JOURNAL OF CHEMICAL PHYSICS

VOLUME 113, NUMBER 7

15 AUGUST 2000

COMMUNICATIONS Dissociative electron attachment to gas-phase 5-bromouracil H. Abdoul-Carime and M. A. Huelsa) Groupe du Conseil de Recherche Me´dicale du Canada en Sciences des Radiations, Faculte´ de Me´decine, Universite´ de Sherbrooke, Sherbrooke, Que´bec, Canada J1H 5N4

F. Bru¨ning and E. Illenberger Institut fu¨r Physikalische und Theoretische Chemie, Freie Universita¨t Berlin, Takustraße 3, D-14195 Berlin, Germany

L. Sanche Groupe du Conseil de Recherche Me´dicale du Canada en Sciences des Radiations, Faculte´ de Me´decine, Universite´ de Sherbrooke, Sherbrooke, Que´bec, Canada, J1H 5N4

共Received 3 April 2000; accepted 13 June 2000兲 We report measurements of dissociative electron attachment 共DEA兲 to gaseous 5-bromouracil 共BrU兲 for incident electron energies between 0 and 16 eV. Low energy electron impact on BrU leads not only to the formation of a long lived parent anion BrU⫺ , but also various anion fragments resulting from endo- and exo-cyclic bond ruptures, such as Br⫺, uracil-yl anions, i.e., 共U-yl兲⫺, OCN⫺, and a 68 amu anion tentatively attributed to H2C3NO⫺. The incident electron energy dependent signatures of either the Br⫺ and (U-yl兲⫺ yields 共at 0, 1.4, and 6 eV兲, or the OCN⫺ and H2C3NO⫺ yields 共at 1.6 and 5.0 eV兲 suggests competing DEA channels for anion fragment formation. The production cross sections, at 0 eV incident electron energy, for BrU⫺, Br⫺, and 共U-yl兲⫺ are estimated to be about 6⫻10⫺15, 6⫻10⫺14, and 1.0⫻10⫺15 cm2, respectively. © 2000 American Institute of Physics. 关S0021-9606共00兲01031-X兴

INTRODUCTION

surements show an increase by a factor of 3 in the neutral fragment desorption yield from BrU modified oligomers, relative to unmodified oligonucleotides. While Br⫺ formation via dissociative electron attachment 共DEA兲 to condensed phase BrU in ice by near 0 eV electrons has been reported in x-ray photoelectron spectroscopy measurements,7 DEA to gas-phase T leads to the formation of various light and heavy anion fragments only above ca. 2 eV.8 Since T, which is replaced in cellular DNA by BrU, differs from the latter only by the methyl group that substitutes the bromine atom at the same five-position 关the structure of BrU is displayed in the inset of Fig. 1共a兲兴, these measurements suggest that DEA to BrU may lead to molecular decomposition over the entire electron energy range between 0 and 20 eV. In order to test this hypothesis and to probe the fundamental mechanisms of electron interactions inducing molecular decomposition, we have performed measurements of DEA to gas-phase BrU at incident electron energies from 0 to 16 eV. Our results, presented here, show that DEA leads to endocyclic as well as exocyclic bond cleavage in BrU, and suggest that at least four different dissociation pathways may be involved in the radiosensitizing action of BrU.

It has been demonstrated in the late 1950s that the replacement of thymine 共T兲 by 5-bromouracil 共BrU兲 in cellular DNA, combined with ionizing radiation, induces a strong enhancement of DNA damage and cell death.1 This work suggests a possible application of BrU in radiation therapy as a treatment enhancing sensitizer.2 However, despite numerous investigations dedicated to the understanding of the mechanism by which this radiosensitizer operates, the early physico-chemical steps in the enhancement processes are still poorly understood. One of the proposed mechanisms involves thermalization 共within ca. 10⫺12 s兲 of the copious 1–20 eV secondary electrons that are created along ionizing radiation tracks.3 It was suggested that, once hydrated, the secondary electrons first reduce the BrU to form BrU⫺, followed by dissociation of the latter into Br⫺ and a reactive uracil-yl, 共U-yl兲•, radical, which then becomes the precursor for DNA damage.4 However, recent studies have demonstrated that resonant mechanisms induced by nonthermal low energy 共0.5–30 eV兲 electrons may also play an important role in damage to either supercoiled DNA5 or BrU containing chemisorbed oligonucleotides which could model a sensitized DNA single strand.6 For the latter, electron stimulated desorption mea-

