J . Chem. Soc., Faraday Trans. 1, 1984,80, 1377-1389 Pulse Radiolysis of p-Hydroxycinnamic Acid in Aqueous Solution BY KRZYSZTOF BOBROWSKI? Radiation Laboratory, University of Notre Dame, Notre Dame, Indiana 46556, U.S.A. Received 20th April, 1983 Using pulse radiolysis with optical detection it is shown that the hydroxyl radical OH reacts with p-hydroxycinnamic acid in acidic and near-neutral solutions predominantly by addition to the unsaturated substituent, forming benzyl-type radicals. Above pH 9 the spectrum obtained is identical to those obtained from the reaction of p-hydroxycinnamic acid with 0-, Cl;, Br; and CH,CHO radicals. It is suggested that OH radicals attack the ionized hydroxyl group under alkaline conditions to give phenoxyl derivatives. In addition, the spectral data presented show a significant influence of the unsaturated substituent on the position of the absorption maxima in the case of phenoxyl-type radicals, i.e.A = 545 and 595 nm with extinction coefficients 1500 and 1800 m2 mol-l, respectively. The hydroxyl radical is known to react with substituted benzenes predominantly by addition to the yielding hydroxycyclohexadienyl radicals, and not by interaction with the substituent. However, little information is available in the literature concerning the reactivity of OH radicals with substituted benzenes containing unsaturated substituents. In this case two possible reaction pathways can occur, i.e. addition to the aromatic ring and/or addition to the double bond. Earlier studies10-14 concerned with the pulse radiolysis of styrene and methylstyrene confirmed both possibilities for the sites of addition but were in disagreement as to the pattern of the distribution. Our previous paper15 has shown a strong preference of the OH radical for attack at the double bond in the substituent in the case of cinnamic acid.Addition of OH radicals to aromatic rings carrying an additional hydroxy group is of particular interest, because of the possibility of elimination of water from the OH adduct to form phenoxyl-type radicals. Such uncatalysed, acid-catalysed and base-catalysed water- elimination reactions have been reported for phenols,ls methylated benzenes,17 and p-dihydroxybenzenes,21 3,4-dihydro~ytoluene,~~ adrenaline,lg9 2o 4-t-butyl- 1,2- dihydr~xybenzene,~~ 4-t-butyl- 1,2-q~inone,~~ adrenal~ne~~ and p-hydroxyphenyl- propionic a ~ i d .~ ~ p ~ ~ The yield of forming phenoxyl radicals can serve as a measure of the OH radical attack on the ring position. The purpose of this investigation was to obtain additional information about the relative extent of ring/double-bond OH addition, and p-hydroxycinnamic acid was chosen as an appropriate compound. EXPERIMENTAL The solutions were irradiated by 5 ns electron pulses from an ARC0 LP-7 linear accelerator. The dose per pulse was such as to produce only 1-3 pmol dm-3 of radicals in order to minimize their second-order decay. Dose effects were measured with pulses of up to 50 ns duration. Optical detection and signal averaging were carried using the computer-controlled pulse- ? On leave of absence (1979-1981) from the Institute of Nuclear Research, 03-195 Warsaw, Poland.13771378 PULSE RADIOLYSIS OF p-HYDROXYCINNAMIC ACID A/nm Fig. 1. Absorption spectrum (corrected for bleaching of the parent compound, dashed line) of the intermediates obtained after pulse radiolysis of aqueous solutions of p-hydroxycinnamic acid mol dm-3), saturated with N,O at pH 6.0 (immediately after the pulse). radiolysis apparatus described previo~sly.