2693
Photochemical Production of Radicals in Solution OShaughnessy. 122nd National Meeting of the American Chemical Society, Atlantic Cky. N.J., 1952, Abstract 1701 are fragmentary but in reasonable agreement with these. (37) Assuming that ([A.]/[R.])([A.] [R.]), f, to be a linear function of [MAN], see eq 6, we have subjected our data taken at loF to a leastc; m X IO7 squares analysis with the following results: f = m[MAN] = 1.03 (0.02), c X IO7 = -0.35 (0.15), the terms in parentheses being
+
+
standard deviations. (38) The reaction conditions were unusual since the polymer was not very soluble in the monomer and tended to precipitate from Solution. (39) A. T. Bullock, G. M. Burnett, and C. M. L. Kerr, Eur. Polym. J.. 7, 1011 (1971). (40) A. A. Bichutinskii. A. I. Prokof'ev, and V. A. Shabalkin, Russ. J. Phys. Chem.. 38, 534 (1964).
Electron Spin Resonance and Pulse Radiolysis Studies of Some Reactions of SO4*-
lg2
Om P. Chawla and Richard W. Fewenden' Radietion Research Laboratories and Department of Chemistry, Mellon Institute of Science, Carnegie-Mellon University. Pittsburgh, Pennsylvania 15213 (Received June 16, 1975) Publication costs assistedby Carnegie-Mellon Universify and the U.S. Energy Research and Development Administration
The in situ photolysis of aqueous solutions of peroxydisulfate (S20s2-) has been investigated as a source of radicals for ESR studies. This initiating system has been found to be quite effective in that the SOP- SO produced reacts readily with a variety of solutes; successful detection of a number of new radicals has been possible. Attention has been focused mainly on radicals resulting from reaction of SO4.- with inorganic ions. Direct observations have been made where possible, and spin trapping with CH2=N02- or fumarate has been used in other cases. Radicals directly observed include C02- from formate, cos- from bicarbonate, H62 from H202, and H2N6 from NH20H. The g factor of c02- was found to be 2.00045 and the line width 2.3 G. No change in parameters was found over the pH range 0.8-4.6. The radical cos- with g = 2.0113 and line width 4.4 G is reported for the first time in aqueous solution. No changes in the spectrum which might give evidence for the pK of this radical were found over the pH range 7.5-11.9. The pK for the dissociation P03H- F? H+ P032-has been determined to be 5.9. Evidence for reaction of SOc- with "3, CN-, H2P04-, and C1- was obtained by trapping the resultant radicals which are believed to be "2, CN, p04e2-, and Cl2.-, respectively. Pulse radiolysis studies confirm the formation of Clr- from C1(kso4-+c1-= 3.1 X los M-l sec-l) by means of its absorption spectrum. Thus the reaction SOY- C1- can be used to provide a source of Cly- in neutral and basic solutions. Conductometric pulse radiolysis studies were carried out on several organic unsaturated compounds to clarify the nature of the Sod.- reaction. It was concluded that SOY- adds to the simple olefins cyclohexene and allyl alcohol while a direct oxidation of benzene to hydroxycyclohexadienyl radical occurs in the time scale of 12) can readily be carried out using the photolysis of solutions of acetone in the presence of isopropyl alcohol and the desired a c ~ e p t o rHigh . ~ radical concentrations are possible because of reasonably efficient absorption of light by the acetone. Where a better reducing agent is desired it is possible to produce eaq- by the photolysis of sulfite: but the production rate is quite low. Oxidation reactions can be initiated in a number of ways. Most common is the use of a flow-mixing system to bring about the reaction of reducing metal ions such as Ti3+ or Fe2+ with H202 to produce OH.6 Alternatively hydroxylamine or peroxydisulfate &Os2-) can be used to produce $-W27,s or SO4.-.9 These oxidizing radicals then react with the desired solute.
