Time-Resolved EPR Investigation of Potential Model Systems for

Apr 17, 2009 - Density functional theory calculations are presented and discussed in support of the experimental data. View: PDF | PDF w/ Links | Full...
0 downloads 0 Views 2MB Size
Download PDF
J. Phys. Chem. B 2009, 113, 6623–6629

6623

Time-Resolved EPR Investigation of Potential Model Systems for Acrylate Polymer Main Chain Radicals Based on Esters of Kemp’s Tri-Acid Natalia V. Lebedeva,† Elena V. Gorelik,‡ Damaris Magnus-Aryitey,† Terence E. Hill,† and Malcolm D. E. Forbes*,† Caudill Laboratories, CB No. 3290, Department of Chemistry, UniVersity of North Carolina, Chapel Hill, North Carolina 27599, and International Tomography Center, Institutskaya 3a, NoVosibirsk 630090, Russia ReceiVed: October 30, 2008; ReVised Manuscript ReceiVed: February 11, 2009

Methyl esters of Kemp’s tri-acid and cyclohexanetricarboxylic acid are structurally similar to acrylate polymers, having the same functionalities and stereoregularities as poly(methylmethacrylate) and poly(methylacrylate), respectively. The photochemistry and free radicals from these model systems have been studied using timeresolved electron paramagnetic resonance spectroscopy with laser flash photolysis at 248 nm. Chemically induced electron spin polarization from the triplet mechanism (net emission) is observed. Well-resolved spectra are obtained at all temperatures for the model system radicals, which are determined to be in the slow motion condition, that is, there is no interconversion of chair conformations. The temperature dependence of the spectra is minimal; some hyperfine lines shift as the temperature increases, but without much broadening. Density functional theory calculations are presented and discussed in support of the experimental data. Introduction In recent years, our laboratory has been studying the photochemistry of acrylic acid, acrylate, and methacrylate polymers.1 We have published several papers on their photodegradation in liquid solution, focusing on the structure and dynamics of the main chain polymeric radicals. The radicals are created by pulsed laser excitation at 248 nm, which results in Norrish I cleavage of the side chains.2 This photochemistry is shown in Scheme 1 for poly(methylmethacrylate) (PMMA) and poly(methylacrylate) (PMA). We have used time-resolved electron paramagnetic resonance (TREPR) spectroscopy to characterize these and many similar radicals from more than a dozen different polymers with different main chain and side chain structures. At room temperature, the polymeric radicals undergo conformational motion that modulates some of the hyperfine interactions, leading to alternating line widths.3 Attempts to model these dynamic effects have not been successful, but qualitatively we have determined that the energies of the polymer conformations near the radical center are a strong function of tacticity.4 The presence of a large number of accessible conformers makes computational modeling cumbersome. Therefore, we have initiated an investigation of small molecule model compounds to guide the development of suitable models for main chain acrylate radical polymer chain dynamics in liquid solution. There are many structural features common to acrylate polymers and the cyclic framework associated with Kemp’s triacid5 (KTA) and cyclohexane tricarboxylic acid (CTA), whose trimethyl esters are shown in Scheme 2 (labeled KTE and CTE, respectively). The KTA molecule is well studied in bioorganic chemistry,6 but no free radicals from this ring structure have ever been reported in the literature. The six-membered ester ring of KTE contains the same chromophore as PMMA. In the KTE radical created by Norrish I R-cleavage of the ester group * To whom correspondence should be addressed. E-mail: [email protected]. † University of North Carolina. ‡ International Tomography Center.

(Scheme 2, right side), the same number of hyperfine couplings and magnetic equivalencies are present as in the PMMA polymeric radical shown in Scheme 1. In a similar fashion, radicals created from the CTE ring are structurally analogous to radicals from PMA. These similarities have previously been recognized in regard to modeling of NMR spectra of acrylic acids using KTA.7 In a separate publication, we have presented a detailed study of the acid radicals from KTA and CTA,8 which are models for poly(methacrylic acid) and poly(acrylic acid), respectively.9 In that work, we determined that the presence of charges on the acid model compounds kept their free radicals locked in a single chair conformation with the charged carboxylate groups in the equatorial position, even at high temperatures. For this reason, the acids were not pursued as models for the conformational motion of their polymer analogs. These compounds showed interesting photochemistry and spin polarization effects that became the main focus of that paper. These findings motivated us to examine the photochemistry, free radicals, and chemically induced spin polarization (CIDEP) phenomena from neutral model compounds, namely the CTE and KTE esters in Scheme 2. The conformational dynamics of heterocyclic radicals from sugars have been studied in great detail, by Giese in particular.10-12 However, reports of EPR parameters for heavily substituted cyclohexyl radicals (with only carbon atoms in the ring) in the literature are sparse,13-15 particularly those with doubly substituted ring carbons. These are interesting structures whose solution chemistry and spectroscopy have not been studied extensively to date, yet they are expected to have a rich chemistry in terms of the reactions that produce them, their possible rearrangements to other radicals, and the products they can form. Scheme 3 illustrates the structural analogy for both polymers and model compounds in more detail, where the main chain polymer radicals from photolysis of PMMA and PMA are directly compared to the KTE and CTE radicals. We use the term “meso” in Scheme 3 to describe these free radicals because