EXPERIMENT

The experiments were carried out at the Berlin Laboratory, in a standard crossed beam apparatus that has been extensively described elsewhere.9 Here, we only give a brief

Author to whom correspondence should be addressed. Tel.: 共819兲-3461110, ext. 14907, FAX: 共819兲-564-5442, electronic mail: [email protected]

a兲

0021-9606/2000/113(7)/2517/5/$17.00

2517

© 2000 American Institute of Physics

Downloaded 09 Oct 2003 to 160.45.20.181. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

2518

Abdoul-Carime et al.

J. Chem. Phys., Vol. 113, No. 7, 15 August 2000

FIG. 1. BrU⫺ 共a兲, Br⫺ 共b兲, and 共U-yl兲⫺ 共c兲 yields produced by electron impact on gas-phase 5-bromouracil 共BrU兲 as functions of incident electron energy. The structure of BrU is also given in 共a兲. For the dotted curves, the yields have been multiplied by a factor of 10 共a兲, 20 共b兲, and 4 共c兲, respectively.

summary. An incident electron beam of well-defined energy 共⬃10 nA, FWHM ⬃0.12 eV兲, generated by a trochoı¨dial electron monochromator,10 orthogonally intersects an effusive molecular beam. The latter emanates from a resistively heated oven containing high purity BrU powder 共Aldrich Ltd.兲. The temperature of the oven is measured by a platinum resistance to be approximately 150 °C, which is far below the temperature of BrU decomposition 共300 °C兲. Similar to experiments on gaseous T,8 at the present low evaporation temperatures, no stable dimer anions were detected in the mass spectra. The experimental chamber, with a base pressure of 3⫻10⫺8 Torr, is maintained at about oven temperatures by two in vacuo infrared lamps during the experiments; this prevents BrU condensation on the surfaces 共plates, chamber wall兲 which may otherwise lead to undesirable changes in contact potentials, and it has been verified that the light of the infrared lamps does not influence the measurements. Anions that are formed via electron-molecule collisions are extracted from the reaction volume by a small electric field 共⬍1.0 V cm⫺1兲 toward a quadrupole mass analyzer, and are detected by single pulse counting techniques. The electron energy scale is calibrated by measuring the formation of SF⫺ 6 , which exhibits a sharp peak of known cross section located at 0 eV.11–13

FIG. 2. Incident electron energy dependence of OCN⫺ 共a兲 and H2C3NO⫺ 共b兲 yields. Solid curves are guides to the eye.

RESULTS

As shown in Figs. 1 and 2, low energy 共0–16 eV兲 electron collisions with gaseous BrU induce formation of various negative ions, such as long lived BrU⫺, (U-yl兲⫺, Br⫺, OCN⫺, and a 68 amu negative ion, as the most intense measurable species in the present experiment. However, as the mass spectrometry measurements give no indication of the structure of the observed negatively charged fragments, the 68 amu negative species is tentatively attributed to the anion fragment that can be formed via the simplest dissociation pathway, namely H2C3NO⫺. Moreover, we do not observe ⫺ measurable yields of either CN⫺, OCNH⫺ , OCNH⫺ 2 , O , ⫺ ⫺ H , or CH2 fragments, which have been reported in previous work on DEA to gas-phase T.8 If at all, these unobserved fragments might be produced in the present case, but well below our detection threshold. The electron energy dependencies of the observed anion yields, shown in Figs. 1 and 2, exhibit peaked structures indicative of resonant processes for negative ion formation, which are well understood within the framework of DEA theory.11,12 A free electron becomes resonantly captured by the neutral target molecule to form a transient molecular anion 共TMA兲, i.e., BrU*⫺. Successful molecular dissociation into a negative ion and one or more neutral radical fragments takes place when the BrU*⫺ lifetime toward autodetachment, ␶ a , is long enough to allow sufficient internuclear bond separation (R⬎R c ) before electron autodetachment occurs; R c is the bond distance beyond