The redox systems, while very useful, suffer from a number of complications as the result of the large number of components and side reactions involving the metal ions.lO*ll The direct photolysis of aqueous solutions of H202 is simpler but without very high concentrations of the organic solute does not give adequate signal intensities.12 In this paper we wish to suggest the photolysis of S 2 0 ~ as ~ -a flexible method of in situ radical generation and to describe some results obtained in this way. The photolysis of S2Os2- has been postulated to result in formation of two Sod.- radicals13
hu
S20F32-
2s04.-
and the flash photolysis study by Dogliotti and Hayon14 identified the optical spectrum o f SOP-. This radical is also formed in radiation chemical systems by reaction of eaq-.15 eaq-
+ SzOs2-
-
SO&-
+ S042-
(2)
Norman et al.9 and Gilbert et a1.16 have studied by ESR some reactions of SOY- in flow systems. They found that The Journal of Physical Chemistry, Vol. 79,No. 24, 1975
2694
SOP- reacts readily with a number of organic compounds in ways similar to those displayed by OH, but that SO4.shows a greater tendency toward direct oxidation by one electron transfer. The present paper will focus more on the reactions with inorganic ions but will also discuss several organic systems previously studied. Some optical absorption and conductometric pulse radiolysis experiments were also carried out to provide supplementary information on several reactions. Experimental Section Photolysis studies were carried out with the lamp and optical system previously described.17 Associated ESR measurements were made with the spectrometer also described earlier.18 The aqueous ESR cell was of conventional design, 1 cm wide and 0.5 mm internal spacing. Hyperfine splittings were determined by means of a field-tracking NMR unit and frequency counter, and g factors were determined directly from field and frequency measurements. Corrections were made for the field difference between the NMR probe and ESR sample. The in situ radiolysis ESR measurements were made with the previously described a ~ p a r a t u sThe . ~ apparatus for pulse c o n d u ~ t i v i f yand ~ ~ kinetic spectrophotometric measurernents2O have also been described. Some of the work for the pulse optical studies used computer control as described; the rest made use of conventional photographs of the oscilloscope traces. The inorganic chemicals were generally Fisher certified or Baker analyzed reagents; many were the same as used in previous work. The Na2S208 was obtained from Sigma Chemical Co. and fumaric acid from Baker Chemical Co. The nitromethane was Eastman Spectrograde. Direct Observation of Inorganic Radicals
The optical absorption spectrum of S20a2- has an extinction coefficient of about 25 a t 250 nm and rises sharply toward shorter wavelengths. Because of this low value it is not possible to get full absorption of the incident light with a 0.5-mm path length. In practice concentrations of 30-50 mM were found to give adequate radical production rates in many systems. Photolysis of such solutions a t various pH values in the absence of other solutes never gave a detectable ESR line from S01.- itself. On the basis of the signals from other radicals to be described below and the recombination rate constant21 2k = 1.4 X lo9 M-l sec-l, it must be concluded that the line width is greater than about 10 G. The SOP- radical can be used to produce radicals from a number of inorganic anions. In many cases the reactions are similar to those of OH22923 and no detailed,discussion will be given. These include production of SO3- from sosZ-,HPOz- from H2P02-, and P032-from HP032-. (In the latter case the acidic form P03H- reported by Gilbert et al.24 was also observed and its pK investigated; see below.) The parameters for these radicals are given in Table I and are found to agree with previous results. HOWever, the line widths for Po32-and HP02- are found to be narrower by about a factor of 2 than reported by Behar and Fessenden.22 Most likely the lines in the latter experiments were broadened by overmodulation although attention was paid to this point. In no instance during these studies was any indication found of a thermal reaction between S2Os2- and the other added substances. Apparently these reactions are sufficiently slow as not to cause problems. Nevertheless, the The Journal of Physical Chemlstry, Vol. 79, No. 24, 1975
Om P. Chawla and Richard W. Fessenden
possibility of such reactions should be kept in mind, particularly with more readily oxidized substances. The Radical POfl-. Gilbert et al.24 reported lines of two different radicals from acid (pH 3-6) solutions of phosphite and identified the radicals as POSH- and P0z2-. Because of interest in the pK values of radicals such as COzand Cor-, it was thought worthwhile to determine that for POSH- for comparison.
