Mechanistic aspects of hydrogen abstraction for phenolic antioxidants

1 downloads 0 Views 567KB Size Report
substituted phenols with hydroperoxyl radical have been determined. ... phenol hydroperoxyl reactions proceed according to a proton coupled electron transfer ...
RESEARCH PAPER

PCCP

www.rsc.org/pccp

Mechanistic aspects of hydrogen abstraction for phenolic antioxidants. Electronic structure and topological electron density analysis Nakul Singh, Patrick J. O’Malley* and Paul L. A. Popelier School of Chemistry, Sackville Site, The University of Manchester, Manchester, UK M60 1QD Received 28th September 2004, Accepted 24th December 2004 First published as an Advance Article on the web 24th January 2005

Density functional calculations using the B3LYP functional are used to provide insight into the hydrogen abstraction mechanism of phenolic antioxidants. The energy profiles for 13 ortho, meta, para and di-methyl substituted phenols with hydroperoxyl radical have been determined. An excellent correlation between the enthalpy (DH) and activation energy (DEa) was found, obeying the Evans–Polanyi rule. The effects of hydrogen bonding on DEa are also discussed. Electron donating groups at the ortho and para positions are able to lower the activation energy for hydrogen abstraction. The highly electron withdrawing fluoro substituent increases the activation energies relative to phenol at the meta position but not at the para position. The electron density is studied using the atoms in molecules (AIM) approach. Atomic and bond properties are extracted to describe the hydrogen atom abstraction mechanism. It is found that on going from reactants to transition state, the hydrogen atom experiences a loss in volume, electronic population and dipole moment. These features suggest that the phenol hydroperoxyl reactions proceed according to a proton coupled electron transfer (PCET) as opposed to a hydrogen atom transfer (HAT) mechanism.

1. Introduction Phenols are the most abundant and widely used natural and synthetic antioxidants. For example, the phenol a-tocopherol (vitamin E) is the most important lipid-soluble, chain breaking natural antioxidant in human blood plasma and low-density lipoprotein. It inhibits the auto-oxidation of organic molecules, a reaction that has been studied extensively.1 The mechanism of action depends primarily on the ability of an antioxidant to donate its phenolic H-atom to a chain carrying peroxyl radical at a rate much faster than that at which the chain-propagating step of the lipid peroxidation proceeds.2 In the reaction, a peroxyl radical ROO removes a hydrogen atom from the antioxidant (ArOH) reaction (1). ROO þ ArOH - ROOH þ ArO

(1)

ROO þ ArO - non-radical products

(2)

DOI: 10.1039/b415075a



614



The efficiency of the antioxidant ArOH depends on the stability of the aryloxyl radical ArO . In most systems the aryloxyl radical terminates the chain and decomposes to non-radical products by reacting with another peroxyl radical, reaction (2). In recent years, it has become feasible to carry out high-level electronic structure calculations to obtain accurate geometries and energies of transition states and compute the activation energies. While this approach is generally successful in a quantitative sense, this approach is difficult because it often involves locating the transition state. These difficulties make the widespread use of electronic structure calculations too expensive for large-scale studies; furthermore various effects that influence the height of the barrier are quite often unknown. The qualitative empirical approach to the estimation of the activation energy Ea has been studied thoroughly.3 The basis of these empirical procedures relies on proper theoretical treatment of electronic effects at the transition state. These approaches, though successful for a large number of simple radical reactions, have failed to reproduce accurate Ea activation energies for hydrogen abstraction for reactions specifically Phys. Chem. Chem. Phys., 2005, 7, 614–619