Downloaded 09 Oct 2003 to 160.45.20.181. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

J. Chem. Phys., Vol. 113, No. 7, 15 August 2000

Electron attachment to Bromouracil

2519

TABLE I. Peak positions 共⫾0.2 eV兲 observed in the anion yields produced by electron impact on gaseous BrU. The estimated cross sections for BrU⫺, Br⫺, and 共U-yl兲⫺ production at 0 eV incident electron energy are given in parenthesis. The values labeled with an asterisk correspond to weak structures in the yield functions. BrU⫺

Br⫺

共U-yl兲⫺

H2C3NO⫺ OCN⫺

0.0 eV 0.0 eV 0.0 eV (0.6⫻10⫺14 cm2) (6.0⫻10⫺14 cm2) (0.1⫻10⫺14 cm2) 0.5 eV 1.3* eV 1.4 eV 1.4 eV 1.6 eV 3.5* eV 5.0 eV 6.0* eV 6.0* eV 6.5* eV

1.6 eV 3.5 eV 5.0 eV 6.5* eV

⫺

which the BrU* energy is below that of the neutral molecule. The DEA cross section may be approximated as ␴ DEA⫽ ␴ capt exp(⫺␶D /␶a), where ␶ D represents the dissociation time to reach R c , and ␴ capt , the electron capture cross section. Although in the present experiment we cannot measure absolute values for ␴ DEA directly, we may nevertheless estimate some anion formation cross sections at 0 eV 共and at 11–13 Under the present 150 °C兲 by comparison to that of SF⫺ 6. experimental conditions, the measured anion signal, S A , is linearly proportional to the number density of the target molecule, ␳ A , and ␴ DEA . If we assume that the proportionality factor, which depends on anion detection efficiency 共i.e., electron beam intensity, ion extraction conditions, transmission of the quadrupole mass analyzer兲, and pumping speed 共the SF6 stream passes through the oven兲, is sufficiently simi14 lar for either SF⫺ 6 or a negative ion produced here, then: 共 ␴ A / ␴ SF6兲 DEA

at 0 eV⫽ 共 S A /S SF6 兲 at 0 eV* 共 ␳ SF6 / ␳ BrU 兲 .

共1兲

Here ␴ A and S A are the DEA cross section and the anion signal, respectively, for a particular anion species A. The ratio of ␳ SF6 / ␳ BrU can be roughly deduced from the pressure measurements in the chamber read by the ionization gauge.14 However, even though the entire chamber is held at oven temperature to prevent the condensation of BrU on the walls, the reading of the BrU pressure at the gauge is likely to be underestimated relative to that of SF6. Furthermore, the true lifetime of the long lived parent anion BrU⫺, relative to that of SF⫺ 6 , is not sufficiently well known, and thus leads to additional uncertainty in the cross sections. Thus the cross sections for formation of BrU⫺, 共U-yl兲⫺, and Br⫺ at 0 eV electron energy and 150 °C, listed in Table I, are approximated to within one order of magnitude. DISCUSSION

