Pulse Radiolysis Study of Chlorpromazine and Promazine Free Radicals in Aqueous Solution BY A. KEITH DAVIES Department of Chemistry and Applied Chemistry, University of Salford EDWARD J. LAND Paterson Laboratories, Christie Hospital and Holt Radium Institute, Manchester AND SUPPIAH NAVARATNAM, BARRY J. PARSONS" AND GLYN 0. PHILLIPS School of Natural Sciences, North E. Wales Institute, Connah's Quay, Clwyd Received 18th January, 1978 .OH radicals may react with chlorpromazine in four different ways, viz. (i) electron transfer to produce the cation radical, (ii) addition to the sulphur atom followed by acid-catalysed OH- elimination to yield the cation radical, (iii) addition to the aromatic rings to produce cyclohexadienyl type radicals, and (iv) abstraction of hydrogen atoms from the -CH2- which is in the cc position to the ring nitrogen.Electron transfer and addition to the sulphur atom each account for 40 % of the *OH radicals. Similar considerations apply to promazine. The hydrated electron may attach itself to either the sulphur atom in promazine or its aromatic system with equal probability. Addition to the sulphur atom probably leads to immediate cleavage of the carbon-sulphur bond whereas addition to the aromatic system probably produces an anion radical. The lifetime of this anion radical is sufficiently long to enable electron transfer between itself and another promazine molecule to occur to produce the promazine cation radical. With chlorpromazine, the hydrated electron probably reacts solely with the aromatic system to eliminate chloride ions by the end of the reaction.The promazine radical which is produced simultaneously may add on to another chlorpromazine molecule to produce a cyclohexadienyl type radical. Chlorpromazine (I) and promazine (CH2)3 1 /*\ CH3 CH3 ( I 1 are known to photoionise monophotonically in aqueous solution when excited by light of 347 nm wavelength. Photoionisation of the parent compound, phenothiazine has also been shown to occur in aqueous micellar and methanolic solutions.2 These compounds find use as drugs in psychiatric treatment where side effects include photosensitisation of the skin and eye ti~sues.~ In micellar systems, they also 22DAVIES, LAND, NAVARATNAM, PARSONS A N D PHILLIPS 23 provide model systems for photosynthesi~.~ A study of the fates of the cation radicals and solvated electrons produced on photoionisation of phenothiazine-based derivatives can be most easily undertaken using pulse radiolysis techniques where difficulties due to overlapping absorption by the triplet excited state of the drug are absent.Radical cations may be generated in aqueous solution using free radicals such as SO;- and *OH radicals as oxidants.6 Hydroxyl radicals, for example, have been shown to react with methoxylated benzenes to produce initially hydroxycyclo- hexadienyl radicals which upon protonation and water elimination yield radical catiom6 In this work, the possibility of direct electron transfer between the drug and either *OH or Bri- was investigated. The reaction of the hydrated electron with halogenated aromatics has been the subject of several studies, and in these cases, has resulted in the cleavage of the carbon-halogen bond to yield halide In view of this, clear differences may be expected in the reactions of promazine and chlorpromazine. EXPERIMENTAL Pure samples of chlorpromazine hydrochloride (ClP) and promazine hydrochloride (PH) were kindly donated by May and Baker and Cohn Wyeth and Brothers, respectively.All other reagents were of AnalaR grade. Solutions were prepared using water which was distilled from alkaline permanganate. The pH of chlorpromazine and promazine solutions was normally in the range 5-6. pH values outside this range were obtained by addition of either perchloric acid or sodium hydroxide solutions. When required, the ionic strength of solutions was adjusted by addition of sodium perchlorate.Solutions were saturated with either argon (Air Products) or nitrous oxide (Air Products). The pulse radiolysis experiments were carried out with the electron linear accelerator