involving phenolic systems. Such failures are explained by invoking low triplet repulsion and inaccurate calculation of bond dissociation energies.4,5 Alternatively, the failure to reproduce accurate Ea can be ascribed to hydrogen bonding prior to hydrogen abstraction.6 The effect of substituted phenol has been linked with the rate of reaction and the bond dissociation energies of the OH bond.7 Recent studies of bond dissociation energies have revealed that electron donating groups substituted ortho and para to the phenolic hydroxyl group lower the bond dissociation energies and increase the rate of hydrogen atom transfer to peroxyl radicals.8–10 A thorough understanding of the mechanism of a hydrogen atom transfer (HAT) can be obtained if one can accurately model the potential energy surfaces of these reactions.11–13 We have investigated a series of reaction profiles of the H-atom transfer of eleven singly substituted and two di-methyl substituted phenols with hydroxyperoxyl radical, Scheme 1. Another property that has been related to the activity is the electron density in the O–H bonding region.14 Electron density analysis, in this report, is based on the theory of atoms in molecules (AIM).15,16 The theory has recently been employed in quantitative structure activity relationship studies of phenolic antioxidants.17,18 Such structure activity relationship studies have proved useful in understanding the mechanism of action.7,19,20 In particular, the use of AIM descriptors will enable us to interpret for the first time the nature of the electronic effects of the hydrogen atom along the potential energy surface, see Scheme 1.

2. Method The geometries of the reactants, intermediates and transitionstate structures have been determined at the UB3LYP/ 6-31G(d) level of theory. All energies are corrected for zero point energies. All transition states were verified with frequency calculations to yield one imaginary frequency. All

This journal is & The Owner Societies 2005

Scheme 1 Reaction profiles for reactions of HOO with mono substituted phenols. A ¼ isolated reactants, B ¼ hydrogen bonded reactant complex, C ¼ transition state, D ¼ hydrogen bonded product complex, and E ¼ isolated products.

calculations were performed using the GAUSSIAN98 program.21 Single point energy calculations were then performed at UB3LYP/6-311G(2df,p) level. Spin contamination is negligible (hS2i less than 0.78) for all reactions. The AIM analysis was performed with MORPHY98.22 The integration errors L(O) all fall below the acceptable threshold of 1  103 au.

3. Results and discussion 3.0 Energetics The effect of substituents on activation parameters for hydrogen abstraction from the phenol hydroxyl group by the hydroperoxyl radical is discussed. The parameters are shown in Table 1 along with all relative energy values along the reaction profile. Comparison of activation energies with experimentally derived values is not possible here. However, for vitamin E derivatives range between 3.8 and 6.9 kcal mol1 for reactions with oxygen centred radicals.23 Compared with phenol, all electron donating groups lower the activation energies with the exception of the meta methoxy substituent. The ortho methyl substituent is also able to lower the activation energy with respect to phenol. The electron withdrawing meta fluoro substituent increases the activation energy of a hydrogen abstraction by 1.2 kcal mol1. Para substituents lower the activation energy considerably even for the highly electron withdrawing fluorine substituent. The

Table 1

Fig. 1 Plot of DEa vs. DH. All values are in kcal mol1. The equation is DEa ¼ 0.57DHr 0.98 with r2 of 0.98.

methoxy group is also electron withdrawing by inductive effects, however, through p electron donation, is capable of delocalising electrons on the para C–O bond. Though the meta and para methoxy substituents have strong electron-releasing effect, the DH values differ, respectively, by 1.2 and 6.1 kcal mol1 and the DEa by 0.6 and 3.4 kcal mol1, respectively, compared with phenol. The ability of an antioxidant to form a stable radical will be reflected in its overall enthalpies. The most thermodynamically favourable reaction is the methoxy substituent at the para position. Both di-methyl substituted groups are able to lower the activation energies relative to phenol because the methyl groups are electron donating. A direct correlation between the activation energy and the overall enthalpy of reaction follows the principle of Evans and Polanyi, Fig. 1.24 The Evans and Polanyi relationship states that within a series of closely related atom-transfer reactions the activation energy displays a linear dependence on the reaction enthalpy. Accordingly, this correlation clearly indicates that the reactions follow the same mechanism. It has been rightly pointed out by Mayer et al.25 that this is an empirical thermodynamic relationship that does not apply to chemical reactions in general or even to other types of radical reactions. The broader applications of this powerful mechanistic relationship are extensively discussed elsewhere.25 In view of the limitations it is rather gratifying that the use of this relationship has been found to aid the development of a new class of