~~ The change in absorbance (AOD) following electron irradiation was followed as a function of time by kinetic spectrophotometry and expressed in terms of an extinction-coefficient parameter (e’) given by the relationship E’ = (AOD x K)/(G x dose), where K is a multiplying factor so chosen that E’ for (SCN), in the same cell is 7600 dm3 mol-l cm-l at 475 nm in N,O-saturated aqueous solutions and G is the yield per 100eV.28 Digitized absorbance data were recorded on magnetic tape and kinetic analyses were made by computer fitting of the curves.2v p-Hydroxycinnamic acid (predominantly in the transform) was obtained from Aldrich and was of the highest purity commercially available.Sodium phosphates (monobasic and dibasic), sodium hydroxide and t-butyl alcohol were analytical grade from Mallinckrodt . Sodium chloride and sodium bromide were from Alfa Division. Ethylene glycol and perchloric acid were obtained from Fischer. In order to minimize thermal oxidation (phenols are extremely sensitive to oxygen in alkaline solutions) all solutions were prepared in the following manner and were not allowed to come into contact with air.First the water was saturated with N,O for 30min and then the appropriate amount ofp-hydroxycinnamic acid was added. During further saturation the KOH was introduced rapidly. Fresh solutions were prepared before each irradiation. The N,O was used in order to remove oxygen and to convert eiq into OH by the reaction N,O+e,, -+N,+OH+OH-. t-Butyl alcohol (0.5 mol dmV3) was used as OH scavenger. The pH of the solution was adjusted with NaOH, HClO,, Na,HPO, and NaH,PO,. Optical absorption spectra of the solutions before irradiation were recorded on a Cary 219 spectrophotometer. The absorption spectra were corrected for the depletion of solute (S) in the wavelength range where the solute absorbed lightK. BOBROWSKI I""'"'''''''''' 4 - ai 5 3- -2 51.D ai 0 OH HC=CH-C< @ +OH OH \ 1379 400 500 600 X/nm Fig. 2. Absorption spectrum (corrected for bleaching of the parent compound, dashed line) of the intermediates obtained after pulse radiolysis of aqueous solutions of p-hydroxycinnamic acid mol dm-3) saturated with N,O at pH 3.0 (immediately after the pulse). (usually < 350 nm). G(p-OHCA),,,,,, was assumed to be 6.0. Owing to the existence of two functional groups (OH and COOH) which can lose a proton, p-hydroxycinnamic acid exists in three forms depending on pH of the solution. A more detailed examination of the optical absorption spectra as a function of pH showed three distinctive regions with pK values pK, = 4.0 and pK, = 9.1 which are in agreement with literature values (Suom.Kemistil. B, 1965, 38, 291). RESULTS REACTION OF OH RADICALS WITH p-HYDROXYCINNAMIC ACID A pulse-radiolysis study of nitrous oxide-saturated aqueous solutions of p- hydroxycinnamic acid (p-OCHA) at pH 3.0-11.5 was carried out using optical absorption for the detection of transients. Under these conditions ca. 90% of the primary radicals are available as OH radicals for reaction with p-hydroxycinnamic acid. The transient absorption spectra obtained at various pH values are given in fig. 1-3. In almost neutral solutions (pH 6.0) a strong sharp absorption band was observed at 0.7 ps after the pulse with an absorption maximum at 3, = 335 nm together with broad shoulder with maximum at ca. 405 nm (fig. 1). The rate of the formation of the transient at 335 nm followed first-order kinetics.From the fitting of the exponential increase of the 335nm absorption after the pulse and the concentration of p- hydroxycinnamic acid (in the range 2 x mol dm-3) a rate constant k = 8.2 x lo9 dm3 mol-l s-l was calculated for the reaction. This value is almost equal -2 xI380 PULSE RADIOLYSIS OF p-HYDROXYCINNAMIC ACID 16 14 12 i 10 e 0- L l r l r l l l l l r l r l r r l l 400 50 0 600 h/nm Fig. 3. Absorption spectrum (corrected for bleaching of the parent compound) of the intermediates obtained after