Radical Addition Rates to Alkenes by Time-Resolved CIDNP: 2

---

9794

J. Phys. Chem. 1996, 100, 9794-9799

Radical Addition Rates to Alkenes by Time-Resolved CIDNP: 2-Hydroxy-2-propyl Radicals S. N. Batchelor and H. Fischer* Physikalisch-Chemisches Institut der UniVersita¨ t Zu¨ rich, Winterthurerstrasse 190, CH-8057 Zu¨ rich, Switzerland ReceiVed: February 20, 1996; In Final Form: April 1, 1996X

A novel time-resolved CIDNP technique is used to measure absolute rate constants for the addition of 2-hydroxy-2-propyl radicals to 10 1,1- and two 1,2- disubstituted alkenes in propan-2-ol-d8 and toluene-d8 solvents and in part in acetonitrile-d3 and cyclohexane-d12, at room temperature. 2-Hydroxy-2-propyl radicals are produced by photolysis of 2,4-dihydroxy-2,4-dimethylpentan-3-one and lead to a multiplet effect CIDNP of the self-termination product, propan-2-ol, which follows a second-order termination kinetics. Addition of alkene partially quenches this polarization through the addition reaction. Quantitative measurement of the time-dependence and its simulation by kinetic equations yields the addition rate constant ka. The technique is applicable to systems with rate constants between 105 and 109 M-1 s-1, a range covered by the alkenes used. The data correlate with the alkene electron affinity and confirm the 2-hydroxy-2-propyl radical as very nucleophilic. A solvent dependence of ka up to 1 order magnitude is observed and is related to the hydrogenbonding ability of the solvent and radical and to the polarity of the solvent. A β-steric effect of the alkene substituents is also noted.

Introduction The addition of carbon-centered radicals to alkenes is an important reaction because it is the basis of polymerization and a prototype case of carbon-carbon bond formation. Due to this, the controlling factors have been the subject of much experimental and theoretical study.1 Steric effects at the attacked carbon of the alkene control the selectivity of the addition and lead to the exclusive addition at the unsubstituted carbon for 1,1-disubstituted alkenes, CH2dCXY. The rate constant is strongly dependent upon the radical and the alkene substituents. This is interpreted as a dependence of the activation barrier on the reaction enthalpy and on partial charge transfer (stabilizing polar interactions) in the transition state. Steric effects of the β-substituents X, Y are thought to be unimportant.1 The polar effect is particularly large for strongly electrophilic or nucleophilic radicals such as perfluoroalkyl2 and dicyanomethyl3 or tert-butyl4 and R-hydroxyalkyl5,6 radicals, respectively. For several other radicals the addition rate is controlled by the reaction enthalpy with only small polar effects.7 In a recent kinetic EPR study, 2-hydroxy-2-propyl radicals were shown to be extremely nucleophilic by correlation of the addition rate constants ka with the electron affinity of 18 alkenes CH2dCXY.6 ka varied between 102 and 108 M-1 s-1 with the alkene substituent at room temperature, but the rate constants above 106 M-1 s-1 could only be estimated since they were beyond the experimental range. Obviously, a proper determination of these higher rate constants is desirable, especially as the reactions of 2-hydroxy-2-propyl radicals have been of much interest in pulse radiolysis.8 Also, little attention has been given to a possible solvent dependence of the additions, which may occur due to the partial charge transfer in the transition state.4c Such a solvent dependence would also be valuable for the comparison of experimental and theoretical data,9 since the latter invariably refer to gas phase additions. In principle, flash photolysis may be applied to measure high rate constants, but the lack of an appropriate chromophore to monitor in many systems requires one to resort to competition X

Abstract published in AdVance ACS Abstracts, May 15, 1996.

S0022-3654(96)00510-2 CCC: $12.00

kinetics relative to a known standard.2,10 This increases the complexity and duration of the experiments, both of which are undesirable when a large number of rate constants are required. Chemically induced dynamic nuclear polarization, CIDNP, spectroscopy has been widely applied to radical reactions.11 The nuclear polarization of radical reaction products arises from nuclear spin dependent singlet triplet mixing in electron spincorrelated radical pairs and their subsequent spin selective reaction. In time-resolved studies with sub-microsecond time resolution, the temporal development of the polarization following laser flash creation of radical pairs may be studied, and the analysis has been shown to give reliable kinetic constants for simple radical recombination reactions.12 Recently, the 2-hydroxy-2-propyl radical has been extensively studied, 13 and the kinetics of the net and multiplet effect polarizations of the products were obtained. The magnitude of the multiplet polarization of propan-2-ol was shown to be proportional to the product concentration and hence to follow a second-order kinetics. Fitting the time-dependence of the multiplet polarization to the appropriate rate law yielded the self-termination rate constant of the radical. In the presence of an alkene the shape of the time dependence and the final magnitude of the polarization should be altered due to the addition of the radical to the alkene. The reaction reduces the 2-hydroxy-2-propyl radical concentration and thereby the amount of self-termination and the multiplet polarization created in this reaction. Accounting for this in the kinetic analysis should yield the addition rate constant. Here, this is used to determine the addition rate constants of 2-hydroxy-2-propyl radicals to 12 alkenes in two solvents, propan-2-ol-d8 and toluene-d8. Furthermore, the rate constants for three alkenes are also measured in acetonitrile-d3 and cyclohexane-d12. Experimental Section The experimental arrangement for time-resolved CIDNP has been described previously.14 In brief, samples are irradiated inside the probehead of a Bruker CXP 200 FT NMR spectrometer, using the 20 ns 308 nm output pulses of a Lambda Physik EMG 100 excimer laser. Post flash, the CIDNP spectra were © 1996 American Chemical Society