+
POSH- P0s2- H+ (3) Experiments were carried out at several pH values where both forms were found (see Table I for the magnetic parameters) and the logarithm of the ratio of concentrations plotted against pH. (The concentration was taken as proportional to the product of the line height and the square of the line width on a first derivative spectrum.) Such an approach is valid because the protonation equilibrium 3 must be faster than the chemical decay of either form. The pK value so determined is 5.9 with an estimated error of f0.2 unit. The Radical c02-. Oxidation of formic acid, or formate, in the Ti3+-H202 system has been reported to produce an ESR line with g = 2.0002 under acid25conditions and a line at g = 2.0003 under basic26 conditions. No line width was stated in either of these studies. These lines have been assigned to CO2H and CO2- by comparison with the data on CO2- in crystals of sodium formate where the isotropic value is 2.0006.27 When 30 mM Na2S208 is photolyzed in the presence of 1-10 mM formate, a single line at g = 2.00045 with width 2.3 G is observed. This line is similarly assigned to COz-. Because of some uncertainty in the pK of this experiments were carried out over a range of pH from 4.6 to 0.8. The g factor remained constant at 2.00045 f 0.00002 over this range while the line width changed from 2.30 f 0.05 only at pH 0.8 where an apparent width of 2.6 was found. Below pH 3, a decrease in intensity occurred so that at pH 0.8 only 25% of the signal for higher pH was observed. Because of this decrease, experiments were not practical a t still lower pH. At pH values below 2, two weak lines appeared a t lower field. Because these decreased markedly in intensity as the flow rate was increased, they must be from secondary products and no further study of them was made. The doublet spectrum reported25 for CO2H a t -pH 1 was not found. From the fact that the parameters do not vary with pH and that the intensity is only somewhat decreased a t pH 2, it seems likely that the pK of COzH is below 2, as found by Buxton and Sellers.28The large line width for this radical is probably the result of spin-rotation interaction and the fact that the moment of inertia about the axis parallel to the 0-0 direction (the OCO bond angle is 127°)27is very small. Because SOP- has been found to decarboxylate simple aliphatic carboxylic acids,9J6 there is a question as to the nature of the reaction with formate. The reaction could be either abstraction of a hydrogen atom to form (Le., H+ S042-)and COz- or direct oxidation leading to a hydrogen atom and COz. In the latter case, the hydrogen atom would most likely abstract from another formate to produce H2 and the C02- observed by ESR. In separate pulse conductometric experiment^^^ for another purpose, it has been found that SO4.- HC02- leads to production of an equivalent of acid, demonstrating the simple abstraction route. The Radical COS.-. The ESR spectrum of COY- in single crystals of KHC03 has been reported by Chantry et
+
+
2695
Photochemical Production of Radicals in Solution TABLE I: ESR Parameters for Directly Observed Radicals
line Source
Radical
HC02HCQB-
eo2-
a, G
width, G
ga
u p = 480.2, a H = 89.8
2.3 4.4 27 0.15
2.00045 (l)b 2.0113 (1) 2.0155 (5) 2.00291 (0.5) 2.00158 (0.5) 2.00156 (0.5) 2.00344 (0.5) 2.0057 (1)
C93:
H20,
H02C
HzP02HP032-
H P03H-die
=
f
651.1
u p = 566.6 0.27 ~ ~ ( 2 ) = 11.61, ~ ~ ( = 4 11.21 ) 0.18 fiZH4: N2H5' H 2NO aN = aH(2)= 12.8 3.5 ",OH a Accuracy (units in the fourth decimal place) is indicated by the figure in parentheses. Previously reported g factor is 2.0003.26 See ref 34 for a review of the parameters for HOz. d Parameters are calculated exactly; see R. W. Fessenden, J. Magn Reson, 1, 277 (1969). Values for HP02- and PO32- are in agreement with those reported by Behar and F e s ~ e n d e n e. ~These ~ radicals have been reported by Gilbert et al.24 f Varies with pH from 0.4 G a t pH 2 to 1.5 G a t pH 6.0. The midpoint of the change is a t pH 4.6. HPOS2-