10.1021/jp809610t CCC: $40.75  2009 American Chemical Society Published on Web 04/17/2009

6624

J. Phys. Chem. B, Vol. 113, No. 19, 2009

Lebedeva et al.

SCHEME 1

SCHEME 2

methods, we often observe chemically induced electron spin polarization (CIDEP) in our spectra and thus the mechanism of the CIDEP processes are also of interest in any new photochemical system. Experimental Section

SCHEME 3

SCHEME 4

of the pseudo-mirror plane of symmetry that establishes magnetic equivalence between members of the two sets of β-methylene protons on each side of the radical center. The presence of this symmetry element is an important factor in dictating the magnitude of the β-hyperfine coupling constants at fast motion (see refs1-4 for a discussion). Note that the CTE and KTE radicals have hyperfine interactions from a single proton and a methyl group respectively, and both model compound radicals have two sets of β-methylene protons that are adjacent to stereogenic centers. This is the same number and position (R or β) of proton hyperfine interactions as in the polymeric PMMA and PMA radicals.1-4 Another interesting feature of the KTE and CTE radical systems is that they have at most two major conformations available to them, both chairs, which are shown for the KTE radical in Scheme 4. Because of its structural and conformational simplicity, the Kemp’s tri-acid framework is an interesting choice for the investigation of dynamic effects such as linebroadening due to conformational motion. As in our study of the cyclic acids, one of the goals of this study is to determine whether KTE and CTE radicals can be used as simpler model systems to mimic the dynamics of the polymer radicals. For this reason, we are interested in the temperature dependence of their TREPR spectra, and whether or not the less substituted CTE radical undergoes more rapid interchange between conformations than the KTE radical structure. Because we study these radicals using TREPR rather than steady state EPR

Materials. Kemp’s tri-acid, cyclohexane tricarboxylic acid, propylene carbonate, methanol, and acetyl chloride were purchased from Sigma-Aldrich and used as received. Methyl esters were formed by refluxing the tri-acids overnight in excess methanol in the presence of acetyl chloride. The reactions were worked up with standard synthetic methods, and the products were recrystallized from ethanol before use. TREPR and High Temperature Flow System. Our X-band TREPR apparatus and flow system for high temperatures is described in detail in several previous publications.16 Computational Methods. All the calculations have been performed using the Gaussian 03 package.17 The structure images and spin density plots were made by the Molekel program.18 The geometry optimization for different conformations of negatively charged KTA and CTA radicals as well as calculation of hyperfine coupling constants was performed at UB3LYP level employing standard 6-31G(d) basis set for geometry optimization, and EPR-III basis set for HFC calculations. No symmetry constraints were applied upon geometry optimizations. No zero point energy corrections were considered; the normal vibration analysis was only performed for the optimized structures to ensure the character of the stationary point found. Results and Discussion KTE Radicals. Figure 1 shows X-band TREPR spectra obtained 0.8 µs after direct 248 nm photolysis of KTE at different temperatures in propylene carbonate solution. The triplet mechanism (TM) of CIDEP19 is the dominant polarization pattern observed for these systems. The spectral width and hyperfine splitting patterns observed in Figure 1 show that the KTE model compound shows similar photochemistry and free radical chemistry to that observed from the PMMA polymer.1-4 It is interesting to note that for all acrylate polymers studied to date, complete net emission from the triplet mechanism is always observed with no multiplet polarization from the radical pair mechanism (RPM)20 of CIDEP. The predicted intensity pattern for the RPM from triplet state precursors is low field emission/high field absorption (E/A). Close examination of the room temperature spectrum in Figure 1 reveals a small superposition of this E/A pattern on the net E polarization from the TM. The RPM contribution decreases at higher temperatures. Both mechanisms depend on viscosity, but the RPM intensity relies on translational diffusion, whereas the TM depends more on rotational tumbling of the ester triplet state. In these systems, the RPM polarization exhibits a strong viscosity dependence compared to the TM polarization. The TM in these esters is very strong at all viscosities. For polymers and model compounds alike, it is one of the strongest TM signals we have ever recorded in our laboratory. For example, in the