Radical Addition Rates to Alkenes

J. Phys. Chem., Vol. 100, No. 23, 1996 9795

recorded using a 0.5 µs NMR detection pulse (flip angle 30°) which detects net and multiplet polarization approximately equally.12b Samples were deoxygenated by repeated freezepump-thaw cycles and sealed under vacuum. 2-Hydroxy-2propyl radicals were produced from 2,4-dihydroxy-2,4-dimethylpentan-3-one, which was synthesized following ref 15 and purified by distillation. All alkenes and deuterated solvents were used in the purest commercially available forms. Optical densities were kept below 0.3 (4 mm path length), and the decomposition of the initial compounds was below 10% to avoid secondary photoreactions. All experiments were carried out at ambient temperatures (21 ( 2 °C). Results and Analysis (1) CIDNP in the Absence of Alkenes. The time-resolved CIDNP spectra obtained after photolysis of 2,4-dihydroxy-2,4dimethylpentan-3-one have been previously reported and explained in detail.13 After excitation, the ketone rapidly intersystem crosses into the triplet state from which R-cleavage occurs, within 50 ns at most. The subsequent geminate reactions are

HO(CH3)2CCOC(CH3)2OHT f HO(CH3)2CC˙ O + (CH3)2C˙ OH (1) HO(CH3)2CC˙ O + (CH3)2C˙ OH f HO(CH3)2CCOC(CH3)2OH (2a) f HO(CH3)2CCHO + CH2dC(CH3)OH

(2b)

f HO(CH3)2CCHO + CH3COCH3

(2c)

Spin mixing within the geminate acyl/2-hydroxy-2-propyl pair produces substantial net CIDNP in the products of reaction 2, the phases of which agree with those predicted by Kaptein’s rules.16 The net polarization is quantified as the absolute geminate polarization per pair, PG/∆m, and determined from the NMR integrals via eq 312a

ICIDNP PB P ) ∆m ∆m INMR

Figure 1. Time dependence of the multiplet effect CIDNP of propan2-ol following laser flash photolysis of 2,4-dihydroxy-2,4-dimethylpentan-3-one in propan-2-ol-d8. The initial radical concentrations were 15.7 × 10-5 M (O), 7.7 × 10-5 M (×), and 3.0 × 10-5 M (*).

However, multiplet polarization is created in the products propan-2-ol and the enol, but by symmetry not in acetone, 2,4dimethyl-2,4-dihydroxybutane, or the escaping 2-hydroxy-2propyl radicals. The polarization is proportional to the buildup of the product concentration and is not influenced by nuclear relaxation or CIDNP cancellation effects.12,17 Here, only the multiplet polarization of propan-2-ol will be considered, its final value, the multiplet polarization produced per reacting F-pair, P∞M/∆m, may be quantified by an equation matching (3),13 where INMR is the difference in intensities of the two adjoining lines and ∆m the concentration of reacting F-pairs. In this case of little geminate reaction the latter is equal to the initial geminate radical pair concentration. The time dependence of the propan-2-ol concentration is given by the solution of eqs 6 and 7

G

(3)

Here, ICIDNP is the sum of the integrals of the net polarized signals immediately after the laser pulse, INMR is the integral of the NMR signal of an arbitrary sample contained in the irradiated volume, PB is the corresponding Boltzmann polarization of this NMR signal, and ∆m is the concentration of the geminate radical pairs, determined from the product yield accumulated per laser pulse by NMR. The geminate reactions occur for only a few percent of radical pairs which reach the singlet state before diffusing apart. The escaping 2,2-dimethylacetyl radicals decarbonylate in several hundred nanoseconds.13

HO(CH3)2CC˙ O f (CH3)2C˙ OH + CO

(4)

Thus, to a good approximation, all subsequent nongeminate (F-pair) reactions involve only 2-hydroxy-2-propyl radicals with an initial concentration of 2∆m. These reactions are as follows:

2(CH3)2C˙ OH f (CH3)2C(OH)C(OH)(CH3)2

(5a)

f (CH3)2CHOH + CH2dC(CH3)OH

(5b)

f (CH3)2CHOH + CH3COCH3

(5c)

Since ∆g ) 0 for the encounter of two hydroxy-2-propyl radicals, no net polarization is produced in these reactions.

d[(CH3)2CHOH]/dt ) λkt[(CH3)2 C˙ OH]2

(6)

d[(CH3)2C˙ OH]/dt ) -2kt[(CH3)2 C˙ OH]2

(7)

which is

[(CH3)2CHOH] ) λ∆m

4kt∆mt ) λ∆mf(t) 1 + 4kt∆mt

(8)

λ is the fraction of radicals that reacts by disproportionation and not combination. The time dependence of the multiplet polarization is then

P∞M PM ) f(t) ∆m ∆m

(9)

Fits of this equation to the experimental results yields the selftermination constant kt. Figure 1 shows a typical fit of the time development of the multiplet effect to eq 9 for several different initial radical concentrations. In propan-2-ol, reaction 4 partially occurs within the geminate pair; hence, in this solvent some initial multiplet polarization is also created. This geminate multiplet polarization, PM,G/∆m was measured as 1.7 × 10-3 and accounted for as an initial offset in the fitting procedure. Table 1 gives the values of PG/∆m, P∞M/∆m, kt obtained for the four solvents, propan-2-ol-d8, toluene-d8, acetonitrile-d3, and cyclohexane-d12. They are in good agreement with the data

9796 J. Phys. Chem., Vol. 100, No. 23, 1996

Batchelor and Fischer

TABLE 1: CIDNP Magnitudes and Self-Termination Constantsa propan-2-ol-d8 toluene-d8 ACN-d3 cyclohexane-d12 2kt/109 M-1 s-1 PG/∆m/10-3 P∞M/∆m/10-3 a

2.2 1.8 18 18 12 12

4.4 3.6 4.8 5.0 1.8 1.7

6.2 3.7 1.9

7.6 7.0 10 8.0 4.5 4.1

Values in italics are from ref 13a.

Figure 3. CIDNP spectrum recorded following laser flash photolysis of 2,4-dihydroxy-2,4-dimethylpentan-3-one in propan-2-ol-d8 containing 0.065 M cyanoethene. The spectrum was recorded 100 µs post flash using 30° pulses. R represents the 2-hydroxy-2-propyl group.

Figure 2. CIDNP spectra recorded following laser flash photolysis of 2,4-dihydroxy-2,4-dimethylpentan-3-one in propan-2-ol-d8 containing various concentrations of cyanoethene. Spectra were recorded 100 µs post flash using 30° pulses. Only part of the spectra are shown.

obtained in a previous time-resolved CIDNP study,13a where kt was obtained from the net effect. However, the termination constants in ACN and propan-2-ol are significantly larger than observed in earlier kinetic EPR studies18 of 2.6 and 1.0 × 109 M-1 s-1, respectively. These solvents are extremely hygroscopic and, consequently, impossible to keep dry unless sealed as done here. Therefore, the difference is most probably due to wet solvents in the EPR experiments, which leads to a reduction of kt by strong hydrogen bonding to the water, for which 2kt is 1.0 × 109 M-1 s-1. A solvent isotope effect may also occur, though it is not expected to have a sufficiently large effect.19 (2) CIDNP in the Presences of Alkenes. Addition of alkene causes a considerable perturbation of the CIDNP spectra, and this is illustrated in Figure 2 for several concentrations of cyanoethene. As the concentration increases the multiplet effect of propan-2-ol is quenched, and new net signals appear in emission (CH3 groups, 1.0-1.5 ppm). These observations are consistent with the addition reaction

(CH3)2C˙ OH + CH2dCXY f (CH3)2C(OH)CH2C˙ XY (10) which leads to a quenching of the propan-2-ol multiplet polarization since 2-hydroxy-2-propyl radicals are lost. The additional products are created from the cross- and selftermination of the adduct radical, reactions 11 and 12

(CH3)2C(OH)CH2C˙ XY + (CH3)2C˙ OH f products 2(CH3)2C(OH)CH2C˙ XY f products

(11) (12)

Little net polarization is produced by these reactions since the g-factor difference between radicals is small for (11)20 and zero for (12). Therefore, the net CIDNP in the product CH3 groups is predominately due to the net emissive polarization of the

methyl protons of 2-hydroxy-2-propyl escaping the geminate reaction 2.12a,13a However multiplet polarization may be created in the new products, as illustrated in Figure 3 for a higher cyanoethene concentration. The new peaks, a1-c, may be assigned to the expected products of reaction 11 and 12, in accord with comparable reactions in the literature.21 These products indicate that little propan-2-ol is generated in crossreactions; i.e., they lead to a negligible contribution to its polarization. Similar CIDNP effects were observed with other alkenes, though, with a different dependence upon the alkene concentration. In order to obtain ka the multiplet effect time profile of propan-2-ol was measured at two radical and/or two alkene concentrations, yielding profiles with different shapes and final polarizations. Assuming that generally negligible propan-2-ol is created in cross-terminations, the time dependencies were then simultaneously fitted by solution of eq 6 combined with

d[(CH3)2C˙ OH]/dt ) -2kt[(CH3)2C˙ OH]2 ka[(CH3)2C˙ OH][CH2dCXY] (13) d(CH2dCXY]/dt ) -ka[(CH3)2C˙ OH][CH2dCXY] (14) The conversion to PM/∆m values was achieved in the same manner as in the analysis without alkene, though now f(t), eq 9, is the numerical solution. In this procedure ka was the only free fitting parameter, since kt was known from Table 1, and the alkene concentration was measured by NMR. The initial radical concentration 2∆m could not be obtained from the product concentration due to the numerous and varied new products. However, the alkene concentrations were sufficiently low so that no quenching of the geminate reaction or the triplet ketone occurred, and the spectrum with zero delay from the laser pulse showed predominately the geminate polarization with little effect of later reactions. From these net polarizations, combined with eq 4 and the known PG/∆m value (Table 1), ∆m was obtained. Typical fits are shown in Figure 4. Alkene and initial radical concentrations were used that gave final polarizations, P∞M/∆m, of 10-90% of the value without alkene (Table 1). In many cases the fit relied heavily upon these final polarizations since little time dependence was observable. The CIDNP spectra could be reliably measured for initial radical concentrations between ∼1 × 10-5 and 2 × 10-4 M. For several alkenes with ka larger than 108 M-1 s-1, low concentrations (10-410-3 M) of alkene were required, which exclude a possible

Radical Addition Rates to Alkenes

J. Phys. Chem., Vol. 100, No. 23, 1996 9797

Figure 5. Addition rate constants verse alkene electron affinities.

TABLE 3: Addition Rate Constants in Various Solvents Figure 4. Time dependence of the multiplet effect CIDNP of propan2-ol following laser flash photolysis of 2,4-dihydroxy-2,4-dimethylpentan-3-one in (i, (ii) propan-2-ol-d8 and (iii, iv) toluene-d8. Solutions contained (i) 3.3 × 10-3 M diphenylethene, (ii) 3.4 × 10-4 M (O), 7.5 × 10-4 M, cyanoethene (*), (iii) 0.062 M (O), 0.16 M (*), diphenylethene, and (iv) 0.052 M methacrylate. The initial radical concentrations were (i) 11.6 × 10-5 M (O), 2.1 × 10-5 M (*), (ii) 10.9 × 10-5 M, (iii) 11.8 × 10-5 M (O), 5.5 × 10-5 M (*), (iv) 11.6 × 10-5 M (O), 6.3 × 10-5 M (*).

TABLE 2: Addition Rate Constants of 2-Hydroxy-2-propyl Radicals to Alkenes CH2dCXY

Ea/eV22

H, Si(OEt)3 Cl, Cl H, CO2Me Me, CO2Me H, Ph Me, Ph H, CN Me, CN H, CHO Ph, Ph trans-XCHdCHX* CO2Me CN a

toluene-d8

propan-2-ol-d8

p/ta

-1.11 -0.76 -0.49 -0.38 -0.25 -0.23 -0.21 -0.17 0.03 0.36

3.2 × 105 6.5 × 106 1.7 × 106 4.1 × 105 1.1 × 105 3.7 × 107 7.3 × 106 6.6 × 107 8.7 × 105

1.7 × 1.3 × 106 2.7 × 107 6.2 × 106 4.2 × 105 1.2 × 105 2.2 × 108 7.7 × 107 3.0 × 108 4.1 × 106

4.1 4.2 3.6 1.0 1.1 5.9 10.5 4.5 4.7

1.25

1.9 × 108 5.5 × 108

3.3 × 108 1.4 × 109

1.7 2.5

105

p/t is the ratio of the values in the two solvents *per CH group.

quenching of the short-lived excited ketone singlet and triplet states. In these cases care was taken that the alkene consumption did not exceed 10%. To ensure the validity of the values given in Table 1, the alkene content was in no case more than 10% by volume (∼0.5 M). This criterion limited the method to rate constants above ∼105 M-1 s-1. Table 2 lists the alkenes and the addition rate constants in toluene and propan-2-ol solvents, together with the gas phase electron affinities of the alkenes.22 For CH2dC(H)Si(OEt)3 in toluene ka was too low to be measured. Errors are estimated to be within (40%, and the main source is the determination of the initial radical concentration, 2∆m. In a previous kinetic EPR study, rate constants for several of the above alkenes in propan-2-ol were reported.6 The data for CH2dC(H)Si(OEt)3 and the estimations for CH2dC(H)CO2Me and CH2dC(H)CN are 8.3 × 104, g1 × 107, and g1 × 108 M-1 s-1, respectively, are in reasonable agreement with the present findings. The earlier values for the two styrenes are considerably larger, which may be due to alkene depletion in the kinetic EPR experiment, a problem for fast reactions.6 Finally, for CH2dCCl2 a ka of 2.2 × 105 M-1 s-1 was reported, which is a factor of 6 lower than found here. There is no clear

CH2dCHX toluene-d8 CO2Me Ph CN

ACN-d3

6.5 × 106 1.8 × 107 4.1 × 105 3.9 × 105 3.7 × 107 8.4 × 107

propan-2-ol-d8 cyclohexane-d12 2.7 × 107 4.2 × 105 2.2 × 108

3.9 × 107 4.5 × 105 1.2 × 108

reason for this discrepancy, though the rate was at the limit of accuracy of the kinetic EPR experiment and errors may have accrued due to this. Some of the addition rate constants have also been measured recently by laser flash photolysis,10 and good agreement with the values of Table 2 is found; e.g., 1.1 × 108 M-1 s-1 was obtained for cyanoethene in methanol. Thus, the CIDNP method yields reliable addition constants. To provide further data on the solvent dependence of the rate constants, experiments were also performed in acetonitrile-d3 and cyclohexane-d12 for three of the alkenes. The results are given in Table 3. Discussion The addition rate constants show a marked variation with alkene substitution. When the high rate constants in propan2-ol-d8 are combined with the data for slower additions measured previously in propan-2-ol,6 a good correlation with the electron affinity of the alkene is observed, Figure 5. This is as expected for a strongly nucleophilic radical and may be quantified by fitting a straight line to the points. However, the phenyl-substituted alkenes have rate constants which are approximately 3 orders of magnitude lower than expected from their electron affinities and were not included in the linear regression. This gave

log10 ka/M-1 s-1 ) 8.28((0.24) + 2.81((0.19)EA/eV (15) with correlation coefficient, r ) 0.97, and a standard error of the estimate σ ) 0.50. The slope is different from than that reported previously,6 log ka/M-1 s-1 ) 6.46 + 1.71EA, when only alkenes with ka < 106 M-1 s-1 were used, and is greater than for other nucleophilic radicals, methyl,7d hydroxymethyl,5 benzyl, and tert-butyl23 with log ka/M-1 s-1 ) 5.72 + 0.90EA, log ka/M-1 s-1 ) 5.57 + 1.53EA, log ka/M-1 s-1 ) 3.36 + 1.14EA, and log ka/M-1 s-1 ) 5.95 + 1.61EA, respectively. Grossly, this reflects the order of the radical’s gas phase ionization potentials, methyl (IP ) 9.8 eV), hydroxymethyl (IP ) 7.56 eV) benzyl (IP ) 7.2 eV), tert-butyl (IP ) 6.7 eV), and 2-hydroxy-2-propyl (IP ) 6.5 eV), with the more easily oxidized radicals showing a stronger dependence upon the alkene EA. However, for the two