Energy values along the potential energy surface. All values are in kcal mol1 relative to its isolated reactants

N

X

HB reactant complexa

Transition stateb

HB product complexc

Enthalpy DHd

DEae

1 2 3 4 5 6 7 8 9 10 11 12 13

H o-Me m-OMe m-Me m-Et m-F p-OH p-OMe p-Me p-Et p-F o,m-Me o,p-Me

4.7 4.7 4.7 4.6 4.6 5.3 4.7 4.7 4.7 4.6 5.0 4.7 4.6

1.9 0.9 2.5 1.7 1.6 2.5 1.3 1.5 0.7 0.8 0.6 0.5 0.2

9.1 9.5 6.3 7.9 7.8 6.2 12.9 13.2 9.4 9.2 9.3 10.1 11.4

1.4 0.4 2.6 1.2 1.3 2.5 3.4 4.7 0.2 0 0.3 1.1 2.1

0 1.0 0.6 0.3 0.4 1.2 3.2 3.4 1.2 1.2 1.0 1.4 2.1

a

HB reactant complex. Allowing for entropic considerations the hydrogen bonded intermediates are expected to have a short-lived existence. Transition state. c HB product complex. d Overall enthalpy DH ¼ E(ArO þ H2O2)  E(ArOH þ  OOH). e DEa ¼ Ea(X) – Ea(H). The absolute Ea for phenol is 1.9 kcal mol1 which is given relative to isolated reactants. Values are commonly quoted to 0.1 kcal mol1.38 b

This journal is & The Owner Societies 2005

Phys. Chem. Chem. Phys., 2005, 7, 614–619

615

antioxidants.26 Another remarkable feature of this correlation is that the reactivity DEa is independent of the hydrogen bond formation. The relatively constant value for hydrogen bond interaction with hydroperoxyl radical provides an explanation why ground state thermodynamic properties like enthalpies and bond dissociation energies, account so well for the reactivity of phenolic antioxidants. The strength of hydrogen bonds are not affected by substituents with variation of 0.7 kcal mol1 for all reactions studied. The similarity of the geometries can be used to explain the relatively constant value of hydrogen bond interaction energy. There are cases where the orientation of the hydrogen bond will predominantly influence the reactivity of phenol with peroxyl or aryloxyl radicals. The strength of these hydrogen bonds will differ from the values calculated here.27,28 The energy of hydrogen bonds between peroxyl radicals and water is experimentally determined to be 4.8 kcal mol1 in close agreement with the energy of hydrogen bonding between peroxyl radicals and phenol.29 3.1 Geometrical parameters Geometrical features of an H-atom donation starting from the hydrogen bonded reactant complex to transition state are discussed here. Substitution of phenols has little effect on the O –H bond length of the hydrogen bonded reactant complex. In the transition state, the O–H bond length is only slightly extended from parent phenol by 12–14% for all substituents. This result is in agreement with results obtained for a theoretical treatment of a similar reaction.30 Accordingly, there is a trend for an ‘‘early’’ transition state for all substituted phenols. The change in C6–O7–H8–O11 dihedral angle from ground state hydrogen bonded form to transition state structures is similar, 37.5–41.71. Similar studies for a hydrogen abstraction from phenol by methyl radical have shown that the hydrogen rotates 901 out of the aromatic plane.31 Another study by Mayer et al.28 for phenoxyl/phenol self exchange reactions have illustrated a distinction where a HAT takes place when the dihedral angle is 901 and proton coupled electron transfer takes place when the dihedral angle is 01 (Table 2). Rotation of the H-atom away from the aromatic plane may therefore mediate the activity of substituted phenols depending on the type of abstracting radical. 3.2 Electron density