Acrylate Polymer Main Chain Radicals

J. Phys. Chem. B, Vol. 113, No. 19, 2009 6625

Figure 1. (Left) Experimental X-band TREPR spectra obtained at 0.8 µs after 248 nm laser flash photolysis of a propylene carbonate solution (∼0.1 M) of KTE at the indicated temperatures. (Right) Simulations using hyperfine coupling constants, and line widths listed in Table 1.

TABLE 1: Radical Structures and EPR Simulation Parameters for Figures 1 and 2

course of our experiments we are able to irradiate our samples (initial optical density ) 0.08) for 45 to 60 min at 60 Hz with 10 mJ pulses without significant loss of signal intensity. This tells us that the quantum yield for radical production from the polymers is probably quite low, but the TM spin polarization is very intense. The onset of absorption for the esters or polymers is about 250 nm, so we are just barely at the edge of their absorbance in the UV when irradiating at 248 nm. The small contribution from RPM polarization is of fundamental interest. We have long suspected that the absence of RPM polarization in our previously reported polymer spectra is a consequence of the two members of the radical pair having drastically different diffusion coefficients. For the systems studied here, this is not the case (both members of the radical pair are small), and as a result the RPM is present in our spectra, at least at room temperature. Of course the superposition of the RPM and TM also depends on the initial concentration of radicals. This issue is outside the scope of the present discussion but it is a subject of current research in our laboratory. Continuing our discussion of Figure 1, TREPR spectra of the KTE radical were collected over the same range (25 to 125 °C) as that for the corresponding polymer, PMMA. Computer simulations, obtained using the parameters listed in Table 1, are shown to the right of the experimental spectra in Figure 1.

The g-factors are typical for carbon-centered free radicals, and the line widths used are typical for rapidly tumbling small organic radicals. The TREPR spectra of the KTE radical exhibit a small variation with temperature. However, at all temperatures there is a large difference between the axial and equatorial β-hyperfine coupling constants. This suggests that the radical is locked in a single chair conformation. It is highly improbable that a radical having a chair conformation with axial ester groups will have the same energy as one having equatorial groups. If the radical in one chair conformation were at slow motion and then moves to the fast motion condition with increasing temperature, we expect to observe significant line broadening over this temperature range. Such broadening is clearly not exhibited by the KTE radical in Figure 1. From Table 1 we can see that the methyl group hyperfine coupling constant for the KTE radical changes very little over this temperature range, which is typical for protons in functional groups undergoing fast rotation. The β-hyperfine coupling constants change by a larger amount going from 25 to 100 °C, but the observed changes are still fairly small compared to the room temperature values. Complete interconversion of the two chair conformations would realize equal or nearly equal values for the β-coupling constants. For example, in the corresponding polymer PMMA the β-hyperfine coupling constants achieve fast

6626

J. Phys. Chem. B, Vol. 113, No. 19, 2009

Lebedeva et al.

Figure 2. (Left) Experimental X-band TREPR spectra obtained at 0.8 µs after 248 nm laser flash photolysis of a propylene carbonate solution (∼0.1 M) of CTE at the indicated temperatures. (Right) Simulations using hyperfine coupling constants, and line widths listed in Table 1.

motion values of 16.4 and 11.7 G, a difference of less than 5 Gauss, at about 100 °C. In the esters, we see differences of ∼20 Gauss even at the highest temperatures accessible with our apparatus (∼120 °C). Our computational results, described in detail in the next section, support the hypothesis that the radicals observed here are locked in one chair conformation. The calculations will also examine the consequence of allowing the ester functional groups to rotate. Propylene carbonate is the same solvent that was used for all of our TREPR experiments at high temperature involving PMMA and PMA. It was initially selected because of its transparency at 248 nm, its high boiling point (this allows for high temperature flow conditions), and its ability to dissolve the polymers in high concentration. While it is not necessary to use it for small molecules in terms of solubility, its boiling point and optical transparency are still desirable features and allow more direct comparison of polymer and model compound spectra. CTE Radicals. Figure 2 shows TREPR spectra of the radical resulting from Norrish I cleavage of the ester group of CTE in propylene carbonate solution. This radical also exhibits strong TM polarization that is net emissive. Additionally, spectral simulation reveals the same large difference in the equatorial and axial hyperfine coupling constants as the KTE radical. This less-substituted cyclohexyl radical seems to also exist at the slow motion condition and locked in a single chair conformation at these temperatures. It is noteworthy that the CTE radical shows a long-range hyperfine interaction to Hδ, which, from