9798 J. Phys. Chem., Vol. 100, No. 23, 1996

Batchelor and Fischer

TABLE 4: Solvent Properties solvent

H-bonding27

28

R/10-24 cm3 28

ACN propan-2-ol toluene cyclohexane

acceptor donor/acceptor acceptor none

37 18 2 2

4.4 7.6 12.3 11.0

hydroxyalkyl radicals the slopes are markedly larger than those of radicals with similar ionization potential such as benzyl and tert-butyl. This is not surprising since hydrogen bonding to the solvent may increase their nucleophilicity more than expressed by the gas phase ionization potentials.24 For nucleophilic radicals the electron affinity effect reflects an electron-transfer from radical to alkene, which lowers the energy of the transition state. For the two easily reduced trans1,2-disubstituted alkenes25 the effect is very large. In spite of the steric hindrance caused by the substituents, the rate constants are 1 order of magnitude larger than for the comparable monosubstituted alkenes. For trans-1,2-dicyanothene the rate constant even approaches the diffusion limit of approximately 1 × 1010 M-1 s-1. As before,6 a correlation of the rate constants has also been attempted with the reaction enthalpy and the ionization potentials of the alkenes. Only very weak correlations were obtained, and no significance was placed in them as there is a matching weak correlation between the electron affinity and the reaction enthalpy and ionization potential.9a,23 Steric effects of β-methyl groups also influence the rate of reaction. Notably, in both propan-2-ol-d8 and toluene-d8 the alkenes CH2dC(Me)X (X ) CO2Me, CN, Ph) react approximately four times slower than CH2dCHX, despite small increases in the electron affinity. Probably, the increased size of the β-substituents provides a small extra steric barrier. Similar effects occur for tert-butyl, where an approximately 2-fold decrease is observed for these alkenes,4 although this has not been noted previously. For small nucleophilic radicals like methyl7d or hydroxymethyl5 no decrease is observed. Thus, the size of the β-substituent significantly alters the addition rate constants only for tertiary nucleophilic radicals of the size of 2-hydroxy-2-propyl (r ) 0.26 nm18). A significant solvent dependence is observed upon the rate constants (Tables 2 and 3) which increase in the series toluened8 < acetonitrile-d3 < propan-2-ol-d8 ≈ d12, except for the styrenes, and the magnitude of the effect varies with alkene. This solvent dependence may arise from several sources.26 Due to the partial electron transfer, the dipole moment of the transition state is likely to be greater than that of the reactants. In more polar solvents then, the transition state will be more stabilized by solvation which lowers the activation energy, and this explained the solvent dependence of tert-butyl radical addition to cyanoethene,4c where the rate constants increased in the series cyclohexane ≈ toluene < propan-2-ol < acetonitrile. Clearly, for 2-hydroxy-2-propyl radicals other factors must contribute. In particular, nonelectrostatic interactions such as dispersion forces and hydrogen bonding may perturb the transition state as well.26 The latter may be particularly important for 2-hydroxy-2-propyl, for which hydrogen bonding effects have been observed on the self-termination rate constants,18 the EPR spectrum,18 and nuclear polarizations.13 To allow discussion, solvent properties are given in Table 4. The importance of hydrogen bonding is suggested by a comparison of the rate constants for toluene and cyclohexane. These solvents have similar dielectric constants and polarizabilites so that the electrostatic and dispersion interactions should

be similar. However, in aromatic solvents 2-hydroxy-2-propyl radicals form hydrogen bonds with the solvent, whereas this does not occur in alkanes.29 This may explain the lower addition rate constant for toluene than for cyclohexane by steric and/or electronic effects. The same may also apply to the rate constants for the solvents acetronitrile and propan-2-ol, which are also lower than for cyclohexane. Intriguingly, little effect of hydrogen bonding to the solvent is seen for additions to styrene, which may indicate that their phenyl rings allow hydrogen bonding to the alkene in the transition state, which could cancel the effect of hydrogen bonding to the solvent. Such a hydrogen bonding between radical and alkene may be inferred from calculated transition state structures for hydroxymethyl radical addition to several alkenes.30 Within the group of hydrogen bonding solvents ka is larger for the more polar ones, acetonitrile and propan-2-ol, than for toluene. This indicates an electrostatic stabilization of the transition state, as found for tert-butyl radicals. The effect may be very roughly estimated from Kirkwood’s formula26,31 which excludes specific solvation.

{

}

2 2 2 1 NA  - 1 µR µAlk µq ln k ) ln k0 + 3 - 3 4π0 RT 2 - 1 r3 rAlk rq R

(16)

Here, k is the rate constant for a solvent of dielectric constant , k0 is the rate constant when  ) 1, µR, µAlk, and µq are the dipole moments of the 2-hydroxy-2-propyl radical, the alkene, and the transition state, respectively, and rR, rAlk, and rq are their radii. Taking the value for µR ) 1.7 D as that of propan2-ol,32 and a radius of 0.26 nm gives µR2/rR3 ) 1.8 × 10-30 A2 s2 m-1. For cyanoethene with µAlk ) 3.4 D32 and rAlk ) 0.27 nm one has µAlk2/rAlk3 ) 6.5 × 10-30 A2 s2 m-1.4c rq may be estimated as the sum of rR and rAlk and µq as the sum of the alkene and radical dipole moments plus an extra contribution due to the partial charge transfer. If in the transition state 60% of an elementary charge is transferred, the extra contribution becomes 7.2 D4c for charges separated by 250 pm9 and µq ) 9.8 × 10-30 A2 s2 m-1. Combining these with k0 ) 1 × 107 M-1 s-1 and  ) 2, 18, and 37 gives k ) 4 × 107, 2 × 108, and 2.3 × 108 M-1 s-1, respectively, in rough agreement with the solvent dependence observed, Table 3, though the experimental ka is smaller in acetonitrile-d3 than in propan-2-ol-d8. If only 20% electron transfer is assumed, the increasing solvent polarity should decrease the reaction rate constant. Hence, the solvent effect points to substantial charge transfer in the transition state. For the styrenes, the transferred charge in the transition state is probably delocalized over the phenyl rings. This decreases the dipole moment of the transition state and a smaller dependence of the rate constant upon solvent polarity should result, as is observed (Table 3). The delocalization also lowers the Coulombic attraction in the transition state, and this explains the lower than expected addition rate constants for the phenylsubstituted alkenes. The fall in rate constant between acetonitrile-d3 and propan2-ol-d8 is not compatible with the polarity. However, the oxidation potential of the 2-hydroxy-2-propyl radical varies with the solvent and is -1.3 V in water33 and -0.7 V in acetonitrile.34 Therefore, the radical may be more easily oxidized in propan2-ol-d8 than in acetonitrile-d3, which could explain the difference in addition rate constants as a reduction of the polar effect. The effect of dispersion forces upon radical addition reactions is unclear. In the Menschutkin reaction, which also involves a transition state with significant charge transfer, the rate has been shown to increase with the solvent polarizability.26,35 Hence,