po32- d , e
al.30 who found g = 2.0112. The unpaired electron in this radical resides mainly on the oxygen atoms. An ESR spectrum for the radical in solution has not been reported. Photolysis of S20s2- with 20 mM NaHCO3 (pH 8.6) gave a single line with g = 2.0112, in good agreement with the data from the solid. The width was 4.4 G . Here there is even more question of the pK of the radical. Chen et al.31 have reported some kinetic data which seem to demonstrate a pK of 9.5 for the equilibrium C63H e C63-
+ H+
(4)
Experiments over the pH range 7.5-11.9, however, showed no change in either line width or g factor as shown in Table 11. In particular, the g factor remains remarkably constant. One is forced t o conclude either that dissociation 4 has no effect on g factor or line width or that the pK for this radical is outside the range studied. Although a change in g factor upon protonation of C 0 3 - seems probable, these data cannot directly contradict the previous finding regarding the pK. The Radical HOP. Tsao and Wilmarth p0stulated3~the reaction of SO&- with H202 to form H62 in both the thermal and radiation chemical reaction of 5 2 0 ~ ~ and - H2Op. The ESR spectrum of H62 is known from flow experim e n t using ~ ~ ~ceric oxidation of H202 and also from experi-, ments with reducing metals under conditions where no complex is formed with the metal ion.34 (In the latter case OH is produced which then reacts with H2Oz.) The g factor for the radical seems to be in the range 2.0140-2.0165.34 Photolysis of 30 mM S20s2- in the presence of 0.4% H202 (0.12 M ) at pH 1.8 gave a broad ESR line. No significant intensity was present when this concentration of H202 was used without S20s2-. This line, assigned to H62, appears at g = 2.0155 f 0.0005 and is 27 G wide, in excellent agreement with Saito and B i e l ~ k iThe . ~ ~same ESR line was evident to about pH 3.7 where loss of intensity and apparent broadening occurred. A decrease in intensity as a result of the dissociation with a pK of 4.8835 is expected, but it seems unlikely that broadening should occur with such a pK.36 The weakness of the signal precluded further study at the higher pH values. &2H4+ and H2NO. The radicals hj,H4+37 and H2NO-3s,39have previously been prepared in ESR experiments by the oxidation of hydrazine or hydroxylamine with Ce'" in flow-mixing experiments. Photolysis of 30 mM S ~ 0 8 in ~ -the presence of 5 mM hydrazine a t pH 2.4 or 0.5 mM hydroxylamine at pH 6.1 led in each case to the spec-
TABLE 11: ESR Parameters of COS.- as a Function of oHa Line width'
PH
Rb
7.50 8.35 9.03 9.31 9.55 10.27 11.85
2.01129 2.01131 2.01128 2.01128 2.01 127 2.01 126 2.01128
3.98 4.48 4.46 4.42 4.53 4.34 4.05
a The solute concentrations were: NaZS208, 30 mM; NaHC03, 7.1 mM; Na2B407, 8.1 mM. b Average of three measurements; the error in one measurement is estimated to be CO.7 X W4. Average of three measurements; estimated error in one measurement is 0.4 G.
trum of the expected radical. The parameters (summarized in Table I) are in agreement with the previous values. Radical Trapping In order t o study other inorganic radicals which cannot be observed directly because of broad lines, it is convenient to resort to radical trapping. Two agents have been used here: which is effective at all pH values but is best where the carboxyl groups of the resultant adduct are dissociated (so that long radical lifetimes are obtained), and the nitromethane aci a n i ~ n , which ~ ~ , ~only ~ exists above about pH 9. In basic solution (necessary in the latter case), the reaction of SOP- with the intended substrate will be in competition with the reaction with base SOP-
+ OH-
-
OH
+ S042-
(5)
with a rate constant of 7.3 X lo7 A4-l sec-' (see the discussion below) so that some care must be taken to distinguish the reactions of SO4.- from those of OH. The main reaction of SO*.- with nitromethane as described by Edge et al.42is oxidation which produces the dimeric species 02NCH2CH2N02-.