evaluated at the BCP for each bond, which provide more chemical meaning than just the electron density alone. The properties evaluated in this study are: (i) rb, the electron density at the BCP; (ii) r2rb, the laplacian, which provides an indicator of whether electron density is locally concentrated or depleted; and (iii) e, the ellipticity at the BCP, which measures the extent to which r is elongated in one direction compared to another. For homopolar bonds, e can be used as a quantitative measure of p-bond character and as a measure of delocalisation.16 However, for heteropolar bonds with significant charge transfer between the atoms, the value of e at the BCP is not indicative of the preferred plane of polarization of the valence density.32 However, for our purposes this concern is not vital because we will only consider changes in the ellipticity of a C–O bond in going from the hydrogen bonded complexes (B) to the transition state (C). We expect such a comparison to be meaningful since the ellipticity is, relatively speaking, sampled at the same point (i.e. the BCP) along the C–O bond in hydrogen bonded complexes and transition states. Other indices of delocalisation may also be suitable.33 Atomic properties are obtained by integrating over the atomic basin O, which is the space the atom occupies.16,34,35 A basin is constructed by tracing gradient paths or paths of steepest ascent in the electron density. The collection of gradient paths originating at infinity and terminating at a nucleus constitutes the (atomic) basin associated with that nucleus. Three atomic properties used in this study are: the atomic volume, n(O) the atomic charge, q(O), and the atomic dipole moment, M1(O). The volume is defined as: Z uðOÞ ¼ dt ð3Þ O

The atomic charge q(O) is defined as: q(O) ¼ M0(O) þ ZO Z M0 ðOÞ ¼  dtrðrÞ

(4) ð5Þ

O

where M0(O) is the electronic population and ZO is the nuclear monopole moment. The atomic dipole moment is defined as: Z M1 ðOÞ ¼  dtrrðrÞ ð6Þ O

A bond critical point (BCP) is defined by AIM as a point at which the gradient of the electron density vanishes, roughly lying between two bonded nuclei. In this study, properties are Table 2 Listed O7–H8 bond lengths for structures HB reactant complex and transition state where distance is measured from oxygen on the parent phenol. The change in dihedral angle from parent phenol to transition state is also listed

616

N

X

RO  H(B)/A˚

RO  H(C)/A˚

D Dihedral(y) C6–O7–H8–O11

1 2 3 4 5 6 7 8 9 10 11 12 13

H o-Me m-OMe m-Me m-Et m-F p-OH p-OMe p-Me p-Et p-F o,m-Me o,p-Me

0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98

1.16 1.14 1.17 1.15 1.15 1.17 1.12 1.12 1.14 1.15 1.14 1.14 1.14

41.5 40.7 38.9 41.4 41.4 37.5 40.3 39.8 40.6 40.2 41.7 40.2 39.9

Phys. Chem. Chem. Phys., 2005, 7, 614–619

where r is a vector centred on the nucleus of the atom. The magnitude of the dipole moment, denoted by M1(O), gauges the what extent the centroid of the atom’s electronic charge cloud is shifted with respect to the nucleus. This is a measure of the (dipolar) polarization of the atom’s charge density. Bonded nuclei are connected by two gradient paths that share a common BCP. These two special gradient paths constitute the so-called bond path (BP). The network of bond paths for a molecule is called its molecular graph. A molecular graph is the network of BPs and it provides an unambiguous definition of a molecular structure. These graphs are generated through an analysis of topology of the charge density using the theory of AIM.35 All BCPs and BPs at the transition state for the H-atom abstraction of phenol by hydroperoxyl radical are displayed in Fig. 2. Also shown is the H-atom basin enclosed by two interatomic surfaces. An interatomic surface is a collection of gradient paths that share a common BCP and separate two bonded atoms. As discussed earlier, atomic properties are obtained by integrating over the atomic basin. A BP between an H-atom on the aromatic ring (H12) and oxygen (O14) on the peroxyl radical, show that the transition state is stabilised by an extra C–H  O bond. The strong curvature of this BP indicates that the interaction is weak and is easily broken.

This journal is & The Owner Societies 2005

Table 3 Atomic properties of transferred H-atom for all substituents. All values are in atomic units

Fig. 2 A molecular graph of the phenol  peroxyl radical transition state. The bond critical points, interatomic surfaces marking the boundary of the transferred hydrogen atom and bond paths are shown. The broken line represents a hydrogen bond.