orbital overlap considerations, we assume to be the Hδ in the equatorial position. Long range hyperfine such as this are normally observed only in rigid bi- or tricyclic radicals (e.g., adamantyl or bicyclooctyl), and there is one report of a previous observation of such a coupling in cyclohexyl radicals.21 The rigidity of the ring system may facilitate this δ coupling. Considering the simulations in Figure 2 and the parameters used to generate them listed in Table 1, it is clear that the β-hyperfine interactions in the CTE radical change only minimally with temperature. There does seem to be some broadening of the transitions in the center, but this is more likely to be a shift in transitions by an amount within the line width, rather than chair-chair interconversion, which should lead to significant broadening and intensity changes over a large temperature range. The rotation of the ester groups in the γ positions of the ring may be responsible for these small shifts in the coupling constants, although we cannot rule out that we are approaching the onset of the broadening process associated with the transition to fast motion. Indeed, at much lower temperatures (-20 °C), TREPR experiments on these radicals show that exactly the same coupling constants are observed as for room temperature. For comparison, Camaioni and Pratt15 observed slow motion spectra for the 2-cyclohexanonyl radical at 93 K and fast motion spectra (with equal β-hyperfines) at 313 K. In the original work on the unsubstituted cyclohexyl radical by Ogawa and Fessenden,22 the transition between fast and slow motion spectra occurred at about 205 K. Therefore, it makes sense that the same transition for the much more heavily

Acrylate Polymer Main Chain Radicals

J. Phys. Chem. B, Vol. 113, No. 19, 2009 6627

TABLE 2: Energies for the Conformations of KTE and CTE Radicals with Different Orientation of Ester Groups as Resulted from UB3LYP/6-31G(d) Geometry Optimization structure

Etotal, au

Erel, kcal/mol

KTE 1ax 2ax 3ax 1eq 2eq 3eq

-808.888 3003 -808.887 5492 -808.891 5451 -808.884 7686 -808.883 6592 -808.884 8222

2.04 2.51 0.00 4.25 4.95 4.22

CTE 1ax 2ax 3ax 1eq 2eq 3eq

-690.943 4573 -690.941 9604 -690.945 3406 -690.948 4843 -690.947 4092 -690.948 2922

3.15 4.09 1.97 0.00 0.67 0.12

substituted CTE and KTE radicals will require much higher temperatures, certainly higher than we can currently access. It is interesting to note that in an NMR study of cyclohexane itself (actually cyclohexane-d11) by Anet and Bourn,23 coalescence was observed at a very similar temperature (230 K). To get a rough idea of the activation barriers involved in our ring inversions, we can use the frequency factor for cyclohexyl radical from ref 22 (5.4 × 1012 s-1) and the difference in slow motion coupling constants. The equatorial and axial constants differ by about 25 Gauss experimentally even at high temperatures. They will average to approximately half-this difference, or about 12.5 Gauss. In frequency units, this is 35 MHz or 3.5 × 107 s-1. If such an averaging process were to take place at 375 K (our highest accessible temperature), the Arrhenius equation gives a value of Ea for this process of about 9 kcal mol-1. Since we do observe averaging at this temperature, this number is a lower limit for the barrier; it must actually be quite a bit higher.