+ CH2=N02CH2NO2 + CH2=N02SOP-
-+
CH2N02 + sod2-
(6)
02NCH2CH2N02-
(7) In addition, we have found what appears to be an adduct of soso- to CH2=N02-. An experiment with 1.5 mM nitro-+
The Journal of Physlcal Chemistry, Vol. 79,No. 24, 1975
2696
Om P. Chawla and Richard W. Fessenden
Flgure 1. Second-derivative ESR spectrum obtained on photolyzing 30 mM Na2S208 solution containing 1.5mM CH3N02 at pH 10.5. Magnetic field increases to the right. Lines of three differentspecies are observed. The stick spectrum marks the position of lines ,of -03SOCH2N02while the lines of t h e OH adduct are indicated by the letter A. The intense triplets are from the dimeric species 02NCH2CH2N02-.
methane at pH 10.5 gave the spectrum in Figure 1. This spectrum shows the lines of 02NCH2CH2N02- as a major species, a very weak set of lines attributable to the OH adduct, and lines of intermediate intensity describable by U N = 23.63 G, aH(2) = 7.15 G, g = 2.00511. (The parameters for adducts to CH2=N02- are summarized in Table 111.) The latter parameters are distinguished by a rather low value of a N and do not match those of any adduct to CH2=N02- known to us. This spectrum is ascribed to the SOP- adduct -03SOCHzN02-. The lines from the two products of sod.- + CH2=N02- are useful indicators of whether a new solute introduced into the solution successfully competes for SOP-. A number of inorganic anions when introduced into the photolytic system mentioned above (namely, 30 mM S~OS~ and - -1 mM nitromethane at -pH 10) gave the same adducts to CH2=N02- as have already been found to result from OH reaction in a radiolytic system.23These include HC02-, HCOs-, S0s2-, HP032-, H2P04-, N3-, As02-, and HS-. No further discussion of these reactions will be given. The basic photolytic system for radical trapping with fumarate contained, typically, 30 mM S 2 0 ~ ~and - 3 m M fumarate. With no other solute present a strong spectrum of the adduct -02CCH(OS03-)CHC02- was observed at pH 6-8 as reported by Norman et al.9 At higher pH (>10.5),a decrease in intensity of these lines occurred, and lines attributable to the OH adduct also appeared. This radical must result from a conversion of SOP- to OH by base (reaction 5 ) . The signal intensity of the SOP- adduct at pH 6.1 is so high as to allow the radicals containing 13C at natural abundance to be detected.43 The parameters for the OH and SO4.- adducts are summarized in Table IV. Reaction of CN-. In the presence of cyanide and CH2=N02- the spectrum is as shown in Figure 2. The main lines are those of the adduct of CONH:! as previously de~cribed.4~ In addition, however, there is a second set showing a small nitrogen triplet of a N = 0.33 and a N = 27.70 G, a H ( 2 ) = 7.68 G, and g = 2.00511. The most obvious radical to account for this spectrum is the . C r N adduct N==CCH2N02-. Although the group which has added to CH2=N02- may contain other nuclei which are nonmagnetic or do not give observable ~ p l i t t i n g sthe , ~ ~pronounced tendency of SO4.- to react by electron transfer oxidation The Journal of Physical Chemistry, Vol. 79, No. 24, 1975