Low energy 共0–16 eV兲 electron impact on gaseous BrU induces formation of BrU⫺ resulting from resonant electron capture by the parent molecule at 0 eV, 0.5 eV, and 1.3 eV, as shown in Fig. 1共a兲. Although electron attachment to BrU at 0 eV, leading to halogen bond dissociation, is an exothermic reaction,7,15 the present observation of BrU⫺ indicates that there exists a parent anion state with a comparatively long lifetime 共i.e., at least ⬃50 ␮s兲 relative to either dissociation or electron autodetachment, sufficient for mass spectrometric detection.9 Near 0 eV, electron attachment to BrU

FIG. 3. Possible dissociation pathways for negative ion production following low energy electron attachment to gas-phase BrU. The neutral radicals in parenthesis may or may not be stable with respect to dissociation. ‘‘A’’ represents a curve-crossing or saddle point for competing fragmentation channels 共see text兲.

may lead a priori to the creation of either ‘‘dipole-bound 共DB兲’’ or ‘‘valence’’ anions. The former anion species corresponds to the excess electron being weakly bound to the closed shell molecule via long range 共dipolar, quadrupolar, polarization, etc.兲 forces, and resides outside the molecular frame.16 It has been demonstrated experimentally,17 as well as from theoretical prediction,18 that subthermal electron attachment to T and uracil 共U兲 共the latter possessing a hydrogen atom at the five-position兲 leads to the formation of DB parent anions. For a pure DB anion, a minimum dipole moment of 2.0–2.5 D is required.16 Here, the dipole moment of BrU is estimated19,20 to be 3.2 D, and thus allows a DB state to be formed. However, valence anions, where the excess electron is bound to a valence molecular orbital, are also predicted to be formed for U.21 It has been shown that, depending on the experimental conditions, either DB or valence anions can be observed.22 BrU⫺ formation at 0.5 eV may arise from a nuclear excited Feshbach resonance, consisting of the free excess electron trapped in the ␲ * molecular orbital. In this case, the anion may further stabilize via bond cleavage within the aromatic ring as observed below 2 eV in gas-phase polycyclic hydrocarbons.23 Negative ion states formed below 1 eV electron energy have already been reported from either electron transmission through gaseous DNA bases,24 and theoretically predicted for T and U. The cross section for BrU⫺ formation estimated from that of SF⫺ 6 共both at 150 °C兲 at 0 eV, via Eq. 共1兲, is about 6.0⫻10⫺15cm2 共Table I兲. The incident electron energy dependence of either the Br⫺ or 共U-yl兲⫺ yields 关Fig. 1共b兲,共c兲兴 exhibits peaks at 0 eV, 1.4 eV, and a weak, broad structure at 6.0 eV, which are typical signatures of DEA, most likely via either shape or Feshbach core excited resonances 共i.e., formation of a singleelectron, or a two-electron–one-hole TMA, respectively兲. The TMA then undergoes unimolecular dissociation into a negative ion and a neutral atom or radical via two competing channels, i.e., Br⫺⫹共U-yl兲• or Br•⫹共U-yl兲⫺, as shown by steps ‘‘b1’’ and ‘‘b2,’’ respectively, in Fig. 3. The weak resonant structure observed at 6 eV incident electron energy is likely to involve electronic lone pair n→ ␲ * and/or ␲ → ␲ * transitions of the valence electron, according to both

Downloaded 09 Oct 2003 to 160.45.20.181. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

2520

Abdoul-Carime et al.