It is interesting to note that in several NMR studies of heavily substituted cyclohexane rings, the transition from slow to fast chair interconversion occurred at very low temperatures (∼230 K) for rings with three or four substituents,24 but at more than 370 K for rings with four or six substituents.25 Unfortunately, these studies were focused on rings with only one substituent per carbon atom; to the best of our knowledge there are no reports of dynamic NMR studies of cyclohexyl rings with doubly substituted carbons. The only dynamic EPR measurements reported previously were of unsubstituted or mono-substituted rings.10-15 While Kemp’s triacid itself favors equatorial methyl groups and axial acid moieties, in the triply deprotonated structure (above pH 9), Coulombic repulsion forces the methyl groups axial so that the charged carboxylates can be as far apart as possible in the equatorial position.6 Our computational results, described in detail below, support the hypothesis that the KTA and CTA radicals are rigidly locked in one chair conformation. The calculations will also shed light on the activation barrier for interconversion between the chairs, vide infra. DFT Calculations. The results of the geometry optimization of the KTE and CTE radicals are summarized in Table 2, and corresponding structures are shown on Figures 3 and 4. In contrast to the charged KTA and CTA radicals where Coulombic interactions predefine equatorial position of carboxilate groups, the present case of neutral KTE and CTE radicals demonstrates prominent stereochemical effects. While the less crowded CTE radical favors conformations with ester groups in equatorial position, the presence of bulky methyl substituents in γ-positions of the ring of KTE radical forces ester groups in axial positions. The energy difference between chair conformations for both radicals appears to be sufficient to prevent chair-chair interconversion within the temperature interval assessed experimentally. The above-mentioned conformational difference between the KTE and CTE radicals causes further distinctions observed in the temperature dependence of the TREPR spectra. Rotation of the ester groups occupying the equatorial positions in the

Figure 3. Ball and stick view of conformations for KTE radical. The structures shown are the result of UB3LYP/6-31G(d) geometry optimization.

6628

J. Phys. Chem. B, Vol. 113, No. 19, 2009

Lebedeva et al.

Figure 4. Ball and stick view of conformations for CTE radical. The structures shown are the result of UB3LYP/6-31G(d) geometry optimization.

TABLE 3: Calculated HFC (in G) for the KTE (Low Energy Conformation 3ax) and CTE (Conformations with Equatorial Ester Groups) Radicals CTE structure

1eq

2eq

HR -CH3 Hβa Hβe Hγ Hδe

-21.2

-21.7

49.2 7.1 -0.8 0.9

46.1 5.7 -1.0 1.1

3eq

KTE 3ax

-21.4 46.0; 49.6 6.4; 6.7 -0.8; -1.0 1.0

24.8 36.5;36.8 2.1; 1.6 1.2

CTE radical leads to some line broadening that was observed while studying the temperature dependence of this radical. In the KTE radical however, the two ester groups are axial and to a large extent occupy the same region of space above the ring, suppressing each other’s rotation. Thus the resulting KTE radical structure turns out to be not only just locked in one chair conformation, but also completely rigid with respect to ester groups rotation. Our final calculated hyperfine coupling constants for both KTE and CTE radicals are summarized in Table 3, and clearly confirm the slow motion condition. Our calculations show that even the energy difference between chair conformations is enough to prevent the interchange of ring conformations. In general, the entropy change in small ring conformational exchanges is close to zero (see ref 23), therefore the Hammond postulate can be used to rationalize this statement further. Taking into account the corresponding activation barriers will not add much to the picture we have described, and would not change our conclusions. Assessing activation barriers is computationally demanding even for much simpler radicals with similar ring structure and in present case are not important for the interpretation of EPR spectra. We did check the unsubstituted cyclohexyl radical at the B3LYP/6-31* level, finding an activation barrier 4.47 kcal/mol, which would place the barriers for the Kemp structures at much higher values, further emphasizing this point and supporting the physical picture. The change of the ester groups conformation is a more subtle question and is strongly substituent dependent. For KTE, where the ground-state conformation has ester groups in the axial

position, rotation of the ester groups is impossible due to steric hindrance. In this case, the ester groups occupy the same spatial region, as is clearly seen in Figure 3. The ground-state conformation of CTE has ester groups in the equatorial position, which opens the way to possible changes of the ester conformations. Small energy differences between these conformations, together with rather low activation barriers, E(TS1eqT3eq) E(1eq) ) 0.15 kcal/mol, suggests effective exchange between conformations of CTE with equatorial ester groups. We can conclude that both KTE and CTE are locked in a certain chair conformation. Ester groups of KTE are in a fixed position due to the steric hindrance, while for CTE we can suppose exchange between conformations with different orientations of the ester groups. These are qualitative conclusions that do not require the use of more precise (and much more expensive) computational methods. Our results support a fixed conformation for KTE, and effective exchange between 3 different equatorial conformations for CTE, as presented in Table 3. Summary The primary UV photodegradation pathway for the KTE and CTE model compounds is Norrish I cleavage, and the heavily substituted cyclohexyl radicals subsequently produced are easily detected on the sub-microsecond time scale by TREPR spectroscopy. Their TREPR spectra show interesting electron spin polarization from both the triplet mechanism and the radical pair mechanism of CIDEP. Their chair conformations, which from DFT calculations have methyl groups equatorial and ester groups axial, remain fairly rigid even at temperatures in excess of 100 °C. The radicals, while interesting structurally due to their high degree of ring substitution, are unlikely to be useful as models for the conformational dynamics of the corresponding polymers, as originally proposed, due to this rigidity unless much higher temperatures can be achieved with our flow system. The rotation of the two remaining ester functionalities is the likely explanation of the rather minor changes in the axial and equatorial hyperfine coupling constants with temperature. Further studies of these radicals as a function of ester side chain structure and solvent are in progress, and we are continuing to