suggests that .C=N is probably formed. The relative intensity of the spectrum of the CONH2 adduct decreases at lower pH, implying that sCONH2 is formed by OH although formation from sod-- is not ruled out. The value of u N for the NO2 nitrogen of the adduct identified as NzCCH2NO2- is unusually large. It has been proposed46 that variations in the NO2 nitrogen splittings are the result of changes in the hybridization at the nitrogen. The triple bond of the CN group is clearly special and may interact rather strongly with the main radical site to increase the bending angle of the NO2 group. Ammonia. Photolysis of the standard S2O&-CH3N02 system in the presence of 0.6 M ammonia at pH 10.0 gave a spectrum which is clearly attributable to the adduct H~NCH~NOZ-. The parameters are a N = 26.07, aH(2) = 9.17, aN(NH) = 1.22, and aH(NH) = 0.41 G; g = 2.00502. These values are similar to those reported by Edge and Norman47 but differ by more than the combined errors. Identical splittings were found in a radiolytic experiment in which OH reacted with "3, so the same radical is clearly involved. The present values are believed to be more accurate. The NH2 radical could also be trapped with fumarate; the parameters of the adduct are given in Table IV. A study of the pH dependence of the intensity of the spectrum showed a strong decrease below about pH 9 with -100 mM NH3 solutions. This result shows that the reaction of SO4.- with NH4+ is much slower than with "3. Phosphate and Pyrophosphate. Although the reactivity of OH toward phosphate ions is low, it has been found possible to produce and characterize radicals of the type HP64-, or P042- and to determine the pK values.48 When phosphate or pyrophosphate was added to the S20s2--CH3N02 photolytic system, an adduct showing a small splitting due to one slP nucleus was found (see Figure 3). The small additional splitting was 0.45 G for phosphate and 1.16 G for pyrophosphate; the other parameters are given in Table 111. These radicals are taken to be adducts of PO42- and P2O73- with bonding through oxygen as suggested by the expected electronic structure with the unpaired electron located mainly on the oxygen^.^^ Bonding to phosphorus would.produce a large 31Psplitting as is seen for the PO32- and HP02- adducts. When phosphate is added to the fumarate-S20& sys-
2697
Photochemical Production of Radicals in Solution TABLE 111: ESR Parameters for Adducts to CHz=N02Source
szo*2-
Adduct
aN
aH(CH2)
23.63 27.70 25.23
7.15 7.68 8.07
a
(other)
g
2.00511 aN(CN) = 0.33 2.00511 a N = 0.51 2.00505 a H ( l )= 0.35 aN = 1.22 2.00502 26.07 9.17 H,NCH~NO~-C 3" ~ ~ ( =2 0.41 ) 25.10 8.64 up = 0.45 2.00504 2-03POCH2N02Pod3' 24.59 7. a5 a' = 1.16 2.00503 3-06P20CH2N02p20T4a Hyperfine constants in gauss, accurate to about k0.03 G; g factors accurate to =t0.00005. Previously reported by Behar and Fessenden.23 Splitting by only one NH2 proton is observed. Previously reported by Edge and Norman.47 CNCN-
'03socH2N02NCCH,NO*-. H,NCOCH~NO~-~
a
TABLE IV: ESR Parameters for Adducts to Fumaratea Source
sod. sod-
Adduct
a orH
a
OBH
-O~CCH(OH)CHCO~-02 C CH (OS03-)6HC02-
20.21 20.38
15.58 8.70
-O~CCH(NH~)CHCO~-~
20.25
9.69
HPO~'c1-
'02 CCH (OP032-)CHCOz-C 'OzCCH(Cl)CHC02'
20.24 20.20
g 2.00320 2.003 19
aN = 3.47 aH(2) = 0.35
3" I
(other)
11.15 6.41
uCl = 13.23'
2.003 17 2.00323 2.00324
Hyperfine constants in gauss accurate to about 3=0.03 G; g factors accurate to zt0.00005. Observed at p H 10.0. C Observed a t pH 6.9; exact state of ionization of the phosphate group is not known. Some suggestion of a further splitting of the lines by -0.1-0.2 G was found. Value for 35C1; that for 37C1 is 11.00.
I
L J L - . L . U L11
4 I
1*oP*045G *OH:864 G4 I
-
~
. aN=Z510G
-
Second-derivative ESR spectrum obtained during photolysis of a 30 m M Na2S208solution containing 5 m M CH3NO2 and 15 mM Na3P04at pH 11.5. ylgure 3.