at the Paterson laboratories lo- Where appropriate, cut-off filters were used to minimise photolysis of the solutions by the analysing light. In all experiments, a single Bausch and Lomb monochromator was used with bandwidths of 10 nm. An E.M.I. 95584 photomultiplier was used in the majority of experiments but for monitoring at wavelengths up to 1000 nm, a U.D.T. PIN10 photo- diode was also necessary. The response times of the two amplifier circuits employed were ~7 and 100 ns. Doses were monitored by a secondary emission chamber calibrated against aqueous potassium thiocyanate l2 taking the yield of the species absorbing at 500 nm to be 2.9 molecules per 100 eV and its extinction coefficient to be 7.1 x lo2 m2 mol-l.The optical path length was 2.5 cm. For kinetics measurements, photographs of the oscilloscope traces were magnified four times and, by using an appropriate computer program, data from these were used to obtain the first-order plots. using pulse lengths up to 100 ns. RESULTS REACTIONS OF HYDROXYL RADICALS WITH PROMAZINE A N D CHLORPROMAZINE Fig. 1 shows the absorption spectra obtained immediately, and 2 5 0 p after an electron pulse was delivered to a nitrous oxide saturated solution of chlorpromazine mol dm-3). These spectra have been corrected for loss of chlorpromazine absorption assuming G(0H) = 5.5. The end of pulse spectrum shows maxima at 850, 770, 525, 440 and <350 nrn (measurements at wavelengths below this were not possible because of the strong absorption of chlorpromazine).The absorptions at 850,770 and 525 nm increased by almost a factor of two over 250 ,us before decaying, whereas the absorptions at 440nm and that at 350nm decreased. In addition a growth in absorption at 950nm was observed over 2ms. These spectra were not affected by changing the chlorpromazine concentration from 5 x to 5 x mol dm-3.24 PULSE RADIOLYSIS OF PROMAZINE FREE RADICALS The spectrum observed immediately ( ~ 1 ps) after the pulse is due to products formed in the reaction of *OH radicals with chlorpromazine. By making observations at shorter time scales, the growth in this absorption was found to follow first-order kinetics and be directly dependent on the chlorpromazine concentration over the range 5 x 10-5-5 x mol dm-3.The second-order rate constant was determined as (8.3 k0.4) x lo9 dm3 mol-l s-l. At 525nm, the subsequent growth over about 250ps also followed first-order kinetics with a rate constant of (1.4k0.2) x lo4 s-l [fig. 2(a)]. Similar observations were also made at 770 and 850 nm. The initial absorption at 350 nm decayed by first-order kinetics yielding a similar rate constant of (1.5f0.2) x lo4 s-l [fig. 2(b)]. 3.c 2.0 m X s I .c 0 I ioo 400 5 0 0 6 0 0 700 800 900 h/nm FIG. 1 .-Transient absorption spectra (after correction for loss of chlorpromazine) at various times after pulsing a nitrous oxide saturated solution of chlorpromazine mol dm-3), (-0-) end of pulse ; (-U--) after 250 ps.Dose = 225 rad, path length 2.5 cm. Neither variation of dose per pulse (80 rad to 1.0 krad) nor chlorpromazine con- centration (5 x mol dm-3) was found to affect the first-order rate constant measured at either 350 or 525 nm. The first-order rate constants at both of these wavelengths were however identically affected by pH. Table 1 summarises these results for 525 nm where the measurements are most accurate and shows that a decrease in the pH leads to an increase in the rate constant. At 440nm, the absorption measured immediately after the pulse showed no dependence on chlorpromazine concentration and decayed according to first-order kinetics but over a longer time-scale than that for the 350 and 525 nm absorptions [fig.2(c)]. The first-order rate constant was found to be (2.850.3) x lo3 s-l. Similarly, at 950 nm, the increase in absorption followed first-order kinetics [fig, 2(d)] yielding a rate constant of (2.8k0.3) x lo3 s-I. Variation of dose per pulse, (200 rad to 1.5 krad) chlorpromazine concentration (5 x to 5 x mol dm-3) and pH had no effect on the first-order rate constant. to 5 xDAVIES, LAND, NAVARATNAM, PARSONS AND PHILLIPS 25 After allowance for the absorption of the cation radical, the spectrum measured at the end of this reaction showed peaks around 440 and 900 nm, respectively. By comparing the absorption spectra obtained after chemical oxidation l3 and photolysis of aqueous chlorpromazine solutions, it would appear that a major portion of the spectra in fig.1 could be assigned to the chlorpromazine cation radical. I -I 0.c 0 0.5 ID - 8 . 0 ~ 100 200 0 I 0 0 5 I .o I .5 I 0 5 0 100 time/p timelps timelms timelms FIG. 2.-First-order plots for changes in absorption after pulsing nitrous oxide saturated solutions of chlorpromazine (lW4 mol dm-3) in a 2.5 cm path length : (a) h = 525 nm, dose = 250 rad ; 6) h = 350 nm, dose = 250 rad; (c) h = 440 nm, dose = 800 rad; (d) h = 950 nm, dose = 800 rad.26 PULSE RADIOLYSIS OF PROMAZINE FREE RADICALS In order to provide further evidence for this, the radical anion, Bri-, was used for the oxidation. The species is a milder oxidising agent than -OH and might be expected to yield a single oxidised species from chlorpromazine by electron transfer.On pulsing a nitrous oxide saturated solution of KBr (1 x mol d ~ a - ~ ) in the presence of mol dm-3 chlorpromazine, the absorption immediately after the pulse at 360 nm (attributable to Br;-)14 disappeared by first-order kinetics with a rate constant of (5.0k0.3) x lo5 s-l. Simultaneously, a first-order growth in absorption was observed at 525 nm with a rate constant of 5.4k0.3 x lo5 s-I. The TABLE EFFECT OF pH ON FIRST-ORDER RATE CONSTANTS FOR GROWTH IN ABSORPTION AT 525 nm k l x 10-4 pH Is-1 (525 nm) 3.75 6.1 4.04 4.3 4.66 3.4 5.46 2.9 Nitrous oxide saturated chlorpromazine solutions, 5 x mol dm-3 ; all solutions were adjusted to an ionic strength of rnol d~n-~). Dose = 275 rad, path length 2.5 cm. second-order rate constant for the reaction of Bri- with chlorpromazine was thus determined as (5.0k0.3) x lo9 dm3 mol-1 s-l.In this solution, no subsequent grow- in of absorption over -250 ps at 525 nm was observed. The spectrum at the end of this reaction is shown in fig. 3. The decay rate of this spectrum was wavelength independent suggesting it corresponds to a single species, the chlorpromazine cation radical. The spectrum of fig. 3 yields an extinction coefficient for the chlorpromazine cation radical at 525 nm of (1.00+0.08) x lo3 m2 mol-l. Using this value, the G- value for the production of the cation radical in N20 saturated solutions at the end hlnm FIG. 3.-Transient absorption spectrum measured 14ps after pulsing a nitrous oxide solution of KBr (1 x mol dm-3) containing rnol dnr3 chlorpromazine.Dose = 215 rad, path length 2.5 cm.27 of the -OH reaction is estimated as 2.2. A further yield of the cation radical (G = 2.2) is found at the end of the slower pH dependent step (lasting 250 zps after the pulse). The reaction of hydroxyl radicals with promazine produces species whose absorp- tion spectra are similar to those obtained with chlorpromazine (fig. 4). The end of pulse (-1 ps) spectrum shows peaks at 850, 760, 515, 440 and ~ 3 5 0 nm. The initial absorptions at 850, 760 and 515 nm, as for chlorpromazine, almost doubled in intensity (following first-order kinetics) over 150 p s after the pulse, whereas those at DAVIES, LAND, NAVARATNAM, PARSONS AND PHILLIPS 2 X 8 h/nm FIG. 4.Transient absorption spectra (after correction for loss of promazine) at various times after pulsing a nitrous oxide saturated solution of promazine mol dm-3), (-0-) end of pulse ; (-Cl-) after 150 ps.Dose = 250 rad, path length 2.5 cm. 440 and 350 nm decreased at a much slower rate. As found for chlorpromazine, the decay in absorption at 440 nm was first-order and matched a growth in absorption at 950 nm. All these first-order rate constants were unaffected by dose per pulse or promazine concentration. The rate constant for the growth in absorption lasting -150 p s at 515 nm was found to be (5.8k0.5) x lo4 s-l. This was measured at pH 5.5 only. The rate constant for the reaction between .OH radicals and promazine was determined as (3.7k0.1) x lo9 dm3 mol-1 s-l. REACTION OF HYDRATED ELECTRONS