J. Chem. Phys., Vol. 113, No. 7, 15 August 2000

theoretical25 and experimental26 studies on similar DNA bases 共T or U兲. However, the resonant structures below 2 eV, which lie at energies too low to involve electronically excited states, are likely to arise from the formation of a shape resonance consisting of the BrU molecule with the excess electron occupying a usually unfilled orbital. Such single particle anion states may have a sufficiently long lifetime at low energy to allow dissociation into an anion and a neutral fragment before electron autodetachment occurs. Furthermore, the branching ratios we estimate for step ‘‘b1’’ or ‘‘b2’’ 共i.e., b1/b2兲 at 0 eV, 1.4 eV, and 6 eV of 44, 6, and 11, respectively, are indicative of the relative predominance of the ‘‘b1’’ dissociative channel, especially near 0 eV incident electron energy. The cross section for Br⫺ production at 0 eV is estimated, via Eq. 共1兲, to be about 6⫻10⫺14 cm2, which is larger than that obtained from soft x-ray irradiation of dry films of BrU (⭐10⫺19 cm2兲17 by about five orders of magnitude, whereas that of 共U-yl兲⫺ is estimated here to be roughly 10⫺15 cm2 共Table I兲. DEA to BrU also induces the formation of OCN⫺ and H2C3NO⫺ 共Fig. 2兲, however, with relative cross sections one to two orders of magnitude smaller than those for the previous anions, which are formed by exocyclic bond cleavage. Here, we tentatively attribute ring fragmentation to reaction steps ‘‘b3’’ and ‘‘b4’’ 共Fig. 3兲, which result in the formation of H2C3NO⫺⫹共Br⫹OCNH兲 and OCN⫺ ⫹共H⫹Br H2C3NO), respectively. The observation of these anions species suggests that, within the lifetime of the transient BrU*⫺, not only single exocyclic bond dissociation but also complex endocyclic multibond cleavages can occur via either stepwise and/or concerted reactions.27 As an example, the reaction pathway ‘‘b4’’ is supported by a thresholdenergy calculation based on average bond energies28 and the electron affinity of OCN.11 We find relatively good agreement between the estimated formation energy of about 1 eV, and the experimental formation threshold of OCN⫺, which lies at around 1 eV incident electron energy 共see Fig. 2兲. The similar signatures for both observed anion species in Fig. 2 with peaks at 1.6 eV and 5.0 eV, and shoulders at 3.5 eV and 6.4 eV, may indicate competing fragmentation channels, and we can estimate the branching ratio of channel ‘‘b4’’ to ‘‘b3’’ to be approximately 6 and 4, at 1.6 eV and 5.0 eV, respectively. Alternatively, the peak observed at about 1.6 eV in the yields of OCN⫺ and H2C3NO⫺ roughly coincides with that of 共U-yl兲⫺, which suggests that the ring fragment anions may also be formed via reaction pathways ‘‘c1’’ and ‘‘c2,’’ as suggested in Fig. 3. Here capture of the incident electron by the BrU leads to a single TMA that first decays along a repulsive BrU*⫺ potential energy surface along the Br-共U-yl兲 coordinate, up to a crossing or saddle point ‘‘A’’ where different dissociative channels become available. Thus we can give an estimate of the branching ratios at 1.6 eV for c1/b2 and c2/b2 dissociation pathways 关 H2C3NO⫺/共U-yl兲⫺ and OCN⫺/共U-yl兲⫺] to be 0.33 and 0.05, respectively. CONCLUSION

We have shown that low energy 共0–16 eV兲 electron impact on gaseous BrU leads to formation of a long lived parent anion, BrU⫺, as well as various fragment anion species

via either single bond or complex multibond dissociation, at energies below 8 eV. The competing fragmentation processes are attributed to DEA via resonant capture of the incident electron into either shape or core excited resonances. Moreover, our results indicate that the most important damage to gas-phase BrU involves slow electrons with energies near 0–2 eV. This is in contrast with measurements of DEA to gas-phase T for which most of the anion fragment production is localized within the 3–10 eV energy range. At 1.4– 1.6 eV incident electron energy, the present observation of 共U-yl兲⫺,OCN⫺ , and H2C3NO⫺ suggests that those anion fragments may arise via either closely lying but distinct resonances, or a single predissociative anion state. The current results show that the radiosensitivity of BrU is much more complex than previously anticipated. Resonant electron mechanisms lead to complex molecular decompositions over the entire electron energy range between 0 and 7 eV, and induce formation of different anion and radical fragments 共compared to T兲 via at least four different dissociation pathways. If formed within DNA, some of these fragments may react and thus lead to lethal clustered damage in addition to that already occurring at electron energies between 3 and 20 eV in unsensitized DNA 共e.g., see damage-to-plasmid DNA5兲. Our present study demonstrates that the radiosensitizing nature of halopyrimidines warrants further theoretical and experimental investigations to better understand their nascent radiolytic effects on DNA, and to thus realize their potential as routine therapeutic agents. Such experimental efforts are currently underway,29 particularly in the condensed phase30 where local effects of the molecular environment may influence the electron-molecule reaction pathways. ACKNOWLEDGMENTS