pulse radiolysis of aqueous solutions of p-hydroxycinnamic acid mol dm-3) saturated with N20 at pH 10.0 (immediately after the pulse). l 4 t I I 2 [H+l/10-5 mol dmV3 Fig. 4. Plot of the observed rate constant kexptl of the fast process against H+ concentration.Inset: pH dependence of the absorption of the intermediate taken 10 ps after the pdse at A = 345 nm in aqueous solution of p-hydroxycinnamic acid saturated with N20.K. BOBROWSKI 1381 to the value obtained for cinnamic acid.15 Similar results were observed at pH 3.0 with a possible absorption maximum located at (or below) 3, = 340 nm (fig. 2). The spectral region below 340 nm becomes inaccessible because the monitoring light is absorbed by the parent compound itself. A broad band was observed in the spectrum at 425 nm together with weak absorption bands at A = 545 and 595 nm. On the other hand transient absorption spectra taken 0.7 ,us after the pulse in alkaline solution (pH 10.0) show broad bands at 470 nm and below 380 nm; again the region below 380 nm was inaccessible because of absorption by the parent compound itself (fig.3). Instead of the very weak absorption bands observed in acid and neutral solution, two intense sharp maxima were observed at 545 and 595 nm, respectively. As one can see, the reaction of OH radicals with p-hydroxycinnamic acid produces absorption spectra which differ in the positions of their absorption maxima as well as in their intensities. Thus a more detailed examination of the effect of varying the pH on the spectra was carried out. The transient absorption at A = 345 nm is shown as a function of pH (inset) in fig. 4. The pK value of the transient observed at this wavelength is ca. 4.9 + O . 1, which is higher than the value of pKl ofp-hydroxycinnamic acid.As in the case of cinnamic acid,15 in a very narrow region of pH (4.4-6.0) the absorption spectrum ofp-hydroxycinnamic acid undergoes a very similar small change during the first 5 ,us after the pulse. In the region 330-345 nm the transient absorption decays fast in a first-order reaction which is independent of dose and solute concentration but slightly dependent on H+ concentration. The experimental rate constants were plotted as a function of [H+] and a straight line was obtained (see fig. 4). Such H+ dependence strongly suggests an equilibrium of the form OH 1 8 HC-CH-c OH I 40 H C-CH-C I OH OH The experimental rate constant kexptl for a reaction of this type is, of course, equal to k,+k,[H+]. From the plot presented in fig. 4 rate constants of 3.8 x 1O1O dm3 mol-1 s-l and 2.9 x lo5 s-l were determined for kf and k,, from which K = k,/k, = 1.33 x mol dm-3 was calculated corresponding to a pK, of 4.9, in excellent agreement with value estimated from the dependence of absorption on pH at A = 345 nm.REACTION OF 0- RADICAL IONS WITH p-HYDROXYCINNAMIC ACID The radical ion 0- was produced by irradiating 1 mol dm-3 NaOH solutions saturated with N,O. Under these conditions essentially all primary radicals (99 %) were converted to 0-. The transient absorption spectrum obtained with irradiated solutions of p-hydroxycinnamic acid (1 x mol dm-3) is presented in fig. 5. The spectrum recorded at pH 14 indicates some contribution of the OH adduct at ca. 470 nm. Sharp maxima at 545 and 595 nm are also present and are similar to those previously observed in solutions of p-hydroxycinnamic acid at pH 10.With phenols the 0- radicals have been ~uggested~O7~l to react by direct electron transfer, and the predominant product is phenoxyl radical. The reaction of 0- with a phenoxide-type ion (-0-C,H,-CH=CH-COO-) was followed directly by examining the buildup of1382 PULSE RADIOLYSIS OF p-HYDROXYCINNAMIC ACID I6 - I4 - 12 - W 10- e 2 D ; 8 - .