WITH PROMAZINE AND CHLORPROMAZINE The rate of reaction of the hydrated electron with promazine was followed at 720 nm by pulsing argon saturated solutions containing promazine and t-butanol (0.5 mol dm-3).The decay of the hydrated electron absorption followed first-order kinetics for the range of promazine concentrations 2.5 x to 5 x mol dm-3. After allowance for other modes of hydrated electron decay at low promazine concentration, the first-order rate constant was found to be directly proportional to promazine concentration yielding a second-order rate constant for the reaction between hydrated electrons and promazine of (5.3 k0.3) x lo9 dm3 mol-' s-l. The spectrum of the product(s) formed at the end of this reaction was measured28 PULSE RADIOLYSIS OF PROMAZINE FREE RADICALS in a rnol dm-3 promazine solution and is shown in fig.5. The effect of proma- zine concentration on this spectrum was investigated and it was found that the absorption band at 440 nm was affected. At mol dm-3 promazine the spectrum at the end of the pulse was approximately half that measured in a mol dm-3 solution. Also shown in fig. 5 is the spectrum at 90 ,us after the pulse which has maxima at 515, 760 and 850 nm. In addition, there is a shoulder at about 450 nm. The growth in absorption at 515nm which occurs for 9Ops after the end of the hydrated electron reaction follows first-order kinetics giving a rate constant of (3.5k0.4) x lo4 s-l at mol dm-3 promazine. Identical kinetics were found for the decays of the absorptions at 350, 925 and 975 nm.Experiments were also carried out using sodium formate instead of t-butanol as a scavenger for OH radicals. These yielded essentially similar results. 16 m 8. X n 0 rr. h/nm FIG. 5.-Transient absorption spectra at various times after pulsing an argon saturated solution of promazine mol dm-3) containing 0.5 mol dm-3 t-butanol, (-0-) end of hydrated electron reaction (- 1 ps) ; (-El-) 90 ps after pulse. Dose = 410 rad, path length 2.5 cm. It is clear from fig. 5 that a major portion of the spectrum which is observed 90 ps after the pulse is attributable to the promazine cation radi~a1.l~ From the shape of the promazine cation radical spectrum observed in the chemical oxidation of promazine l3 and also by analogy with the chlorpromazine cation radical spectrum generated in this work using Bri- (fig. 3), it is also apparent that the promazine cation radical absorption should contribute little ( N 10 %) to the 440 nm absorption observed 90 p s after the pulse. However, since there is relatively little change in absorption at 440 nm between the end of the hydrated electron reaction and 90 ,us later, it would, therefore, appear that the kinetic behaviour at this wavelength is different from that observed at 350, 525, 925 and 975 nm.In order to investigate further the conditions governing the formation of the promazine cation radical, the effect of promazine concentration on its yield and rate of formation was studied. Fig. 6(a) shows the effect of promazine concentration (5 x 10-5-5.0 x rnol dm-3) on the first-order rate constant for the cation radicalDAVIES, L A N D , NAVARATNAM, PARSONS A N D PHILLIPS 29 formation.It is apparent from this figure that an increase in the promazine con- centration results in an increase in the rate constant, and, that there is a small but detectable value when the plot is extrapolated linearly to zero promazine concentration. From a consideration of all possible errors, this value was estimated as (3.5 10.5) x lo3 s-l. The second-order rate constant obtained from the slope of this plot was determined as 3.3 x lo7 dm3 mol-1 s-l. The dose per pulse was = 100 rad in these experiments which was considered to be sufficiently low to avoid unwanted radical- radical recombination. From the same experiments, it was also observed that the amount of cation radical formed at the end of this reaction also increased with n I X E: [promazine] x 103/mol dm-3 X n 0 Y [promazinel-I x 10-4/dm3 mol-I FIG.6.