This study was supported by the Medical Research Council 共MRC兲 and the National Cancer Institute 共NCI兲 of Canada, and the Deutsche Forschungsgemeinschaft. One of us 共F.B.兲 acknowledges a post-doctoral fellowship from the Ministe`re de l’E´ducation du Que´bec 共Program Que´becois de Bourse d’Excellence兲. S. Zamenhof, R. DeGiovanni, and S. Greer, Nature 共London兲 181, 827 共1958兲; W. Szybalski, Cancer Chemotherapy Reports, 58, 539 共1974兲, and references cited therein. 2 D. J. Buchsbaum, M. B. Khazaeli, and M. A. Davis, Cancer 共N.Y.兲 73, 999 共1994兲. . 3 T. Cobut, Y. Fongillo, J. P. Patau, T. Goulet, M. J. Fraser, and J. P. Jay-Gerin, Radiat. Phys. Chem. 51, 229 共1998兲. 4 L. Ling and J. F. Ward, Radiat. Res. 121, 76 共1990兲; M. Katouzian-Safadi and M. Charlier, 94, 326 共1997兲. 5 B. Boudaiffa, P. Cloutier, D. Hunting, M. A. Huels, and L. Sanche, Science 287, 1658 共2000兲. 6 H. Abdoul-Carime, P. C. Dugal, and L. Sanche, Radiat. Res. 153, 23 共2000兲. 7 D. V. Klyachko, M. A. Huels, and L. Sanche, Radiat. Res. 151, 177 共1999兲. 8 M. A. Huels, I. Hahndorf, E. Illenberger, and L. Sanche, J. Chem. Phys. 108, 1309 共1998兲. 9 共a兲 O. Ingo`lfsson, F. Weik, and E. Illenberger, Int. J. Mass Spectrom. Ion Processes 155, 1 共1996兲; 共b兲 I. Hahndorf and E. Illenberger, ibid. 167Õ168, 87 共1997兲. 10 A. Stamatovic and G. J. Schulz, Rev. Sci. Instrum. 41, 423 共1970兲. 11 E. Illenberger, in Gaseous Molecular Ions, Topics in Physical Chemistry, Vol. 2, edited by H. Baumga¨rtel, E. U. Frank, and W. Gru¨nbein 共Steinko1

Downloaded 09 Oct 2003 to 160.45.20.181. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