- * id W - 6 - 4 - 2- L I 1 1 1 1 1 1 " 1 1 " 1 1 I I 400 500 600 h/nm Fig. 5. Absorption spectrum (corrected for bleaching of the parent compound, dashed line) of the intermediates obtained after pulse radiolysis of aqueous solutions of p-hydroxycinnamic acid ( mol drnp3) saturated with N,O at pH 14.0 (40 p s after the pulse). transient absorption at A = 595 nm.The rate constant determined from the reaction periods observed over the concentration range 5 x 10-4-5 x mol dm-3 was 3.1 x lo8 dm3 mol-l s-l (after correction for the reaction of OH with p-hydroxycin- namic acid). This rate constant was found to be one order of magnitude greater than that expected for addition of 0- to the aromatic ring.31 These facts suggest that this reaction also occurs through transfer of an electron from the phenoxide anion to 0-. However, the spectrum obtained is quite different from the spectra of phenoxyl-type radicals; usually they absorb in the region 380-420 nml6* 3 0 ~ 32-37 with well defined maxima at ca. 390 and ca. 410 nm. The shape of the observed spectrum resembles that of the phenoxyl radical but is red-shifted by ca.170 nm. REACTIONS OF OTHER OXIDIZING RADICALS (Br;, Cl; AND CH,CHO) WITH p-HYDROXYCINNAMIC ACID It is necessary to determine if the spectrum obtained through the reaction of 0- withp-hydroxycinnamic acid is due to a phenoxyl-type radical or to a different species. It appeared to us that observation of the spectra formed by the reaction of other oxidizing radicals with p-hydroxycinnamic acid could throw some light on the origin of the spectrum. Of the oxidants mentioned above, Br; is the most reactive after C1, and seemed the most convenient to study because it could be produced over a wide pH range.32 However, even in this case the reaction rate with undissociated phenols is too slow, so studies could be made only above the pK, of the phenol. To provide other oxidizing radicals38 studies were also made with C1; (at pH 3) and with theK.BOBROWSKI 1383 I I I I I 400 500 600 X/nm Fig. 6. Absorption spectrum of the intermediates obtained after pulse radiolysis of aqueous solutions containing I x mol dmW3 ofp-hydroxycinnamic acid and 1 mol dm-3 KBr at pH 11.5 and saturated with N,O. Inset: plots of the change in relative absorbance with time monitored at the maximum of the absorption band at 595 and 385 nm (near the maximum of the absorption band of Br;). formylmethyl radical, CH,CH0,39v 40 which has received some recent attention. Use of the SO, radical*l as a direct oxidant of carboxylated phenoxide ions and for the generation of C1; above pH 3 was unsuccessful because of the thermal reaction between sodium persulphate and p-hydroxycinnamic acid.Reaction of Br; with carboxylated phenoxide, -O--C,H,-CH=CH' 0 '0- (pK, 9. lo), was carried out at pH 11.5 in N,O-saturated solutions containing 1 mol dmF3 KBr. At wavelengths > 500 nm the optical spectrum was essentially identical to that found with 0- as a reactant (fig. 6). Below 500 nm there was much less absorption than in the previous cases, since any OH radical formed reacts with Br- (because of the high concentration of the latter and the high rate constant for its reaction with OH). Thus there was no possibility of a contribution by the OH adduct to the benzene ring in such conditions. Clearly direct oxidation of carboxylated phenoxide ions by Br; radicals was observed at A = 595 nm (see inset in fig. 6). As can be seen, the dependences of the two normalized plots of the change in relative absorbance with time monitored at both wavelengths gave symmetrical curves with the same rate constant, k = 9.9 x lo8 dm3 mol-l s-l.Similar experiments were performed at pH 2.9 in N,O-saturated solutions contain- ing 1 mol dm-3 Cl- in