Acrylate Polymer Main Chain Radicals refine our flow system to access the highest possible temperatures in order to access fast motion dynamics for all of these structures. Acknowledgment. We thank the National Science Foundation for their continued support of this work through Grant CHE0809530. References and Notes (1) Harbron, E. J.; McCaffrey, V. P.; Xu, R.; Forbes, M. D. E. J. Am. Chem. Soc. 2000, 122, 9182. (2) McCaffrey, V. P.; Forbes, M. D. E. Macromolecules 2005, 38, 3334. (3) McCaffrey, V. P.; Harbron, E. J.; Forbes, M. D. E. Macromolecules 2005, 38, 3342. (4) McCaffrey, V. P.; Harbron, E. J.; Forbes, M. D. E. J. Phys. Chem. B 2005, 109, 10686. (5) Kemp, D. S.; Petrakis, K. S. J. Org. Chem. 1981, 46, 5140. (6) (a) Rebek, J., Jr.; Marshall, L.; Wolak, R.; Parris, K.; Killoran, M.; Askew, B.; Nemeth, D.; Islam, N. J. Am. Chem. Soc. 1985, 107, 7476. (b) Menger, F. M.; Ladika, M. J. Am. Chem. Soc. 1988, 110, 6794. (7) (a) Lujan-Upton, H. A.; Okamoto, Y. Chem. Lett. 1997, 1253. (b) Lujan-Upton, H. A.; Okamoto, Y. J. Fluorescence 1998, 8, 355. (8) Lebedeva, N. V.; Gorelik, E. V.; Prowatzke, A. M.; Forbes, M. D. E. J. Phys. Chem. B 2008, 112, 7574. (9) Lebedeva, N. V.; Forbes, M. D. E. Macromolecules 2008, 41, 1334. (10) Dupuis, J.; Giese, B.; Rueegge, D.; Fischer, H.; Korth, Hans G.; Sustmann, R. Angew. Chem. 1984, 96, 887. (11) Korth, H. G.; Sustmann, R.; Groeninger, K. S.; Witzel, T.; Giese, B. J. Chem. Soc., Perkin Trans. 2 1986, 9, 1461. (12) Korth, H. G.; Sustmann, R.; Giese, B.; Rueckert, B.; Groeninger, K. S. Chem. Ber. 1990, 123, 1891. (13) Lloyd, R. V.; Causey, J. G. J. Chem Soc., Perkin Trans. 2 1981, 8, 1143.

J. Phys. Chem. B, Vol. 113, No. 19, 2009 6629 (14) Camaioni, D. M.; Walter, H. F.; Jordan, J. E.; Pratt, D. W. J. Am. Chem. Soc. 1973, 95, 7978. (15) Camaioni, D. M.; Pratt, D. W. J. Am. Chem. Soc. 1972, 94, 9258. (16) Forbes, M. D. E. Photochem. Photobiol. 1997, 65, 73. (17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision D.01; Gaussian, Inc.: Wallingford, CT, 2004. (18) Portmann, S.; Lu¨thi, H. P. CHIMIA 2000, 54, 766–770. (19) (a) Atkins, P. W.; Evans, G. T. Chem. Phys. Lett. 1974, 25, 108. (b) Wong, S. K.; Hutchinson, D. A.; Wan, J. K. S. J. Chem. Phys. 1973, 58, 985. (20) Harbron, E. J.; Forbes, M. D. E. Chem. Phys. Phys. Chem. 2001, 2, 1389. (21) Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin Trans. 2 1992, 2025. (22) Ogawa, S.; Fessenden, R. W. J. Chem. Phys. 1964, 41, 994. (23) Anet, F. A. L.; Bourn, A. J. R. J. Am. Chem. Soc. 1967, 89, 760. (24) Jancke, H.; Werner, H. J. Prakt. Chem. 1980, 322, 247. (25) Werner, H.; Mann, G.; Jancke, H.; Engelhardt, G. Tetrahedron Lett. 1975, 24, 1917.

JP809610T