J. Chem. Phys., Vol. 113, No. 7, 15 August 2000 pff, Darmstadt, Springer, New York, 1992兲, Part III, and references cited therein. 12 L. G. Christophourou, in Electron-Molecule Interactions and Their Applications, Vol. 1, edited by L. G. Christphourou 共Academic, New York, 1984兲; T. F. O’Malley, Phys. Rev. 150, 14 共1966兲. 13 D. Smith, N. G. Adams, and E. Alge, J. Phys. B 17, 461 共1984兲; O. J. Orient and A. Chutjian, Phys. Rev. A 34, 1841 共1986兲. 14 Y. LeCoat, R. Azria, M. Tronc, O. Ingo`lfsson, and E. Illenberger, Chem. Phys. Lett. 296, 208 共1998兲. 15 P. C. Dugal, H. Abdoul-Carime, and L. Sanche, J. Phys. Chem. 共accepted for publication兲. 16 D. C. Clary, J. Phys. Chem. 92, 3173 共1988兲; C. Desfranc¸ois, H. AbdoulCarime, N. Khelifa, and J. P. Schermann, Phys. Rev. Lett. 73, 2436 共1994兲; H. Abdoul-Carime and C. Desfranc¸ois, Eur. Phys. J. D 2, 149 共1998兲. 17 C. Desfranc¸ois, H. Abdoul-Carime, C. P. Schulz, and J. P. Schermann, Science 269, 1707 共1995兲; C. Desfrancois, H. Abdoul-Carime, and J. P. Schermann, J. Chem. Phys. 104, 7792 共1996兲; J. H. Hendricks, S. A. Lyapustina, H. L. deClercq, and K. H. Bowen, ibid. 108, 1 共1998兲. 18 R. N. Compton, Y. Yoshioka, and K. D. Jordan, Theor. Chim. Acta 54, 259 共1980兲; N. Oyler and L. Adamowicz, J. Phys. Chem. 97, 11122 共1993兲. 19 S. LeCaer, A. FilaliMouhim, and J. P. Jay-Gerin 共private communication兲. The dipole moment is estimated by calculating the BrU atomic partial charges via semi-empirical methods developed in Refs. 20a,b. This method has shown to provide dipole moment values for DNA bases and FluoroUracil which are in relatively good agreement with Hartree–Fock calculations 关Ref. 20共c兲兴.

Electron attachment to Bromouracil

2521

共a兲 J. Gasteiger and M. Marsili, Tetrahedron 36, 3219 共1980兲; 共b兲 M. Marsili and J. Gasteiger, Croat. Chem. Acta 53, 601 共1980兲; 共c兲 J. Li, J. Xing, C. J. Cramer, and D. G. Truhlar, J. Chem. Phys. 111, 885 共1999兲. 21 M. D. Sevilla, B. Besler, and A. O. Colson, J. Phys. Chem. 99, 1060 共1995兲. 22 R. N. Compton, H. S. Carman, C. Desfranc¸ois, H. Abdoul-Carime, J. P. Schermann, J. H. Hendricks, S. A. Lyapustina, and K. H. Bowen, J. Chem. Phys. 105, 3472 共1996兲. 23 B. G. Zykov, Y. V. Vasil’ev, V. S. Fal’ko, A. N. Lachinov, V. I. Khvostenko, and N. G. Gileva, JETP Lett. 64, 439 共1996兲. 24 K. Aflatooni, G. A. Gallup, and P. D. Burrow, J. Phys. Chem. 102, 6205 共1998兲. 25 J. Lorentzon, M. P. Fulscher, and B. O. Roos, J. Am. Chem. Soc. 117, 9265 共1995兲. 26 L. B. Clark, G. G. Peschel, and I. Tinoco Jr., J. Phys. Chem. 69, 3615 共1965兲; T. Yamada and H. Fukutome, Biopolymers 6, 43 共1968兲. 27 C. P. Andrieux, A. LeGorande, and J. M. Saveant, J. Am. Chem. Soc. 114, 6892 共1992兲; M. Stepanovic, Y. Pariat, and M. Allan, J. Chem. Phys. 110, 11376 共1999兲; S. Jacobsen, H. Jensen, S. U. Pedersen, and K. Daasbjerg, J. Phys. Chem. 103, 4141 共1999兲. 28 T. H. Lowry and K. S. Richardson, in Mechanism and Theory in Organic Chemistry, 3rd ed. 共Harper Collins, 1987兲, Chap. 2, p. 159. 29 H. Abdoul-Carime, M. A. Huels, E. Illenberger, and L. Sanche 共unpublished兲. 30 M. A. Herve du Penhoat, M. A. Huels, P. Cloutier, J. P. Jay-Gerin, and L. Sanche 共unpublished兲. 20

Downloaded 09 Oct 2003 to 160.45.20.181. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp