears that the G(H,) values measured in the presence of styrene represent -60% of the values measured in benzene containing mixtures, which could indicate that 40% of the hydrogen has free radical precursors.B. MIXTURES OF STYRENE WITH BENZENE This system is of special interest. Earlier studies on the radiation-induced polymerization of styrene in benzene solutions showed that the data could be inter- preted quantitatively on the basis of a linear relationship of free radical yields with respect to the composition of the binary mixture.,, G(R*) values were found to be 0.69 for styrene and 0.74 for benzene in good agreement with other scavenger It can thus be taken that no significant energy transfer processes nor any other complications occur in these mixtures, each component being radiolysed irrespective of the presence of its partner.The values of G(H2) and G(C2H2) measured in the mixtures are shown respectively in fig. 2 and 3. It can be seen that for both products FIG. 2.-Hydrogen yields in binary mixtures of styrene and benzene (mole fractions). the yields are much lower than would be expected on the basis of a linear relationship. Mole fractions are used in the plot which, in this system, are equal to electron fractions. The steep drop of G(H,) and G(C,H,) of benzene upon addition of styrene again suggests that styrene efficiently scavenges free radical precursors of H2 and C2H2. The fact that G(C2H,) decreases much steeper with styrene content than does G(H,) is significant. It indicates that the competition kinetics are more severe for Ha atoms than for the radical precursor of acetylene.4. ESTIMATE OF “MOLECULAR” YIELDS IN BENZENE In a pure monomer most radicals R are assumed to add to the double bonds In [reaction (3)] provided the competition with reactions (6) and (7) is cancelled out.138 “MOLECULAR” PRODUCTS I N THE RADIOLYSIS OF VINYL MONOMERS 1 0.0075 styrene 25 50 75 100 benzene (mole %I FIG. 3.-Acetylene yields in binary mixtures of styrene and benzene (mole fractions). such a case all low molecular (gaseous) products are expected to have a non-radical origin. It should be noted, however, that at the dose-rate of 1700 rad min-l used in these experiments the rate of polymerization of styrene no longer obeys the classical square root law,24 hence some of the primary radicals escape scavenging by the double bonds and are involved in radical-radical reactions (6) and (7).In reaction (7) the radical R becomes attached to a polymeric chain RM;, and cannot, therefore, generate any low molecular weight product. In contrast, reaction (6) may be responsible for the formation of gas. A series of experiments was carried out at a much lower dose-rate: 68 rad/min in order to check the influence of reaction (6). The G(H,) and G(C,H2) values were found to be very close to those obtained at the higher dose-rate. It can therefore be assumed that most radicals R react with the double bonds of styrene and that the gases result from non-scavengeable precursors. This view is most likely to hold in the case of acetylene. It is less obvious for hydrogen, since the scavenging of the radical precursors (H* atoms) by the double bonds has to compete with still another process : hydrogen abstraction H* + RH --+ H2 + R .(8) Reaction (8) usually requires a higher activation energy than the addition to the double bond and is, therefore, less likely to occur particularly at low temperatures. Some influence of this reaction cannot be ruled out, however, a priori. In the case of styrene, reaction (8) occurs with a lower probability than with an aliphatic hydro- carbon. Moreover, the addition of H* to the double bond of styrene is favoured by the formation of a highly resonance-stabilized benzyl radical. It is, therefore, justified to assume that the formation of H, on behalf of thermal H* atoms via reaction (8) can be neglected. The values of G(H,) and G(C2H2) measured in styrene can thus be taken as yields of “ molecular ” products.On the other hand, the structure of styrene being very close to that of benzene, the “ molecular ” yields in both compounds are expected to be similar. Considering fig. 2 and 3 and assuming that the steep drop of product yields in benzene uponA . CHAPIRO, A . M . JENDRYCHOWSKA-BONAMOUR A N D G . LELIEVRE 139 addition of styrene is indeed caused by scavenging of free radical precursors by the vinyl monomer, it appears that the " molecular " yields in benzene can be obtained by extrapolating the linear portion of the curves for styrene-rich mixtures towards the pure benzene end. In the case of acetylene the linear portion of the curve is well defined and the extrapolation leads to G(C2H2) = 0.0075. This value is close to the G(C2H2) in styrene and represents -1/3 of the actual G(C2H2) measured in benzene.It can thus be assumed that the radiolytic event which ruptures the benzene ring generates one molecule of acetylene and the radical precursor of two more molecules of acetylene : C6&, ---+ C2H2 + 2*C2H; or --+ C2H2 + C4Hi the total yield of this process being g(-C(&) = 0.0075. A similar treatment of the G(H2) values is less satisfactory since the linear portion of the curve in fig. 2 is less accurately defined. The extrapolation shown by the broken line in the figure leads to Gmolecular (H2) = 0.0195 which would suggest that 59% of the hydrogen evolved from irradiated benzene arise by " molecular " processes while 41% are formed via free radical intermediates. The " molecular " hydrogen could result from a molecular detachment from a highly excited species or from hydrogen abstraction by " hot " hydrogen atoms.5. GAS YIELDS FROM METHYL METHACRYLATE The radiolysis of methyl methacrylate generates a gaseous mixture of at least six constituents. The total gas yield was foilnd to be linear with dose over the range of 3000 to 30 000 rad with G(tota1 gas) 0.51. The identification of the various constituents is in progress. Methane was found to accumulate linearly with dose over the range of 100 000 to 260 000 rads with G(CH4) = 0.05. These values are significantly lower than the yields of gas evolution from poly(methy1 methacrylate) which are shown in table 2. TABLE 2.-GAS YIELDS (MOLECULES PER 100 ev) IN THE RADIOLYSIS OF POLY(METHYL METHA CRY LATE) present work ref.(25) G(H*) - 0.21 G(CH4) 0.44 0.54 W O ) 0.45 G(C02) 0.32 G(tota1 gas) 1.6 Here again the much lower G-values measured in the monomer are assumed to result from scavenging of free radical precursors. The data presented above show that interesting information about " molecular " yields can be gained from a study of gaseous products in the radiolysis of concentrated monomer solutions. Further work along these lines is in progress. A. Chapiro, Radiation Chemistry of Polymeric Systems (Interscience, New York, 1962). V. S. Ivanov, Radiatsionnaya Polimerizatsiya (Khimiya, Leningrad, 1967). F. Williams, Principles of Radiation Indirced Polymerization in Fundamental Processes in Radio- tion Chemistry, ed.P. Ausloos (Interscience, New York, 1968), pp. 515-98.140 “MOLECULAR” PRODUCTS IN THE RADIOLYSIS OF VINYL MONOMERS A. Chapiro, Radiation Induced Reactions in Encyclopedia of Polymer Science and Technology (John Wiley, N.Y., 1969), vol. 11, pp. 701-60. Ref. (l), pp. 132-140 and pp. 301-307. E. Collinson, F. S. Dainton and D. C. Walker, Trans. Faraday SOC., 1961, 57, 1732. V. S. Ivanov, Yu. V. Medvedev, V. F. Vasilenko, A. Kh. Breger, V. B. Csipov and V. A. Goldin, Vysokomol. Soed., 1963, 5, 1255. * V. S. Ivanov, Yu. V. Medvedev, Khou-Gui and A. A. Taran, Zhur. Obshch. Khim., 1964, 34, 3852; see also ref. (2), pp. 107-108. J. C. Muller, J. Chim. Phys., 1968, 65, 567. lo A. Chapiro and G. Lozach, Int. J. Radiation Phys. Chem., 1972,4, 285. l1 A. Chapiro, M. Lahav and G. M. J. Schmidt, Compt. rend., 1966,262, 872. l2 S. K. Ho and G. R. Freeman, J. Phys. Chem., 1964, 68, 2189. l3 J. Y. Yang and I. Marcus, J. Chem. Phys., 1965,42, 3315. l4 W. G. Burns and C. R. V. Reed, Trans. Faraday SOC., 1963,59, 101. l5 T. Gaumann and R. H. Schuler, J. Phys. Chem., 1961 , 65, 703. l6 Ref. (l), p. 164. l7 J. P. Manion and M. Burton, J . Phys. Chem., 1952,56, 560. l9 J. F. Merklin and S. Lipsky, J. Phys. Chem., 1964,68, 3297. 2o R. A. Holroyd, Organic Liquids in Fundamental Processes in Radiation Chemistry, ed. P. 21 J. HoignC, Aromatic Hydrocarbons in Aspects of Hydrocarbon Radiolysis, ed. T. Gaiumann and 22 Ref. (l), pp. 255-260. 23 Ref. (21), p. 94. 24 Ref. (l), pp. 162-164 and 172-175. 25 L. A. Wall and D. W. Brown, J. Res. Natl. Bur. Stand., 1956, 57, 131. G. R. Freeman, J. Chem. Phys., 1960,33,71. Ausloos (Interscience, New York, 1968), pp. 496-99. J. Hoign6 (Academic Press, London, 1968), pp. 123-25.

ISSN:0301-7249    

DOI:10.1039/DC9776300134

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Formation of Radical Cations and Zwitterions versus Demethoxylation in the Reaction of OH with a Series of Methoxylated Benzenes and Benzoic Acids. An Example of the Electrophilic Nature of the OH Radical BY P. O'NEILL, D. SCHULTE-FROHLINDE AND S. STEENKEN Institut fiir Strahlenchemie im Max-Planck-Institut fur Kohlenforschung, Stiftstrasse 34-36, D-4330 Mulheim a.d. Ruhr, Germany Receiued 3rd December, 1976 Using spectrophotoinetric and conductometric pulse radiolysis, in situ e.s.r. and product analysis techniques, the reaction of OH radicals with mono-, di- and tri-niethoxylated benzenes and benzoic acids has been further studied in aqueous solution. It is found that, in addition to radical cations or zwitterions, phenoxyl radicals and methanol are formed. With the methoxylated benzenes the yield of phenoxyl radical is the same as the yield of methanol, and the sum of the yields of radical cation and of phenoxyl radical (or methanol) is N 1000/, of G(0H).The ratio of the yield of radical cation to the yield of phenoxyl radical depends on the number and on the positions of the methoxyl groups relative to each other. The yields of phenoxy) radical are low when the methoxyl groups are meta to one another, whereas the yields are high for substrates with methoxyl groups in an ortho or para relation. A similar dependence on substrate structure of the distribution of products, i.e., radical zwitterions, phenoxyl radicals and methanol, is found for methoxylated benzoic acids. The results are quantitatively interpreted in terms of electrophilic addition of OH to the aromatic ring, the positions of attachment being governed by the ortho-para directing effect of the electron- donating methoxyl groups.H-abstraction from the methyl groups is negligible. The radical cations or zwitterions are derived from OH adducts formed by addition of OH to ring positions not occupied by methoxyl groups, whereas the phenoxyl radicals are produced by elimination of methanol (k - 4 x lo4 s-' at pH 7) from OH adducts formed by attachment of OH to ring carbons carrying methoxyl groups. Recently we reported1 that the yields of radical cation in the reaction of OH at pH < 7 with a series of isomeric di- and tri-methoxybenzenes depend on the positions of the methoxyl groups relative to each other. With 1,3-dimethoxy- and 1,3,5-tri- methoxy-benzene the radical cation yields are >90% of G(0H) as compared with -65% for 1,2- and 1,4-dimethoxygenzene and -50% for 1,2,3- and 1,2,4-trimethoxy- benzene.Completely analogous observations were made on radical zwitterion yields from di- and tri-methoxylated benzoic acids. The reason for the differences in the yields was not clear. We have now found that, in addition to radical cations or zwitterions, methanol and phenoxyl radicals are formed, the yields of which have been measured. Complete material balance is obtained insofar as the fate of the OH radical is concerned. On this basis, an explanation is now presented for the reaction of OH in terms of (a) preferential addition of OH to those ring positions which are activated by the ortho-para-directing methoxyl groups, (b) formation of radical cations or zwitterions from those OH adducts where OH is attached to ring carbons which do not carry a methoxyl group and (c) elimination of methanol to yield phenoxyl radicals from those OH adducts where OH is attached to ring cargons substituted by a methoxyl group. An analogous reaction scheme is proposed3 for the reaction of OH142 FORMATION OF RADICAL CATIONS AND ZWITTERIONS with methoxylated phenols.Oxidative replacement of the methoxyl group by OH is a new example for a reaction4 which involves loss of a substituent X by a 1,l- elimination mechanism to yield -0. and HX. EXPERIMENTAL N20 saturated aqueous solutions containing 5-10 mmol dm-3 substrate were irradiated with 6oCo y-rays using doses from 2.0 x 1 0 I 8 to 1.5 x 10'' eV cmw3 at a dose rate of 8.1 x lo1' eV C M - ~ min-'.10 mm3 samples of the irradiated solutions were analysed for methanol by gas chromatography using the double column technique with back-flushing5 (pre-column: length 1.5 m, packed with 10% Marlophen on Teflon; main column: length 3 m, packed with P 4000; both columns were operated at 85OC). The detection limit corresponded to -0.05 mmol dm-3 methanol. The 3 MeV van de Graaff accelerator and the optical and conductivity6 detection systems have been described. Conductivity measurements were limited to 3.5 < pH < 11 for methoxylated benzenes and to 6 < pH < 8 for methoxylated benzoic acids. Solutions were irradiated at 20 f 2" C with electron pulses of 1 ps duration.The pH of the solutions was adjusted using HC104 or NaOH. Dosimetry was performed using either mol dm-3 FeS04 in 0.8 mol dm-3 H2SO4 or C(NO2)+ The in situ radiolysis e.s.r. experiments were carried out using the method described by Ei ben and Fessenden. The substrates were obtained from Fluka or Aldrich. They were of the highest purity available and were used as received. G(0H) was taken to be 5.5. RESULTS AND DISCUSSION (i) PRODUCTION OF RADICAL CATIONS AND RADICAL ZWITTERIONS The formation of radical cations from OH adducts of methoxylated benzenes has previously been rep0rted.l The radical cations were identified by e.s.r. and their yields determined using optical and conductometric pulse radio1ysis.l On the basis of the complete material balance for reaction of OH with the substrates obtained in this study (see later) it is now possible to assign as precursors of the radical cations those OH adducts, which are formed by addition of OH to ring positions not occupied by methoxyl groups [reaction scheme (l), step A].Protonation of these OH adducts (step B) and elimination of water (step C) leads to radical cations. In the case of methoxylated benzoic acids radical zwitterions such as OCH, are produced2 by an analogous mechanism. The yields of radical cation formed at -pH 1, expressed as a percentage of G(OH), are presented in table 1. It is seen that the yields1 are high for anisole and for meta substituted inethoxylated benzenes but lower for ortho and para substituted benzenes. A similar dependence of yields2 on substrate structure is observed with respect to zwitterion yields from methoxylated benzoic acids (table 2).TABLE l.-PRODUCT YIELDS FROM REACTION OF OH WITH METHOXYBENZENES IN AQUEOUS SOLUTION 1 2 3 4 5 6 7 % radical cation G(radica1 G(phenoxy1 rad.)/G(OH) columns 1 and 2, 3 or 4 % phenoxyl rad.C yields from substrate cation)/G(OH) G(CH,OH)/G(OH) by con- ductivity by e.s.r. 1 + 2; 1 + 3; 1 + 4 calculated e calculated gcH3 OCH j I OCH, I CH,O B O C H , t90 60 93 64 51 53 91 4 28 13 36 51 50 16 <6 3 <94; 29 31 88 ; <6 5 106; 31 34 100; 39 47 102; 41 45 103; t 9 7 107; t 9 6 ; 89 ; <99; 95 ; 90; 94 ; < 100; 93 1oo:o 97:3 91 67: 33 67:33 98 1oo:o 94: 6 98 67:33 67: 33 98 50: 50 50: 50 98 50: 50 50: 50 98 1oo:o 91:9 ~~~ ~~~~~~ Expressed as a percentage of G(0H); the error in the yields is estimated to beflO%; “determined at pH N 1 [see ref.(l)]; calculated on the basis of a ratio of 10:1 for determined at pH 3 ; calculated on the basis of 100% selectivity of OH (see text); addition of OH to activated as compared with non-activated positions.144 FORMATION OF RADICAL CATIONS AND ZWITTERIONS TABLE 2.-PRODUCT YIELDS "' FROM REACTION OF OH WITH METHOXYBENZOIC ACIDS IN AQUEOUS SOLUTION substrate 2-MBA 3-MBA 4-MBA 2,3-DMBAf 3,4-DMBA 2,4-DMBA 2,6-DMBA 3,5-DMBA 2,3,4-TMBA' 3,4,5-TMBA 2,4,5-TMBA 2,4,6-TMBA G(radica1 zwitterion)/G(OH) (@ (h) (h) (h) (h) 90 91 92 -40 53 46 92 G(CH,OH)/G(OH) 5 10 4 26 19 11 15 12 24 36 47 -25 a Expressed as a percentage of G(0H); the error in the yields is estimated to be &lo%; MBA = methoxybenzoic acid; not measurable due to overlap of radical zwitterion spectra with those of H adducts and phenoxyl type radicals; methanol yield possibly increased due to reaction' of radical zwitterion with H,O followed by demethoxylation ; J dependence of methanol concentration on dose is not linear.determined at pH 1 [see ref. (2)]; DMBA = dimethoxybenzoic acid; determined at pH 3-7; TMBA = trimethoxybenzoic acid; (ii) PRODUCTION OF METHANOL On 6oCo y-irradiation of N20 saturated 5-10 mmol dm-3 substrate solutions in the pH range 3-7, methanol is formed as a reaction product. The concentration of methanol formed is linearly dependent on dose in the range 2.0 x 1018-(7-15) x loz8 eV ~ m - ~ , which corresponds to a conversion of substrate <25%.The yields of methanol determined from plots of methanol concentration against dose are presented in table 1. It is evident that the methanol yields depend on the positions of the methoxyl groups relative to each other and on their absolute numbers. Among substrates with the same number of methoxyl substituents, those possessing methoxyl groups in an ortho or para relation to each other give appreciably higher yields than substrates with methoxyl groups meta to each other. With respect to G(OH), the methanol yields and the radical cation yields are complementary to each other, i.e., the yields add up to -100%. With the methoxylated benzoic acids the same trends are observed: the substrates with methoxyl groups in an ortho or para relation to each other yield more methanol than those which carry methoxyl groups meta to one another (table 2).On the whole, the methoxybenzoic acids tend to yield somewhat less methanol than the methoxybenzenes. This effect is particularly pronounced with 3,4-dimet hoxy- and 2,3,4-trimet hoxy-benzoic acid. The formation of methanol is proposed to proceed via addition of OH to ring positions substituted by methoxyl groups (@so reaction step A'). The OH adducts thus formed (type 11) are suggested to decay by acid catalysed (steps D-E), base catalysed (steps G-H) or uncatalysed (step F) elimination of methanol. The methanol formation is of the 1,l-elimination type which has previously been described4 for substituents such as F, C1, Br and NO2, which are good leaving groups. The acid catalysed decomposition of OH adduct 11, steps D-E, is proposed for reasons of analogy with corresponding OH adducts from methoxylated phenols, and because I1 is of the semi-acetal type which should be acid labile.The demethoxylation reaction also explains the formation' of phenol from reaction of OH with anisole. A yield ofP . O'NEILL, D. SCHULTE-FROHLINDE AND s. STEENKEN 145 Q + O H - OCH, products " K - . v phenol equal to 5% of G(0H) was ~btained,~ which, within experimental error, is in good agreement with the 4% yield of methanol found in this study (table 1). This result renders unnecessary the hypothesis lo that the phenol is produced from a radical formed by H-abstraction from the methyl group. (iii) IDENTIFICATION AND YIELDS OF PHENOXYL RADICALS A. OPTICAL DETERMINATIONS The phenoxyl radicals expected by steps A' and D-H are difficult to identify directly using optical pulse radiolysis since many of them absorb in a region (-400 nm) where there is considerable contribution from OH adducts and, at pH < 7, from radical cations1 or, in the case of methoxybenzoic acids as substrates, from radical zwitteriom2 However, with 1,4-dimethoxy- and 1,2,4-trimethoxy-benzene and 2,4,5-trimethoxybenzoic acid absorptions are observed l e 2 at 420 nm which are identified as due to phenoxyl radicals.This assignment is confirmed by the observa- tion that the rate of decay of the intermediates absorbing at 420 nm is appreciably enhanced in the presence of ascorbate, which has been shown to react specifically with phenoxyl radicals l1 [see reaction (2)].Under these conditions absorptions at 360 nm, characteristic of the ascorbate radica1,ll grow in as the absorptions at 420 nm decay. However, due to contributions from OH and H adducts of the substrates to the absorption of ascorbate radical at 360 nm it is not possible to determine G-values with satisfactory accuracy when applying this method. Using 0.5 mmol dm-3 solutions saturated with N20, the pH dependence of the yields of phenoxyl radical from the substrates mentioned was measured in the pH range 7-1 1 by observing the absorption at 420 nm. In order to be able to correct3 for reduction of the yields due to competition between steps J and those yielding146 FORMATION OF RADICAL CATIONS AND ZWITTERIONS phenoxyl radical, a dose rate variation (0.2-1 krad/pulse) was performed at each pH value.The yields obtained by extrapolating to zero dose/pulse were found to be independent of pH between pH 7 and 11. In the case of 1,4-dimethoxybenzene it was possible to estimate from the absorption at 420 nm the G-value of phenoxyl formation since the extinction coefficient of 4-methoxyphenoxyl radical can be obtained using oxidation of 4-methoxyphenol with Tl(I1). On the basis12 of E (4-methoxyphenoxyl),,, nm = 6360 dm3 mol-1 cm-l, the yield of phenoxyl radical from 1,4-dimethoxybenzene is (25 -J= 4)% of G(OH), in reasonable agreement with the yield of methanol found. In the case of 1,3-dimethoxy- and 1,3,5-trimethoxy-benzene the 3-methoxy- and 3,5-dimethoxy-phenoxyl radicals, respectively, are expected as transients from the demethox] lation reaction.These radicals absorb at 430 and 510 nm, respectively.12 From the absence of absorptions at these wavelengths on reaction of OH with 1,3-dimethoxy- and 1,3,5-trimethoxy-benzene it is concluded that the yields of phenoxyl radical are (10% of G(OH), in agreement with the methanol yields. The rate of formation of the phenoxyl radicals, as determined from their absorption at 420 nm, follows first order kinetics and increases with increasing pH. The rate constant increases from -5 x lo4 s-l at pH 7 to >lo6 s-l at pH 11; this indicates that OH adducts of type I1 undergo base catalysed demethoxylation (reaction 1, steps G-H). B. CONDUCTIVITY AND E.S.R. MEASUREMENTS If phenoxyl radicals react with ascorbate at pH values greater than the pK, values of the corresponding phenols (pK,'s - lo), a decrease in conductivity results lla since the proton formed is removed by the excess OH- present.0 0 + $Oo+ H@ ( 2 ) CHOHCH20H The yields of phenoxyl radical may, therefore, be determined by conductivity provided the reaction is sufficiently fast that the phenoxyl radicals are quantitatively scavenged. As seen from table 3, the rate constants for reaction (2) are of the order of 3 x lo8 TABLE 3.-RATE CONSTANTS a FOR REACTION OF METHOXYLATED PHENOXYL RADICALS WITH ASCORBATE substrate 2-MP 3.3 x lo8 3-MP 4.2 x lo8 4-MP 2.2 x lo8 2,3-DMPC 1.8 x 109 2,6-DMP 1.1 x lo8 3,5-DMP 7.7 x lo8 k(phenoxy1 radical + ascorbate)/dm3 mol-1 s-1 a Determined at pH 10.8-11.0 and 20 f 2°C; MP = methoxyphenol; DMP = dimethoxyphenol.dm3 mol-1 s-l, a value similar to those for otherllb phenoxyl radicals and large enough for trapping phenoxyl radical almost quantitatively if >O. 1 mmol dm-3 ascorbate is used.P . O’NEILL, D. SCHULTE-FROHLINDE AND s. STEENKEN 147 On pulse irradiation of 5-10 mmol dm-3 solutions of methoxylated benzene solu- tions saturated with N,O and containing 0.5 mmol dm--3 ascorbate at pH 11, a decrease of conductivity was observed -50 ps after the pulse (-200 rad) due to neutralisation of H+ produced in reaction (2). Under these conditions, no conducti- vity changes occur in the absence of ascorbate. Using the mobility values p(OH-) = 18.4 x lop4 V-l cm2 s-l and p(pheno1ate) = 3.62 x V-l cm2 s-l and assuming p(ascorbate) = p(ascorbate radical), G(phenoxy1) was calculated from the decrease in conductivity.The yields obtained (table 1, column 3) are slightly lower than the yields of methanol. The formation of ascorbate radical on reaction of OH with methoxylated benzenes and benzoic acids in the presence of ascorbate can also be shown to occur using the in situ radiolysis e.s.r. technique. Experiments were performed by irradiating N20 saturated 5 mmol dm-3 substrate solutions containing 0.5 mmol dm-3 ascorbate at pH 11. In all cases the characteristic e.s.r. s p e c t r ~ m ~ ~ * ~ ~ of the ascorbate radical was observed. Substrates with methoxyl groups in an ortho or para relation to each other always yielded appreciably higher stationary concentrations of ascorbate radical than those with methoxyl groups meta to one another.The yields listed in table 1, column 4, were obtained by dividing the amplitudes of ascorbate radical observed in the presence of the substrates by the amplitude measured on irradiation of a 5 mmol dm” phenol solution containing 0.5 mmol dm-3 ascorbate, i.e., by assuming that in the reaction of OH with phenol at pH 11 G(phenoxy1) = G(0H). The yields obtained are in good agreement with those based on the conductivity method and those of methanol. On in situ irradiation of solutions of methoxybenzoic acids containing ascorbate as scavenger for phenoxyl radicals the spectrum of the ascorbate radical was observed. However, the yields of phenoxyl radical, as determined from the stationary con- centration of ascorbate radical, tended to be lower than the yields of methanol from the same substrates. This is thought to be due to incomplete scavenging of the carboxylated phenoxyl radicals by the ascorbate anion.Rate constants for this scavenging reaction have been shownllb to depend on the position of the radical site with respect to the carboxyl group. Determination of exact yields using the ascorbate method is, therefore, not possible without additional information on rate data. Qualitatively, however, the same dependence of phenoxyl yields on structure is found as in the case of the methoxybenzenes. (iv) DISTRIBUTION OF OH ADDITION AND OF YIELDS OF PRODUCTS AS A FUNCTION OF THE NUMBER AND THE POSITIONS OF METHOXYL GROUPS The differences in the yields of methanol or of radical cation/zwitterion from the individual substrates are interpreted to be the result of the electrophilic nature15 of the OH radical.Due to its electrophilicity, OH attaches preferentially to those ring positions which are activated by the ortho-para directing methoxyl groups. With substrates characterized by methoxyl groups meta to each other none of the activated positions is substituted by a methoxyl group, whereas with substrates carrying methoxyl groups in ortho or para relation all ring positions are activated including those occupied by methoxyl groups. Assuming that the products are formed accord- ing to reaction scheme (1) and that, to a first approximation, OH adds only to activated positions, the product distribution may be calculated. On this basis it is predicted that the yield of methanol or phenoxyl radical from anisole, 1,3-dimethoxy- and 1,3,5-trimethoxy-benzene should be zero, since with these substrates the ring carbons carrying methoxyl groups are not activated.For 1,2- and 1,4-dimethoxybenzene or148 FORMATION OF RADICAL CATIONS AND ZWITTERIONS 1,2,3- and 1,2,4-trimethoxybenzene the yields of demethoxylation predicted amount to 2/6 or 3/6 of G(OH), since all ring positions are activated. It should be noted, however, that with these compounds, which all contain methoxyl groups in an ortho or para relation, the same values (2/6 or 3/6) are obtained if it is assumed that OH reacts unselectively. Although the simple picture described leads to a good repre- sentation of the data (compare columns 3 and 4 with 6, table l), it is obvious that the assumption of complete selectivity of OH addition is oversimplified. In order to obtain a better estimate of the selectivity, the following procedure is used : a variable r is introduced which is defined as the ratio of the rate of attachment of OH at activated positions to the rate of OH attachment at non-activated positions.It is thereby assumed that (a) the rates of addition of OH to singly and multiply activated positions are the same and (p), the presence of a methoxyl group at a particular ring position has no effect on the rate of OH attachment to that position. Under these conditions the best simulation of the product distribution is obtained for r = 10 : 1 (see column 7, table 1). If condition (a) is not fulfilled, i.e., if the rate of OH addition to multiply activated positions is larger than that to singly activated positions, the situation becomes more complicated.In order to obtain a value for r, it is then necessary to specify quanti- tatively the increase in the rates of OH addition to multiply as compared with singly activated positions. This results in a corresponding decrease for r. Calculations based on this procedure, however, do not result in an improved representation of the experimental product distribution from which it may be concluded that assumption (a) is not unreasonable. The results obtained with the methoxybenzoic acids (table 2) are interpreted in a way similar to that described for methoxybenzenes. In summary, it is shown that the addition of the OH radical to the substrates is governed by the electrophilic nature of OH and by the ortho-para directing methoxyl groups. This finding is similar to those obtained with other aromatics containing electron-donating substit~ents.~~~*’~ In contrast, the addition of OH to ring positions of aromatics substituted by electron withdrawing groups has been ~ h o w n ~ ~ * ~ * to be rather unselective. l P. O’Neill, S. Steenken and D. Schulte-Frohlinde, J. Phys. Chem., 1975, 79, 2773. P. O’Neill, S. Steenken and D. Schulte-Frohlinde, J. Phys. Chem., 1977, 81, 31. S. Steenken and P. O’Neill, J . Phys. Chem., 1977, 81, 505. For a review see P. Neta, Adv. Phys. Org. Chem., 1976, 12, 223. D. R. Deans, Chromatographia, 1968, 1, 18. H. G. Klever, Ph.D. Thesis (Ruhr-Universitat Bochum, 1974). K. Eiben and R. W. Fessenden, J. Phys. Chem., 1971,75, 1186. R. B. Moodie and K. Schofield, Accounts Chem. Res., 1976,9,287. J. H. Fendler and G. L. Gasowski, J. Org. Chem., 1968,33, 2755. lo J. Holcman and K. Sehested, J . Phys. Chern., 1976, 80, 1642. l1 R. H. Schuler, (a) Radiation Res., 1975, 62, 539; (b) 1977, in press. l2 P. O’Neill and S. Steenken, to be submitted for publication. l3 G. P. Laroff, R. W. Fessenden and R. H. Schuler, J. Amer. Chem. SOC., 1972,94, 9062. l4 S. Steenken and G. Olbrich, Photochem. Photobiol., 1973, 18,43. l5 M. Anbar, D. Meyerstein and P. Neta, J. Phys. Chem., 1966, 70, 2660. l6 M. K. Eberhardt and M. I. Martinez, J . Phys. Chem., 1975,79, 1917. l7 G. W. Klein, K. Bhatia, V. Madhavan and R. H. Schuler, J. Phys. Chem., 1975,79, 1767. M. K. Eberhardt, J. Phys. Chem., 1975, 79, 1913.

ISSN:0301-7249    

DOI:10.1039/DC9776300141

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Formation of Carbonium Ions, Carbanions and Carbanion-Pairs in Irradiated Solutions, and the Reactivity of these Intermediates BY LEON M. DORFMAN, YING WANG, HSIEN-YIEN WANG AND RICHARD J. SUJDAK Department of Chemistry, The Ohio State University, 140 West 18th Avenue, Columbus, Ohio 43210, U.S.A. Received 24th November, 1976 Pulse radiolysis studies of a variety of compounds containing the phenyl group, in solution in 1,2-dichloroethane, have led to the observation of the arylcarbenium ions, PhCH?, Ph2CH+ and Ph3C+. Rate constants for electrophilic reactions of these carbocations with a variety of nucleophiles have been determined. The trends in the rate constants reflect the steric and electronic effects. The observed reactivities provide some guidance regarding the role of these intermediates in radiation chemical systems and in organic mechanisms more generally.In a recent series of investigations, the results of some of which have already been rep~rted,l-~ we have shown, by direct observation using the pulse radiolysis method 6*7 with selected solute-solvent systems, that carbocations and carbanions are formed in irradiated systems. We have determined the absolute reactivity, in many reactions, of these intermediates which play such a broad, pervasive role ' v 9 in organic chemistry. Our attention, to date, has focused on aromatic carbanions and carbocations, namely the benzyl carbanion, PhCH,, and a series of arylcarbenium ions: the benzyl cation, PhCH:, the benzhydryl cation, PhzCHf, and the trityl cation, Ph3C+. The results reported will be limited to the presentation of as yet unpublished data on the formation and reactivity of these arylcarbenium ions.FORMATION AND IDENTIFICATION OF CARBANIONS The formation of the carbanions occurs through the action of the solvated electron, as primary reducing species in liquids such as tetrahydrofuran, where a dissociative attachment to an appropriate solute, such as dibenzyl-mercury, will occur : e,- + (PhCH2)2Hg --+ PhCH? + PhCH2Hg. (1) If an alkali metal cation is present, with which the solvated electron may be paired,lO*l' the analogous reaction of this ion-paired species will occur: Na+, e; + (PhCH2),Hg -+ PhCH,, Na+ + PhCH2Hg (2) to form the carbanion in a cation-paired state. Alternatively, depending on the concentrations of the various reactants, the ion-paired species may be formed by combination of the free carbanion with the metal cation: PhCH,' + Na+ 3 PhCH;, Na+.(3) Since the optical absorption bands for benzylsodium l2 and for benzyllithium l3 are150 FORMATION OF CARBONIUM IONS, CARBANIONS AND CARBANION-FAIRS known, the identity of the free benzyl carbanion, PhCH;, was established from the fact that the known cation-paired species were formed from it by reaction with the alkali metal cations. FORMATION AND IDENTIFICATION OF CARBOCATIONS The carbocations, on the other hand, are formed by the action of the solvent cation, as primary oxidizing species in liquids such as 1 ,Zdichloroethane, in promoting a dissociative ionization of appropriate compounds such as dibenzylmercury, benzyl bromide, dibenzyl sulphide and various other compounds : RCl? + PhCH,Br -+ PhCH: + Br + RCl. (4) The electron, in the halogenated hydrocarbon, will be localized on a chloride ion formed in a dissociative attachment to the solvent.Accordingly, no carbanion is formed in dichloroethane. This mode of formation of the arylcarbenium ions is quite similar to our earlier work on aromatic cation radical^,^^*^^ with the important distinction that formation of the radical ions involved a non-dissociative ionization. The arylcarbenium ions were identified in two cases from their known spectra. These two are the benzhydryl16 cation and the trityl cation.17 PhCHz, which had not been observed previously, was identified not only from its analogous mode of formation, but also from the fact that a common optical absorption band3 was formed from a variety of benzyl compounds, Since the natural lifetime of the arylcarbenium ions under our conditions was in the order of several ps, absolute rate constants for a variety of their electrophilic reactions could be readily determined.This was done by observing the decay of the carbocations, in the presence of a sufficient concentration of the nucleophilic reactant to make the observed half-time on the order of tens of nanoseconds. EXPERIMENTAL The source of the electron pulse was a Varian V-7715A electron linear accelerator delivering 3-4 MeV electrons in pulses ranging from 20 to 1500 ns duration. Pulse current was about 300 mA for pulse duration of 100 to 1500 ns and about 600 mA for pulse duration of 80 ns or less.Electron pulses ranging from 40 to 200 ns were used in this work. Transient absorption spectra were determined using an RCA 1P28 or RCA 7200 photomultiplier as detector. The time resolution of the detection system was about 5 ns with the 1P28, so that it was feasible to observe the kinetics on a submicrosecond time scale. A Bausch and Lomb grating monochromator, type 33-86-25, j73.5 was used. Appropriate Corning filters were used to eliminate second-order components from the analysing light beam. Our standard 20-mm reaction cells, equipped with high-purity silica windows, were used in all experiments with, for the most part, a double pass of the analysing light beam. All the data were accumulated at 24 & 1°C. The solvent used exclusively was 1 ,Zdichloroethane (1,2-DCE), reagent grade, from Matheson, Coleman and Bell.It was purified by a scheme detailed e1~ewhere.l~ Just prior to each experiment, the desired quantity of 1,ZDCE was distilled in vacuo into the reaction cells from a storage bulb, and the amount distilled was determined by weight difference. In one case it was dried over PzOs and fractionally distilled. A variety of compounds was used as carbonium ion precursors: dibenzylmercury from Alfa Inorganics ; bromodiphenylmethane, technical grade, from Chemical Samples Co. ; chlorodiphenylmethane, technical grade, from Chemical Samples Co. ; triphenylmethanol from Aldrich Chemical; and triphenyImethy1 bromide from J. T. Baker and Co. Di- benzylmercury was recrystallized from absolute ethanol and stored in the dark in uacuu until used.Chlorodiphenylmethane was purified by a series of fractional freezing cycles. It wasL . M . DORFMAN, Y . WANG, H-Y. WANG AND R . J . SUJDAK 151 dried with barium oxide and stored in a closed vessel until used. Bromodiphenylmethane was purified by vacuum sublimation. Triphenylmethanol was recrystallized from absolute ethanol, the filtrate being treated with activated charcoal. Triphenylmethyl bromide was recrystallized from DCE. The following compounds were used as reactants : ammonia from Matheson; water, triply distilled from our standard still ; triethylphosphine and triethylarsine, obtained from Orgmet, Inc., were degassed from storage in a nitrogen atmosphere and used without further purification.Ammonia solutions in DCE were made up from amounts of ammonia determined by pressure measurement in the gas phase, the solution concentration being determined as follows. Since the solubility of ammonia in 1,2-DCE at room temperature is 8 mole%, and a cell was used in which the liquid volume amounted to 80% of the total, we calculate, assuming that Henry’s Law holds, that only 1 % of the measured amount of NH3 is in the gas phase. Ammonia concentrations ranging from to rnol dmb3 were used. The cation rate curves were observed at the optical absorption maxima of the respective cations, 363 nm for benzyl, 445 nm for benzhydryl, and either 415 or 439 for trityl. RESULTS ELECTROPHILIC REACTIONS OF ARYLCARBENIUM IONS In a recent p~blication,~ we reported the values of the rate constants for the electrophilic reactions of the three arylcarbenium ions with three tertiary alkyl amines, the simplest of which was triethylamine.The reaction observed is presumably the formation of the quaternary ammonium ion. The values, ranging from 2.0 x lo9 to 7.0 x lo6 dm3 mo1-l s-l, showed a decreasing trend with increasing substitution of the arylcarbenium ion, and a slight decrease with increasing size of the alkyl group in the tertiary amines. It was suggested4 that the observed trends simply reflected the combined steric and electronic effects. In the present investigation we were interested in the effect of different central atoms in the nucleophile on the specific rate. The rate constants were thus deter- mined for the reaction of benzyl cation with triethylphosphine and with triethylarsine, both of which could be compared with our earlier value4 for triethylamine.The source of benzyl cation in these runs was dibenzylmercury (3 x lom4 to 7 x mol dm-3). The concentrations of the reactants ranged from 1 x to 8 x mol dm-3 for triethylphosphine, and from 4 x to 2 x mol drn-3 for tri- ethylarsine. In all the runs, the rate curves were found to fit a first-order rate law by plotting In D, against time. The pseudo-first order rate constants thus obtained were plotted against reactant concentration to give the straight lines shown in fig. 1, from the slopes of which the values of the bimolecular rate constants were obtained. The values, which have an uncertainty of about -&15%, are shown in table 1, together with the earlier value for triethylamine.The rate constants for the electrophilic reaction with ammonia, presumably to form the quaternary ammonium ion, were determined for all three arylcarbenium ions. Benzyl cation was formed from dibenzylmercury, benzhydryl cation from bromodiphenylmethane and trityl cation from triphenylmethanol. The concentra- tion of ammonia was varied in the range lov4 to mol dm-3. The cation decay curves, in all three cases, were found to fit a first-order rate law. Plots of the pseudo- first order constant against ammonia concentration, shown in fig. 2 for benzyl and benzhydryl, gave straight lines. The rate constants, obtained from the slopes, are presented in table 1. the first-order rate constants showed a non-linear dependence on alcohol concentration which was interpreted by The uncertainty is about &15y0.With methanol and ethanol in the earlier152 FORMATION OF CARBONIUM IONS, CARBANIONS AND CARBANION-PAIRS [ ( E t I3P ] I 1 d'rn o I d m-3 W (D C a 0 I 5.0 t 2 4.0 (0 7 3.0 CD 0 2.0 2 2 3 1.0 : Om m, 0 - [ ( E ~ ) ~ A s ] /10-3mo~ dm-3 FIG. 1 .-Pseudo-first order rate constant for the reaction of benzyl cation, in 1 ,2-DCE at 24C, with triethylphosphine (0) against concentration of triethylphosphine, and with triethylarsine (0) against concentration of triethylarsine. .- -LT ammonia concentration/10-4moI FIG. 2.-Pseudo-first order rate constant for the reaction of arylcarbenium ion (0, -- for benzyl cation; 0,- for benzhydryl cation) with ammonia in 1,ZDCE solution at 25"C, plotted against ammonia concentration.TABLE 1.-RATE CONSTANTS FOR THE REACTIONS OF ARYLCARBENIUM IONS WITH NUCLEOPHILES IN 1,2-DCE AT 25°C. (dm3 m0l-l S-'). PhCH? 8.1 x 109 2.2 x 109 2.0 x 109 4.2 x 109 1.8 x 1 0 7 6 x 10' arylcarbenium ion PhzCH+ PhSC+ - - - - 4.3 x 109 2.4 x 107 1.3 x lo6 - 1 x lo8 - 9.1 x 1Olo 8.0 x 1OloL. M . DORFMAN, Y . WANG, H - Y . WANG AND R . J . SUJDAK 153 involving the monomer-dimer equilibrium of the alcohol. The rate constant was higher for reaction with the dimer than with the monomer. A similar phenomenon has now been observed in the reaction of benzyl cation and of benzhydryl cation with water. Benzyl cation in these runs was formed from dibenzylmercury ; benzhydryl cation was from bromodiphenylmethane. The concentration of water was varied from 1 x to 6 x mol dm-3.Benzyl cation decay curves were found to fit a first order rate law. Benzhydryl decay curves were of mixed order, and the first order constant in this case was obtained by curve fitting on an analogue computer. The curves obtained by plotting these rate constants against water concentration, were found to be non-linear, as shown in fig. 3, for the benzyl cation reaction. The i4t E 0 0 1:lo.20 40 60 water concentration /lO%ol dm-3 FIG. 3.-Kinetic plots for the reaction of benzyl cation with water in 1,ZDCE at 25°C. The right- hand ordinate (0,-) pertains to a plot of the pseudo-first order rate constant, k’, against con- centration of water. The left-hand ordinate (I,--) refers to a plot of k’/[H,O] against con- centration of water.curve could be linearized on the basis of the following assumption. exists in a dimer-monomer equilibrium in solution : If the water 2H20 3 (H20), (5) with kd > k,, where kd and k, are the rate constants for reaction with the dimer and monomer respectively, we may write the differential rate expression as : where the quantity in parentheses represents the observed first order constant. This involves the additional assumption that the dimer concentration is sufficiently low so that the monomer concentration may be represented by the concentration of water. In this case, a plot of the first order constant divided by the water concentration, against the water concentration, should give a straight line with intercept k, and slope kd& Such a plot is shown in fig.3. The values for k,, with an uncertainty of about &20%, are given in table 1. From a value1* of Ke = 0.54 (see Discussion) we estimate the values for kd, with an uncertainty about &40%, shown in table 1.154 FORMATION OF CARBONIUM IONS, CARBANIONS A N D CARBANION- PAIRS The rate constant for the reaction of benzhydryl cation with chloride ion was determined from the natural decay curve of the cation. This curve was found to fit a second order rate law, which is consistent with the interpretation that the natural decay is the combination of the benzhydryl cation with the chloride ion in solution. Since the molar extinction coefficient of Ph,CH+ is known,lg 3.8 x lo4 dm3 mol-l crn-l, the rate constant was calculated on the basis of the differential rate expression: with the further assumption that this combination is the only fate, or at least the major fate of chloride ion, namely that [Ph,CH+], = [C1-lt.The value obtained, from k = SEZ, where s is the slope of the line in the second order plot and / is the optical path length, is 9.1 x lolo dm3 mo1-1 s-l. In the earlier work,4 rate constants for the reaction with halide ions had been determined under pseudo-first order conditions, namely with halide ion in excess. FORMATION CONSTANTS FOR ARY LCARBENIUM IONS Formation rate curves were observed for all three arylcarbenium ions from various precursor compounds present in solution at appropriate concentrations. All of the observed formation curves exhibited the following characteristics.There were two kinetic regimes, a fast increase in absorption during the pulse (which ranged from 40 to 80 ns), followed by a slower formation. In all cases, the slower formation was found to fit a first order rate law for the precursor concentrations selected. A typical example is shown in fig. 4. While we cannot say that the formation curve after the pulse represents only a single reaction, it is instructive to plot ln(D, - D,) against time for these data. A straight line is obtained in each case, from the slope of which a first order constant was obtained. The bimolecular rate constants in table 2 were obtained from the slopes of the straight lines which resulted from plots of these first order constants against concentration of precursor compound. If the formation does involve only a single process, the constants in the table may be identified with that process.TABLE 2.-FORMATION CONSTANTS OF ARYLCARBENIUM IONS FROM VARIOUS PRECURSOR COMPOUNDS IN 1,Z-DCE AT 24"~. cation precursor formation formed compound constant/dm3 mol-l s-' PhCH; (PhCHJzHg 1.3 x 1O1O PhZCH+ PhzCHBr 1.6 x 1Olo Ph3C+ PhjCCl 4.0 x lo8 The rate constants for this slower portion range from the diffusion controlled limit to lower values. At the time of writing (October, 1976), experiments are underway which, it is hoped, will serve to explain the nature of the fast portion of the rate curve which includes a component due to a solvent species. Ph3C+ Ph3CBr 8.4 x 109 Ph3C+ PhjCOH 5.7 x lo8 DISCUSSION With regard to the reaction of the arylcarbenium ions with water, for which the data have been interpreted in terms of separate reactions with water monomer andt- 2004 t i m e / n s FIG. 4.-Rate curve for the formation of benzhydryl cation, observed at 445 nm, in a solution of bromodiphenylmethane in 1,ZDCE at 24"C, following an 80 ns electron pulse.The sweep time is 200 ns per division. [To face page 155L . M . DORFMAN, Y . WANG, 13-Y. WANG AND R. J . SUJDAK I55 with water dimer, it is important to make the following points. First, there is independent evidence from both infrared spectroscopic measurements 2o in carbon tetrachloride solution, and from activity coefficient determination l8 in 1,2-DCE solution, that water, in these halogenated solvents, exists in a monomer-dimer equilibrium. K, for this equilibrium is 2.21 mol-1 in carbon tetrachloride20 and 0.54 in 1,2-DCE1' at 25°C.This value, with the experimental value for Kekd, leads to a reasonable value for kd, rather lower than the diffusion-controlled limit, over the particular range of water concentration we have used. The value 0.54 gives water dimer rate constants which are 30 to 70 times larger than k,. It is also consistent with the assumption that monomer concentration is approximately equal to the water concentration. As was pointed out in the case of the alcohol^,^ the subsequent deprotonation in this case of dimer, involves detachment of H30+, a solvent stabilized proton, rather than a proton itself. It has also been pointed out21 that molecular orbital calculations on linear water dimers indicate a higher negative charge on the oxygen of the proton donor molecule than exists on the water monomer.The rate constants for the reaction with ammonia may be compared with the earlier data4 for triethyl-, tripropyl- and tributyl-amine, which showed a slightly increasing trend (somewhat less than a factor of two) with decreasing size of the alkyl group. It was suggested that the steric effect in this series dominated slightly over the electron-donating effect of the alkyl group. The rate constant for ammonia, where very little in the way of a steric effect would be anticipated compared with the more bulky alkylamines, is only twice as large with benzyl cation, and three times as large with benzhydryl cation, as it is for triethylamine. And with trityl cation, the value for ammonia is actually lower than for triethylamine.These trends apparently reflect the steric and electronic effects, which here are opposite. Changing the central atom in the nucleophile (phosphine, arsine, amine) does not lead to any simple correlation with steric effects. Since the size of the central atom is in the order N < P < As, one might expect rate constants in the opposite order, from the steric effect of the three ethyl groups. Such is not the case, the phosphine having a larger value. With regard to the formation rate constants in table 2, we have pointed out that these constants may not necessarily be identified with a single solvent cation species. The rate constants observed are a property of the precursor compound which is involved in the charge transfer.Evidence is being sought to establish the validity and nature of the apparent fast process in the carbonium ion formation. Two general conclusions may be stated from our observations. For systems of the type studied, it is clear that carbonium ions are formed in the radiation chemical processes, and that these carbonium ions exhibit reactivities toward various nucleo- philes which require us to include thein as important intermediates in product formation. With regard to the role of these important intermediates in organic mechanisms, the reactivities (for which no quantitative model of any predictive useful- ness exists) seem to show trends of which the direction, for steric and electronic effects, is in accord with conventional thinking.We gratefully acknowledge the support of this work by the United States Energy Research and Development Administration. B. Bockrath and L. M. Dorfman, J. Amer. Chem. SOC., 1974,96, 5708. B. Bockrath and L. M. Dorfman, J. Amer. Chem. SOC., 1975,97, 3307. R. L. Jones and L. M. Dorfman, J. Amer. Chem. Soc., 1974,96, 5715. R. J. Sujdak, R. L. Jones and L. M. Dorfman, J. Amer. Chem. Soc., 1976,98,4875. L. M. Dorfman, R. J. Sujdak and B. Bockrath, Accounts Chem. Res., 1976,9, 352.156 FORMATION OF CARBONIUM IONS, CARBANIONS A N D CARBANION-PAIRS M. S. Matheson and L. M. Dorfman, Pulse Rudiolysis (M.I.T. Press, Cambridge, Mass, 1969). L. M. Dorfman in Techniques of Chemistry, ed. G. G. Hammes (Wiley-Interscience, N.Y., 1974), p. 463. D. J. Cram, Fundamentals of Curbanion Chemistry (Academic Press, N.Y., 1965). G. A. Olah and P. V. R. Scheyer, Curbonium Ions (Interscience, N.Y., 1968). lo B. Bockrath and L. M. Dorfman, J. Phys. Chem., 1973,77, 1002. l1 B. Bockrath and L. M. Dorfman, J. Phys. Chem., 1975,79, 1509. l2 R. Asami, M. Levy and M. Szwarc, J. Chem. SOC., 1962, 361. l3 R. Waack and M. A. Doran, J. Amer. Chem. SOC., 1963, 85, 1651. l4 S. Arai, H. Ueda, R. F. Firestone and L. M. Dorfman, J. Chem. Phys., 1969,50, 1072. l5 N. E. Shank and L. M. Dorfman, J. Chem. Phys., 1970,52,4441. l6 V. Gold and F. L. Tye, J. Chem. SOC., 1952,2172. l7 M. S. Newman and N. C. Deno, J. Amer. Chem. SOC., 1951,73,3644. l8 W. L. Masterton and M. L. Gendrano, J. Phys. Chem., 1966,70,2895. l9 G. A. Olah, C. U. Pittman, Jr., R. Waack and M. Doran, J. Amer. Chem. Soc., 1966,88,1488. 2o L. B. Magnusson, J. Phys. Chem., 1970,74,4221. 21 J. E. Del Bene, J. Chem. Phys., 1971,55,4633.

ISSN:0301-7249    

DOI:10.1039/DC9776300149

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Electron Spin Resonance Studies of Electron Attachment to Fluorocarbons and Related Compounds BY A. HASEGAWA,? M. SHIOTANI~ AND F. WILLIAMS" Department of Chemistry, University of Tennessee, Knoxville, Tennessee 37916, U.S.A. Received 29th November, 1976 Radical anions of three perfluorocycloalkanes and several halogenofluoromethanes have been detected and identified by e.s.r. studies following Y irradiation at 77 K of solid solutions containing up to 5 mol % of the parent co:npound in neopentane or tetramethylsilane. The isotropic e.s.r. spectra of c-C3Fa, c-C4FC and c-C5F10 are photobleached by visible light and show the second-order structure characteristic of 6, 8 and 10 equivalent fluorines, respectively, the total 19F coupling being approximately the same value (1 170 & 20 G) in each case.Identical e.s.r. spectra were generated in photoionization experiments using tetramethyl-p-phenylenediamine, confirming the radical anion identifications. The equivalence of the fluorines indicates that the unpaired electron is delocalized over the entire molecular framework in an orbital of high symmetry. The e.s.r. spectra of the CF3X- radical anions (X = C1, Br, I) were anisotropic and showed clear evidence for axially sym- metric hyperfine interactions with three equivalent fluorines and the unique halogen. On this basis, a matrix diagonalization program was used to calculate the line positions and the best-fit e.s.r. parameters obtained. Confirmation of the CF3X- identifications was achieved by parallel photo- ionization experiments and by studies showing that the decay of the CF3X- spectrum in neopentane above 100 K was accompanied by a growth in the spectrum of the CF3 radical.The spin density distributions calculated from the e.s.r. parameters of these congeneric radical anions suggest that the unpaired electron resides in an al (a*) antibonding orbital which is composed largely of thep orbitals from carbon and the unique halogen which lie along the C3" symmetry axis of the radical anion. Consistent with this proposal, the spin densities in the s andp oribtials of the unique halogen increase along the series C1, Br, I, which is the order expected for the effect of decreasing halogen electronegativity . Electron capture is one of the most subtle and important of the fundamental processes that contribute to the detailed mechanism of radiation effects in liquids and solids. It is now well established that for tetrahedral molecules such as C12S02 and Cl3P0 where the central atom is a second-row element,lP2 electron attachment results in the formation of hypervalent radical anions congeneric with the important class of neutral phosphoranyl (PX,) radical^.^'^ Since these radicals possess one more electron than can be accoinmodated by valence-bond structures based on the Lewis octet rule, considerable attention has been directed to their geometrical and electronic structure^.^-^ The starting point for the present work was the realization that d-orbital participation is not crucial to the bonding description of these hypervalent radicals,6 the experimental spin distribution being in accord with the involvement of only s and p orbitals in the symmetry-allowed combination of atomic orbitals that constitutes the HOMO.Thus it was reasoned that even in the case of saturated molecules where the central atom is derived from first-row elements, stable radical anions might be formed at low temperatures and identified by e.s.r. spectroscopy. Although the logic behind this theoretical prognostication seems indisputable, the t Permanent address: Faculty of Science, Hiroshima University, Hiroshima 730, Japan. 2 Permanent address: Faculty of Engineering, Hokkaido University, Sapporo 060, Japan.158 ELECTRON SPIN RESONANCE STUDIES OF ELECTRON ATTACHMENT pre-existent facts were hardly encouraging. Thus, there is considerable prima facie evidence to suggest that electron attachment to a saturated carbon compound con- taining halogen substituents is followed immediately by dissociation into a neutral carbon-centred radical and a halide ion, the intermediate molecular ion having only a transitory existence." Even in the case where an intermediate stage in the process of dissociative electron attachment at a saturated carbon atom was revealed by the e.s.r.detection of the weakly-bound species CH3* - - -Br- and CH3* - - -I- derived from methyl halides in an acetonitrile matrix," it could justifiably be argued that the shallow minimum in the potential energy curves of these radical-anion pairs or adducts derives principally from the constrictive cage effect of the rigid crystalline lattice.An effect of this type involving the rigid cage conversion of a repulsive electronic state into a " bound " electronic state has been proposed to explain the emission spectrum which is observed following the photodissociation of perfluoroalkyl iodides in rare gas matrices.12 The impetus for the present study of fluorocarbons resulted from the recent discovery of the hypervalent F3NQ- radical anion formed by electron attachment to trifluoramine oxide.13 Since F,NO- is isoelectronic with CF4-, this finding immediately suggested that radical anions of saturated fluorocarbons might also possess sufficient stability. A study of various perfluorocycloalkanes and chloro- fluorocarbons also seemed appropriate because it has been established that these compounds possess extremely large rate constants for thermal electron attachment in the gas phase.l4-I7 Moreover, radiation chemical studies have shown that perfluoro- cycloalkanes compete very effectively against nitrous oxide for the electrons released in the radiolysis of liquid cyclohexane," although the subsequent chemistry appears to be incapable of an unequivocal interpretation." Therefore we hoped, among other things, that our e.s.r.studies could provide an answer to the unsolved problem of dissociative or non-dissociative electron attachment to perfluorocycloalkanes in the condensed phase. EXPERIMENTAL Perfluorocyclobutane and perfluorocyclopentane were obtained from Peninsular Chem Research (PCR). The purity of the perfluorocyclobutane (Freon C 318) was assayed by the manufacturers and found to exceed 99% since no other component was detectable by gas chromatography using three different columns.Perfluorocyclobutane was also supplied by Air Products and Chemicals, Inc. and gave identical e.s.r. results to those obtained with the material from PCR. A sample of perfluorocyclopropane was kindly provided by Prof. R. W. Fessenden from original material suppliedlSb by Dr. K. Hartman. Trichlorofluoro- methane (Freon-1 l), dichlorodifluoromethane (Freon- 12), chlorotrifluoromethane (Freon-1 3), and bromotrifluoromethane (Freon-1 3B1) were obtained from Matheson Gas Products, and the iodotrifluoromethane from PCR, Inc. The other chemicals used in this study were neopentane (2,2-dimethylpropane) from Matheson Gas Products, tetramethylsilane (TMS) from either Mallinckrodt or the Norell Chemical Co.and 2-methyltetrahydrofuran (MTHF) from Eastman Organic Chemicals. N,N,N',N',-Tetramethyl-p-phenylenediamine (TMPD) was supplied by Eastman as the dihydrochloride. The general methods of sample preparation, y irradiation, and e.s.r. measurement have been described in previous publications from this l a b o r a t ~ r y . ' ~ * ~ ~ - ~ ~ In brief, the radicals of interest were generated by y irradiation at 77 K of solid solutions containing up to 5 mol % of the fluorocarbon, the solid matrices being neopentane, TMS and MTHF. The y irradiation dose was usually less than 1 Mrad and the e.s.r. spectra were recorded after irradiation at temperatures up to the disappearance points of the various radicals.In the perfluorocycloalkane systems studied here, the e.s.r. spectra were essentially isotropic in the neopentane and TMS matrices at sufficiently high temperatures. On the other hand, theA . HASEGAWA, M . SHIOTANI AND F. WILLIAMS 159 spectra of the radical anions derived from the trifluoromethyl halides in each of the three solids consisted of anisotropic powder patterns. However, the excellent resolution of the line components in the TMS matrix allowed these spectra to be analysed in detail, as described below. In order to confirm the radical anion identifications, attempts were made to generate these anions by an independent and unambiguous method based on the 320 nm photoionization of TMPD in the various solid The experimental technique has been described in connection with the use of this photo-ionization method to generate trapped electrons in hydrocarbon glasses.25 The samples were irradiated at 77 K with U.V.light (Pyrex and Corning No. 9863 filters) from a 1 kW mercury arc lamp, the e.s.r. signal intensities generally reaching a stationary value after irradiation times of 10 to 20 min. The magnetic field strengths in the electron resonance region were determined using a proton magnetic resonance probe (Walker/Magnetrics Precision NMR Gaussmeter, Model G-502), the oscillator frequencies being measured with a Systrom-Donner counter (Model 1037 with plug-in transfer oscillator ACTO-Model 1255 A) which also served to monitor the X-band microwave frequency. The e.s.r. parameters for the perfluorocycloalkane radical anions were calculated using expressions derived from the second-order solution of the appropriate spin Hamilt~nian.~~J~ For the trifluoromethyl halide and analogous radical anions, the parameters were derived with the aid of a matrix diagonalization computer program written by Dr.C. M. L. Kerr of Southampton University. This program is capable of handling up to three groups of non- equivalent magnetic nuclei with a product of spin multiplicities t50. One of the authors (A.H.) modified the program to operate for the case of axially symmetric hyperfine and g tensors with coincident axes. It was found that suitable trial parameters could be readily derived from the e.s.r. spectrum by using second-order t h e ~ r y . ~ ~ , ~ ’ These trial parameters were then refined until the best fit was obtained between the calculated and experimental line positions.A Varian X-band spectrometer (V-4502-15) was used to record the e.s.r. spectra. RESULTS It should be emphasized at the outset that the successful observation of radical anions in this investigation depended not only on the fundamental chemistry of electron attachment reactions as mentioned in the Introduction, but also on the practical need to obtain e.s.r. spectra with sufficient resolution for definitive radical identifications. The latter requirement was met by the use of neopentane and TMS as matrices. Apparently, these plastically crystalline solids 29 are similar to SFs30 and adamantane3’ in possessing a rotator phase which excites a considerable motion of the trapped radical, especially if the radical is small and highly symmetrical.This is illustrated by recent reports describing the isotropic, or nearly isotropic, e.s.r. spectra of SiF;,32 PFj-,33 c-C,F, 22 and BF,’ 23 in y-irradiated solid solutions of the parent compound in neopentane or TMS. Even in the case of the trifluoromethyl halides where the e.s.r. spectra of the radical anions were found to be anisotropic (see below) in neopentane and TMS, the spectral resolution of the powder pattern in the TMS matrix was greatly superior to that observed in a rigid glass such as MTHF. PERFLUOROCY CLOALKANE RADICAL ANIONS As reported in a preliminary communication,22 the e.s.r. spectrum of a y-irradiated solid solution of c-C,F, in neopentane at 130 K showed second-order structure characteristic of a large hyperfine interaction with eight equivalent 19F ( I = 1/2) nuclei, the isotropic parameters being a@) = 147.7 & 0.6 G* and g = 2.002 1 -j= * G (gauss) = lop4 T (tesla).160 ELECTRON SPIN RESONANCE STUDIES OF ELECTRON ATTACHMENT 0.0017.A better resolved spectrum was subsequently obtained in TMS and is shown in fig. 1. In this intense spectrum, the weak outermost (MI, I = &4,4) lines are observable, the resolution of the four second-order components in each of the MI = + 1 and MI = -1 groups is exemplary, and the five central MI = 0 com- ponents are detectable although two of them ( I = 2 and I = 0) are strongly over- lapped by the strong signals from the * CH,SiMe, (matrix) radical. These details i XI25 (12.5 FIG.1 .-First-derkative e.s.r. spectrum of a y-irradiated solid solution of 5 mol % octafluorocyclo- butane (Freon C 318) in tetramethylsilane recorded at 130 K, v = 9104.3 MHz. The line diagram shows the expected positions of the second-order hyperfine components calculated from the para- meters given in table 1 for the C-C~FC radical anion possessing eight equivalent fluorines. completely verify the analysis in terms of hyperfine interaction with eight equivalent fluorines. While it is evident that the lineshapes of the components in the outer linegroups are distorted by anisotropic broadening, the basic spectrum can be regarded as isotropic, and accurate parameters were derived from the field positions of the sharp components belonging to the three inner linegroups.The assignment of the nine-linegroup spectrum of fig. 1 to the negative ion of octafluorocyclobutane is strongly supported by the generation of the same spectrum by TMPD photoionization, as shown in fig. 2. Although the signal-to-noise level of this spectrum is considerably less than that shown in fig. 1, the identical nature ofA . HASEGAWA, M. SHIOTANI A N D F . WILLIAMS 161 FIG. 2.-First-derivative e.s.r. spectrum of a u.v.-irradiated solid solution containing 1 niol yo TMPD and 5 mol "/I octafluorocyclobutane in tetramethylsilane recorded at 130 K, v = 9097.2 MHz. the two spectra is clearly established by a comparison of the overall patterns and of the well-resolved MI = +1 linegroups. Further evidence for the anionic nature of the radical is provided by the results of photobleaching experiments.Since all the line components assigned to c-C,Fr were removed cleanly by exposure of the sample to unfiltered light from a tungsten lamp, a mechanism of photoelectron detachment seemed likely to be involved.22 This suggestion has recently been confirmed by experiments showing that electrons can be transferred from c-C,Fg to another electron scavenger (SF6 or methyl halides) by photobleaching when one of the latter compounds is also incorporated in the TMS matrix.34 It was of obvious interest to extend this study to other perfluorocycloalkanes. In fig. 3 and 4 are presented the e.s.r. spectra of c-C3F; and c-C5Ffi obtained in experi- ments similar to those described for perfluorocyclobutane, these spectra showing the anticipated second-order structure for hyperfine interaction with six and ten equivalent fluorines, respectively. In each case the spectrum assigned to the radical anion was entirely photobleached by visible light.The e.s.r. spectrum of c-C3F; was also generated by TMPD photoionization in the neopentane matrix, thereby confirming the radical anion assignment. Similar experiments with solid solutions of perfluorocyclohexane in neopentane and TMS did not yield highly resolved e.s.r. spectra. However, the observation of eight broad features which were symmetrically positioned and evenly spaced (115 & 10 G) about the centre of the spectrum could be consistent with a binomial distribution of at least nine unresolved linegroups, but any further attempt at analysis would be speculative at this stage.A similar e.s.r. pattern was produced by TMPD photo- ionization in the corresponding perfluorocyclohexane solutions, suggesting that the unresolved spectrum is indeed that of the radical anion. In a previous e.s.r. study of162 ELECTRON SPIN RESONANCE STUDIES OF ELECTRON ATTACHMENT A x 3 * 4 I I 111 w /(I( FIG. 3.-First-derivative e.s.r. spectra of a y-irradiated solid solution of 5 mol % hexafluorocyclo- propane in neopentane, v = 9121.7 MHz. The upper and lower spectra were recorded before and after exposure to unfiltered light from a tungsten lamp. The line diagram shows the expected positions of the second-order hyperfine components calculated from the parameters in table 1 for the c-C3Fg radical anion possessing six equivalent fluorines.y-irradiated neat perfluoromethylcyclohexane at 77 K,35 several broad features with a total spectral width of 1080 G were tentatively attributed to the perfluoromethylcyclo- hexane negative ion. A summary of the isotropic e.s.r parameters for c-C,F,', c-C4F8, and c-C,Fi is given in table 1. The most interesting feature of these results is the very close agree- TABLE 1 .-ISOTROPIC E.S.R. PARAMETERS" OF PERFLUOROCYCLOALKANE RADICAL ANIONS radical anion matrix T/K g aFlG %/G c - C ~ F ~ neopentane 125 2.003 1 198 1188 c - C ~ F ~ TMS 147 2.0028b 1 97b 1182 c - C ~ F ~ neopentane 130 2.0021 148 1184 C-C~F; TMS 130 2.0024 147 1176 c-CSFG TMS 167 2.0028 115 1150 Error limits are estimated to be f0.0006 for g factors and f 1 G for uF values; determined from the measured parameters gll = 2.0040, gl = 2.002 2 and Ail = 176 G, A , = 208 G for a spectrum showing residual anisotropy.A .HASEGAWA, M . SHIOTANI A N D F . WILLIAMS 163 FIG. 4.-First-derivative e.s.r. spectra of a y-irradiated solid solution of 5 mol yo decafluorocyclo- pentane in tetramethylsilane at 167 K, Y = 9102.8 MHz, recorded before (upper spectrum) and after (lower spectrum) exposure to unfiltered light from a tungsten lamp. The line diagram shows the expected positions of the second-order hyperfine components calculated from the parameters in table 1 for the e.s.r. spectrum of c-C5FG possessing ten equivalent fluorines. ment between the values of n a&), where n is the number of equivalent fluorines, for the three radical anions.This finding has obvious implications with respect to the electronic structure of these radical anions and this aspect will be discussed later in this paper. Even from an empirical point of view, these results are extremely gratifying because they establish a clear relationship in this family of novel radical anions. RADICAL ANIONS OF TRIFLUOROMETHYL HALIDES AND RELATED COMPOUNDS The identification and structural characterization of these hypervalent radical anions was largely achieved through the use of TMS as a matrix for studies of their anisotropic e.s.r. spectra. Invariably the spectral resolution attained with TMS samples was superior to that for the corresponding spectra in neopentane and MTHF solid solutions. In addition, e.s.r. studies of the CF,X- radical anions could be carried out in TMS to higher temperatures than was possible in the other matrices.A third advantage of the TMS matrix for these studies is connected with the intense signal of the matrix radicals which masks the centre of the spectrum; relative to the164 ELECTRON SPIN RESONANCE STUDIES OF ELECTRON ATTACHMENT other matrices (neopentane and MTHF) this spectral region is appreciably narrower in TMS, thereby exposing more structural detail of the radical anion spectra. (a) CF3C1-, CF3Br-, AND CF31- Except for the intense central lines originating from the matrix radical CH,SiMe,, the e.s.r. spectrum shown in fig. 5 is largely attributable to CF3C1-. The weak features in the wings are typical of parallel components whereas the stronger lines inside the H-atom doublet can be assigned to the corresponding perpendicular features of a spectrum originating from a radical with axially symmetric chlorine and fluorine FIG.5.-First-derivative e.s.r. spectrum of a y-irradiated solid solution of 5 mol % chlorotrifluoro- methane (Freon 13) in tetramethylsilane observed at 101 K, v = 9109.1 MHz. The stick plots show the calculated positions of the perpendicular (upper) and parallel (lower) Components for CF3 35Cl- and CF3 37Cl- according to the parameters given in table 2. hyperfine coupling tensors. Hyperfine interaction with one chlorine is clearly indicated by the presence in the wing features of two quartets having relative intensities of 3 : 1, the stronger lines being attributable to 35Cl (I = 3/2) and the inner components to ,'Cl (I = 3/2).Further inspection of the spectrum shows evidence of second-order structure in the inner components of the perpendicular spectrum, as would be expected for interaction with three equivalent 191F ( I = 1/2) nuclei. The equivalence of the three fluorines implies that the CF,Cl- radical possesses C,, symmetry and that the CF, group is rotating rapidly about the symmetric C-Cl axis in the TMS matrix at 101 K. In accordance with the above analysis, the matrix diagonalization program for axially symmetric hyperfine interactions with coincident axes was used to calculate the line positions. As shown by the correspondence between the spectrum and the line diagrams in fig. 5, an excellent fit was obtained for the e.s.r. parameters which are listed in table 2.The excellent agreement between the observed and calculated second-order 19F splittings for the inner Perpendicular components provides a par- ticularly sensitive test of the e.s.r. analysis since these splittings depend on the values of both the parallel and perpendicular I9F hyperfine couplings.A . HASEGAWA, M. SHIOTANI AND F. WILLIAMS 165 TABLE 2.-E.S.R. PARAMETERS AND SPIN DENSITIES FOR CF3X- AND CF2C1; RADICAL ANIONS radical anion T/K gl gll nucleus CFJCl- 101 2.0070 2.002 1 35Cl 19F CF3Br- 121 2.021 2 2.0036 79Br 81Br 9F CF3I- 98 2.048 3 2.000 2 1271 19F CF2C1; 103 2.010 5 2.002 0 35Cl (a) 37c1 35c1 (t) 19F A,IG 17.7 14.7 87.4 105.9 114.2 79.2 178.8 67.2 16.0 16.0 42.2 AldG 43.2 36.0 197.3 244.6 263.6 176.4 373.1 144.6 47.7 26.9 149.5 P S " PPa 0.015 6 0.167 0.015 6 0.167 0.007 2 0.019 5 0.201 0.019 5 0.201 0.006 5 0.033 3 0.285 0.005 4 0.015 9 0.207 0.011 7 0.004 5 a Spin densities were derived using the magnetic parameters listed by B.A. Goodman and J. B. Raynor, A h . Inorg. Chem. Radiochern., 1970, 13, 135; these parameters were calculated from the wave functions given by C. Froese, J . Chern. Phys., 1966,45, 1417. Confirmation of the CF3Cl- identification was achieved by additional experiments. First, as shown in fig. 6, an identical e.s.r. spectrum was generated by TMPD photo- FIG. 6.-First-derivative e.s.r. spectrum of a u.v.-irradiated solid solution containing - 1 mol yo of TMPD and 5 mol % of chlorotrifluoromethane in tetramethylsilane observed at 100 K, v = 9 107.6 MHz.[The sample was irradiated at 77 K for N 20 min with filtered light from a BH-6 mercury arc lamp (Corning 7740 and 9863 filters).] ionization in a CF,Cl/TMS solid solution at -100 K. Secondly, electron capture by the solute is indicated by the fact that the signal from the trapped electron25 was not observed in the e.s.r. spectrum of a y-irradiated MTHF glass containing CF3Cl, the outer portions of the spectrum resembling that in fig. 5 and 6 except for broader linewidths. Finally, as shown by the spectral changes in fig. 7, the e.s.r. spectrum of the CF3 identified by the parameters g = 2.003 1 and 4 3 ) = 144.7 G was observed to grow in during the decay of the CF3C1- spectrum in neopentane above 100 K. These experiments provide independent support from chemical evidence that the CF3C1- radical anion is produced in these systems and, therefore, they confirm the e.s.r.identification of CF3Cl- beyond doubt. A similar e.s.r. study of the y-irradiated CF3Br + TMS solution was carried out at166 ELECTRON SPIN RESONANCE STUDIES OF ELECTRON ATTACHMENT n I b j H I H H I FIG. 7.-First-derivative e.s.r. spectra of a y-irradiated solid solution of 5 mol % chlorotrifluoro- methane in neopentane recorded at 94 K before (a) v = 9109.5 MHz, and after (b) v = 9108.0 MHz, annealing for 10 min at 123 K. The components belonging to the e.s.r. spectrum of the CF3. radical are indicated by the stick diagram. temperatures up to 125 K, and the spectrum obtained at 121 K is shown in fig. 8. The features in the wings of the spectrum have the typical appearance of parallel components and it is noticeable that they occur as distinct pairs of almost equally intense lines.This type of repeating pattern is expected for hyperfine interaction with a single bromine nucleus owing to the presence of the "Br ( I = 3/2; natural abundance = 49.43%) and 79Br (I = 3/2; natural abundance = 50.57%) isotopes with slightly different nuclear magnetic moments (p81/,u79 = 1.078). Starting from the positions of the wing features outside the H-atom lines, the spectrum corresponding to the parallel components is readily analysed in terms of hyperfine coupling to one bromine and three equivalent 19F nuclei, the construction of the stick diagram in fig. 8A . HASEGAWA, M . SHIOTANI A N D F . WILLIAMS 167 showing the *lBr and 79Br quartets of quartets.Although the total spread of the perpendicular features is much less than that of the parallel features, several of the outer perpendicular components are recognised from their characteristic lineshapes and the rest of the pattern is easily constructed. The line positions were calculated using the matrix diagonalization programme for axially symmetric hyperfine and g tensors as in the case of CF,Cl-. As can be seen from fig. 8, excellent agreement L II LLLLLo-II 1, II :1 ' II I 9 l(79) I 1 I d 1 L 1 1 !! 1 ! 1 1 1 //(81) I 1 1 r i L L i 1 ' ' 1 ! 1 1 I //I791 FIG. g.--First-derivative e.s.r. spectrum of a y-irradiated solid solution of 5 mol yo bromotrifluoro- methane in tetramethylsilane observed at 121 K, v = 9107.5 MHz.The diagrams show the calculated line positions of the perpendicular and parallel components for CF3*IBr- and CF3"Br- according to the parameters given in table 2. was obtained between the observed and calculated field positions for both sets of components, and the best-fit e.s.r. parameters are listed in table 2. Considerable confidence can be placed in the uniqueness of the fit due to the wealth of structural detail resulting from the presence of the two bromine isotopes. Finally, confirmation of the CF,Br- assignment was obtained by employing the TMPD photoionization method to produce a spectrum similar to that in fig. 8. The identification of the CF31- radical anion was established by e.s.r. and chemical experiments similar to those described for CF3C1- and CF3Br-.The spectral analysis will be given elsewhere but it should be mentioned that the e.s.r. spectrum of the irradiated CF31 + TMS samples showed the orientation dependence which is typical of partially aligned radica1s.l" Actually, this effect can be exploited to enhance the signal intensities of one set of components (parallel or perpendicular) relative to the other. Thus, it was possible to obtain essentially the individual e.s.r. spectra of the parallel and perpendicular components at optimized orientations differing by a 90" rotation about the tube axis. The e.s.r. parameters for CF31- were derived from these spectra and are given in table 2. Preliminary observations indicate that the e.s.r. spectra of these isostructural radical anions show a temperature dependence for the hyperfine coupling tensor of the unique halogen.For example, the value of AII(81Br) for CF,Br- decreased from 272.1 G to 263.6 G as the TMS solution was warmed from 91 to 121 K, although the other e.s.r. parameters remained practically unchanged. Since both CF,Cl- and168 ELECTRON S P I N RESONANCE STUDIES OF ELECTRON ATTACHMENT CF,I- were observed to dissociate into the CF3* radical at temperatures close to 100 K in neopentane, the bond between the carbon and the unique halogen in these radical anions must be extremely labile. Thus, it is not surprising to find this thermal lability reflected by a perturbation of the spin density in the po orbital of the unique halogen. (b) CF2Clz AND CFCl,’ Only a brief summary will be given here of the results obtained for CF2C12 and CFC13.The e.s.r. measurements on a y-irradiated solution of CF2C12 in TMS showed pronounced orientation effects similar to those previously described for the spectrum of CF31-, the best resolution being obtained for the spectrum consisting mainly of parallel components. A careful analysis of this spectrum revealed evidence for hyperfine interaction with two equivalent fluorines and two nonequivalent chlorines designated (a) and (t). Accordingly, the spectrum was mainly interpreted in terms of contributions from three distinct hyperfine patterns resulting from the chlorine isotope combinations 35(a) - 35(t), 35(a) - 37(t), and 37(a) - 35(t). The e.s.r. parameters are given in table 2 and it was concluded that CF2Clz possesses pseudo-C,, symmetry such that chlorine (a) having the larger hyperfine tensor component A,, occupies the axial position and chlorine (t) is in one of the trigonal sites.This conclusion is consistent with the finding that the CF2C1. radical is produced on dissociation of the CF2C1F radical anion, the chloride ion being extruded from the axial position as in the dissociation of CF3Cl-. In contrast to the well-resolved spectra described previously in this paper, the e.s.r. spectrum of a y-irradiated solid solution of CFC13 in TMS at 81 K consisted of a broad featureless absorption extending over 400 G. A similar spectrum was observed for a y-irradiated CFC13 + MTHF solution. On annealing the TMS solution above 115 K, the broad spectrum decayed out and the isotropic spectrum of the CFC12* radical36 grew in until only the latter was observable at 137 K.An examination of the spectrum taken at 81 K after this annealing treatment revealed that the initial broad absorption had decayed irreversibly and did not correspond to the anisotropic spectrum of the CFCl; radical. Thus, we attribute the broad spectrum to the likely chemical precursor of the CFC12- radical, namely, the CFClg radical anion. Further- more, if it is assumed, in accordance with this chemical evidence, that a chlorine atom occupies the dissociable axial position in a pseudo-C,, structure for CFCl?, the total width of the broad spectrum (-400 G) agrees reasonably well with that calculated using the parallel values reported for the 19F, 35Cl (a) and 35Cl (t) hyperfine tensor components in CF3C1- and CF2Cl;: (table 2).To summarize, evidence has been obtained that each of the chlorofluoromethane radical anions dissociates by the loss of a chloride ion giving the corresponding neutral radical. The e.s.r. parameters of the CF3*, CF2C1 and CFC12* radicals produced in these systems are collected in table 3. DISCUSSION Together with previous ~ t ~ d i e s , ~ ~ , ~ ~ * ~ ~ * ~ ~ this work shows the usefulness of the solid neopentane and TMS matrices for the e.s.r. detection of radical anions generated by electron attachment during y irradiation or photoionization at 77 K. It is, there- fore, interesting to recall that high-mobility electrons 37*38 and large free-ion yields 39 are associated with the irradiation of these particular compounds in the liquid state.Moreover, since photoionization studies using TMPD give similar re~ults,2~* these solvent effects are not peculiar to the use of high-energy radiations. If the electronsA . HASEGAWA, M . SHIOTANI A N D F. WILLIAMS 169 TABLE 3.-E.S.R. PARAMETERS FOR CF3, CFzCl AND CFCIz RADICALS radical T/K g i s o nucleus CF3 1 20 2.0031 19F (3) CF3' 2.00287 19F (3) CFzCl 146 2.0046 19F (2) CFzClb 134 2.0052 19F (2) CFCl2 137 2.0068 19F CFCl2' 19F 35c1 35c1 35c1 (2) 35c1(2) aiso/G precursor matrix 144.7 CF3X- neopentane (X = Cl,Br,I) 145.3 111.0 CFzClT TMS 16.6 109.1 16.2 84.7 CFCl, TMS 10.2 84.6 10.5 Ref. (28); generated by y irradiation of CFzClz in the SF6 matrix (M. Shiotani and F. Williams, unpublished results); ref. (36). released into neopentane and TMS in the solid state also have high mobilities, this would contribute to the efficiency of electron scavenging. However, we are dis- inclined to attribute the entire success of these e.s.r.experiments to the electronic properties of the matrix molecules in facilitating the process of charge separation. In our opinion it is much more likely that the crucial matrix properties are those characteristic of rotator solids which permit the ions to remain trapped in the crystal- line lattice while they execute considerable local motion. It is this latter feature, of course, which differentiates the present work from similar experiments in rigid glasses 25 and allows e.s.r. spectra of high resolution to be obtained. PERFLUOROCYCLOALKANE RADICAL ANIONS The equivalence of the fluorines in the e.s.r.spectra of c-C,F,, c-C4F; and c-C,FG suggests that the unpaired electron is delocalized over the entire molecular framework in an orbital of high symmetry. Before considering the nature of this semi-occupied orbital, we should emphasize that this observation of equivalent fluorines does not require stereochemically-rigid planar (Dnh) ring-structures for c-C4F, and c-C,F;. For example, even if the limiting geometry of the c-C,Fr radical were that of the puckered D,, structure found for the neutral molecule in the gas phase,40 a rapid ring inversion would interchange the " axial " and " equatorial " fluorines of the extreme structures and bring about a time-averaged D4,, symmetry. Moreover, this internal motion combined with a rotation about the principal (C,) axis would result in an averaging of the 19F hyperfine tensor components.Thus, it is not necessary to have complete tumbling of the c-C4F, and c-C5F~ radicals to explain the nearly isotropic spectrum which is observed. This lack of complete tumbling would also explain why the e.s.r. spectrum of the smaller but rigid c-C3F; radical in TMS shows residual anisotropy (table 1) as manifested by well-defined lineshapes which are characteristic of those observed in powder spectra for axially symmetric hyperfine tensors. A similar anisotropy has been observed for the e.s.r. spectrum of the perfluorobenzene radical anion,,, and may be associated with rigid planar ring- structures. Since the perfluorocycloalkanes are saturated molecules, the molecular orbital occupied by the unpaired electron in these radical anions is of exceptional interest.In addition to the fact that the total number n of fluorines in each of the radical anions170 ELECTRON SPIN RESONANCE STUDIES QF ELECTRON ATTACHMENT are magnetically equivalent, we find that the isotropic coupling constant a,(n) diminishes with increasing ring size such that the total I9F coupling represented by na,(n) remains sensibly constant (table 1). These results clearly support the concept of complete electron delocalization in the perfluorocycloalkane radical anions. In generating the complete molecular orbital, it is useful to examine the localized antibonding orbitals of a >CF2 group in Czu local symmetry, as shown in fig. 9. 0; ) b; FIG.9.-The localized antibonding orbitals of a CFz group (C2” symmetry). The representations in parentheses refer to the orbitals of corresponding symmetry for planar (D4*) perfluorocyclobutane. These a; and 6; group orbitals are symmetric and antisymmetric, respectively, with respect to the symmetry plane containing the other two bonds from carbon. Since the a; orbital corresponds closely to the half-occupied orbital in phosphoranyl radicals,6 it is expected to be of lower energy than the b,* orbital. From these group orbitals, two antibonding orbitals can be constructed for c-C4F, in D4h symmetry by FIG. 10.-A schematic drawing of the a;, molecular orbital of perfluorocyclobutane. taking the in-phase combinations around the ring, the representations becoming a;, (from a;) and afu (from br).the results of INDO calculations on c-C4F; in D,,, symmetry do suggest that the unpaired electron An illustration of the a;, orbital is shown in fig. 10. Although this simple approach has neglected the C-C 0A . HASEGAWA, M. SHIOTANI AND F. WILLIAMS 171 occupies either the a;, or the a;, orbital, depending on the precise geometry. Using the bond distances determined for the neutral an energy minimum was found for an FCF angle of 110" with the unpaired electron in the at,, orbital, whereas FCF angles of 120" and larger resulted in a crossover to the a;, orbital. However, the calculated spin density in the fluorine 2s orbital was relatively insensitive to this change, values of 0.0129 and 0.0112 being obtained in general agreement with the experimental value of 0.0087.Therefore we cannot choose between these two orbitals on the basis of our results and the resolution of this question must await an experimental determination of the isotropic 13C hyperfine coupling constant. Although intuition would seem to favour the choice of the totally symmetric a;, orbital,42 our preliminary INDO calculations appear to be rather equivocal on this point. The demonstration of electron attachment to c-C3F6 is of particular interest because the rate constant for electron attachment to this compound in the gas phase is less than that for c-C,F, by a factor of 4 x lo5 01' more.15b Another noteworthy point is the structural similarity between the perfluorocycloalkane and the permethyl- cyclosilane radical anions, the latter species having been prepared many years ago by West and his co-w~rkers.~~ If the compositions of the semi-occupied molecular orbitals are comparable for these two series of radical anions, the role of silicon d orbitals seems unlikely to be a crucial factor in the stabilization of the cyclopolysilane negative ions.HALOGENOFLUOROMETHANE RADICAL ANIONS The spin density distributions calculated from the e.s.r. parameters of the trifluoro- methyl halide radical anions are included in table 2. Despite the absence of I3C data, the results suggest that the unpaired electron resides in an a, (o*) antibonding orbital which is composed largely of the orbitals from carbon and the unique halogen Z 0.1 67 0.01 6 FIG. 11.-The a; ( D * ~ - ~ ~ ) molecular orbital of CF3C1- showing the experimental values of the spin densities in the chlorine (3s and 3p) and fluorine (2s) orbitals.172 ELECTRON SPIN RESONANCE STUDIES OF ELECTRON ATTACHMENT which lie along the C30 symmetry axis of the radical anion. Consistent with this proposal, the spin densities in the s and p orbitals of the unique halogen increase along the series C1, Br, I, which is the order expected for the effect of decreasing halogen electronegativity .In fig. 11, the composition of the semi-occupied molecular orbital for CF3C1- is illustrated, the numerical values being the experimental spin densities in the chlorine (3s and 3p) and fluorine (2s) orbitals. These results can be compared with those in table 4 which were obtained by a CND0/2 calculation.Although the calculated spin TABLE 4.-sPIN DENSITIES FOR THE CF3Cl- RADICAL ANION CALCULATED BY THE CND0/2 METHOD' atomic orbitalsb atom S P x PY Pz C 0.277 1 -0.012 8 -0.012 8 0.196 0 - 0.000 2 0.387 6 c1 0.031 3 -0.000 2 FIC 0.003 6 0.000 8 0.014 9 0.027 0 FZ 0.003 6 0.011 0 0.003 2 0.027 0 0.003 6 0.011 0 0.003 2 0.027 0 F3 ~~~ ~ The following molecular parameters were obtained for the energy-optimized geometry: rc-cl = 194 pm, Y ~ - ~ = 135 pni, and cc = 24", where a is the angle between the C-F bond and the plane perpendicular to the CJ ( z ) axis; F1 is the fluorine in the y z plane. the coordinate system is shown in fig. 11; densities in the chlorine orbitals are larger than the experimental values by about a factor of two, the CNDO calculation confirms the e.s.r.findings that the CF3Cl- radical anion possesses C3v symmetry with the unpaired electron in an a, (a*) anti- bonding orbital mainly between carbon and chlorine. The CNDO calculations also predict that the C-Cl bond distance increases from 175 pm for the neutral CF3Cl molecule to 194 pm for the radical anion, whereas the corresponding change for the C-F distance from 133 to 135 pm is much less significant. Therefore, we can expect the occupation of the o& orbital by the unpaired electron to lengthen and weaken the C-C1 bond so that the radical anion is much more easily dissociated than the neutral molecule, as observed. In fact, the chemical evidence presented in this paper for the dissociation of the CF3Cl-, CF2C1; and CFCl; radical anions by loss of a chloride ion offers very strong support for the structural conclusions reached from e.s.r.spectroscopy and theoretical calculations. As far as we are aware, the CF3X- radical anions are the first examples of 33 valence-electron radicals in which carbon is the central atom and the radical assumes C3" symmetry. It is especially significant that the spin density in the po orbital of the bromine in CF,Br- (table 1) is about a factor of four greater than the corresponding spin density observed for the weakly-bound species CH3----Br-.l1 There is a similar difference between the results for CF31- and CH3*- - -I-,11 so it seems appro- priate to describe the CF3X- species as true radical anions rather than as radical- anion adducts. These findings are consistent with molecular beam studies 44 which have shown that whereas CF31- is produced as a transient negative ion by the reaction of hyperthermal alkali atoms with CF31, only I- is observed in the reaction with CH31.Furthermore, these gas phase show that the CF31- ion decomposes almost exclusively to form CF3* + I-, which is the dissociation path observed in the present study.A . HASEGAWA, M. SHIOTANI AND F . WILLIAMS 173 We thank Dr. C. M. L. Kerr for providing us with a copy of her computer program for matrix diagonalization calculations. This work was supported by the Division of Physical Research, U.S. Energy Research and Development Administration. (a) C. M. L. Kerr and F. Williams, J. Amer. Chem. SOC., 1971, 93, 2805; 1972, 94, 5212; (b) T. Gillbro and F. Williams, Chem.Phys. Letters, 1973, 20, 436. (a) A. Begum and M. C. R. Symons, J. Chem. SOC. A , 1971,2065; (6) C. M. L. Kerr and F. Williams, J. Phys. Chem., 1971, 75, 3023. (a) R. W. Fessenden and R. H. Schuler, J. Chem. Phys., 1966, 45, 1845; (b) W. Nelson, G. Jackel and W. Gordy, J. Chem. Phys., 1970,52,4572; (c) G. F. Kokoszka and F. E. Brinckman, J . Amer. Chem. SOC., 1970,92, 1199. (a) P. J. Krusic, W. Mahler and J. K. Kochi, J. Amer. Chem. SOC., 1972, 94, 6033; (b) A. G. Davies, D. Griller and B. P. Roberts, J.C.S. Perkin II,1972,993; 1972,2224; (c) K. U. Ingold, J.C.S. Perkin 11, 1973, 420. J. Higuchi, J. Chern. Phys., 1969,50,1001. T. Gillbro and F. Williams, J. Amer. Chem. SOC., 1974, 96, 5032. ' A. J. Colussi, J. R. Morton and K. F. Preston, J. Phys. Chem., 1975,79, 651, 1855.D. J. Nelson and M. C. R. Symons, J.C.S. Dalton, 1975, 1164. A. Hasegawa, M. Ohnishi, K. Sogabe and M. Miura, Mol. Phys., 1975, 30, 1367. (a) E. D. Sprague and F. Williams, J. Chem. Phys., 1971, 54, 5425; (b) S. P. Mishra and M. C. R. Synions, J.C.S. Perkin 11, 1973, 391; (c) Y . Fujita, T. Katsu, M. Sat0 and K. Taka- hashi, J. Chem. Phys., 1974, 61, 4307. L. E. Brus and V. E. Bondybey, Chem. Phys. Letters, 1975,36,252; J. Chem. Phys., 1976,65, 71. lo W. E. Wentworth, R. George and H. Keith, J . Chem. Phys., 1969, 51, 1791. l3 K. Nishikida and F. Williams, J. Amer. Chem. SOC., 1975, 97, 7166. l4 (a) R. K. Asundi and J. D. Craggs, Proc. Phys. SOC., 1964,83,611; (b) B. H. Mahan and C. E. l5 ( a ) R. W. Fessenden and K. M. Bansal, J. Chem. Phys., 1970,53,3468; (b) K.M. Bansal and l6 (a) W. T. Naff, C. D. Cooper and R. N. Compton, J. Chem. Phys., 1968,49, 2784; (b) F. J. l7 L. G. Christophorou, D. L. McCorkle and ID. Pittman, J . Chem. Phys., 1974, 60, 1183. " ( a ) L. A. Rajbenbach, J. Amer. Chem. SOC., 1966, 88, 4275; (b) L. A. Rajbenbach and U. Kaldor, J. Chem. Phys., 1967,47,242; (c) N. H. Sagert, Canad. J. Chem., 1968,46,95; (d)L. A. Rajbenbach, J. Phys. Chem., 1969,73, 356. l9 H. Sutcliffe and I. McAlpine, Fluorine Chem. Rev., ed. P. Tarrant (M. Dekker, New York, N.Y., 1973), vol. 6, chap. 1, p. 1. 2o K. Nishikida, F. Williams, G. Mamantov and N. Smyrl, J. Amer. Chem. SOC., 1975,97, 3526. 21 K. Nishikida, F. Williams, G. Mamantov and N. Smyrl, J. Chem. Phys., 1975, 63, 1693. 22 M. Shiotani and F.Williams, J. Amer. Chern. Soc., 1976,98,4006. 23 R. L. Hudson and F. Williams, J. Chern. Phys., 1976, 65, 3381. 24 (a) A. C. Albrecht, Accounts Chem. Res., 1970,3,238; (b) R. C. Jarnagin, Accounts Chem. Res., 1971, 4, 420. 25 (a) J. Lin, K. Tsuji and F. Williams, J. Chem. Phys., 1967, 46, 4982; J . Amer. Chem. Soc., 1968,90,2766; (b) K. Tsuji and F. Williams, J.Phys. Chem., 1968, 72, 3884; (c) R. M. Keyser, K. Tsuji and F. Williams, The Radiation Chemistry of Macromolecules, ed. M. Dole (Academic Press, New York, 1972), vol. 1, chap. 9, p. 145. 26 (a) R. W. Fessenden, J . Chem. Phys., 1962, 37, 747; (b) R. W. Fessenden and R. H. Schuler, J. Chem. Phys., 1965, 43, 2704. 27 C. M. L. Kerr, K. Webster and F. Williams, J. Phys. Chem., 1975,79,2650. R. W. Fessenden, J. Magnetic Resonance, 1969, 1, 277. 29 E. F. Westrum, Jr., J. Chern. Educ., 1962, 39, 443. 30 R. W. Fessenden and R. H. Schuler, Advances in Radiation Chemistry, ed. M. Burton and J. L. Magee (J. Wiley, New York, 1970), vol. 2, chap. 1 , p. 1. 31 M. B. Yim and D. E. Wood, J. Amer. Chem. SOC., 1955, 97, 1004, and references therein; see also the recent e.s.r. study of fluorinated benzene anions in the adamantane matrix, M. B. Yim and D. E. Wood, J. Amer. Chem. SOC., 1976, 98, 2053. Young, J. Chem. Phys., 1966,44,2192. R. W. Fessenden, J. Chem. Phys., 1973,59, 1760. Davis, R. N. Compton and D. R. Nelson, J. Chem. Phys., 1973,59,2324. 32 J. R. Morton and K. F. Preston, Mol. Phys., 1975, 30, 1213. 33 A. R. Boate, A. J. Colussi, J. R. Morton and K. F. Preston, Chem. Phys. Letters, 1976,37, 135. 34 R. McNeil, M. Shiotani and F. Williams, unpublished results.174 ELECTRON S P I N RESONANCE STUDIES OF ELECTRON ATTACHMENT 35 C. Chachaty, A. Forchioni and M. Shiotani, Canad. J. Chem., 1970, 48, 435. 36 A. J. Bowles, A. Hudson and R. A. Jackson, Chem. Phys. Letters, 1970,5, 552. 37 P. H. Tewari and G. R. Freeman, J. Chem. Phys., 1968,49,4394. 38 W. F. Schmidt and A. 0. Allen, J. Chem. Phys., 1969,50, 5037. 39 W. F. Schmidt and A. 0. Allen, J. Chem. Phys., 1970,52,2345. 40 C. H. Chang, R. F. Porter and S. H. Bauer, J . Mol. Struct., 1971, 7, 89. 41 For drawings of the molecular orbitals of planar cyclobutane in symmetry, see W. L. Jorgensen and L. Salem, The Organic Chemist’s Book of Orbitals (Academic Press, New York, 1973), p. 222. 42 J. F. Liebman, J. Fluorine Chem., 1973/74, 3, 27. 43 R. West and E. Carberry, Science, 1975, 189, 179, and references therein. 44 (a) P. E. McNamee, K. Lacmann and D. R. Herschbach, Faraday Disc. Chem. SOC., 1973,55, 318; (b) S. Y . Tang, B. P. Mathur, E. W. Rothe and G. P. Reck, J. Chem. Phys., 1976,64,1270.

ISSN:0301-7249    

DOI:10.1039/DC9776300157

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

Mechanisms for Electron-Capture Processes' BY SHUDDHODAN P. MISHRA AND MARTYN C. R. SYMONS Department of Chemistry, The University, Leicester, LE 1 7RH Receiced 17th December, 1976 Electron-capture processes resulting from exposure to ionizing radiation have been studied by e.s.r. spectroscopy at low temperatures. Intratnolecular competitions between electron-addition and dissociative electron capture are illustrated with particular reference to organophosphorus compounds, which can add electrons to give phosphoranyl radicals or lose halide ions to give alkyl or phosphoryl radicals. Similar competition experiments in which halide ion loss competes with electron addition to a cyanide group are also described and discussed. INTRODUCTION Solid-state radiolyses seem to be dominated by electron-loss and electron-capture processes, although homolyses, probably induced by electron-return to give excited- states of the parent molecules sometimes also occur.A simple example of the former is the radiolysis of potassium nitrate at 74 K, giving NO, and NO:- as the primary products detected by e.s.r. spectroscopy.2 Such products are only expected to be trapped provided there are considerable changes in nuclear configurations relative to the parent molecules or ions. Thus for nitrate, NO:- is clearly different because it is pyramidal, so that electron-trapping is efficient. Also, NO,, although planar, is thought to have a distorted structure,, in which case again, trapping should be efficient. This result contrasts with the behaviour of many materials, especially in the semi-conductor class, for which electron-transfer occurs more rapidly than the required distortion, so that there is no net reaction. Primary products in low-temperature solids are often different from those obtained from the same materials in fluid solution.This is partly because of the different balance between entropy and enthalpy, but also reflects the role of the solvent. Thus if protic solvents are used solvation is rapid and strong for anions, including electrons, so that ejected electrons (e,) after thermalisation are very rapidly solvated (e;), and thereby stabilised. In contrast electrons in rigid media may not be solvated at all and, even in such media as glassy alcohols, are solvated relatively ~lowly,~ and are far more likely to be captured by reactive solutes prior to being deeply trapped than in fluid solution. Again, after solvation in fluids they remain mobile and hence most electron capture processes occur cia e; centres, whereas in solids e; centres are not mobile and so subsequent reactions depend upon the solutes being fortuitously close enough for electron-transfer to occur.Although this is not always accepted, our view is that e6 is likely to be more reactive than e; (the margin being close to the solvation energy). One apparently good example of this difference is that, in the liquid-phase, t-butyl alcohol (2-methyl propan-2-01) exhibits no tendency to react with e;, and is indeed often used as a specific scavenger for *OH radicals. In contrast, the pure alcohol, or its solution in water or methanol, interacts efficiently with e, to give Me&* radicals at 77K:' Me,COT-i $- e- -+ Me,C* + OH-.176 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES In competitive reactions, electron-capture can be remarkably specific in the solid- state, suggesting that at least in certain media, the electron migrates over large distances prior to capture.One example is the efficient formation of NO:- rather than COi- in carbonate crystals doped with nitrate ions.6 This, and similar reactions7 are strongly favoured by the environment because of the charge deficiency of the substrate. A most striking example is that of KPFs doped with ions.' Even at very low levels of concentration, the reaction + e- + *AsF,- + F- (2) takes precedence over PFs- + e- -+ *PF5- + F-, (3) despite the fact that step (3) is known from liquid-phase studies to be efficient. processes, which can follow a wide variety of paths.We have recently turned attention to intramolecular competitive electron-capture EXPERIMENTAL All materials were of the highest grade available and were used as supplied since ultra-high purity is not normally required. Samples were irradiated as fine powders or as dilute solutions in CD,QD prepared in the form of small spherical beads, after removal of dissolved oxygen. They were irradiated at 77 K in a Vickrad 6oCo pray cell at a dose rate of -2.5 MRad h-I for up to 2 h. E.s.r. spectra were measured on a Varian E3 spectrometer at 77 K. Samples were routinely annealed by decanting the liquid nitrogen from the insert Dewar flask and monitor- ing the e.s.r.spectra as the samples warmed up. Whenever significant changes in the e.s.r. spectra were noticed samples were re-cooled to 77 K and the new spectra were recorded. INTRAMOLECULAR COMPETITIVE ELECTRON-CAPTURE We classify electron-capture processes in a molecule ABC, containing two in- dependent potential capture centres, A and C in the following manner: process (i) Direct capture to give *ABC and/or ABED. (ii) Dissociative electron capture (d.e.c.) to give A- + *BC, A- + BC-, AB* + C- or AB- + *C. (iii) Any admixture of (i) and (ii). It is important to stress that, although (i) and (ii) may be in direct competition in Consider the the initial process, (ii) may not follow (i) as a subsequent thermal step. steps ABC + e- -+ ~ B C (4) and ABC + e- -+ A- + *BC.( 5 ) It is frequently supposed that step ( 5 ) is preceded by (4) the full mechanism for (5) being ABC + e- -+ *ABC -+ A- + *BC. (6) However, we have frequently observed that, although the products of both (4) and (5) appear in the e.s.r. spectra after irradiation at 77 K, nevertheless, on heating there is no evidence for the second stage of step (6).9 Thus we prefer to write step (6) as ABC + e- --+ (ABC-)* -+ A- + *BC, (7)SHUDDHODAN P . MISHRA A N D MARTYN C . R . SYMONS 177 with (ABC-)* --f -ABC as an alternative pathway. Then the reaction *ABC --f A- + *BC (9) may or may not prove to be one of the thermal reactions on annealing. The anion (ABC-) * is a " Franck-Condon " anion in which the electron is weakly " captured " prior to the necessary nuclear relaxation to give the stable configuration of the anion.SOME EXAMPLES OF COMPETITIVE ELECTRON CAPTURE PROCESS (i) Good examples of competitive addition are found in the reactions of certain aryl- phosphorus compounds.1° Thus C6K5PO(OH)H was found to undergo electron addition at the benzene ring and at phosphorus, giving *C;H5PO(OH)H and c6H,b(oH)H-. However, C~IH,PO(OH)~ and (c~H,),Po(oH) gave only the aryl anions, there being no evidence for addition to phosphorus. One factor is thought to be that the electron-affinity of tetrahedral phosphorus is low, and the time taken for distortion to the generally accepted " trigonal bipyramidal " structure (I) relatively L I ,J- I L - ' P i L I long except for the first of these compounds which contains a P-H bond.In this case movement of the hydrogen is fast and so addition at phosphorus is more com- petitive. There is probably only relatively minor movement of nuclei when an extra electron is added to the benzene ring. However, ring addition can be " fixed " by protonation to give the corresponding cyclohexadienyl radicals which were indeed detected on annealing.lO In no instance could we detect electron transfer from the aryl ring to phosphorus on annealing. Thus it is the electron-affinity of the un- distorted centres that must be considered when such reactions are compared, although the ease of distortion is also clearly an important factor. PROCESS (ii) When alkyl halides react with electrons they do not usually form negative ions, but undergo d.e.c.directly. (Exceptions to this statement are the formation of CBr4- ,11 and certain cyclic perfluorides such as C,F,.12) Also, such dissociation only goes in one direction, to give the alkyl radical and the halide anion. Thus in a dihalide there are only two possible routes rather than the four given above. We have recently studied the mixed chloro-bromo-hydrocarbons BrCH,Cl, BrCHCl, and BrCCl, in order to probe the relative tendencies for loss of chloride or bromide ions.13 In non-polar media the results are unambiguous: bromide is preferentially lost. Thus the radicals detected by e.s.r. spectroscopy were *CH2Cl, *CHC12 and CCl,. We suggest that a major factor determining this route is the size of the anion. This is supported by our results in glassy methanol.In this solvent, BrCH,Cl gave weak In the absence of polar solvation Br- i s more stable than C1-.178 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES *CH,Cl features and strong features for *CH,Br radicals. Similarly, BrCHC.1, gave CHC1, and CHClBr radicals, whilst C1,CBr gave CCl, and CC1,Br radicals. These differences must be connected with the fact that the solvation energy of chloride in methanol is considerably greater than that of bromide. It might be supposed that at 77 K glassy methanol is incapable of anionic solvation, but this is not the case. Thus electrons trapped in vacancies in glassy alcohols seem to be fully solvated at 77 K,14 although this is not the case at 4.2 K.15 Furthermore, this is not a special effect confined to electrons, since, when superoxide ions are generated from oxygen in such media at 77 K, e.s.r.studies reveal that solvation is as efficient as it is in fluid solution.16 Even so, we need some mechanism whereby such induced solvation can affect the pathway of the d.e.c. process. It seems to us unlikely that the parent molecules are significantly hydrogen bonded by methanol. Dilute solutions of methanol in such solvents have infrared spectra that are characteristic of free molecules, so if there is any hydrogen bonding it must be extremely weak. It also seems unlikely that solvent inolecules will become oriented in the required manner during the bond-stretching process of a one-step d.e.c. reaction. We, therefore, suggest that there is an unstable anionic intermediate formed, having a long enough life to become solvated, and that such solvation is directed more towards chlorine than bromine.A similar, but even more dramatic solvent effect was found for electron-capture by Cl,C-NO, molecules. Of the three reasonable paths for d.e.c., that giving *NOz and CC13- was observed in inert media, whilst that giving *CC12 and NOz- was the only route detected in glassy methanol.’’ Again, we postulate the formation of an intermediate anion which is solvated primarily at the oxygen atoms thus favouring nitrite formation. An interesting varient of d.e.c. was found for vic-dihalides such as C1CH2CH2C1.l8 Exposure to 6oCo y-rays gave high yields of C l 2 - anions, which are clearly identified by their e.s.r. spectra. Mixtures of ClCH2CH2C1 and BrCH,CH,Br gave C l Z - and OBI-,-, but ClCH,CH,Br gave only *ClBr,- anions.It is tempting to postulate a one-step extrusion following electron-capture, i.e., but a two-stage cage process remains possible. ClCH2CH2Cl + e- -+ CH, = CH, + *C12-, (10) ELECTRON-ADDITION TO PHOSPHORUS Scheme I summarises the types of electron-capture reactions expected for 4- Reactions of these various types have been coordinated phosphorus compounds. L,* + PL3- L1- + *PL3 studied in a range of inorganic and organic phosphate^.'^*^^ In the present study we have selected two compounds for which various d.e.c. routes are available as well as electron-capture to give the corresponding phosphoranyl radical anions.SHUDDHODAN P. MISHRA AND MARTYN C. R . SYMONS 179 REACTION OF Cl2P(O)CH2C1.The two chloride ligands should strongly favour addition to give 11, and both the pure compound and its solutions in methanol (CD,OD) gave well defined features characteristic of a phosphoranyl radical with two axial chloride ligands20 (fig. 1 and table 1). On annealing above 77 K these radicals were lost, but features for a FIG. 1.-First derivative X-band e.s.r. spectrum (9.1 15 GHz) for CI2P(O)CH2Cl after exposure to 'To y-rays at 77 K and partial annealing, showing features assigned to *PCI,(O)CH,Cl- radicals (A) and *PC1(0)CH2C1 radicals (B). Only the parallel features ( 2 ) for (A) are marked. phosphoryl radical containing one chloride ligand (111) grew in : *PC12(0)CH2C1- -+ *PCl(0)CH2C1 + Cl-. (1 1) .p---o rC' CHZCl \ m However, methanolic solutions also gave clear features for C12- at 77 K, although these radicals were not detected in the pure compound.This is reminiscent of the reaction of SO,Cl, with electrons, which also gave features for C12-.21 The results180 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES TABLE ~.-E.s.R. DATA FOR RADICALS FORMED BY RADIOLYSIS OF ClCH2P(O)C12 AND c13cp(o)(oH)2 compound radical ClCH2P(O)C12 Hz&P(O)C12 'H 31P 31P P(O)Cl,(CH,Cl)- 31P ClbHP(0)Cl 'H 35c1 35c1 35c1 Cl,CP(O)(OH)2 CCI, 35c1 P( O)Cl(CH,Cl) lP hyperfine coupling/G" A x A, A, Aiso 22 b 40:almost isotropic 21 b 40:almost isotropic 21 b b 980, 980," - b 60, 60, - 590, 490, 490, 523" -20 19, 0, 0, 6.3 17.5 b 50 -50 g-values X Y Z 2.003 :almost is0 tropic 2.003 b b 2.007 -2.006 2.007, 2.01, 2.01 2.002, b (I G = T.Features insufficiently defined to provided accurate parameters. Estimated using the Breit-Rabi Equation. TABLE 2.-E.S.R. DATA FOR RADICALS FORMED BY RADIOLYSIS OF VARIOUS ORGANIC CYANIDES compound radical ClCHzCN ~ H ~ C N lH ClCHCN IH 35c1 ClCH2CH2CNb H26CH2CN lH, lH8 Br2CHCN BrdHCN lH 14N I4N 81Br Br26CN *lBr NCCH2C02Hd N=C(H)CH2CO,H 'H 14N NCCCH2CONH2 N=C(H)CH2CONH2 IH 14N suggest that extrusion of C12 - occurs after electron-capture, but before complete distortion (la) to give the stable phosphoranyl radical has occurred, since C12- ions were not formed during the annealing process. In addition to these reactions which involve electron capture by phosphorus, the radical H,CP(Cl,)(O) was detected, but only in methanolic solution [fig. 2(b)] Cl,P(O)CH,Cl + e- -+ C12P(0)CH2 + C1-.(12) This is, in our view, a surprising result in the light of the relative stability of the phosphoranyl anion. Clearly, solvation is again involved, but, as stressed above, this implies some life-time for an intermediate anion.SHUDDHODAN P . MISHRA AND MARTYN C. R . SYMONS 181 (The major electron-loss centre detected was C1CHP(0)C12 [fig. 2(a) and table 11. This novel species had e.s.r. features typical of a-chloro radicals,22 together with a doublet splitting typical of a-phosph~rus.~~) RE A c T I o N s o F CI,CP(O)( OH), In marked contrast, this compound gave no evidence for electron-capture at phosphorus, despite the fact that phosphoranyl and phosphoryl radicals are readily detected by e.s.r. spectroscopy. The pure compound gave features characteristic of CC13 radicals,24 no other features being clearly identifiable [fig.3(a) and table 11. However, in CD,OD, CC13 radicals were not clearly detected, the dominating species being identified as CI,CP(O)OH), [fig. 3(b) and table 11. Thus electron capture can take two routes: Cl,C-P(O)(OH), + e- -+ CCI3 + H2PQ- (13) or CI,C-P(O)(OH), + e- -+ Cl,CP(O)(OH), + C1-. (14) Once again, the large anion is favoured in a non-polar environment, whilst the small anion with high solvation energy is favoured in methanol. This result nevertheless makes a curious contrast with the reactions of CI3C-NO2 discussed above. We stress that reactions (13) and (14) can only be said to dominate the reactions of the pure compound and its solutions in methanol respectively, since the two species have strong, 1 3230 G FIG.2. (a)182 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES 1 3230 G (b) FIG. 2.-First derivative X-band e.s.r. spectra (9.1 15 GHz) for Cl2P(O)CH2C1 after exposure to 6oCo y-rays at 77 K. (a) Central region for the pure compound, showing features assigned to ClkHP(0)Cl2 radicals, and (b) for a solution in CD30D, showing features assigned to H2CP(O)CI2 radicals. (Although there is clear anisotropy, this is difficult to apportion between that for the a- protons and that for 31P. Probably the a-protons dominate. Note the absence of coupling to 35Cl or 37Cl.) broad features in the same spectral regions so that minor contributions of features from the alternative species would be hidden. We stress that absence of any phosphoryl radicals means that reaction b is far more favourable than c in Scheme I.This must reflect the stability of the CCl, radical and the fact that this species is pyramidal,24 so that bond-breaking as in b does not involve major concomitant movement of nuclei. The electron loss centre for C1,C-P(O)(OH), was probably C1,C-P(O)(OH)O*, formed by loss of e- and H+. This radical is expected to exhibit a broad asymmetric doublet in the free-spin region.25 Such doublets were probably present, but no definitive parameters could be deduced from the spectra. SOME REACTIONS OF ORGANIC CYANIDES The cyanide group is an efficient electron-capture centre, but the resulting anion, RCN-, is a strong base and is usually rapidly protonated to give R(H)C=N radicals.SHUDDHODAN P.MISHRA AND MARTYN C . R . SYMONS 183 .1 3230 G FIG. 3. (a) These are relatively stable,26 and are characterised by a large almost isotropic proton hyperfine coupling and hence are easy to detect. For the molecule CNCH2C02H, there is competition between addition at the -CN group and addition at the -C02H group. Our results for this acid, and for its solutions in methanol showed that addition to the -CN group dominates.27 Similarly, CNCH,CONH, was found to give the radical (H2NCOCH2)CH=N in high yield.28 Our present results for the halogen substituted molecules ClCH,CN, C1CH2CH2CN and Br2CHCN show conclusively that d.e.c. takes precedence over addition to CN. Thus the pure materials gave no features for the RCN- anions or their conjugate acids, but clear features for CH,CN, CH2CH2CN and BrCHCN 29 radicals, respec- tively.For the pure compounds, these features overlapped strongly with those for the electron loss centres, which were identified as CleHCN, ClCHCH,CN and Br2CCN respectively. However, for solutions in CD,OD, where the electron-loss centres were primarily CD20D having only a narrow central component, their features were well defined (fig. 4). We conclude that electron addition to cyanide cannot compete effectively with halide ion ejection, at least at 77 K. This may arise in part because RCN- radicals are strongly bent in their ground-states. (CJ for the isostructural HCO radical, 8 = 123.3".) Thus the competition is between bond bending and bond stretching.184 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES I 3250 G (b) FIG.3.-First derivative X-band e.s.r. spectra for CI,C-P(O)(OH), after exposure to 6oCo y-rays at 77 K. (a) For the pure material, showing features assigned to CCI3 radicals (note central distortion from OPL3 doublet), and (b) for a methanolic (CD30D) glass, showing features assigned to Cl,CeL, radicals. (The x and y features, although typical for R k I 2 radicals, are not well enough defined to warrant extraction of e.s.r. parameters.) These results make an interesting contrast with our results for electron addition to BrCN.30 We had expected to observe a bent anion, BrCN-, having a structure similar to HCN-,3' or to FC032 and ClC0.33 In fact, however, the radical proved to be more comparable with VK centres, having a linear, cr* structure with high spin- density on bromine.Thus again, bond-stretching rather than bond bending was favoured. Whilst checking the reactions of NCCH2CONH2, we have obtained results for solutions in CD30D which are of particular mechanistic interest (fig. 5). The major species, N=C(H)CH2CONH2, gave 4.5 G triplets from the methylene protons, but freshly prepared solutions in CD30D gave well defined doublets (fig. 5), the major 85 G proton coupling being retained. After storing the solution prior to irradiation for several hours at room temperature, these doublets gave way to singlets, and, for the first time, features assignable to NC=(D)R radicals were also obtained. These results mean that the process we had anticipated, namely electron addition followed by protonation by the solvent: NCCH,CONH, + e- + MeOH + N=C(H)CH,CONH, + MeO-, (15)3300 G 4 J 9 FIG.4. (a) 3250 G FIG. 4. (6)186 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES J ! I 1 )3220 G (c) FIG. 4.-First derivative X-band e.s.r. spectra for a range of organic cyanides after exposure to 6oCo prays at 77 IS. (a) ClCH2CN in CD30D, showing features assigned to *CH2CN radicals, which appear to be undergoing restricted rotations at 77 K, together with weak outer features ( a ) assigned to ClCHCN radicals (9.005 GHz). (b) BrzCHCN in CDJOD showing features assigned to BrCHCN radicals (only the x features are indicated) (9.050 GHz) and (c) NCCH2CH20P03Ba, after slight annealing, showing features assigned to N=C(H)CH20P03*- radicals (9.145 GHz). is not the major source of these radicals. However, the following scheme fits the data for solutions in CD30D satisfactorily : NCCH2COND2 + e- -+ I%kH2COND2 N=C(H)CHCOND, + CD,OD -+ N=C(H)CH(D)COND,. (1 6) NkCH2COND2 -+ I%=C(H)CHCOND, (17) (18) We also need to postulate slow exchange of the activated methylene protons to give NCCH(O)COND, and NCCD2COND2 in order to explain the formation of N=C(H)CD2COND2 and N=C(D)CD,COND,.This intramolecular 1,2-proton shift is novel, so far as we are aware. The ease with which it occurs reflects the marked basicity of the RCN- group and the relatively high acidity of the methylene protons. The electron loss centres formed together with the radicals formed by loss of halideSHUDDHODAN P. MISHRA A N D MARTYN C. R . SYMONS 187 t A 3230 G I I d d FIG.5.-First derivative X-band e.s.r. spectra for cyano-acetamide after exposure to 6oCo y-rays at 77 K. (a) Showing features assigned to N=C(H)CH2CONH2, (b) freshly prepared solution in CD30D showing (parallel) features assigned to N=C(H)CHDCOND, and (c) after partial equilibration in CD30D, showing features assigned to N=C(H)CD2CONDz and N-C@)CD2CONDz (a) (9.055 GHz). ions from the pure materials have been difficult to identify with certainty because of extensive overlapping of e.s.r. features. However, they seem to be formed by loss of hydrogen a- to halogen. Thus features for ClCHCN appeared outside those for CH,CN, and were typical of a-chloro radicals. Similarly, the species Br,CCN was readily identified by comparing the spectrum with those assigned to a range of RCBr2 radicals.We stress that all these e.s.r. spectra for radicals containing a-halogen atoms have well defined (x) features that enable us to be sure of identification, but generally the y and z features are too poorly defined to warrant any attempt to extract e.s.r. data. Finally, we have selected an ion (NCCH2CH2P032-) that contains both a -CN and a phosphate group in order to study competitive electron addition to these two centres. In the absence of the CN group we would expect to detect the formation of phosphoranyl or phosphoryl radicals and alkyl radicals,34 but in its presence, we in fact only detected addition to the cyanide group, the resulting adduct appearing in its protonated form [fig. 4(c)]. S . P. M. thanks the Chemistry Department, Banaras Hindu University, India, for study leave.Taken as Radiation Mechanisms. Part 16. S. P. Mishra and M. C. R. Symons, J. Chem. Research, (submitted). P. W. Atkins and M. C. R. Symons, The Structure of Inorganic Radicals (Elsevier, Amsterdam, 1967). G. W. Chantry, A. Horsfield, J. R. Morton and D. H. Whiffen, Mol. Phys., 1962, 5, 589. D. R. Smith and J. J. Picroni, Cnnad. J. Chem., 1967, 45,2723. K. V. S. Rao and M. C. R. Symons, Chem. Phys. Letters, 1973,20, 555. R. S. Eachus and M. C. R. Symons, J. Chem. SOC. A , 1968,790. R. S. Eachus and M. C. R. Symons, J. Chem. SOC. A , 1968,2433; M. D. Bloom, R. S. Eachus and M. C. R. Symons, J. Cheni. SOC. A , 1970, 1235. See Part 15.188 MECHANISMS FOR ELECTRON-CAPTURE PROCESSES M. C. R. Symons, Int. J. Radiation Phys.Chem., 1976, 8,643. I. S. Ginns, S. P. Mishra and M. C. R. Symons, J.C.S. Dalton, 1973, 2509. lo S. P. Mishra and M. C. R. Symons, J.C.S. Perkin 11, 1976, 21. l1 S. P. Mishra and M. C. R. Symons, J.C.S. Chem. Comm., 1973, 577. l2 M. Shiotani and F. Williams, J. Amer. Chem. SOC., 1976, 98,4006. l3 S. P. Mishra and M. C. R. Symons, submitted for publication. l4 M. J. Blandamer, L. Shields and M. C. R. Symons, J. Chem. SOC., 1965, 1127. l5 H. Yoshida and T. Higashimura, Canad. J. Chem., 1970,48, 504. l6 G. Eastland and M. C. R. Symons, J. Phys. Chem., submitted. l7 S . P. Mishra, M. C. R. Symons and B. W. Tatteshall, J.C.S. Faruday I, 1975,71, 1772. l8 S. P. Mishra and M. C. R. Symons, J.C.S. Perkin 11, 1975, 1492. l9 T. Gillbrow and F. Williams, J. Amer. Chem. SOC., 1974, 96, 5032; T. Berclaz, M. Geoffroy and E. A. C. Lucken, Chem. Phys. Letters, 1975, 36, 677; M. C. R. Symons, Chem. Phys. Letters, 1976, 40, 226. 'O D. Nelson and M. C. R. Symons, J.C.S. Dalton, 1975, 1164. 21 C. M. L. Kent and F. Williams, J. Amer. Chem. SOC., 1971,93,2805. 22 S. P. Mishra, G. W. Neilson and M. C. R. Symons, J.C.S. Faraday 11, 1973, 69, 1425. 23 A. Begum, S. Subramanian and M. C. R. Symons, J. Chem. SOC. A , 1970, 1334. 24 S. P. Mishra and M. C. R. Symons, Int. J. Radiation Phys. Chem., 1975, 7, 617. 25 A. Begum, S . Subramanian and M. C. R. Symons, J. Chem. Sac. A, 1970, 1323. 27 S. P. Mishra and M. C. R. Symons, J.C.S. Perkin 11, 1973, 394. 29 S . P. Mishra, G. W. Neilson and M. C. R. Symons, J.C.S. Faraday 11, 1974, 70, 1165. 30 S. P. Mishra, G. W. Neilson and M. C. R. Symons, J.C.S. Faraday II, 1974, 70, 1280. 31 K. D. J. Root, M. C. R. Symons and B. C. Weathesly, MoZ. Phys., 1966, 11, 161. 32 F. J. Adrian, E. L. Cochran and V. A. Bowers, J. Chem. Phys., 1966,44,4626. 33 F . J. Adrian, E. L. Cochran and V. A. Bowers, J. Chem. Phys., 1972,56,6251. 34 D. J. Nelson and M. C. R. Symons, J.C.S. Perkin 11, 1976, in press. M. C . R. Symons, Tetrahedron, 1973, 29, 615. W. C. Lin, N. Cyr and K. Toriyama, J. Chem. Phys., 1972,56, 6272.

ISSN:0301-7249    

DOI:10.1039/DC9776300175

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

GENERAL DISCUSSION Dr. A. Singh (Pinawa, Manitoba) said: Is there any likelihood of the rate of re- combination of ions changing on application of the magnetic field? Dr. B. Brocklehusst (Shefield) said: An effect of magnetic field on the recombina- tion rate would invalidate the interpretation of the results. However, it is difficult to see how such an effect could occur. Also measurements1 made at Atomic Energy of Canada Laboratories, Pinawa, show that the field-induced enhancement saturates at 0.1 T and remains constant up to 0.7 T; effects on the distribution of ion recombina- tion times would surely continue to increase in this range. Dr. G. A. Salmon and Dr. K. Selby, (Lee&) (partly communicated): Although a great deal of evidence has been presented that the majority of excited states of aroma- tic solutes in aliphatic solvents are formed in ion-recombination processes [for refer- ences see ref.(2)] an increasing amount of evidence suggests that other processes must be important. In particular, Beck and Thomas observed that the formation of sing- let excited states of aromatic solutes was complete in <750 ps, whereas no triplet states were detected in this period, and furthermore little or no decay of solute-ions was observed. We have also carried out experiments which indicate differences in the mode of formation of singlet and triplet states of solutes. The intensity-against-time profile of the fluorescence from 1 mmol dm-3 solutions of either anthracene or PPO (2,5- diphenyloxazole) in decalin was observed when these solutions were subjected to 10 ns pulses of 3 MeV electrons.As shown in fig. 1 the yield of fluorescence was reduced by about 25-30% on lowering the temperature from 15 to - 130 "C, in which temperature range the viscosity of the solvent changes from -5 CP to -lo9 cP. On the other hand, only a very slight broadening of the intensity-against-time profile of the fluorescence signal ensued over this range of temperatures. More particularly, the expected displacement of the maximum in these time-profiles to longer times and the eventual " freezing out " of the fluorescence due to the slowing down of ionic- recombination processes was not observed (see fig. 2). Observations have also been carried out on the effect of temperature on the formation of triplet excited states of anthracene and on the behaviour of the anthra- cene radical-ions for solutions containing 2.5 x mol anthracene dm-3.Although there is some overlap in the absorption spectrum of triplet anthracene (Amax = 425 nm) with that of the radical-ions, the absorption due to the ions is much weaker and can largely be neglected in assessing the yield of triplet excited anthracene. At room temperature about 75% of the anthracene triplets are formed very rapidly, but the remaining 25% is formed over -300 ns, over which time the majority of the absorption due to the ions (A = 720 nin) also decays. As the temperature is reduced the yield of rapidly formed triplets decreases, and the rate of formation of the growing portion R. S. Dixon et al., Canud. J. Chern., in press ; F.P. Sargent et nf. , J. Phys. Chem., in press. G. A. Salmon, Int. J . Radiation Phys. Chem., 1976, 8, 13. G. Beck and J. K. Thomas, J. Phys. Chem., 1972,76,3856.190 GENERAL DISCUSSION t 1 I 1 l I I 1 t20 -20 -60 -100 -141 TI0 c FIG. 1 .-Dependence of fluorescence yields of 1 mmol dm-3 solutions of 2,5-diphenyloxazole (PPO) (0) and anthracene (A) in decalin (measured at 365 and 405 nm respectively) on temperature. Yields normalized at room temperature. The dependence of the viscosity of the solvent on temperature (--) is included for comparison. 1 1 I l I I I I I 1 I I I I 0 10 20 30 0 20 40 t Ins tlns la) 161 FIG. 2.-Normalized fluorescence signals of (a) 1 mmol dm-3 2,5-diphenyloxazole (PPO) and (b) 1 mmol dm-3 anthracene in decalin at $20 "C (-) and -130 "C (- --).The Cerenkov signal . . . . .), and the fluorescence signal assuming instantaneous singlet formation (calculated by con- volution of the Cerenkov signal) (-. -. -) are included for comparison. decreases until at the lowest temperature (- 135 "C) no growth of absorption was de- tected over 4500 ns, but there remained a yield of rapidly formed triplets amounting to approximately 50% of that at room temperature. The absorption maximum of the anthracene triplet state was observed to shift from 425 nm at room temperature to 430 nm at lower temperatures, but preliminary results indicate that all of the yieldGENERAL D1 S C U S SI ON 191 of triplet anthracene remaining at - 135 "C can be attributed to intersystem crossing from the singlet excited states which are still formed at the low temperature.The yield of anthracene radical-ions changed very little over this range of temperature, but their rate of decay became slower as the temperature was lowered until at the lowest temperature no decay was observed over 4500 ns. Thus, whereas a large fraction of the excited singlet states of solutes are still formed at lower temperatures where the viscosity is high ( -lolo cP), the initial formation of triplet excited states seems to be eliminated when the solvent viscosity is such that recombination of the molecular ions of the solute is negligible. Prof. J. Kroh and Dr. J. Mayer (Eddz', Poland) said: No experimental evidence for the initial, primary production of triplets exists either for liquids1 or for glasses.2 Particularly, investigating singlet and triplet formation of 9,lO-diphenylanthracene (DPA) in 3-methylpentane glass (3MP) we could not observe, on the microsecond time scale at 87 K, any triplet formation under conditions when singlet production was still observed. The end-of-pulse absorption spectrum of the irradiated DPA- 3MP system at 87 K is entirely due to DPA ions.The arguments of Brocklehurst as to the primary formation of triplet ion pairs are based mainly on the magnetic field enhancement ratio (R,) which never reaches the theoretical value 2. The question is, however, whether the above value could not be obtained for much stronger magnetic fields. Perhaps the plateau of magnetic effect reported in ref. (3) has a limited width and the effect increases again for higher fields.Dr. B. Brocklehurst (Shefield) (partly communicated) : Concerning the remarks of Kroh and Salmon about the initial formation of triplet states, it does not seem possible to account for all the observations at the present time. [For further discussion and references, see ref. (4).] However, let me emphasise that other mechanisms of excita- tion are not likely to interfere with the limiting enhancement ratio:excitation of singlets by Cerenkov radiation or by energy transfer from the solvent will die out within 30 ns. If my conclusions are in error, they must be due to misinterpretation of the limiting value of R,. The limiting high field value of ps, the singlet density in the ion-pair wave-function, will always be close to one half, but the zero field value does not necessarily reach one quarter : this is due to degeneracy of some of the stationary state^.^ In the simple case of one type of proton only on each centre, if l i s large, ps approaches 1/3 and R, would be only 1.5.Anything which splits the degenerate states will decrease ps and increase R,: one factor is the presence of protons with different a values on one centre, as in all the systems studied to date; exact calculations of their effect has not yet been made. Another factor would be the residence of one charge (positive) on two centres (sol- vent and solute) for a significant time. Again one can make an exact calculation for identical protons on each centre: if both times are, on average, long compared with the oscillation times, and I is large, the limiting value of ps is 5/18, giving R, = 1.8.In this connection, the results of Ceulemans, summarised below, are very interesting as a values for alkane cations are not yet known. However, the line-widths he has l G . Beck and J. K. Thomas, J. Phys. Chern., 1972, 76, 3856. J. Kroh, J. Mayer, J. L. Gebicki and J. Grodkowski, Int. J. Radiation Phys. Chern., 1976, 8, R. S . Dixon and F. P. Sargent et ul., unpublished work. B. Brocklehurst, Nature, 1977, 265, 613. B. Brocklehurst, J.C.S. Furuduy 11, 1976, 72, 1869. 433.192 GENERAL DISCUSSION obtained so far correspond to intermolecular interaction which would average out in the liquid state. Since the molecules used contain three or four types of protons, the figures just given suggest that failure to reach ps = 1/4 is not the explanation for the low values of R,, though it may account for the differences observed between alkanes (charge transfer is very rapid in cyclohexane etc.).Other possibilities have been discussed [ref. (5)] but they do not provide satisfactory explanations either. Dr. Jan Ceulemans (Leuuen) (communicated) : Brocklehurst states in his paper that little is known about the magnetic properties of alkane radical cations. I would like to point out that this situation is improving very rapidly, now. We have recently developed a method that enables one to obtain fairly pure e.s.r. spectra of alkane radical cations. The method consists in irradiating the hydrocarbon studied at 77K in a pentane matrix, to which a suitable electron acceptor, for instance 2-chloropen- tane, is added to stabilize the negative charge.A substantial number of the positive ions formed by irradiation are then trapped as radical cations of the hydrocarbon additive. The e.s.r. spectra are as such uninterpretable, at least as far as the contribu- tion of the alkane radical cations is concerned, but the alkane radical cations may be removed fairly selectively by illumination with light of suitable wavelength. Conse- quently, fairly pure e.s.r. spectra of those radical cations may be obtained by taking the difference, by computer techniques, between the e.s.r. spectra before and after removal of the radical cations. The e.s.r. spectrum of alkane radical cations appears to consist of a single un- resolved line. The width of the line is mainly due to unresolved Fermi-contact hyperfine interaction with hydrogen (or deuterium) nuclei of the radical cation and, to an important extent, of adjacent matrix molecules, the latter being due to the forma- tion of a charge transfer complex between alkane radical cations and matrix molecules. The interesting point, as far as Brocklehurst’s work is concerned, is that the total spread of the e.s.r.spectrum is rather small, (about 26 G, for instance, in the case of non- deuterated octane ions in non-deuterated pentane). Considering the fact that part of this spread is due to hyperfine interaction with matrix nuclei, it may be concluded that the hyperfine coupling with hydrogen nuclei in alkane radical cations is rather small; the unpaired electron appears to reside in a carbon-carbon orbital and not in a carbon-hydrogen orbital. The total spread of the e.s.r.spectrum varies consider- ably, depending on the alkane radical cation considered. For instance, in all cases studied so far, the width of the e.s.r. singlet appeared to be smaller for branched hydrocarbon ions, than for the corresponding (same number of carbon atoms) normal ions. A systematic study of the e x . characteristics of different alkane radical cations is currently in progress. Dr. P. W. Percival (Ziirich) said: We have been measuring muonium atom rate constants for aqueous solutions as part of our programme of muonium chemistry at the Swiss Institute for Nuclear Research.l I would like to present some of our results for comparison with the hydrogen atom data obtained by Fessenden and co-workers.First, a few words about muonium, since not everybody here will be acquainted with this exotic species. Muonium is an H-like atom consisting of a positive muon ( p + ) and an electron. Some important properties are given in table 1, and from these it can be seen that the chemistry of Mu (as predicted by atomic size and ionization potential) should be very similar to that of H. Thus muonium may be considered as P. W. Percival, E. Roduner, H. Fischer, M. Camani, F. N. Gygax and A. Schenck, Chem. Phys. Letters, 1977, 47, 1 1 .GENERAL DISCUSSION 193 TABLE 1 .-MUON AND MUONIUM PROPERTIES ~~~~ ~ ~~ Muon (positive) p+ spin 3 mass 1/9 mass of proton magnetic moment 3.2 proton magnetic moment lifetime 2.2 ,us Muoni~cm, Mu p+e- - mass 1/9 mass of H Bohr radius 0.532 A ionisation potential 13.539 eV lifetime limited by that of p+ a light isotope of hydrogen, in contrast to positronium, whose reduced mass and atomic structure are quite different.That Mu is only 1/9 of the mass of H is expected to give rise to important kinetic isotope effects, and this is just what I wish to demon- strate. Muonium can be directly detected in water,l and we have measured the first-order decay of its psr-signa12 as a function of solute concentration for various aqueous solutions. After allowance for spontaneous muon decay the " relaxation " rates were fitted to the expression T,-l = Tz(O)-l + k [ X ] and the concentration independent relaxation time T2(0) and the second-order rate constant k extracted.Some of our results are compared in table 2 with data taken from the present paper by Fessenden and from earlier publications by Fessenden, Neta and Schuler. 3--5 Large isotope effects are evident, muonium reacting both faster and slower than H, depending on the nature of the other reactant. Observations of long-lived muonium signals in pure methanol and ethanol as well as in aqueous solutions provide un- equivocal evidence for the slowness of the H-abstraction reaction by Mu. Estimates from preliminary measurements on aqueous solutions of isopropanol and sec-butanol indicate that there is a trend of reactivity in the alcohol series, as is found for H. In sharp contrast to abstraction we find that addition to a carbon-carbon double bond is faster for muonium than for H.The abstraction results can be rationalized in terms of the additional zero-point vibration energy in the transition state for the muonium case. The addition reaction results are more difficult to explain, but even in the absence of a full understanding our present experience may be applied to make informed guesses about other reactions. So, for example, we find that muonium reacts faster than H with acetone, and deduce that abstraction is not an important reaction pathway. Finally it is satisfying to note the absence of an isotope effect in the reaction with Mn04, this being consistent with the belief that this is a diffusion controlled reaction. P. W. Percival, H. Fischer, M. Camani, F. N. Gygax, W. Ruegg, A. Schenck, H. Schilling and J.H. Brewer, K. M. Crowe, F. N. Gygax and A. Schenck in Muon Physics VoZ. 111: Chemistry P. Neta, R. W. Fessenden and R. H. Schuler, J. Phys. Chern., 1971,75, 1654. P. Neta and R. H. Schuler, Radiation Res., 1971,47, 612. P. Neta and R. H. Schuler, J. Phys. Chem., 1972,76, 2673. H. Graf, Chem. Phys. Letters, 1976, 39, 333. and Solids, ed. V. W. Hughes and C. S. Wu (Academic Press, New York, 1975).194 GENERAL DISCUSSION TABLE 2 .-COMPARISON OF HYDROGEN AND MUONIUM RATE CONSTANTS (AQUEOUS SOLUTION) reactant kH/dm3 molV1 s-l k,,/dm3 mol-I s-l kH/kMu abstraction : addition : miscellaneous : methanol ethanol isopropanol sec-butanol formate (pH > 7) maleic acid (pH 1) fumaric acid (pH 1) acetone (pH 1) NO; MnOi 2.5 x lo6 t 3 x 104 :>go 2.1 x 107 < 3 x 105 :>70 6.8 x lo7 -1 X lo6 -70 1.3 x lo8 -1 x lo6 -130 1.2 x 10' (7.8 f 0.7) x lo6 16 G x lo8 (1.1 i- 0.1) x 1O1O 0.05 9 x lo8 (1.4 f 0.2) x 1O1O 0.06 2.8 x lo6 (8.7 rt 0.7) x lo7 0.03 2.4 x 1Olo (2.5 f 0.3) x lolo 1 9 x 106 (1.5 i- 0.1) x 109 0.006 Dr.D. Meyerstein (Israel) said: I just wonder as a chemist what the products of reaction of p-atoms with alcohols, double bonds and permanganate are? If the rates of reaction are compared with those of hydrogen atoms the corresponding products should be observed. Dr. P. W. Percival (Ziirich) said: As I mentioned before, the ionization potential and Bohr radius of muonium are almost exactly equal to those of the hydrogen atom, so we can really think of Mu as a light isotope of hydrogen. Thus the product of an H-abstraction reaction would be Mu-H, and that of addition to a carbon--carbon double bond a muonic free radical.Isotope effects will be important in determining reaction rates but the mechanisms are expected to remain unchanged. Unfortunately we have no direct proof. In principle it should be possible to detect and characterize the muonic free radicals formed in an addition reaction-in practice we have been un- able to do so. The experimental method used in our muonium chemistry studies involves monitor- ing the time evolution of muon spin polarization in a magnetic field transverse to the initial spin directi0n.l This is a limitation of the method, because the magnitude of the muon spin polarization is not always a good measure of the probability of existence of a muonic species.(Fessenden has to consider such effects in his paper where he seeks the time-dependence of H atom concentration in a system exhibiting chemically induced electron polarization.) For muons in diamagnetic molecules the muon spin polarization precesses at the nuclear Larmor frequency. This can be thought of as very low resolution n.m.r. Four frequencies are theoretically predicted for muonium, but only two of these are low enough to be experimentally accessible, and in suffi- ciently low fields they become degenerate. Following our earlier analogy we can consider this situation as low-field e.s.r. The muonium theory also applies to muon- substituted free radicals, but because the muon-electron hyperfine coupling is smaller in a radical the splitting of the two lower frequencies will be greater.Our failure to observe such characteristic radical frequencies could be due to one or more of the fol- lowing causes. J. H. Brewer, K. M. Crowe, F. N. Gygax and A. Schenck in Muon Physics VoZ. 111: Chemistry and Solids, ed. V. W. Hughes and C. S. Wu (Academic Press, New York, 1975).GENERAL DISCUSSION 195 (i) Muon spin polarization may be lost in the Mu + radical reaction as a result of a reaction rate less than, or equal to, the muonium freq~encies.l-~ Our direct rate measurements on maleic acid solutions, however, show that for this system, at least, reaction is fast enough. (ii) Muon spin polarization may be lost as a result of spin correlation effects in radical pairs (such as those discussed by Brocklehurst in his Paper, and Fessenden in his) or in the Mu -+ radical reaction.2 (iii) The probability of muonium existing long enough to react to radical may be influenced by the high reactant concentrations necessary to overcome the potential " depolarization " mentioned in (i).Such inhibition is well known in positronium chemistry4 and we have experimental evidence that the muonium fraction is strongly reduced in large concentrations of sodium nitrate. Dr. D. Meyerstein (Israel) said: The rate of the reaction must be lower than the literature value of k = 7.4 x 10l2 dm3 mol-1 s-l. This rate of reaction must obviously be smaller than koH- +Haof = 1.4 x loll dm3 mo1-I s-l. (It is known by electrochemists that the method used by I. Brdicka for the determina- tion of the former rate constant is not applicable for reactions which approach the diffusion controlled limit.) Therefore, naturally 5 x mol dm-3 CrOf- has nearly no effect on the con- centration of H30f in the spurs.Thus the conclusions concerning the mechanism of formation of Cl, in the radiolysis of neutral solutions have to be reconsidered. Dr. R. M. Sellers (CEGB, Berkeley) said: The determination of radiation chemical yields by the addition of solutes to scavenge the radiolytic intermediates etc., followed by measurement of G( - solute) or G(product) is a familiar method, widely employed by radiation chemists. Where the scavenging reaction is reversible, however, con- siderable caution is required in equating G(-solute) or Gbroduct) with G(intermedi- ate). This is because at equilibrium some of the intermediates may remain un- scavenged, and it is important either to establish that the fraction of species which remain unscavenged is negligible, or to make an appropriate correction.In the case of scavenging of protons by chromate ion, eqn (1) as described by Pikaev et al., it is not immediately apparent that the chromate does indeed scavenge all, or essentially all, of the protons. At the pH values employed (6.5-6.9) the relative concentrations of CrO$- and HCr04 before radiolysis are in the range 1 : 1 to 4: 1, since pKa (HCrO,) = 6.5 (at 25 "C and I = O).5 Follow- ing the radiation pulse H+ ions are formed with a concentration of - J. H. Brewer, F. N. Gygax and D. G. Fleming, Phys. Rev. A., 1973,8,77. P. W. Percival and H.Fischer, Chem. Phys., 1976, 16, 89. W. E. Fischer, Helu. Phys. Acta, 1976, 49, 629. S. J. Tao, Appl. Phys., 1976, 10, 67. (a) Chemical Society Special Publication No. 17 (The Chemical Society, London, 1964); (b) Chemical Society Special Publication No. 25 (The Chemical Society, London, 1971).196 GENERAL DISCUSSION mol dm-3 at the doses employed by Pikaev et al., and these react according to eqn (1). The equilibrium is presumably rapidly established (the value reported for kl is 7.4 x 10l2 dm3 mol-l s-l;l this value seems absurdly high, and is probably in the range 1O1O - 1011 dm3 mo1-I s-l). The percentage of €if ions not reacting can be calculated knowing pKa (HCrOr), the initial concentration of chromate, and the amount of H+ formed (Le., the dose).The results of such calculations are shown in the table, and indicate that under Pikaev et aZ.’s conditions a considerable fraction of TABLE.-~ALCULATED VALUES OF THE PERCENTAGE OF HYDROGEN IONS UNREACTED WITH CHROMATE AT EQUILIBRIUM [HI +puls//mol dm-3 [CrV1]total/m~l dm-3 % H+ unreactedSb 10-5 9 5 x 10-5 20 10-5 10-5 46 10-5 5 x 10-5 2.5 1 0 - 4 1 10-5 1 0 - 4 1 a Concentration of H+ formed during pulse. A concentration of mol dmW3 corresponds to a dose of 2 x loi7 eV and G(H+) = 3. b defined as ( [H+lequil* - [H+li”it.) x 100 where [H+Iequil. and [H+Ilnit. are the concentrations [H+lgulse of H+ when eqn (1) has reached equilibrium, and before radiolysis respectively. No account is taken of reactions such as (2) and H+ + OH- + HzO. the H+ values are not scavenged at the lower [Cr”’] employed, but that there is practically complete scavenging at the higher concentrations. Dimerisation of Her04 to give dichromate ion, eqn (2), is too slow kinetically to play any part on microsecond time scales at the pH values used.2 It is also important in experiments involving the bleaching of an absorption to allow for other species absorbing at the monitoring wavelength.In the case of chromate these will be, (i), the hydrogen chromate ion, formed in reaction (1) and, (ii), the small yield of CrV ions produced by reduction of the Crvl by the H and eLg not scavenged by the O2 present in the solutions. CrV certainly absorbs at the monitor- ing wavelength (370 nm)394 and although present in small yield (-0.5), may make a relatively large contribution if its extinction coefficient is large.Bleaching of the chromate absorption by reaction with H202 and 0, is presumably too slow kinetically to be of importance on the timescales of the experiments described. The radiolysis of water produces not only hydrogen ions, but also hydroxide ions with GOH- = 0.7.’ These react rapidly with HCr0; according to (3) with k3 = -6 x lo9 dm3 mo1-l s-l at 20 “C and I = 0.2.(j Under the conditions of Pikaev et aZ.’s experiments this equilibrium will be rapidly established, and therefore the value of G(H+) measured can, at best, only be equated with (GH+ - GOH-) HCr04 + OH- + CrOi- + HzO. (3) I. Brdicka, Z . Elektrochem., 1960, 64, 16. J. R. Pladziewicz and J. N. Espenson, hzorg. Chem., 1971, 10, 634. J. H. Baxendale, E.M. Fielden and J. P. Keene, Proc. Roy. SOC. A., 1965,286,320. R. M. Sellers, Ph.D. Thesis (University of Leeds, 1972). J. Rabani, M. Gratzel, S . A. Chaudhri, G. Beck and A. Henglein, J . Phys. Chem., 1971, 75, M. Eigen and G . Schwarz, unpublished data quoted in M. Eigen, W. Kruse, G. Maass and L. de 1759. Maeyer, Progr. Reactbn Kinetics, 1964, 2, 285.GENERAL DISC US SI ON 197 Prof. A. K. Pikaev (Moscow) said: Taking into consideration the criticism of Brdicka's value' of rate constant of reaction (1) (7.4 x 10l2 cm3 mol-l s-l) during the Discussion we performed the redetermination of kl H30+ + CrOi- --+ HCr0; + H,Q (1) by the method which has been used in our work for the measurements of rate constants of reactions between H30+ and pH indicators. As follows from our preliminary data, k, is -5 x 10" dm3 mol-' s-l.From this value and pK, for respective equili- brium (6.5 at 25 "C) it is easy to find that the rate constant of the reverse reaction (2): HCr04 + H20 --j H30+ + CrOi- (2) is -1.6 x lo5 s-l, i.e., the half-life for the dissociation of HCr04 is -4.4 ps. Hence the percentage of hydrogen ions unreacted with chromate ions is small. Owing to such a value of kl it is necessary to introduce small changes in our inter- pretation of experimental data on G'( - CrOi-). Chromate ion concentrations of -5 x mol cm-3 react with H30+ predominantly in the bulk of solution (the reactions in spurs at respective times are almost completed). Under these conditions G'(-CrOi-) z G(H,O+) + GH, i.e., G(H30+) w 3.1-3.2. This value coincides within the limits of experimental error with recent value by Klever et aL2 The values of G'(-CrOi-) at CrOi- concentrations of mol dm-3 seem to be slightly lowered because of the decrease of CrOi- concentration during the pulse and the competition of reactions (1) and (3) H30+ + OH- -+ 2H20.(3) The contribution of other radiolysis products has been taken into account. In the calculations of G(-CrO;-) the formation of HCr04 has been taken into considera- tion by using the E value at 370 nm, which was E(CrOi-) - E(HCrOq) = 3.4 x lo2 m2 mol-' at 25 "C. The contribution of the absorption of CrV which can be formed with G w 0.7 under the conditions employed is very small since the values of G'- (-CrOi-) for both aerated and oxygenated solutions are the same within the limits of experimental error.The contribution of reaction (4) is negligible HCr04 + OH- -+ CrOi' + H20 (4) since the half-life for this reaction under the conditions employed is 2 1 2 ps, i.e., much more than the pulse duration. We have measured G(-CrO$-) immediately after the 2.3 ps pulse. Dr. D. R. McCracken (CEGB, Berkeley) said: Much of Pikaev's argument on the mechanism of CI, production in neutral solution hinges on the value of the rate con- stant for reaction (1). CrOi- + H30+ HCrOc + H20 (1) The literature value' of 7.4 x 10l2 dm3 mol-' s-' is clearly wrong, as can be seen by comparison with the values of other diffusion-controlled protolytic reactions I. Brdicka, 2. Elektrochem., 1960, 64, 16. H. Klever, L. Toth, B. Wagner and D. Schulte-Frohlinde, Ber.Bunsenges. phys. Chem., 1976, M. Eigen, W. Kruse, G. Maass and L. DeMaeter, Progr. Reaction Kinetics, 1964, 2, 285. 80, 1265.198 GENERAL DISCUSSION (lolo - 1011 dm3 mo1-l s-l). In particular, the rate constant for the reaction between the two high mobility species H+ and OH- is 1.4 x 10" dm3 mol-1 s-l, correspond- ing to an encounter distance of 8 A, and that for reaction (2) (2) k Ht + SO:- = HSO, is 1.0 x loll, and we would not expect much difference between the rates of protona- tion of chromate and sulphate. A rate constant of 7.4 x 10l2 requires either an anomalously large diffusion coefficient or an enormous encounter distance. Taking a reasonable value kl = 1 x 1011 dm3 mol-1 s-l, then for the highest chromate concentration used (lov4 mol dmW3) the half life of protons is about s and competition with spur reactions involving protons is impossible.It follows that G(H02) in the aerated chromate system will be -0.5, and there will always be a con- tribution to G(H,O+) of 0.5 from the dissociation (3) H02 --f H+ + 0,. G(-CrOi-) = GHsO+ + GH - GOH- +fG; wherefG, is the removal due to reaction of electrons with CrOi- rather than 02, and the yield of OH- is subtracted to allow for the fast reaction (4)' (3) The removal yield of chromate is, therefore OH- + HCr04 --f CrOi' + H20. (4) For a chromate concentration of mol dm-3 this gives GHsO+ = 3.8 subject to fG,- = 0.7. If G(H) were 0 then GH30+ would only be 3.1. The experimentally undetectable effect of temperature on both this system and the chloride system is in accord with other studies2 on the effect of temperature on funda- mental yields.Below 100 "C changes are small and only detectable by steady state methods. The decrease in the extinction coefficient of C1, radical ion with tempera- ture explains the anomalous apparent decrease of G(C1;) with temperature observed by Ogura and Hamill.3 The lack of effect of chromate addition on G(C1,) shown in fig. 6 can be explained by the fact that there is insufficient chromate to compete with spur reactions of C1- (taking kl = 1 x loll dm3 mol-' s-'), or by mechanistic complications such as the reaction of a chloride intermediate,4 e.g., chloride/hydroxyl radical complex ClOH - as follows : OH + C1- + ClOH- ClOH- + HCr04 = C1+ CrOi- + H20 ClOH- + H30+ --+ Cl.In any case, one cannot argue that H30+ is not involved in the process of C1, forma- tion. It can be seen in this system that following a large initial decrease of G(-CrO:-), the sum of G(C1,) + G(-CrOi-) is nearly constant over a ten fold factor of C1- concentration. This is given in the following table compiled from the information in graphs 6 and 7: M. Eigen, W. Kruse, G. Maass and L. DeMaeter, Progr. Reaction kinetics, 1964, 2, 285. (a) C. H. Hochanadel and J. A. Ghorrnley, Radiation Res., 1962, 16, 653; (6) I. Balakrishnan and M. P. Reddy, J. Phys. Chem., 1972, 76, 1273; (c) R. W. Matthews, Radiation Res., 1973, 55, 242. H. Ogura and W. H. Hamill, J. Phys. Chem., 1973,77,2952. ( a ) M . Anbar and J. K. Thomas, J. Phys. Chem., 1964,68,3829; (b) G. G. Jayson, B.J. Parsons and A. J. Swallow, J.C.S. Faraday I, 1973,69, 1597.GENERAL DISCUSSION 199 KCl/mol dm-3 G(C1;) + G(-CrOf-) 0 0.05 0.1 0.2 0.4 0.5 0.6 4.0" 3.3 3.18 2.98 2.95 3.08 3.17 * See earlier discussion concerning this value. The conclusion could be drawn that both:species C1, and HCr04 have at least in part a common precursor, but the numbers are rather puzzling. The system appears to be more complex than the mechanism suggested in the paper. The effect of addi- tives on G(C1;) is not well understood and there is disagreement in the literat~rel-~ over mechanisms and magnitudes. Even the value of &(C12)340 = 1.25 x lo3 m2 mol-1 2 ( n ) used by the authors of the paper to measure G(C1;) is disputed since others2@) measured E = 0.88 x lo3 m2 mol-'.Prof. A. K. Pikaev (Moscow) said: As noted in our answer to comments by Sellers, we redetermined the value of k,. It is -5 x loll dm3 mol-1 s-l (such a value may be caused by proton tunnelling). Hence, for the highest chromate concentration used s and competition of reaction (I) with reactions of hydrogen ions in spurs is partially possible. The equation for G(-CrOi-) given by McCracken in his comments is wrong, For example it is necessary to exclude GoH- from the right hand side of the equation because the half-life for reaction (4) is >, 12 ps, i.e., much more than the pulse duration (for further discussion see our answer to comments by Sellers). Our determination of yields of H30+ ions at different temperatures adds the literature data [see ref. (2a-c) in McCracken's comments] on the temperature dependences for the other primary products of water radiolysis.The hypothesis on [H30+ . . . OH] was primarily developed for the interpretation of the dependence of G(C1,) in neutral aqueous solutions of alkaline metal chlorides on the nature of the alkaline metal ~ a t i o n . ~ Obviously it is impossible to explain this dependence only by the hypothesis on the formation of ClOH- as an inter- mediate product of reaction between C1- and OH. Besides the analysis of the shape of observed optical absorption spectrum of C1, shows that it belongs to one species (see also our answer to question d by Symons). mol dm-3) the half-life of hydrogen ions is -2 x Prof. M. C. R. Symons (Leicester) said: (a) It seems to me that the life-time of the hydrogen-bonded adduct H30 + - - -OH is unlikely to be appreciably longer than that of H,O+ in view of the high mobility of protons in water.Am I right in thinking that the mechanism proposed in eqn (5)-(7) require a long life-time for this adduct? (b) Monoprotonated chromate has an absorption spectrum with band maxima H. Ogura, and W. H. Hamill, J. Phys. Chem., 1973,77, 2952. (a) M. Anbar and J. K. Thomas, J. Phys. Chem., 1964,48,3829; (b) G. G. Jayson, B. J. Parsons and A. J. Swallow, J.C.S. Faraday I, 1973,69, 1597. M. M. Fisher and W. H. Hamill, J. Phys. Chem. 1973,77, 171. S. A. Kabakchi, A. A. Zansokhova and A. K. Pikaev, DokIady Akad. Nauk S.S.S.R., 1975, 221, 1107.200 GENERAL DISCUSSION at 340 and 460 nm, including a considerable absorption at 370 nm.l Do you make allowance for this absorption and have you looked, for example, in the 460 nm region to see if you can detect HCr04 directly? (c) The HCr04 ion is certainly present in neutral solution prior to irradiati0n.l Is allowance made for this in your calculations? ( d ) In studies of rigid aqueous glasses containing chloride we have found e.s.r. spectra characteristic of ClOH- as well as C1; after exposure to It is suggested that ClOH- could be a significant intermediate in the forma- tion of C1, though, of course, we would not be able to detect C1 atoms by e, s.r. spectroscopy . (e) Finally, a general comment. It is not always appreciated that one should think about dry anions as well as dry electrons in studies at very short time intervals, e.g., we find that when 0,- is gener- ated from dissolved oxygen in alcoholic media at 4.2 K, no e.p.r.signal is obtained. A characteristic signal, however, grows in on warming and is clearly shown at 77 K immediately after radiolysis. Prof. A. K. Pikaev (Moscow) said: (a) It is only an hypothesis to stress the high rate of reaction of H20+ with H20 in liquid water, but it gives an opportunity to explain the considerable effect of the nature of alkaline metal cation on G(C1,) in neutral aqueous solutions of C1- ions. In our view, the life time of [H30+ . . . OH] may be more than 10- s. Cations which have a destructuring action on water decrease the life time. The respective experimental data and their explanation were described in our previous paper.3 (b, c) Yes, we have observed the formation of an additional band in the optical absorption spectrum of the solution after the pulse.It has been found4 that the maximum of this band is at 353 3 nm. Obviously this band belongs to the acid form of chromate. Her04 has been taken into account in the E value used. (d) There are indirect data5 on the existence of ClOH- as an intermediate product of C1; formation in liquid neutral aqueous solutions of C1- ions. It is possible to suggest that in our case this species may be formed as a result of reaction: [H30+ * OH] + C1- -+ ClOH- + H30f. However we have carried out the study of Cl,, optical absorption spectrum in neutral 3 mol dm-3 aqueous solutions of LiCl, NaCl and CsCl after microsecond pulses. We have not found any inflection points in this spectrum (after substraction of e z spectrum).The optical absorption band of C1, has a Gaussian shape in the range of 400-320 nm and a Lorentzian shape in the range of 320-270 nm. The shape of the band does not depend on the temperature within the range of 20-90 "C. These data show that in our experiments we have observed the optical absorption spectrum belonging only to one species. Dr. A. Singh (Pinawa, Manitoba) (communicated): (i) What is the lifetime of N. Bailey, A. Carrington, K. A. K. Lott and M. C. R. Symons, J. Chem. SOC., 1960,290. I. Marov and M. C. R. Symons, J. Chem. SOC. A , 1971,201. S. A. Kabakchi, A. A. Zansokhova and A. K. Pikaev, Dokludy Akad. Nuuk S.S.S.R., 1975, S. A. Kabakchi, A. A. Zansokhova and A. K. Pikaev, Khimiyu vysokikh energii, 1974,8,255.G . G. Jayson, B. J. Parsons and A. J. Swallow, J.C.S. Furaday I, 1973, 69, 1597. 221, 1107.TTTTTTGENERAL DISCUSSION 201 (H30+ . . . OH)? Can (Na+ . . . OH), (K+ . . . OH) also play a significant role in aqueous solutions? (ii) The concentration of chromate ions used in these studies is quite low (5104 mol dm-3). It seems unlikely that at these concentrations chromate will be an efficient scavenger of H& in spurs. Has any work been done with higher concentra- tions of chromate ions? (iii) In this paper, you have not discussed the reaction H20* + Cl; + H20 + C1+ e; (1) though you referred to it briefly a few minutes ago (during discussion). Grossweiner and Matheson (J. Phys. Chem., 1957, 61, 1089) observed the forma- tion of dihalide ions (X2-) when solutions containing halide ions (X-) were flash photolysed, which suggests that the following reactions take place : X i + hv + X + e; X + X i 4 XZaq.(2) (3) The lifetime of H20* is s, as discussed earlier. So, analogous reactions on energy transfer from H20* should occur in radiolysis of concentrated aqueous solutions. H20* + X; --+ H20 + X + e; (4) (5) X + X i + XTaq. It is relevant to mention that the reactions (4) and (5) seem to have been over- looked by most other workers, e.g., see discussion of X;sol formation in Hunt’s recent review [J. W. Hunt, Advances in Radiation Chemistry, ed. M . Burton and J. L. Magee (John Wiley, New York, 1976), vol. 5, chap. 3, p. 1861. Prof. A. K . Pikaev (Moscow) (communicated): I would like to comment on some communicated remarks by Singh.(i) As in our answer to Symons, the life-time of [H30+ - * OH]. in our opinion, may be more than s. We believe1 that [H30+ * OH] is formed in reaction (1): H20+ + H20 --+ [H30+ * . * OH]. (1) Because of it the species [Na+ * * * OH}, [K+ * * OH] can not appear and play any role in the radiolysis of aqueous solutions. Besides, H30+ and OH are connected by a hydrogen bond. There is no case for the species proposed by Singh. (ii) In our paper we have used Brdicka’s value2 of rate constant of reaction (2) (7.4 x 10l2 dm3 mol-l s-’) for the interpretation of H30+ + CrOi- -+ HCr04 + H20 (2) G( - CrOi-) values measured. However, after the Discussion we redetermined the value of k,. It is -5 x loll dm3 mol-l s-l (see also our answers to comments by Sellers and McCracken).Hence for the highest chromate concentration employed niol dmd3) the half-life of hydrogen ions is -2 x lo-* s and a Competition of reaction (2) with reactions of hydrogen ions in spurs is still partially possible. It is difficult to use higher concentrations since chromate has strong optical absorption (we S . A. Kabakchi, A. A. Zansokhova and A. K. Pikaev, Doklady Akad. Nauk S.S.S. R . , 1975, 221, 1107. ’ 1. Brdicka, Z. Efektrochern., 1960, 64, 16.202 GENERAL DISCUSSION have measured the decrease of CrOi- concentration as a result of the action of the pulse). Besides, in this case the contribution of the reaction between e i and CrOi- to the values of G(-CrOi-) will be more considerable. However, such work is now planned.(iii) Another possible hypothesis is the participation of H20* in the formation of C1,. To explain the considerable effect of the nature of alkaline metal cation on G(C1,) it is necessary to suggest that hydrogen bonding favours energy migration (as has been proposed 15 years ago by Voevodsky).l Earlier2 in our laboratory the formation of C1, was studied in pulse radiolysed neutral concentrated aqueous solutions of LiCl (up to -14 mol dm-3). To interpret the results obtained it was suggested that one of the C1 (and then C1,) sources was the reaction (3): (Cl-)* + OH -+ C1 + OH-. (3) The precursors of (Cl-)* seem to be subexcitation electrons3 and/or excited water. It has been worked out that the yield of the precursors is -1.6. Prof. A. Henglein (Berlin) said: A comparison may be made between the electrochemical properties of H and its electron transfer reactions in aqueous solutions.Electron transfer can be observed in reactions of H with metal ions. The reaction H + Ag+ +Ago + H+ can be observed by recording the simul- taneous increase in Ago absorption at 360 nm and the increase in conductivity of the solution. Since the redox potential of the system Ag+/Ag" is about - 1.8 V in aque- ous solution and that of H+/H not very much lower, the high observed rate constant of 3.1 x 1O1O dm3 mol-1 s-l cannot be explained by a simple electron transfer. One has to postulate an intermediate AgH+ that rapidly dissociates within s. Oxidizing effects of the H-atom have hitherto been observed only if H+ ions partici- pated.The oxidation of ferrous ions, for example, proceeds via H + Fe2+ --f FeH+ FeH+ + H+ --+ Fe3+ + H2. However, in the case of a substrate of more negative redox potential, such as of ab- normal valency states of metals, the oxidation of the metal ion by H-atoms without participation of protons is possible. Example: Ni+ + H -+ Ni2+ + H- (followed Several reactions of this type, which occur with rate constants of the order of lo9- by H- + H20 + H2 + OH-). dm3 mol-l s-l have recently been observed in our laboratory. Dr. A. Bernas (Orsay) (communicated) : The experiments Henglein has just men- tioned pertain to rigid matrices, y-irradiated at 77 IS, and I doubt that our results can apply to the liquid solutions Symons is interested in. At any rate, our solid phase experiments were designed to identify negatively charged radiolysis products ; the technique used was to bleach stabilized electrons and thus stimulate neutralization luminescence after the y-irradiation was com- pleted.In such conditions, and whatever the emitting centre, the luminescence V. V. Voevodsky, Trudy 11 Vsesoyuznogo sovestchaniya PO radiatsionnoi khimii (Izd-vo Akad. T. P. Zhestkova and A. K. Pikaev, Khim. vysokikh energii, 1975, 9, 165. R. L. Platzman, Radiation Res., 1955, 2, 1. Nauk S.S.S.R., Moskva, 1962), p. 102.GENERAL DISCUSSION 203 excitation (or " stimulation ") spectrum giving the neutralization luminescence in- tensity as a function of the bleaching wavelength, reflects an electron photodetachment efficiency and characterizes trapped negative species.From y-irradiated crystalline ices, glassy alkaline ices and glassy ethanol samples, stimulation spectra have shown, besides the intense solvated electron band, an extra band peaking at 1050 nm. It was first suggested that the latter band might be corre- lated with H-l and more recent results on ethanol + water mixtures seem to sub- stantiate this interpretati0n.l In effect, it is clear that, if an electron is photodetached from a negative species X-, the latter can be regenerated only when X is matrix-trapped. After a complete bleach- ing of the 1050 nm band, its restoration has been proved to be feasible only for some EtOH + H20 mixtures where H atoms are known to be stabilized at 77 K, whereas trapped H(HJ are not observed in either pure component.Moreover, the intensity of the regenerated stimulated band is found to be maximum for the mixture composi- tion which leads to the highest H, yield.2 One should add that the red limit of the H- stimulated band is close to the elec- tron affinity value of the isolated H atom; this would imply that, contrary to expecta- tion, the solvation energy of H- in these rigid matrices is quite low. Dr. D. Meyerstein (Israel) said: I would like to point out the kinetic implications of the results described in the paper of Toffel and Henglein. The observation that the overvoltage required both for the reduction and oxidation of H atoms and *CH3 radicals is >2 V implies that similar activation energies are required also for redox processes involving these radicals which follow the outer sphere mechanism.Thus due to the major rearrangements required for the solvation shells of the particles involved, all redox reactions of H atoms and *CH3 radicals proceed via a " bridged " mechanism. A possible exception might be reactions of a very strong oxidant or reductant with these radicals, e.g., Cd$ + H -+ Cd,2,+ + H i where the free energy gain is >2 V. However, the fact that Henglein et al. (personal communication) do not observe CdH+ in the latter reaction might also be due to the short life time of the CdH& complex. This observation is correct for many other redox reactions involving free radicals. Thus the free energy gain in the reaction OH + Fe(CN)& -+ OH,, + Fe(CN):a, is about 40 kcal mol-l. However, the hydration energy of OH- is about 110 kcal mo1-1 and thus an enormous rearrangement of the hydration sphere of the OH radical prior to the electron transfer is required. The observation that the rate of this reaction approaches the diffusion controlled limit, though a simple outer sphere reac- tion without a large activation energy due to solvent rearrangement, would be endo- thermic suggests that a short-lived intermediate is formed during the reaction.It is difficult to describe the exact chemical nature or to estimate the life time of this inter- mediate, though intermediates similar to those observed in reactions of halides with the hydroxyl radical but with shorter lifetimes can be postulated. The observation of Zehavi and Rabani that the rate of reaction (2) OH + H,Fe(CN);- -+ OH- + H,Fe(CN); (2) is lower by one order of magnitude than that of reaction (1) can be explained by a lower stability for the intermediate involved in reaction (2).A similar explanation might be T. B. Truong, Chem. Phys. Letters, 1975, 35, 426. A. Bernas and T. B. Truong, Canad. J. Chem., in press.204 GENERAL DISCUSSION given to the very low rate of the reaction 0' + Fe(CN)%' 3 20H& + Fe(CN):-, though the free energy gain in all these reactions is similar. Similar arguments might explain the observation of Cohen and Meyerstein (J.C.S. Dalton, in press) that the reduction of Co(NH3)3,+ by CH20H radicals does not follow a simple outer sphere mechanism. This reaction has a free energy gain of about 20 kcal mol-l when written as Co(NH3);' + CH,OH -+ Co(NH3);' + CH20 + H30&.(3) However, it is endothermic if the product is the partially hydrated CH20H+ which would be formed in a simple outer sphere electron transfer reaction. On the other hand, the reaction which has a similar free energy gain, might follow the outer sphere mechanism as the required rearrangements are much smaller. Thus it is suggested that many of the redox reactions of free radicals which are commonly considered as outer sphere electron transfer reactions cannot, due to the major solvation shell rearrangments required, follow this mechanism. Instead, it has to be concluded that some kind of short-lived intermediates must be involved in these reactions. The nature and stability of these intermediates might have a major influence on the kinetics of redox reactions of free radicals.Prof. A. Henglein (Berlin) said: Redox reactions in which H+ or OH- are con- sumed or formed in aqueous solution will always require a lot of free activation energy (or overpotential at an electrode) because of the strong structural changes in the aqueous medium. Such reactions will generally not be fast unless the free energy of reaction is large. Protons are often formed or consumed in redox reactions of free radicals in aque- ous solutions. Such reactions will often occur in two steps: Addition of the radical to the substrate and subsequent reaction of the complex with protons or with water. As it has been shown for CH3 in table 1, several protentials for oxidation and reduction may be attributed to an organic free radical, depending on whether it reacts with the substrate in a single electron transfer step or via the two steps mentioned above.A free radical may thus act as a strong reductant towards certain substrates, and as an oxidant towards others. Dr. C. von Sonntag (Miilheirn) said: With regard to Chapiro's paper, the species C2H2* is, in fact, either triplet or singlet (ground state or excited state) acetylene. It is difficult to see how the radical scavenger styrene prevents acetylene formation from such a precursor. Prof. K.