-(a) Plot of the first-order rate constant for the formation of the promazine cation radical against promazine concentration. Dose 100 rad, path length 2.5 cm. (b) Plot of the reciprocal optical density (measured at 515 nm at the end of the cation radical formation) against reciprocal promazine concentration. Dose = 100 rad, path length 2.5 cm. increase in promazine concentration. Fig. 6(b) summarizes this data in the form of a plot of the reciprocal optical density due to the promazine cation radical at 515 nm against reciprocal promazine concentration. The slope of this plot was determined to be 1.1 x mol dm-3.From fig. 6(b) it can be deduced that the maximum optical density due to the cation radical at 515 nm that could be expected under these conditions at high promazine concentration would be 3.7 x Assuming similar extinction coefficients for the promazine and chloropromazine cation radicals at their respective wavelength maxima, it can be calculated that the maximum G-value for its production is 1.43. By assuming therefore that the remainder of the hydrated electron yield (G = 2.75-1.43 = 1.32) accounts for the species formed at the end of the hydrated electron reaction which absorbs at 440 nm, and by assuming also that30 PULSE RADIOLYSIS OF PROMAZINE FREE RADICALS the decay of this species is slow compared with cation radical formation and allowing for the contribution of the cation radical to the absorption at 440 nm, the extinction coefficient of the 440 nm species can be calculated as 4.1 x lo2 m2 mol-l. The reaction of the hydrated electron with chlorpromazine is substantially different from that for promazine.The rate of the reaction was studied by following the decay of the hydrated electron absorption at 720 nm in aqueous, deaerated solutions A/nm FIG. 7.-Transient absorption spectra measured at various times after pulsing an argon saturated solution of chlorpromazine mol dm-3) containing 0.5 mol dm-3 t-butanol ; (-0-1 2 ps ; (-El--) 20 ps after pulse. Dose = 930 rad, path length 2.5 cm. containing chlorpromazine and 0.5 mol d r r 3 t-butanol. The decay followed first- order kinetics over the range of chlorpromazine concentration 5 x 10-5-2.5 x mol dm-3 yielding a second-order rate constant of (2.2k0.1) x 1O1O dm3 mol-1 s-l.It was difficult to measure the spectrum at the end of the hydrated electron reaction because of the electron’s relatively large absorption and because of a subsequent build-up in absorption which occurred over several microseconds. Fig. 7 therefore shows the spectrum at 2 p s after the pulse, at which time there may be a substantial contribution from subsequent reactions. Also shown in fig. 7 is TABLE Z-EFFECT OF CHLORPROMAZINE CONCENTRATION ON OPTICAL DENSITY AT 440llm MEASURED AT 20pS AFTER PULSING ARGON SATURATED SOLUTIONS OF CHLORPROMAZINE CONTAINING 0.5 mOl dm-3 f-BUTANOL O.D. (440 nm) [ClPI /mol dm-3 / x 102 5x 1.77 2~ 10-4 1.32 1 x 10-4 0.79 Dose = 1.8 had, path length 2.5 cm.DAVIES, LAND, NAVARATNAM, PARSONS A N D PHILLIPS 31 the spectrum measured 20 ps after the pulse when the subsequent build-up is complete.Both of these spectra have essentially the same shape with one prominent maximum at 440 nm in the range 350-1000 nm. It was observed that an increase in chlorpromazine concentration led to an increase in the amount of absorption measured 20 ,us after the pulse at 440 nm. The dose was considered low enough to minimise radical-radical reactions on this time scale and table 2 summarises the data. It was difficult, for the reasons mentioned above, to observe the effect of increasing chlorpromazine on the rate of formation of this transient and this aspect cannot therefore be discussed.DISCUSSION REACTIONS OF HYDROXYL RADICALS WITH PROMAZINE A N D CHLORPROMAZINE From this work in which chlorpromazine has been oxidised using Bri-, previous chemical l3 and photolytic oxidation experiments, it has been shown that the chlorpromazine cation radical exhibits absorption maxima at 525, 770 and 850 nm. A major portion of the end of pulse spectrum in fig. 1 is therefore attributed to this species, the yield, as argued earlier, being given by G = 2.2. The reaction can be written thus : Fig. 1 also shows that a further yield (G = 2.2) of chlorpromazine cation radical is obtained at 250 ps after the pulse. This is most probably achieved uia addition of the hydroxyl radical to the sulphur atom followed by an acid-catalysed elimination of OH- : *OH + C1P 3 ClP" + OH-.