order to minimize the competition by p-hydroxycinnamic acid1384 PULSE RADIOLYSIS OF p-HYDROXYCINNAMIC ACID [p-OHCA]/10-3 mol dm-3 Fig. 7. Plot of the observed rate constant kexpt, of the formation of 595 nm absorption against concentration of p-hydroxycinnamic acid in aqueous solutions containing 1 mol dm-3 of ethylene glycol at pH 1 1.5, saturated with N,O. Inset: experimental trace followed at A = 595 nm in aqueous solutions of p-hydroxycinnamic acid (2 x rnol dm-3) at pH 1 1.5 containing 1 rnol dm-3 ethylene glycol and saturated with N,O.for hydroxyl radicals. In this case the only reactive species which can oxidize p-hydroxycinnamic acid is the C1; radical. The intense absorption of C1; below 380 nm was observed immediately after the pulse. Simultaneously with the decay of the C1, radical an absorption spectrum was formed which was essentially identical to those found with Br; and 0- as reactants (at wavelengths > 500 nm). From the plot of the pseudo-first-order rate constants for the decay of Cl, as a function of p-hydroxycinnamic acid concentration, the second-order rate constant for the reaction of C1; with p-hydroxycinnamic acid was derived as 2.9 x loa dm3 mol-1 s-l. the formylmethyl radical can be produced from ethylene glycol via H abstraction As previously HOCH2CH20H + OH --+ HOCH,cHOH + H20 followed by a rapid elimination of water in alkaline solutions: OH- -OH- HOCH,cHOH - HOCH,CHO- - cH2CH0.The resultant radical was found to oxidize various phenols by direct electron t r a n ~ f e r . ~ ~ ~ * ~ In our case the formylmethyl radical produced by the reaction of OH with 1 mol dm-3 ethylene glycol at pH 11.5 was allowed to react with p-hydroxycin- namic acid [(0.5-2) x mol dm-3]. The inset in fig. 7 shows the experimental trace obtained for the p-hydroxycinnamic acid + ethylene glycol system in water monitored at 1 = 595 nm. In this case a build-up of absorbance is observed due to the formation of the oxidation product of p-hydroxycinnamic acid. By using a large excess of ethylene glycol (1 mol dm-3) the direct reaction of OH radicals withp-hydroxycinnamicK.BOBROWSKI 1385 1 I 1 I 400 450 500 h/nm Fig. 8. Absorption spectrum at pH 1 of the hydrogen adduct to p-hydroxycinnamic acid mol dm-3) saturated with N, and containing 0.5 mol dm-3 t-butyl alcohol to remove OH radicals. acid was prevented. A plot of kexpt against p-hydroxycinnamic acid concentration yielded a straight line (fig. 7). From the slope of this line kp.0CHA+CH2CH0 was estimated as 7.7 x lo7 dm3 mol-l s-l. This value is typical of such an electron-transfer reaction (values are usually in the range 107-109 dm3 mol-1 s - ' ) . ~ O The specttal absorption maxima (i.e. 545 and 595 nm) of the transient were found to be the same as those of the transients formed in previous systems.REACTION OF H ATOMS WITH p-HYDROXYCINNAMIC ACID The reaction of H atoms withp-hydroxycinnamic acid was studied after the removal of most of the OH radicals through reaction with t-butyl alcohol at pH 1. The spectrum observed immediately after the pulse has an absorption maximum around 2 = 425 nm (fig. 8). The spectral region below 360 nm becomes inaccessible because of the absorption of the monitoring light by the parent compound itself. DISCUSSION SPECTRUM OF THE PI-ENOXYL-TYPE RADICAL The absorption spectra (fig. 5 and 6) represent transients which are formed by direct electron transfer. The similarity between the optical absorption spectra at il > 500 nm observed with 0-, Br,,Cl, and CH,CHO radicals as oxidants strongly suggests that direct oxidation of the phenoxide ion from p-hydroxycinnamic acid is involved.In each case the spectrum is similar in shape to that of phenoxyl radical but with a strong 46 FAR 11386 PULSE