Radical Addition Rates to Alkenes dispersion forces may significantly stabilize dipolar transition states. Here, no correlation of the addition rates is seen with the solvent polarizability, although it may also explain the difference between propan-2-ol and acetonitrile solvents. Conclusion A new method for the measurement of radical addition rate constants to alkenes, based upon time-resolved CIDNP experiments, has been presented. By application to the addition of 2-hydroxy-2-propyl radicals to various alkenes the method is shown to be capable of measuring rate constants between 105 and 109 M-1 s-1. The results are in good agreement with those obtained by other methods. The addition of 2-hydroxy-2-propyl radicals to alkenes represents a case of extreme polar effects; the rate constants correlate with the alkene electron affinity and vary by over 8 orders of magnitude with alkene substitution. Solvent effects of up to 1 order of magnitude are observed. In part they are attributed to changes in steric and/or electronic effects caused by hydrogen bonding of the radical to the solvent and to a solvent polarity effect. References and Notes (1) For reviews see: (a) Giese, B. Angew. Chem., Int. Ed. Engl. 1983, 22, 573. (b) Tedder, J. M. Angew. Chem., Int. Ed. Engl. 1982, 21. (2) (a) Avila, D. V.; Ingold, K. U.; Lusztyk, J.; Dolbier, W. R.; Pan, H.-Q. J. Am. Chem. Soc. 1993, 115, 1577; (b) J. Am. Chem. Soc. 1994, 116, 99. (c) Rong, X. X.; Pan, H.-Q.; Dolbier, W. R.; Smart B. E. J. Am. Chem. Soc. 1994, 116, 4521. (3) Riemenschneider, K.; Bartels, H. M.; Dornow, R.; Drechsel-Grau, E.; Eichel, W.; Luthe, H.; Michaelis, Y. M.; Boldt, P. J. Org. Chem. 1987, 52, 205. (4) (a) Mu¨nger, K.; Fischer, H. Int. J. Chem. Kinet. 1985, 17, 809. (b) Fischer, H. In Substituent Effects in Radical Chemistry; Viehe, H. G., Janousek, Z., Me´re´nyi, R., Eds.; Reidel: Dordrecht; 1986. (c) Salikhov, A.; Fischer, H. Appl. Magn. Reson. 1993, 5, 445. (5) Wu, J. Q.; Fischer, H. Int. J. Chem. Kinet. 1995, 27, 167. (6) He´berger K.; Fischer, H. Int. J. Chem. Kinet. 1993, 25, 913. (7) (a) He´berger, K.; Fischer, H. Int. J. Chem. Kinet. 1993, 25, 249. (b) Walbiner, M.; Wu, J. Q.; Fischer, H. HelV. Chem. Acta 1995, 78, 910. (c) Wu, J. Q.; Beranek, I.; Fischer, H. HelV. Chem. Acta 1995, 78, 194. (d) Zytowski, T.; Fischer, H. J. Am. Chem. Soc. 1996, 118, 437. (8) See: Landolt-Bo¨rnstein. Radical Reaction Rates in Liquids; Fischer, H., Ed.; New Series; Springer Verlag: Berlin, Heidelberg, New York, 1995; Volume IIB, pp 223-275. (9) (a) Wong, M. W.; Pross A.; Radom L. J. Am. Chem. Soc. 1994, 116, 6284. (b) Zipse, H.; He, J.; Houk, K. N.; Giese B. J. Am. Chem. Soc. 1991, 113, 4324. (c) Houk, K. N.; Paddon-Row, M. N.; Spellmeyer D. C.; Rondan, N. G.; Nagase, S. J. Org. Chem. 1986, 51, 2874. (d) Fueno, T.; Kamachi, M. Macromolecules 1988, 21, 908. (e) Gonzales C.; Sosa, C.; Schegel, H. B. J. Phys. Chem. 1989, 93, 2435; (f) 1989, 93, 8388. (g) Arnaud, R.; Vidal, S. New J. Chem. 1992, 16, 471. (h) Clark T. J. J. Chem. Soc., Chem. Commun. 1986, 1774. (i) Tozer, D. J.; Andrews, J. S.; Amos, R. D.; Handy, N. C. Chem. Phys. Lett. 1992, 199, 229. (j) Schmidt, C.;