(1) OH I. OH I R I R A similar mechanism has been proposed in the formation of the radical cation of simple aliphatic and cyclic monosulphides. In those studies, dimer cation radicals were also formed at sulphide concentrations in the range 10-4-10-2 mol dm-3. In this work, no such dependence on chlorpromazine concentration (5 x 5 x mol d w 3 ) could be found for the species having absorption maxima at 440 and 525 nm. However, for either steric or resonance reasons, dimer cation radicals are not formed for R2S when R is a t-butyl group.' Stabilisation of the monomeric cation radical may also be achieved intramolecularly via electron pairs on nitrogen and oxygen.16 It would seem, therefore, that coordination of the cation radical of chlorpromazine to chlorpromazine itself is not necessarily a requisite for its stability.32 PULSE RADIOLYSIS OF PROMAZINE FREE RADICALS Our results also suggest that the *OH adduct to the sulphur atom absorbs principallj at 350nm and below.The conversion of this species into the radical cation can occur via two pathways depending upon the pH (see table 1). From a plot of the rates given in table 1 against hydrogen ion concentration, the acid independent pathway, presumably elimination of OH- has a rate constant of (2.9k0.5) x lo4 s-I whereas the acid catalysed step occurs with a rate constant of (1.2k0.3) x lo8 dm3 mol-I s-l . An alternative pathway for the production of CIP+' during the 250 p s after the pulse would be the addition of *OH to the aromatic rings followed by OH- elimination.However, the production of cation radicals via *OH addition to the sulphur atom at diffusion controlled rates appears to be a universal reaction for monosulphides whereas cation radical formation from hydroxycylohexadienyl radicals has only been demonstrated for a few aromatics having electron-donating groups which are expected to stabilise the cation radicak6 However, addition of *OH radicals at the hydrogen containing positions in the aromatic ring is likely to account for the fate of some of the remaining -OH radicals. It is also likely that some will abstract hydrogen atoms from the -CH2- group a to the ring nitrogen from comparison with studies on triethy1amine.l' The total G-value for radicals produced from these latter two processes is expected to be the residue, 1.1.The slower rates of reaction of =OM radicals with protonated tertiary amines compared with the unprotonated forms also makes attack at any other side chain position unlikely. It is proposed therefore that the species absorbing at 440 nm and formed at the end of the *OH reaction with chlorpromazine are hydroxycyclohexadienyl type radicals, e.g. The decay of this absorption at 440 nm follows first-order kinetics and is not dependent on either pH or chlorpromazine concentration. A possible explanation for this observation would be the elimination of water (and possibly HC1 as a minor process) to yield a phenyl type radical, e.g. : This reaction would therefore have a rate constant of (2.8 k0.3) x lo3 s-l. The production of cyclohexadienyl radicals absorbing at 440nm has also been observed in preliminary experiments on the reaction of hydrogen atoms with chlor- promazine. Similarly, this type of radical is produced in the reactions of the hydrated electron with promazine and chlorpromazine which are discussed below. In these instances, the amount of absorption is dependent on the substrate concentration. In order to maintain consistency therefore, with the hydrated electron reaction schemesDAVIES, LAND, NAVARATNAM, PARSONS A N D PHILLIPS 33 discussed below, it would appear that the phenyl type radical produced in reaction (5) must add onto another chlorpromazine molecule to produce a dimeric chlorpromazine cyclohexadienyl type radical.Since the spectrum at the end of reaction (5) also shows an absorption