Atomic properties are listed in Table 3. Hydrogen atom volumes decrease on going from free antioxidant to hydrogen bonded complexes. Values range from 23.04 au in the free phenols to 9.46 au in the transition state. The positive charge on the hydrogen atom increases upon hydrogen bond formation and is further increased at the transition state. In other words, in this process the hydrogen increasingly loses electronic population, which is transferred to the oxygen atoms. The dipole moment of the hydrogen atom at the transition state for each hydrogen abstraction reaction has an unusually low value and nearly vanishes. In all the systems studied, H undergoes a reduction in polarisation upon H-bond formation. At the transition state the polarisation declines by almost an order of magnitude. Distinction between pure HAT and PCET for hydrogen abstraction is often blurred. A recent attempt at such a classification has been put forward recently by Mayer et al.28 HAT is the familiar mechanism of organic free radical chemistry and involves hydrogen atom (electron accompanying proton) transfer. PCET involves separate proton and electron transfer. As a means of distinguishing the mechanism operative in the reactions of Scheme 1 we have looked at the AIM parameters for two well characterised systems which are known to operate by both mechanisms. The first involves toluene/benzyl radical hydrogen exchange and the second is phenol/phenoxyl hydrogen exchange. The toluene/benzyl radical is a classic example of pure HAT whereas phenol/phenoxyl has been shown to proceed via a PCET mechanism. We analysed the electronic wavefunction using the AIM theory for these systems in order to clarify the mechanism that takes place for reactions studied here. The results are shown in Table 4. The HAT mechanism, as exemplified by toluene/benzyl, proceeds with a large atomic volume and a small positive charge for the transferring hydrogen in the transition state structure. The PCET mechanism as exhibited by phenol/phenoxyl is contrastingly characterised by low atomic volume and high charge value. Therefore the reactions studied here, characterised by hydrogen atom volumes and charges similar to those found for the phenol/phenoxyl system, can be said to proceed via a PCET mechanism. AIM analysis, which should not be confused with a mere population analysis, may be an important and convenient means of distinguishing between these two mechanisms.

1

H

2

o-Me

3

m-OMe

4

m-Me

5

m-Et

6

m-F

7

p-OH

8

p-OMe

9

p-Me

10

p-Et

11

p-F

12

o,p-Me

13

o,m-Me

u(O) q(O) M1(O) u(O) q(O)) M1(O)) u(O)) q(O)) M1(O)) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O) u(O) q(O) M1(O)

A

B

C

D

22.95 0.5593 0.1611 22.99 0.5593 0.1614 22.90 0.5599 0.1608 22.97 0.5588 0.1613 23.00 0.5583 0.1617 22.76 0.5635 0.1592 23.01 0.5579 0.1616 23.03 0.5575 0.1617 23.04 0.5584 0.1615 23.04 0.5585 0.1614 22.86 0.5609 0.1603 22.99 0.5585 0.1618 22.89 0.5592 0.1613

15.61 0.5921 0.1315 15.37 0.5928 0.1307 15.83 0.5918 0.1323 15.48 0.5918 0.1311 15.78 0.5911 0.1323 15.46 0.5949 0.1304 15.61 0.5909 0.1316 15.53 0.5908 0.1314 15.61 0.5916 0.1317 15.77 0.5911 0.1322 15.61 0.5927 0.1312 15.80 0.5910 0.1325 15.61 0.5920 0.1323

9.60 0.6156 0.0200 9.56 0.6173 0.0250 9.48 0.6181 0.0183 9.58 0.6156 0.0220 9.53 0.6149 0.0226 9.46 0.6203 0.0169 9.55 0.6171 0.0334 9.55 0.6171 0.0334 9.45 0.6165 0.0253 9.57 0.6167 0.0246 9.57 0.6163 0.0233 9.55 0.6178 0.0296 9.54 0.6149 0.0267