Flgure 2. Low-field third of the secondderivative ESR spectrum obtained during photolysis of a 30 mM Na2S208solution containing 5.5 mM NaCN and 5 mM CH3N02 at pH 11.2. Lines of the CN and 6 0 N H 2 adducts are indicated.
tem, a new adduct is formed at pH values around 6.5. This adduct is believed to be -02CCH(OP032-)CHC02- with parameters as shown in Table IV. At higher pH, -9-10, no such lines are found. The higher pH shifts the phosphate equilibrium toward HP042- and the phosphate radical equilibrium toward H P O C - . ~Either ~ the reaction of sod.with HP042- is less than that with HzPO4- or the addition of HP04.- to fumarate is slower than that of HzPO4.. Halides. Oxidation of Br- or C1- is expected to lead to the radicals Br2- or Cl2- which can oxidize CH2=N02and produce OzNCH2CHzN02-, as was found for Brz- in the radiolytic system. Addition of either of these ions decreased the intensity of the SOc- adduct (see Figure l)and increased that of 02NCH2CH2NO2- in accord with this view. More positive evidence for an intermediate produced from C1- was obtained with fumarate as a trap. When 9 mM C1- was introduced into the standard system, the spectrum shown in Figure 4 was observed. Two sets of lines
4re evident which can be assigned to the 35Cland 37Cladducts to fumarate -02CCH(CI)CHC02-. The measured hyperfine constants of 13.23 and 11.00 G are in the proper ratio as expected from the magnetic moments (0.8314 observed, 0.8322 expected). The complete disappearance of the lines of the SO4.- adduct shows that the reaction of so4- with C1- competes effectively with addition of Sodsto the fumarate. There is no doubt that some intermediate dhlorine radical is produced (see below). The net result of the reaction of that radical with fumarate is the addition of a chlorine atom. With Br- the lines of the SOP- adduct disappeared, but no new lines were detectable. It is possible either that an adduct was formed but that the lines were broad as a result of relaxation caused by the nuclear quadrupole moment or that no further reaction occurred and the active species remained as Brz-. SO4.- Reaction with Organic Compounds. The reactions of SOP- with some carboxylic acids were also studied during the course of this work. Reaction of SOP- with acrylate (1 mM, pH 6.9) gave an adduct with the parameters aHcl= 20.74, aHp = 24.95, and g = 2.00306. This result is in agreeThe Journal of Physical Chemistry, Vol. 79, No. 24, 1975
2698
Om P. Chawia and Richard W. Fessenden
W
-OOC-CH-CH-COOI c133
1
U
I
I
aC6323 G __I(
Figure 4. Second-derivative ESR spectrum obtained during photolysis of a 30 mM NaZS208 solution containing 9.0 mM CI- and 3 m M fumarate at pH 6.8. Lines from radicals containing the two chlorine isotopes are indicated. ment with the previous findings by Norman et al.9 The reactions at near-neutral pH of CH3C02- (20 mM), CH2(C02-)2 (6 mM), and -02CCH2CH2C02- (10 mM) all gave radicals resulting from decarboxylation, namely CHs (aH = 22.65 G, g 2.00250), CH2C02- (aH 21.17 G, g = 2.003221, and CH2CH2C02- ( a H a = 20.37, aHp = 23.60, g = 2.00324). At higher pH (about IO), lines of the H abstraction products became evident as a result of conversion of Sods- to OH. Production of CH3 from acetateg and of CH2CH2CO2- from succinate (as determined by CH2=N02- trapping)16 have been reported earlier. The present direct observations are in complete accord with those findings. The observation of only CH3 when SOU- reacts with CH&O2- shows that little abstraction occurs because the lines of CH2C02- are narrower than those of CH3 and are much more readily detected. Only a few percent abstraction would have produced observable lines of CH2C02-. This behavior can be compared with that of OH which decarboxylates malonic acid about 10% of the time but does not do so with malonate to a detectable extent (