maximum at 440 nm, the rate constant for the production of the dimeric species must be > 3 x lo8 dm3 mol-1 s-l.A rate constant for the analogous reaction of the phenyl radical with iodobenzene of 7 x lo7 dm3 mol-1 s-l has been determined and provides support for this mechanism. Further support is provided by other work,21 in which it was concluded that phenyl radicals add onto halogenated benzenes to form dimeric cyclohexadienyl type radicals. The reaction of .OM radicals with promazine shows identical features to those observed for chlorpromazine and hence the above discussion would also apply to this system. REACTIONS OF HYDRATED ELECTRONS WITH PROMAZINE A N D CHLORPROMAZINE From the kinetic behaviour of the absorption spectra shown in fig.5, it is apparent from the kinetics already detailed that two species contribute to the end of pulse spectrum in the range 350-1000 nm, i.e. one absorbing principally at 440 nm whose amount is promazine concentration dependent, the other having detectable maxima at 925 and 975 nm and accounting also for most of the absorption at 350 nm. It was concluded from the .OH reaction scheme that cyclohexadienyl type radicals of chlorpromazine and promazine absorb at 440 nm. To produce such a radical at the end of the hydrated electron reaction whose amount depends on promazine con- centration, therefore, it is necessary to generate a precursor which can add on to the aromatic positions of promazine itself. A suitable precursor would again be a phenyl type radical. It is suggested that this species is formed via addition of the hydrated electron to the sulphur atom followed by cleavage of the carbon-sulphur bond... .. 0 0 1 R I R :. 0 .. (6J;B - (7) I R I R Reactions (6) and (7) might be considered analogous to the elimination of halide ion after addition of the hydrated electron. In fact, the electronegativety of sp3 hybridised sulphur atoms is greater than, for example, chlorine atoms while the bond strength of a carbon-sulphur bond is some 55 kJ mol-1 less than that of a carbon-chlorine b0nd.l The phenyl type radical produced in reaction (7) is expected to add on to promazine itself to produce a cyclohexadienyl type radical. The rate of this reaction and 1-234 PULSE RADIOLYSIS OF PROMAZINE FREE RADICALS reaction (7) must be of the same order (or larger) as the initial reaction of the hydrated electron if the rate-determining step in the production of the 440 nm species is the rate of loss of the hydrated electron itself.The remainder of the end of pulse spectrum of fig. 5 is therefore presumably attributable to a promazine anion radical where the hydrated electron has added on to the aromatic system of the molecule. Absorption spectra extending as far as 1000 nm are a common feature of aromatic and heterocyclic radical anions 2o and this fact provides support for the assignment. The decay of the absorption due to this species matches that of the growth of the cation radical absorption. The dependence on promazine concentration of the rate of the growth and yield of the cation radical shown in fig.6(a) and (b) indicate that at low promazine concentrations a first-order reaction begins to compete with the direct reaction between the radical anion and promazine itself. The latter reaction has been shown to generate the promazine cation radical, a plausible reaction being, e.g. : I R I R 1 R + I R The possibility that t-butanol readicals can oxidise promazine to form the cation radical was ruled out from the results of experiments in which formate was used instead of t-butanol as a scavenger for -OH radicals. The first-order reaction of the radical anion is presumably hydrolysis, e.g. : R R The rate constants for reactions (8) and (9) can be determined from fig. 6(a) and are (3.3k0.5) x lo7 dm3 mol-1 s-' and (3.5k0.5) x lo3 s-l, respectively.The slope of the plot in fig. 6(b) gives the ratio k9/k8 (= 1.1 x mol dm-3) whose value is in agreement with that obtained from fig. 6(a). From the maximum yield of the promazine cation radical obtained at high promazine concentrations, it can be estimated that ~ 5 