13.79 0.6019 0.1211 13.69 0.6027 0.1206 13.97 0.6015 0.1215 13.63 0.6029 0.1205 13.72 0.6022 0.1209 14.16 0.5996 0.1224 13.44 0.6039 0.1194 13.34 0.6049 0.1189 13.79 0.6026 0.1207 13.79 0.6026 0.1207 13.65 0.6016 0.1208 13.82 0.6031 0.1205 13.89 0.6025 0.1209

The moment M1 is determined by the average value of r over the atomic electron distribution and its value is therefore sensitive to the amount of electron density relatively far from the nucleus.36 The substituents with the largest Ea value have the lowest M1 value. As a result, the polarisation of the electron density of the hydrogen atom is decreased for substituted phenols with higher activation energies. The electron distribution becomes less distorted and the charge distribution is more spherical for the hydrogen atom. An excellent correlation between dipole moment, M1 (H), and activation energy at the transition state (C) is shown in Fig. 3. However, from Fig. 4 it is clear that the dipole moment of the H atom does not correlate with the activation energy at the reactant state(A). According to Brinck,10 for a ground state property to be Table 4 The atomic properties of the hydrogen atom for two self exchange type reactions listed Self exchange reaction

Atomic properties

Phenol/phenoxyl radical (PCET)

u(O) q(O) M1(O) u(O) q(O) M1(O)

Toluene/benzyl radical (HAT)

This journal is & The Owner Societies 2005

10.39 0.6386 0.0015 30.56 0.1302 0.0006

Phys. Chem. Chem. Phys., 2005, 7, 614–619

617

Table 5 Bond critical point properties evaluated at the C6–O7 bond. Calculated change [transition state (C) – hydrogen bonded reactant complex (B)] in AIM properties. All values are in atomic units

Fig. 3

Plot of DEa vs. M1(O) at the transition state (C) with r2 of 0.92.

successfully linked to activation energy (or more precisely O–H bond dissociation energy, with which it correlates highly) it must capture the delocalisation of the unpaired electron after hydrogen abstraction. The dipole moment of hydrogens at the ground state, M1(H) does not satisfy this criterion. This is why it yields a poor correlation with the activation energy. Electron transmission through the ipso-carbon (C6) to the oxygen (O7) is thought to play an important role in the mechanism of antioxidant activity.20,31 The delocalisation of the oxygen lone pair on phenol is found to account for the polar effect on the bond dissociation energy of substituted phenols.37 In order to estimate this quantity on the activation energy, BCP properties are evaluated at the C6–O7 bond. Table 5 lists the changes in the BCP properties in going from the hydrogen bonded reactant complex (B) to the transition state (C). The extent of the lone pair delocalisation, as measured by De, indicates that meta substituted phenols are better able to delocalise the oxygen lone pair on the phenol. The increase in ellipticity in going from the hydrogen bonded complex to the transition state is on average 20% larger for the meta substituents compared to the para substituents. For para subtituents the electron delocalization of the unpaired electron is more important than that of the lone pair. This link can only be made under the assumption that De is more influenced by the lone pair localization than the delocalization of the unpaired electron. Secondly, based on their De values, the para fluoro substituent exhibits a similar lone pair delocalising effect to the para methoxy substituent but their activation energies differ by 2.4 kcal mol1. In order to explain this discrepancy, one is inclined to invoke the spin delocalisation of the unpaired electron in the transition state.12

Fig. 4 Plot of DEa vs. M1(O) at the isolated reactant state (A). 618

Phys. Chem. Chem. Phys., 2005, 7, 614–619

1 2 3 4 5 6 7 8 9 10 11 12 13 a a

H o-Me m-OMe m-Me m-Et m-F p-OH p-OMe p-Me p-Et p-F o,m-Me o,p-Me r2

Dr

Dr2r

De

0.035 43 0.035 93 0.035 36 0.035 35 0.035 22 0.035 82 0.035 01 0.034 74 0.035 45 0.035 38 0.035 25 0.0352 0.035 75 0.306

0.029 93 0.033 41 0.032 93 0.032 22 0.032 17 0.035 67 0.020 88 0.017 98 0.028 55 0.0298 0.028 44 0.030 12 0.027 88 0.780

0.015 0.014 0.016 0.015 0.015 0.016 0.009 0.012 0.014 0.014 0.012 0.012 0.013 0.707

Ea 14 86 99 57 40 16 72 72 44 47 18 85 45

6.6 5.6 7.2 6.3 6.2 7.8 3.4 3.2 5.4 5.4 5.6 5.2 4.5

Correlation coefficient with respect to DEa.