RADIOLYSIS OF P-HYDROXYCINNAMIC ACID 0 4 HC=CH-C 'OH t OH /p '0- HC=CH-C @ + OH It -H* I I+H* 0 4 9 + OH \ 0- +Hi0 0- U - 2 OH- Scheme 1. r e d - ~ h i f t ~ ~ in the position of the two maxima. The extinction coefficients of the phenoxyl-type radical can be determined directly from absorption spectra produced in a solution co,ntaining 1 mol dm-3 Br- and 1 mmol dm-3 ofp-hydroxycinnamic acid. Under these conditions any OH radical should react with Br- because of the high concentration of the latter and the high rate constant for its reaction with OH. As was mentioned earlier, Br; is quite reactive and direct oxidation of phenoxide-type ions by Br; should proceed with the yield G(Br;) = G(0H) = 6.0.The extinction coefficients calculated on the basis of complete reaction between Br; and p- hydroxycinnamic acid are E,,, = 1500 and E,,, = 1800 m2 mo1-l. PROPOSED REACTION MECHANISM The experimental results are explained in terms of three processes: (a) addition of OH to the unsaturated substituent leading to formation of benzyl-type radicals, (b)K. BOBROWSKI 1387 addition of OH(H) to the benzene ring leading to the formation of OH(H) adducts (in the first case these may decay by elimination of water, yielding phenoxyl-type radicals) and (c) direct attack of OH at the ionized phenolic group, giving also phenoxyl derivatives (see scheme 1).The wavelength maxima and approximate relative intensities of the transient bands obtained under different conditions are summarized in table 1. The observed bands Table 1. Transient spectra obtained by pulse radiolysis in aqueous solutions of p-hydroxycinnamic acid (1 0-3 mol dm-3) additive PH transient absorption maxima"/nm none none none with with with with 1 mol dm-3 NaOH 1 mol dm-3 KBr 1 mol dm-3 KCl 1 mol dm-3 ethylene glycol 3.0 < 360b (23) 6.0 330 (102) 10.0 < 380* (14) 14.0 < 380b (10) 1 1.5 absorption due to Br; 2.9 absorption due to C1; - 11.5 42Y (13.5) 41OC (13.5) 47OC (8.0) 47OC ( 10.5) no absorption no absorption no absorption 545 (2.5) 595 (3.0) 545 (2.5) 595 (3.0) 545 (10.0) 595 (12.5) 545 (1 3.0) 595 (14.5) 545 (1 5.0) 595 (18.0) 545 (13.0) 595 (14.5) 595 (13.0) 545 (1 1 .O) a The number in parenthesis indicates the relative band height.Not determinable due to the strong absorption of the analysing light by the parent compound. It is possible that the wavelength region corresponds to both the cyclohexadienyl and hydroxycyclohexadienyl derivatives, but the strong overlapping between OH addition and H addition products makes the estimation of the yield very uncertain. fall into three groups. The strong absorption near 340 nm for p-hydroxycinnamic acid corresponds to the characteristic benzyl-type radicals. In the pH region where both functional groups are protonated the initial attack of hydroxyl radicals is by addition to the unsaturated substituent [reaction (1 b)], and a relatively sharp and strong absorption band in the region 360 nm is assigned to species 16.Very weak absorption in the regions 410-470 and 540-600 nm testifies to a low yield of OH addition to the aromatic ring, if any, followed by acid-catalysed elimination of the water molecule and generation of phenoxyl-type radicals [reaction (1 a)]. In the pH region where only the carboxyl group is deprotonated the spectrum obtained is practically identical to the previous one. This shows that in acidic and near-neutral solutions OH addition to the unsaturated substituent, leading to benzyl-type radicals, is the predominant process between the substrate and the initial radiolysis products of water. The fast decay observed during the first 5ps after the pulse represents (as in the case of cinnamic acid)15 an equilibration process between species I b and I1 6, i.e.by