0 % of the initial hydrated electron reaction involves addition to the sulphur while the other 50 % add on to the aromatic s ys tem . The rate constant for the reaction between the hydrated electron and chlor- promazine is approximately four times greater than that for promazine and obviously reflects the presence of the chlorine atom in the molecule. Although the spectrum of the product of this reaction cannot be given with any certainty, it is apparent that the spectrum measured at 20 ps after the pulse is due to a subsequent reaction of thisDAVIES, LAND, NAVARATNAM, PARSONS A N D PHILLIPS 35 product.Again, the spectrum shows a maximum at 440 nm which is dependent on chlorpromazine concentration. A scheme which takes account of these facts and maintains consistency with the other mechanisms detailed above would involve addition of the hydrated electron to the aromatic system followed by elimination of a chloride ion. Such eliminations are often found with halogenated aromatic and heterocyclic The reaction can be described as : CIP- + P* + Cl- and probably occurs at a faster rate than 4 x lo5 s-l since no absorption similar to that attributed to the radical anion shown in reaction (8) could be detected. A reaction analogous to reaction (8) does not occur for chlorpromazine presumably due to this short lifetime of ClP-*.The promazine radical, P-, probably has little absorption in the range 350-1000 nm and is expected to add on to chlorpromazine itself to produce a cyclohexadienyl type radical. This radical probably accounts for the absorption at 440nm formed 2 0 p after the pulse. The dependence of this absorption on chlorpromazine concentration suggests that either an alternative pathway for the decay of the promazine radical exists at low chlorpromazine concentration or that the promazine radical and the cyclohexadienyl radical are in equilibrium. Since the only other feasible reaction of the promazine radical, P-, would be protonation to yield the cation radical, the equilibrium suggestion is considered more likely.(10) This work was supported by grants from the Cancer Research Campaign and the Medical Research Council. The authors thank Dr. A. J. Swallow for helpful discussions. S. Navaratnam, B. J. Parsons, G. 0. Phillips and A. K. Davies, J.C.S. Faraday I, 1978,74,1181. S . A. Alkaitis, G. Beck and M. Gratzel, J. Amer. Chem. SOC., 1975, 97, 5723. L. W. C. Massey, Canad. Med. Assoc. J., 1965, 92, 186. J. A. Bassara, J. C. Newton and J. C. Saunders, J. Amer. Med. ASSOC., 1965, 193, 10. G. Porter and M. D. Archer, Interdisciplinary Sci. Rev., 1976, 1, 119. P. O’Niell, S. Steenken and D. Schulte-Frohlinde, J. Phys. Chem., 1975, 79, 2773. ’ K. M. Bansal, L. K. Patterson and R. H. Schuler, J. Phys. Chem., 1972, 76,2386. L. K. Patterson and K. M. Bansal, J. Phys. Chem., 1972,76,2392. R. Koster and K.-D. Asmus, J. Phys. Chem., 1973,77,749. lo J. P. Keene, Quaderni dell’ Area di Ricerce dell’ Emilia-Romugna, 1972, 1, 49. J. P. Keene, Quaderni dell’ Area di Ricerca dell’ Emilia-Romagna, 1972, 1, 63. l2 G. E. Adam, J. W. Boag, J. Currant and B. D. Michael, Pulse Radiolysis, ed. M. Ebert, J. P. Keene, A. J. Swallow and J. H. Baxendale (Academic Press, London, 1965), p. 117. l 3 D. C. Borg and G. C. Cotzias, Proc. Nat. Acad. Sci., 1962,48,617. 14B. Cercek, M. Ebert, C. W. Gilbert and A. J. Swallow, in Pulse Radiolysis, ed. M. Ebert, l5 M. BonifgciC, H. Mockel, D. Bahnemann and K.-D. Asmus, J.C.S. Perkin I l , 1975, 675. l6 K.-D. Asmus, D. Bahnemann, M. BonifiiciC and H. A. Gillis, Favaday Disc. Chem. SOC., 1977, 63, 213. A. J. Swallow, in Free Radical Reactions, Organic Series 1, ed. D. H. Hey and W. A. Waters (Butterworth, MTP International Science), vol. 10. J. P. Keene, A. J. Swallow and J. H. Baxendale (Academic Press, London, 1965). l8 B. Cercek and M. Kongshaug, J. Phys. Chem., 1970,4,4319. l9 J. E. Huheey, Inorganic Chemistry (Harper and Row, London), p. 161 and 699. 2o A. Habersbergerova, I. Janovsky and P. Kourim, Radiation Res. Rev., 1972,4, 123. 21 A. MacLachlan and R. L. McCarthy, J. Amer. Chem. SOC., 1962,84,2519. (PAPER 8/085)