-D. Asmus (Berlin) said: I have two questions for Schulte-Frohlinde. In 2,4- and 2,6-dimethoxybenzoic acid the 1 -position is also activated. Addition of OH- radicals should lead to a HO COOH \ / C / \ configuration. What kind of information is available about the stability of this con- figuration with respect to decarboxylation? Is the stability affected by the ionization of the carboxyl group, i.e., -COOH + -COO- + HA?GENERA L D I S CUSS ION 205 The rate constant for the OH* radical addition at a certain carbon atom is increased if the site of attack is activated by suitable functional groups, e.g., methoxy groups.Multiple activation does not, however, seem to result in any additional enhancement of the rate of reaction as compared to singly activated centres. What is to be expected if multiple activation occurs simultaneously through groups of opposite effect (e.g., electron donating and electron withdrawing groups) so that the overall effect mathematically cancels? Will the rate of reaction be the same as for a molecule without any activating group or will the rate be enhanced as for singly and multiply activated systems described in this paper ? Prof.D. Schulte-Frohlinde (Miilheim) said: In our e.s.r. work with methoxylated benzoic acids we have not found any indication for decarboxylation so far. Instead we observed that the OH adduct " " y q 0 C H 3 HO COO' H30 Q0CH3 coo0 - OH' + OCH, OC H, eliminates OH- leading to radical zwitterions. OH- elimination is observed also in the case of OH adducts of anilines and phenolates. Under preparative conditions, which means higher concentrations of starting material and higher irradiation doses, we observed with 2,4,5-trihydroxybenzoic acid strong elimination of COz. I would expect that in benzenes with electron donating and electron withdrawing substituents certain positions are respectively activated or deactivated and that there should be a cancellation of the effects if they are acting on the same positions.How- ever, the problem is rather complex since the same substituents may have more than one mechanism by which activation or deactivation is produced. Dr. D. Meyerstein (Israel) said : The observation that the rates of reaction of OH radicals with the 1,2,3- and 1,3,5-trimethoxybenzenes are equal and not much higher than that with benzene is of interest in view of the high selectivity of the position of attachment of the OH radicals. If it is not assumed that the methoxyl groups have a large deactivating effect on the meta positions it has to be concluded that a short-lived intermediate complex between the OH and the aromatic compound is formed prior to the addition to a specific site.The high selectivity would be attributed in this case to the rate of rearrangement of this intermediate complex. The much lower reactivity of 00- radicals towards aromatic compounds could be explained by the lower stability of the intermediate complex. It is also plausible to suggest that the fact that H atoms and OH radicals abstract with a high selectivity the whydrogen to a functional group of an aliphatic compound is not due only to resonance stabilization of the product radical but also to the forma- tion of an intermediate complex between the H/OH radicals with the low lying or- bitals of the functional group. Prof. D.Schulte-Frohlinde (Miilheim) said: The rate constants for OH addition to benzene and to several methoxybenzenes are indeed not very different. The k values vary between 5 and 8 x lo9 dm3 mol-1 s-l. The reason for the high selectivity may therefore be found in the formation of an intermediate. Such an intermediate could be a n-complex with polar character which then converts to the observed a-complex.206 GENERAL DISCUSS ION However, the rate constants are smaller than diffusion-controlled. For comparison the rate constants for the addition of OH to phenols and different methoxylated phenols are 2 to 4 times higher (P. O'Neill and S. Steenken, Ber. Bunsenges. phys. Chem., in press). Thus a part of the selectivity may be due to an activated process. Furthermore, the k values are not exactly constant, which also indicates some differ- ences in the activation processes.I don't think that our results demonstrate whether or not an intermediate exists. Prof. L. Dorfman (Columbus) said: Since submitting the manuscript, several addi- tional experiments bearing on two points, the reactivity with water and the formation process, have been carried out.* The additional data for the reactivity of beiizyl cation with water in DCE indicate the following. The temperature dependence of the rate constant for this reaction has been determined. The observed first order rate constant [the term in parentheses in eqn (6)] decreases with increasing temperature. k,, determined as explained in the text, decreases from 1.8 x lo7 dm3 mol-l s-' at 25°C to (0.8 &- 0.2) x lo7 dm3 mol-1 s-l at 73.5 "C.The activation energy is roughly - 1 kcal mol-l. Since K, decreases with increasing temperature, the value18 being 0.54 dm3 mol-l at 25 "C, and 0.6 dm3 mol-l or possibly slightly higher at 5 "C, the dimer concentration will decrease with increasing temperature. The boiling point of DCE, incidentally, is 83.5 "C. We can- not, however, determine an accurate temperature coefficient for kd since the temperature coefficient of K, is not known with sufficient accuracy. Several additional experiments on the formation process have also been carried out. These were designed to elucidate the nature of the portion of the rate curve during the pulse, which, as indicated, contains a contribution from a solvent species. The magnitude of this contribution of the solvent species was estimated, thus allowing us to define the extent of any contribution to carbonium ion formation by a non- diffusive process such as resonance charge transfer at room temperature. These experiments involved pure DCE, as well as two separate solutes.A forma- tion rate curve, such as the one in fig. 4, was corrected for the solvent absorption. It was assumed there was no change in the decay rate of the solvent species when solute was added. In this case we can account for as much as 0.9 of the observed optical density at the end of the pulse without invoking additional formation process. Alternatively, we may include in this estimate the effect of decay of the solvent species. This was done from observations of the effect of solute upon the absorption of the solvent species at a wavelength of minimal overlap.We conclude that at least 0.7 (and probably more) of the observed optical density at the end of the pulse may be accounted for by the normal bimolecular formation process plus the absorption of solvent species. In a subsequent experiment, carried out since the Discussion, we have observed the formation of p-terphenyl cation radical in DCE at 960 nm. At this wavelength the absorption of the solvent species is negligible. The formation curve observed indicates no evidence of a fast formation process during the pulse, and all the formation is accounted for by the normal bimolecular processes. Prof. A. K. Pikaev (Moscow) said: Recently we have used the method of pulse radiolysis for preparation of fluorinated benzyl carbanions in tetrahydrofuran and study of their properties.' It has been found that irradiation of tetrahydrofuran soh- See I.P. Beletskaya, G. A. Artamkina and A. K. Pikaev, Izuest. Akad. Nauk. S.S.S.R., ser. khim., 1976, 2403; A. K. Pikaev, G. A. Artamkina and I. P. Beletskaya, Doklady Akad. Nauk. S.S.S.R., 1977, 232, 634. * References as in Dorfman's paper.GENERAL DISCUSSION 207 tions of (C,F,CH,),Hg and (CF,C,H,CH,),Hg leads to the formation of respective radicals RHg and carbanions R-. We have measured optical absorption spectra of these species [Ibm,, for RHg and R- are respectively at 320 and 360 nm in the case of (CF3C6H4CH2),Hg and 325 and 350 nm in the case of (C6F5CH,),Hg]. The half- lives of carbanions are 1.3 x s for C6F,CHt, CF,C6M,CHt and C6H5CHF respectively.This coincides with pK, values for re- spective hydrocarbons (they are -25, 28 and 35). Alcohols and water react with fluorinated carbanions more slowly than with benzyl carbanion. In the presence of NaBPh,, ion pairs R-Na+ are formed. Their half-lives are higher by approximately one order of magnitude. In addition to the data of Dorfman, our results show the great power of the pulse radiolysis method for the study of carbanions which play an important role in organic chemistry. 3.5 x low6 and <2.3 x Prof. L. Dorfman (Columbus) said: I should point out that we have, of course, a substantial amount of data on the reactivity of the benzyl carbanion. These data have not been presented in this Faraday paper since they were published two and three years ago [ref.(1) and (2) of this paper]. But it may, nevertheless, be appropriate for me to refer to the absorption spectra of benzyl cation and benzyl anion [see ref. ( 5 ) of this paper for a summary]. They are very nearly identical, with A,,, at 363 and 362 respectively. With regard to the reactivity of PhCH,’ and of PhCH,, Na+ in protonation reactions, our data indicate that the latter species (the ion-paired form) is more reactive, in THF, by roughly one order of magnitude than is PhCh, [see ref. ( 5 ) for summary]. Prof. K.-D. Asmus (Berlin) said: One of your compounds, namely dibenzyl sulphide, attracted my attention. According to your mechanism [reaction (4)] the reaction of (qCH,),S with RCI+ would lead to qCH; cations which show absorptions in the 400-500 nm range.On the other hand it is known* that the oxidation of or- ganic sulphides, e.g., by OH* radicals in aqueous solutions or by radical cations in DMSO and hydrocarbons, leads to relatively stable radical cation complexes of the type R2S :. SR, and R,S :. X (+I with X being a neutral molecule or atom or an anion. The absorptions of these species are also found to be in the 400-500 nm range. In addition a molecular cation yCH2-S-CH2-q absorbing below 350 nm might exist as a short lived transient. My question is: Did you observe any absorptions during the study of dibenzyl sulphide which would indicate the formation of the above mentioned transient radical cations and which possibly could be precursors of yCH,+ cations? (f’- Prof.L. M. Dorfman (Columbus) said: This is a good question, to which I can respond only in part. First let me correct your citation of the absorption region of PhCH,f. It has Lmax at 363 nm, not in the region 400-500 nm. With regard to your question about dibenzyl sulphide, the spectrum obtained with this compound is complex, and exhibits complex dynamics with concurrent formation and decay. The absorption band of PhCH,f is obtained from a difference spectrum. In our first paper on the carbocations [ref. (3) of this paper] we reported other transient bands in the region beyond 460 nm, at 480 and 540 nm. And indeed, we * See paper by Asmus et al. at this Discussion and references cited therein.208 GENERAL DISCUSSION suggested that these “ may be either PhCH2S+ or (PhCh2)2St, possible precursors of benzyl cation”, but we have not yet had time to sort out the species involved in con- current decay and growth.Prof. A. Henglein (Berlin) said: Have you considered the energetics of the reaction of benzyl cation with water? Prof. L. M. Dorfman (Columbus) said: We have not considered the energetics of the reaction, but have experimental data providing the actual rate constants, which is indeed more informative. Henglein will note that our results indicate that the reac- tivity toward dimer, (Hz0)2, is greater than toward monomer. I point out again that the deprotonation step involves H+ in the case of monomer and H30+ in the case of dimer. Prof. M. C. R. Symons (Leicester) said: (a) It is interesting to compare Dorfman’s reaction (I), which gives the radical *Hg-CH,Ph, and the results of our study of the radiolysis of Me,Hg.l This species captured electrons to give Me,Hgs, having two equivalent methyl groups and a very large hyperfine coupling to Ig9Hg.The e.s.r. data suggest that the unpaired electron is in a p* orbital with about an equal distribu- tion on the three groups. There was no tendency for this to give CHy + *HgCH3, but this is not to be expected since CH, is far more nucleophilic than PhCH:. (b) The precise mechanism for the formation of carbocations is of some interest. Consider the reaction of benzyl bromide: initial electron loss could be either from a 3p non-bonding orbital on bromine, or from the phenyl n-system. I suspect that it will be the latter, to give the cation I, 8 r + I which is expected to have the preferred structure shown.2 Bromine could then readily migrate into the ring, where it would be weakly bound and from which it could readily dis~ociate,~ leaving the required cation, PhCH;.Prof. A. Hummel (Delft) said: A number of years ago, just after we had obtained evidence for the migration of positive charge in cyclohexane (the hole) on a time scale of a few tens of nanoseconds, we were very puzzled when a paper appeared by Capellos and Allen4 that showed growth kinetics of the triphenyl methyl carbonium ion (TPM+) in solutions of triphenyl methyl chloride (TPMCl) in cyclohexane that was very similar to the growth kinetics of the carbonium ions in 1,2-dichloroethane as is shown now in Dorfman’s paper.We expected a much faster growth of the absorp- tion of the TPM+ ion since we knew from product analysis that the reaction of the TPMCl with the hole was very efficient, and we had observed such a fast growth for example in the formation of the biphenyl positive ion in biphenyl solutions. It was the slow part of the growth that puzzled us. Erika Zador, from Budapest, looked into this controversy in some detail. From B. W. Fullam and M. C. R. Symons, J.C.S. Dalton, 1974,1086. A. R. Lyons, G. W. Neilson, S. P. Mishra and M. C. R. Symons, J.C.S. Faraday 11, 1975, 71, S. P. Mishra and M. C. R. Symons, J.C.S. Chem. Comm., 1973, 577. C. Capellos and A. 0. Allen, J. Phys. Chem., 1969,73, 3264. 363.GENERAL DISCUSSION 209 this work it was concluded that the hole is captured on the TPMCl with a large rate constant (-loll dm3 mol-I s-l), forming the TPMCl+ ion.(Later this rate con- stant was measured by means of the microwave absorption method, and was found to be 2.8 x 10" dm3 rno1-I s-I.) It was concluded that the TPMCl+ ion decays in two ways: by geminate recombination with the Cl- ion and by a decomposition reac- tion in which the TPM+ ion is formed. An absorption at 360 nm was found that was assigned to the TPMCl+ ion. It is due to the geminate decay kinetics of the TPMCl+ ion that the growth of the TPM+ absorption shows the fast rise followed by a slowly growing part. It would seem that a similar reaction mechanism may be operative in the formation of the carbonium ions in 1,2-dichloroethane.It would be interesting to check this by studying the hole movement and the reactions of the holes in the 1,2- dichloroethane by means of conductance measurements. Prof. L. Kevan (Detroit) said: Whereas in 77 K radiolysis of CH3CI the primary radical formed is CH3 consistent with the dominance of dissociative electron capture, we have found that 77 K radiolysis of single crystals of CH2C12 leads mainly to CHCI,.l In fact, we could not clearly detect CH,Cl, which would be expected if dissociative electron capture were dominant. Of course, CHC12 could be formed by a hole reaction or by electron-hole recombination but the difference between the type of radical formed from CH3Cl and CH,CI, is striking. Does Williams see any evidence for molecular anion formation in any alkyl halides containing H bonded to a central carbon? Would he think such anions might be stabilized at very low temperature, perhaps below 4 K? Prof.F. Williams (Tennessee) said: To date, we have not carried out a systematic e.s.r. study of radical anion formation from the CH,X4-, halomethanes. Regarding the use of the em-. method to draw conclusions about electron attachment in radiation chemistry, I should like to stress that the unambiguous identification of a radical anion from its anisotropic e.s.r. spectrum is not a routine procedure. Thus, the complexity of the e.s.r. pattern for a radical with low symmetry might well defy a successful analysis, and hence negative results by the anisotropic e.s.r. method de- scribed in our paper can rarely be used to show that anion formation has not occurred.Prof. M. C. R. Syrnons (Leicester) said: It is of interest to link Williams' novel results for the anions of F3C-Cl, F,C-Br and F3C-I with those for a range of similar 8 radicals which seem to have comparable parameters. These include the species I to IV, discussed in ref. (1) to (4) respectively, as listed on the next page, /i' R-C \* -1, ha" N=CABr' / 0 R-C \\ F F I 11 I11 I V A. Lund, T. Gillbro, D. F. Feng and L. Kevan, Chem. Phys., 1975,7,414.210 GENERAL DISCUSSION the data for I being especially convincing in that the 14N hyperfine parameters tied in most satisfactorily with those for C1, Br and 1.l The question is, why are these species CT* whilst the alkyl halides all seem to dissociate to give R* + hal-.We know that under certain circumstances these fragments remain in close contacts*6 so it would seem that there is no minimum along the dissociative pathway. I suggest that the difference lies in the shape of the fragment R*. For species I-IV, there is no tendency for any major shape change as hal- is lost, nor is there any major orbital re-hybridization at carbon. For F3C -!- hal- the same applies, since the radical F3C. is known to be strongly pyramidal. However, CH3, and alkyl radicals in general are almost cer- tainly planar. I envisage that this planarity is achieved as the bond stretching pro- gresses, so that the system moves smoothly from an orbital (q2, q * I ) system that is near sp3 on carbon, to one in which the unpaired electron is confined to a 2p orbital on carbon.Prof. L. Kevan (Detroit) said: Although the necessity for a geometrical change undoubtedly decreases the efficiency of electron capture reactions, the time of inter- action is also important. If this time is sufficiently long, as may perhaps be expected in many condensed phase reactions, Franck-Condon restrictions may be minimized and electron capture may occur even with a simultaneous geometrical change. Symons mentioned that C02 does not attach electrons in the gas phase because COT is bent. However, COT is formed in the gas phase by Cs + COz -+ Cs+ 4- COT where the collision or interaction time is much longer. Prof. F. Williams (Tennessee) said: It appears that electron attachment reactions involving large changes in molecular geometry can take place more readily in solid solution than in the gas phase.A recent illustration is provided by the successful e.s.r. observation of BF, in a tetramethylsilane m a t r i ~ , ~ the anion being pyramidal, whereas the neutral molecule is planar. In contrast, the BF, negative ion could not be observed either by direct electron attachment or through ion-molecule reactions in the gas phase.' Perhaps the crucial factor here is the ease of vibrational energy transfer to the surrounding molecules in the condensed phase, a process which would lead to the collisional relaxation of the Franck-Condon state of the negative In the gas phase, other reaction channels such as unimolecular dissociation and auto- ionization would be strongly favoured. The latter process might also explain the low efficiency of direct electron attachment in the gas phase to molecules such as CF3C1 and c-C3F6 where large geometrical changes would not be expected.Prof. F. Williams (Tennessee) said: Mishra and Symonsg have drawn attention to the diversity of electron-attachment processes in certain molecules, and we concur with their suggestion that certain dissociation products are more likely to result from a (Franck-Condon) vibrationally excited state rather than from a thermally relaxed state of the radical anion. However, it is also interesting to consider the alternative G. W. Neilson and M. C. R. Symons, J.C.S. Faraday 11, 1972,58, 1682; Mol. Phys., 1974,27, 1613. S. P. Mishra, G. W. Neilson and M. C. R. Symons, J.C.S. Faraday ZI, 1974,70 1280. D.Nelson and M. C. R. Symons, Chem. Phys. Letters, 1977,47,436. M. C . R. Symons, J.C.S. Chem. Comm., 1977,408. E. D. Sprague and F. Williams, J. Chew. Phys., 1971, 54, 5425. S. P. Mishra and M. C. R. Symons, J.C.S. Perkin IZ, 1973, 391, ' R. L. Hudson and F. Williams, J. G e m . Phys., 1976,65, 3381. * J. A. Stockdale, D. R. Nelson, F. J. Davis and R. N. Compton, J. Chern. Phys., 1972,56,3336. S. P. Mishra and M. C. R. Symons, this Discussion.GENERAL DISCUSSION 21 1 process whereby an electronically excited state of the radical anion might be the im- mediate precursor of the dissociation products. In the absence of direct evidence, the application of orbital symmetry rules might be helpful in distinguishing between these two processes. For example, it was found’ that electron attachment to S02C12 in various glasses resulted in the dissociation products SO, + C12 and to SO2 + C ~ F , as well as to the undissociated radical anion S02C1;.In this case, the SOMO- LUMO orbital interaction method2 suggests that both sets of dissociation products are allowed from the electronic ground state of S0,Cl;. A similar conclusion was ten- tatively reached3 on the basis of Pearson’s rules,4 and therefore there is no need to invoke an excited-state precursor in this case. Although I am unaware of any evi- dence for the participation of electronically excited states of radical anions in the radiation chemistry of the condensed phase, it would seem that the physics of electron attachment does not rule out such a process. Prof. J. Kroh and Dr.A. Plonka (Eddi, Poland) said: There are several arguments suggesting that water absorbed in porous materials freezes in an amorphous glassy state.’ Some disorder of a perfect crystalline lattice, like that in glassy systems, is neces- sary to form suitable traps for mobile electrons. So far, however, efficient electron trapping has only been reported for aqueous solution glasses with high content of inorganic salts and soluble organic substances or, to some extent, in an amorphous ice prepared by condensation of water vapour at very low temperatures. Here we report, for the first time to our knowledge, the very effective electron trapping in gel-frozen water y-irradiated at 77 K. The white, opaque samples of frozen polydextran gels (Sephadex G-25, Pharmacia) turn intensely blue under y-irradiation at 77 K in the dark.The dark blue colour is stable for a few days in the dark at 77 K. It disappears, however, in seconds, in sunlight and/or upon warming. Subtracting the e.s.r. spectrum left after the photobleaching from the initial spec- trum one obtains the singlet with half-width of about 14.5 & 1 G and g equal to about 2.0018. If heavy water is embedded into the polydextran gel one obtains by the same pro- cedure the singlet with half-width of about 3.8 & 1 G, and g equal to 2.0018. The maximum of the optical absorption is at about 550 nm, which corresponds to 2.27 eV. The radiation yield of electrons in fully swollen polydextran gel, containing about 73% of water, is equal to about 0.42 as estimated from the e.s.r.comparison of the amount of electrons trapped in gel to the amount trapped under the same conditions of y-irradiation in 8 M NaOH aqueous glass, for which G(e;),,, = 2.1 was taken. Dr, D. Meyerstein (Israel) said: The high selectivity observed in dissociative elec- tron capture is in principle identical to the large HID isotope effect ion the formation of hydrogen atoms in the reactions eFq + H30+, eLq + NHZ and e; + H2P0r.6 Dr. W. A. Bernhard (Rochester, N. Y.) said: I would like to offer an example, rele- vant to radiation biology, which fits very nicely the framework for electron dissociative C. M. L. Kerr and F. Williams, J. Amer. Chem. Soc., 1972, 94, 5212. K. Fukui, Accounts Chem. Res., 1971,4, 57. T. Gilbro and F. Williams, Chem. Phys. Letters, 1973, 20, 436. R. G . Pearson, J. Amer. Chern. Soc., 1972, 94, 8287. L. Burlamacchi, J.C.S. Firaduy 11, 1975,71, 54. M. Anbar and D. Meyerstein, J. Phys. Chem., 1965, 69, 698.212 GENERAL DISCUSSION capture by halides, as outlined by Symons. The pyrimidine halide 5-bromouracil (BrU) when incorporated into DNA in place of thymine sensitizes cells to ionizing radiation. It is known from aqueous radiolysis of BrU that electron capture leads to loss of Br-. Furthermore, Sevilla has shown that this process also occurs in frozen aqueous glasses. Yet, Hiitterman and co-workers have observed that 5-bromodeoxy- uridine (BrUdR) X-irradiated in the solid state does not necessarily lose Br- after electron capture but, instead, can go on to capture a proton at c6 yielding a free radical due to the net gain of a hydrogen atom at c6, the C,-H-addition radical. This illus- trates Symons’ point of dehalogenation. In aqueous solution and low temperature glasses the chemical environment is able to trap Br- but in crystalline BrUdR, where such traps are not readily available, Br‘ release is not a necessary consequence of electron capture ; instead, the alternative reaction of anion protonation occurs. This is of interest in radiation biology because some regions of brominated DNA, in uiuo, may have a chemical environment unsuitable for Br- elimination. Dr. A. Miiller-Broich (Regensburg) said: As to the biological significance of single crystal studies I should like to point out that they generally supply evidence comple- mentary to pulse radiolysis and studies of vitreous systems. The structure of nucleo- protein in higher organisms is largely unknown, but certainly not much less remote from a highly ordered organization than from a dilute aqueous solution. The co- crystallization of water or other small molecules with crystals of the pure substances is used as an approach to understand environmental effects on highly ordered struc- tures. Prof. M. C. R. Symons (Leicester) (communicated) : I would like to stress that the route taken by a relaxing electron excess centre after capturing the electron is likely to be one that is kinetically controlled rather than thermodynamically controlled. In the particular case of Cl,CP(O)(OH),, the expected phosphoranyl radical : OH may be unable to form sufficiently rapidly because of strong hydrogen-bonding to neighbouring molecules. Hence bond-stretching is faster than bending in this instance, and the phosphoranyl radical is not detected. Prof. M. Magat (Orsay) (communicated): I do not know if this state does exist in the case of biological polymers, but in the case of synthetic polymers there exists a state called coacervate which can be described as being a solution of the solvent in the polymer. Should such a state exist in the case of DMA and RNA it is probably the most similar one to the situation in a living cell and it is I feel in this state that the in vitro experiments should be made.

ISSN:0301-7249    

DOI:10.1039/DC9776300189

出版商:RSC    

年代:1977

数据来源: RSC