reactions with constants kf and k,. This conclusion is supported by the fact that the decay is dependent on H+ concentration, the agreement in the values of K, obtained from the kinetic treatment of the equilibration process and the dependence of the optical density observed at il = 345 nm as a function of pH. As in the case of fully protonated acid 46-21388 PULSE RADIOLYSIS OF p-HYDROXYCINNAMIC ACID molecule a very low yield of phenoxyl-type radicals is observed at this pH. Thus using values of the extinction coefficient for the phenoxy radical calculated in this work one can estimate that reactions ( l a ) and (2a) account for ca. 15% of the total hydroxyl radical yield.On increasing the pH to ca. 9, i.e. above pK2 (fig. 3), intense absorption was observed at A 2 500 nm which was very weak in acid and neutral solutions. From the similarity of the optical absorption spectra to those observed in systems containing Br;, Cl;, CH,CHO and 0- it is concluded that phenoxyl-type radicals from p- hydroxycinnamic acid are formed. In alkaline solution (at pH lo), in which the hydroxy group in p-hydroxycinnamic acid is ionized, OH radicals may both add to the benzene ring and react with the OH group simultaneously. In this context note that the water-elimination reaction16 via the hydroxycyclohexadienyl radical occurs with uncatalysed elimination rate constants of lo3 s-l. However, such reactions are acid/base catalysed with subsequent elimination rate constant k = lo5 s-l.If reactions (1 a) and ( 3 a) were significant then the yield of phenoxyl-type radicals in the acid and alkaline regions should be very close. The large difference in yields of the phenoxyl-type radical in both acid/neutral and alkaline solutions is taken to mean that the transient phenomena observed in both regions make the occurrence of reactions ( 1 a), (2a) and ( 3 a) followed by the elimination of water very unlikely in this case. Our observations regarding the large increase in the yield of phenoxyl-type radical on going from low (fig. 2) to high pH (fig. 3 ) clearly show that OH radicals attack the ionized hydroxy group under alkaline conditions to give phenoxyl derivatives. Such a reaction was proposed by Chrysoch0os~~9 26 for p-hydroxyphenylpropionic acid, tyrosine and polytyrosine in aqueous media.The wavelength region corresponding to benzyl-type radicals indicates that these are formed to some extent, but the strong absorption of the analysing light below 380 nm by the parent compound makes an estimation of the absorption maximum impossible. However, based on the known extinction for the phenoxyl-type radical and the absorption value of phenoxyl radicals formed at pH 10, the proposed direct electron-transfer reaction between the ionized hydroxy group and hydroxyl radical accounts for ca. 70% of the total hydroxyl radical yield. The for- mation of phenoxyl radicals through step (3a) can also be excluded for the following reason. If step (3a) were to be an important path, one should observe predominantly the formation of the hydroxycyclohexadienyl radical followed by the delayed formation of the phenoxyl radical with simultaneous decay of the hydroxycyclohexadienyl derivative (as was observed in some systems).On the contrary, in our case absorption due to phenoxyl radicals occurs at the same time as very weak absorption in the region 410-470 nm. We therefore conclude that formation of phenoxyl radicals occurs mainly through direct electron transfer between the OH radical and the ionized hydroxy group in the ring and not via a hydroxycyclohexadienyl derivative. CONCLUSIONS Three conclusions may be derived from these results. First, the addition of an unsaturated side chain can cause a large red shift in