J. Phys. Chem., Vol. 100, No. 23, 1996 9799 Warken, M.; Handy, N. C. Chem. Phys. Lett. 1993, 211, 272. (10) Farley, R. D.; Fischer, H. Manuscript in preparation. (11) For reviews see: (a) Chemically Induced Magnetic Polarisation; Muus, L. T., Atkins, P. W., McLauchlan, K. A., Pedersen, J. B., Eds.; Reidel: Dordrecht, 1977. (b) Spin Polarisation and Magnetic Field Effects in Radical Reactions; Salikhov, K. M., Molin, Yu. P., Sagdeev, R. Z., Buchachenko, A. L., Eds.; Elsevier: Amsterdam, 1984. (12) (a) Vollenweider, J. K.; Fischer, H.; Hennig, J.; Leuschner, R. Chem. Phys. 1985, 97, 217. (b) Vollenweider, J. K.; Fischer, H. Chem. Phys. 1986, 108, 365. (13) (a) Batchelor, S. N.; Fischer, H. J. Phys. Chem. 1996, 100, 556. (b) Salzmann, M.; Tsentalovich, Yu. P.; Fischer, H. J. Chem. Soc., Perkin Trans. 1994, 2119. (14) Schaffner, E.; Kweton, M.; Vesel, P.; Fischer, H. Appl. Magn. Reson. 1993, 5, 127. (15) Faworsky, F.; Umnova, A. J. Prakt. Chem. 2. 1912, 88, 679. (16) Kaptein, R. Chem. Commun. 1969, 732. (17) Morozova, O. B.; Tsentalovich, Yu. P. Yurkovskaya, A. V.; Sagdeev, R. Z. Chem. Phys. Lett. 1995, 246, 499. (18) Lehni, M.; Fischer, H. Int. J. Chem. Kinet. 1983, 15, 733. (19) Tyrell, H. J. V.; Harris, K. R. Diffusion in Liquids; Butterworth Series; Cambridge University Press: Cambridge 1984. (20) Landolt-Bo¨rnstein. Magnetic Properties of Free Radicals; Fischer, H., Hellwege, K. H., Eds.; New Series; Springer Verlag: Berlin, Heidelberg, New York, 1965-1986; Volumes II/1, II/9b, II/17b. (21) Hammond, G. S.; Wu, C. H. S.; Trapp, O. D.; Warkentin, J.; Keys, R. T. J. Am. Chem. Soc. 1960, 82, 5394. (22) Lias, S. G.; Bartmess, J. E.; Liebmann, J. F.; Holmes, J. L.; Levin, R. D.; Mallard W.G. J. Phys. Chem. Ref. Data 1988, 17, Suppl 1. (23) He´berger, K.; Walbiner, M.; Fischer, H. Angew. Chem. 1992, 104, 651; Angew. Chem., Int. Ed. Engl. 1992, 31, 635. (24) Klemt, R.; Roduner, E.; Fischer, H. Chem. Phys. Lett. 1994, 229, 524. (25) Petrovich, J. P.; Baizer, M. M.; Ort, M. R. J. Electrochem. Soc.: Electrochem. Sci. 1969, 116, 743. (26) Reichardt, C. SolVents and SolVent Effects in Chemistry; VCH: New York, Weinheim, Cambridge, 1988. (27) Connors, K. A. Medium Effects; VCH: New York, Weinheim, Cambridge, 1990. (28) Handbook of Chemistry and Physics; Weast, R. C., Ed.; CRC Press: Boca Raton, 1985. (29) Effects of this are observed when the diffusion-controlled selftermination constants of the similar sized 2-hydroxy-2-propyl and tert-butyl radicals are compared. In benzene the self-termination rate constants for 2-hydroxy-2-propyl are much smaller than tert-butyl radicals and have a different temperature dependence, but in n-heptane they are alike: Schuh, H. H; Fischer, H. HelV. Chim. Acta 1978, 61, 2463; ref 18. (30) Supplementary material for ref 9a. (31) (a) Kirkwood, J. G. J. Chem. Phys. 1934, 2, 351. (b) Glasstone, S.; Laidler, K. J.; Eyring, H. The Theory of Rate Processes; McGraw Hill: New York, London, 1941; pp 400ff. (32) McClellan, A. L. Tables of Experimental Dipole Moments; W. H. Freeman: New York, 1963. (33) Lilie, J.; Beck, G.; Henglein, A. Ber. Bunsengs. Physik. Chem. 1971, 75, 458. (34) Griller, D.; Wayner, D. D. M. Pure Appl. Chem. 1989, 61, 717. (35) Reinheimer, J. D.; Marley, J. D.; Meyers, W. W. J. Org. Chem. 1963, 28, 5394.

JP9605109