From Table 5 it is clear that the Dr2r values of the C6–O7 bond are positive for all substituents. The r2r values themselves (not shown) are all negative in the hydrogen bonded complexes (B) and transition states (C), which is a signature of a so-called shared interaction. One can interpret the observed increase in r2r in going from B to C (i.e. less negative value) as a weakening of the shared interaction. Going further, one could argue that the r2r values are on their way to become positive, which is a feature of an intermediate interaction. Carbonyl bonds are well-known examples of this type of bond. Of course, in the reaction profile this limiting case is never reached but the increase in r2r might be attributed to quinoid character, which would be consistent with an observation made before.12 This suggested link is strengthened by the observed bond length shortening (ca. 0.05 A˚ for all reactions). There is little correlation between rb and activation energy. However, BCP data represent a concerted movement of electronic density as bonds are broken and formed.

4. Conclusion Substituent effects reveal that para substituents influence the activation energies of hydrogen abstraction more strongly than meta substituents. An excellent linear relationship between the activation energy and the enthalpy is found. Such a relationship may be used to understand the mechanistic behaviour for hydrogen abstraction from phenolic compounds. However, the behaviour of this relationship is not general for all radical reactions. When hydrogen bonding is involved, an empirical formula estimating Ea for phenolic antioxidants must include terms to account for hydrogen bonding. While hydrogen bonding is important in retrieving the correct energy profile it is shown that ground state thermodynamics like the enthalpy can be used to estimate the reactivity of phenolic antioxidants. Our electron density analysis reveals that there is satisfactory correlation between r2r, e on one hand and the activation energies on the other. Phenol oxygen lone pair delocalisation can be used to describe substituent effects. Atomic properties reveal that the hydrogen atom experiences a loss of charge, volume and also a vanishing dipole moment during the transition from hydrogen bonded reactant complex to transition state. The topology of the electron density allows a better understanding of electronic properties. The straightforward application of AIM not only recovers the classical bonds but it also reveals new information. Based on the low atomic volume and high charge of the transferring hydrogen the reactions can be said to proceed via proton coupled

This journal is & The Owner Societies 2005

electron transfer (PCET) as opposed to a pure hydrogen atom transfer (HAT) mechanism.

Acknowledgements We thank Olivier Lamarche, Michel Rafat and Michael Devereux for their contribution to and comments on this work and the EPSRC for their financial support.

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

G. W. Burton and K. U. Ingold, Acc. Chem. Res., 1986, 19, 194. G. W. Burton, Y. L. Page, E. J. Gabe and K. U. Ingold, J. Am. Chem. Soc., 1980, 102, 7791. B. P. Roberts, J. Chem. Soc., Perkin Trans. 2, 1996, 2719. A. A. Zavitsas and C. Chatgilialoglu, J. Am. Chem. Soc., 1995, 117, 10645. A. A. Zavitsas, J. Chem. Soc., Perkin Trans. 2, 1998, 499. B. P. Roberts and A. J. Steel, J. Chem. Soc., Perkin Trans. 2, 1994, 2155. G. W. Burton, T. Doba, E. J. Gabe, F. L. Lee, L. Prasad and K. U. Ingold, J. Am. Chem. Soc., 1985, 107, 7053. J. S. Wright, D. J. Carpenter, D. J. Mckay and K. U. Ingold, J. Am. Chem. Soc., 1997, 119, 4245. J. S. Wright, E. R. Johnson and G. A. DiLabio, J. Am. Chem. Soc., 2001, 123, 1173. T. Brinck, M. Haeberlein and M. Jonsson, J. Am. Chem. Soc., 1997, 119, 4239. O. Tishchenko, E. S. Krayachko and M. T. Nguyen, J. Mol. Struct., 2002, 615, 247. P. J. O’Malley, Chem. Phys. Lett., 2002, 364, 318. K. Tanaka, S. Sakai, S. Tomiyama, T. Nishiyama and F. Yamada, Bull. Chem. Soc. Jpn., 1991, 64, 2677. F. G. Bordwell, X. M. Zhang, A. V. Satish and J.-P. Cheng, J. Am. Chem. Soc., 1994, 116, 6605. R. F. W. Bader, Atoms in Molecules. A Quantum Theory, Oxford University Press, Oxford, 1990. P. L. A. Popelier, Atoms in Molecules. An Introduction, Pearson Educ., Harlow, UK, 2000. J. Ruiz and R. Pouplama, Quant. Struct. -Act. Relat., 2002, 21, 605. C. D. Selassie, R. Garg, S. Kapur, A. Karup, R. P. Verma, S. B. Mekapati and C. Hansch, Chem. Rev., 2002, 2585. S. A. B. E. Van Acker, L. M. H. Koymans and A. Bast, Free Rad. Biol. Med., 1993, 15, 311.