the absorption spectrum of phenoxyl-type radicals.Secondly, as in the case of cinnnamic acid, addition of hydroxyl radicals to the unsaturated substituent can effectively compete for addition to the benzene ring at both acid and neutral pH. Thirdly, phenoxyl-radical formation from p-hydroxycinnamic acid is negligible unless the phenol is ionized, thus allowing electron transfer to take place.K. BOBROWSKI 1389 Helpful discussions with Dr P. Neta are gratefully acknowledged. The research described herein was supported by the Office of Basic Energy Sciences of the Department of Energy. This is document no. NDRL-2352 from the Notre Dame Radiation Laboratory. The paper was presented in part at the 28th IUPAC Congress held in August 1981 at Vancouver, B.C., Canada. P. Neta and L. M.Dorfman, Adv. Chem. Ser., 1968, 81, 222. C. R. E. Jefcoate and R. 0. C. Norman, J. Chem. SOC. B, 1968, 48. P. Neta, M. Z. Hoffman and M. J. SimiC, J. Phys. Chem., 1972,76, 847. N. V. Raghavan and S. Steenken, J. Am. Chem. SOC., 1980,102, 3495. N. Selvarajan and N. V. Raghavan, J. Phys. Chem., 1980,84, 2548. 0. Volkert, W. Bors and D. Schulte-Frohlinde, Z. Naturforsch., Tie1 B, 1961, 22, 480. P. Neta, Radiat. Res., 1973, 56, 201. S. Steenken and P. O’Neill, J. Phys. Chem., 1978,82, 372. A. Samuni and P. Neta, J. Phys. Chem., 1973,77, 1629. lo A. J. Swallow, Adv. Chem. Ser., 1968, 81, 499. 0. Brede, W. Helmstreit and R. Mehnert, J. Prakt. Chem., 1974, 316, 402. l 2 0. Brede, J. Bos and R. Mehnert, Radiochem. Radioanal. Lett., 1979, 39, 259. l3 0. Brede, J. Bos, W.Helmstreit and R. Mehnert, 2. Chem. (Leipzig), 1977, 17, 447. l4 C. Schneider, Adv. Chem. Ser., 1967, 91, 219. l5 K. Bobrowski and N. V. Raghavan, J. Phys. Chem., 1982,86,4432. l6 E. J. Land and M. Ebert, Trans. Faraday SOC., 1967, 63, 1181. K. Sehested, H. Cofitzen, H. C. Christensen and E. J. Hart, J. Phys. Chem., 1975, 79, 310. W. Bors, M. Saran, C. Michel, E. Lengfelder, C. Fuchs and R. Spottl, Int. J. Radiat. Biol., 1975, 28, 353. l9 M. Gohn, N. Getoff and E. Bjergbakke, Int. J. Radiat. Phys. Chem., 1976, 8, 533. 2o M. Gohn, N. Getoff and E. Bjergbakke, J. Chem. SOC., Faraday Trans. 2, 1977, 73, 406. 21 G. E. Adams and B. D. Michael, Trans. Faraday SOC., 1967, 63, 1 171. 22 M. Gohn and N. Getoff, J . Chem. Soc., Faraday Trans. I , 1977, 73, 1207. 23 H. W. Richter, J. Phys. Chem., 1979, 83, 1123. 24 W. Bors, M. Saran and C. Michel, J. Phys. Chem., 1979,83, 2447. 25 J. Chrysochoos, Radiat. Res., 1968, 33, 465. 26 J. Chrysochoos, J. Phys. Chem., 1969, 73,4188. 27 L. K. Patterson and J. Lillie, Int. J. Radiat. Phys. Chem., 1974, 6, 129. 28 R. H. Schuler, L. K. Patterson and E. Janata, J. Phys. Chem., 1980,84, 2088. 29 R. H. Schuler and G. Buzzard, Int. J. Radiat. Phys. Chem., 1976,8, 563. 30 P. Neta and R. H. Schuler, J. Am. Chem. SOC., 1975, 97, 912. 31 P. Neta and R. H. Schuler, Radiat. Res., 1975, 64, 233. 32 K. M. Bansal and R. W. Fessenden, Radiat. Res., 1976, 67, 1. 33 K. Kemsley, J. S. Moore, G. 0. Phillips and A. Sosnowski, Acta Vitaminol. Enzymol., 1974, 28, 263. 34 J. L. Redpath and R. L. Willson, Int. J. Radiat. Biol., 1975, 27, 389. 35 E. J. Land, G. Porter and E. Strachan, Trans. Faraday SOC., 1961, 57, 1885. 36 H. Joschek and L. I. Grossweiner, J. Am. Chem. SOC., 1966,88, 3261. 37 P. K. Das and S. N. Bhattacharyya, J. Phys. Chem., 1981,85, 1391. 38 K. Hasegawa and P. Neta, J. Phys. Chem., 1978,82, 854. 39 S. Steenken, J. Phys. Chem., 1979,83, 595. 40 S. Steenken and P. Neta, J. Phys. Chem., 1979, 83, 1134. 41 H. Zemel and R. W. Fessenden, J. Phys. Chem., 1978,82, 2670. 4 2 K. M. Bansal, M. Gratzel, A. Henglein and E. Janata, J. Phys. Chem., 1973, 77, 16. 43 S. Steenken and P. Neta, J. Phys. Chem., 1982, 86, 3661. (PAPER 3/634)