20 S. Tomiyama, S. Sakai, T. Nishiyama and F. Yamada, Bull. Chem. Soc. Jpn., 1993, 66, 299. 21 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheesman, V. G. Zakrzewski, J. A. J. Montgomery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, V. Barone, J. Tommass, M. Cossi, R. Cammi, B. Mennuci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachani, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanaykara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle and J. A. Pople, GAUSSIAN98, (Revision A.7), Gaussian Inc., Pittsburgh, PA, 1998. 22 MORPHY98, a program written by P. L. A. Popelier with a contribution from R. G. A. Bone; UMIST: Manchester, UK, 1998. 23 S. Nagaoka, A. Kuranaka, H. Tsuboi, U. Nagashima and K. Mukai, J. Phys. Chem., 1992, 96, 2754. 24 M. G. Evans and M. Polanyi, Trans. Faraday Soc., 1935, 31, 875. 25 J. M. Mayer, Acc. Chem. Res., 1998, 31, 441. 26 M. Foti, E. R. Johnson, M. R. Vinquist, J. S. Wright, L. R. C. Barclay and K. U. Ingold, J. Org. Chem., 2002, 67, 5190. 27 M. Foti, K. U. Ingold and J. Lusztyk, J. Am. Chem. Soc., 1994, 116, 9940. 28 J. M. Mayer, D. A. Hvrovat, J. L. Thomas and W. T. Borden, J. Am. Chem. Soc., 2002, 124, 1142. 29 Z. K. Mainyes, G. E. Zarkov and L. M. Andronov, Russ. J. Phys. Chem., 1967, 41, 590. 30 J. Espinosa-Garcia, J. Am. Chem. Soc., 2003, 126, 920. 31 Y. M. Sun and C. B. Liu, Eur. J. Org. Chem., 2004, 120. 32 J. R. Cheesman, M. T. Carroll and R. F. W. Bader, Chem. Phys. Lett., 1988, 143, 450. 33 M. Tafipolsky, W. Scherer, K. O¨fele, G. Artus, B. Pederson, W. A. Herrmann and S. G. McGrady, J. Am. Chem. Soc., 2002, 124, 5865. 34 F. M. Aicken, PhD Thesis, University of Manchester Institute of Science and Technology, 2000. 35 R. J. Gillespie and P. L. A. Popelier, Chemical Bonding and Molecular Geometry: From Lewis to Electron Densities, Oxford University Press, New York, 2001. 36 R. F. W. Bader and M. T. Carroll, Mol. Phys., 1988, 3, 695. 37 F. G. Bordwell and J.-P. Cheng, J. Am. Chem. Soc., 1991, 113, 1736. 38 J. B. Foresman and A. Frisch, Exploring Chemistry with Electronic Structure Methods, Gaussian Inc., Pittsburgh, PA, 2nd edn., 1998.

This journal is & The Owner Societies 2005

Phys. Chem. Chem. Phys., 2005, 7, 614–619

619