Mechanistic analysis of the photochemical carboxylation of o-alkylphenyl ketones with carbon dioxide

Abstract

The mechanisms for photochemical carboxylation reactions are studied theoretically using two model systems: o-methylbenzophenone and o-methylacetophenone, with the M06-L and the 6-311G(d,p) basis set. Three reaction routes are used to ascertain the actual photochemical reaction mechanism for the o-alkylphenyl ketone molecules. The structures of the intersystem crossing, which play a central role in these photocarboxylations, are determined. The intermediates and transition structures for the ground singlet states and the excited triplet states are also computed to allow a qualitative interpretation of the entire reaction path. The theoretical findings show that when o-alkylphenyl ketone molecules are produced in the T₁ state by photoexcitation at 365 nm, intersystem crossing to the S₀ surface is the most probable pathway for deactivation. When they relax to the S₀ state, the o-alkylphenyl ketone molecules can either react with CO₂ to undergo the insertion reaction, which produces the carboxylic acid, or revert to the reactants. In particular, the theoretical investigations strongly suggest that the insertion of CO₂ in these photocarboxylation reactions for o-alkylphenyl ketones occurs on the S₀ surface, rather than on the excited T₁ surface.

---

Very recently, Murakami and co-workers were the first to report that o-alkylphenyl ketones can undergo carboxylation reactions with carbon dioxide (CO₂) by forming a C–C bond to yield o-acylphenylacetic acids, under simple UV irradiation (Scheme 1).¹ These photochemical reactions are distinctive and straightforward and only light energy is used to obtain the carboxylation products for organic molecules. No transition metal catalysts² or Grignard reagents³ are involved. These clean and transition-metal-free carboxylation reactions via photo-irradiation should be favorable. CO₂ has long been recognized as an essential component of greenhouse gases, which have caused a significant rise in atmospheric temperature and an abnormal change in the global climate.⁴ The use of CO₂ is an attractive alternative to compounds that are derived from either coal or petroleum as a starting material for the synthesis of useful chemicals.⁵ The use of UV light or even solar light to incorporate CO₂ into organic compounds as a chemical feedstock not only reduces the amount of CO₂, which alleviates serious environmental problems, but also generates potentially useful ingredients for a greener environment.⁶

Scheme 1 Photocarboxylation of o-alkylphenyl ketones with CO₂.

Murakami and co-workers offered a mechanism for these photo-assisted carboxylation reactions for o-alkylphenyl ketone with CO₂ (ESI†).¹ However, to the authors' best knowledge, this proposed mechanistic scenario has not been confirmed either experimentally nor theoretically. This stimulates interest to determine the photochemical pathways using theoretical computations. A theoretical study of the photocarboxylation reactions of o-methylbenzophenone (Rea-1-S₀) and o-methylacetophenone (Rea-2-S₀), as shown in eqn (1), was performed. It is hoped the combination of the experimental results and theoretical explanations will allow greater comprehension of the excited state reaction for the light-driven carboxylation of o-alkylphenyl ketone with CO₂.

The density functional theory (M06-L)⁷ was used in conjunction with the 6-311G(d,p) basis set. In order to verify the properties of the stationary points that were calculated at the M06-L/6-311G(d,p) level of theory, vibrational frequency analysis was used to describe either the minima (the number of imaginary frequencies (NIMAG) = 0) or transition states (NIMAG = 1). Since this study involves the excited triplet behavior of o-alkylphenyl ketones, the minimum energy crossing points between the triplet and singlet potential energy surfaces were examined using the computer code that was developed by Harvey et al.⁸ All of the computations were executed using the GAUSSIAN 09 package of programs.⁹ The Cartesian coordinates and the energies computed for the various points are obtainable as ESI.†

Earlier works concerning the photochemical reactions of o-alkylphenyl ketones¹⁰ experimentally reported that the overall reaction proceeds via abstraction of the triplet γ-hydrogen, which generates a triplet 1,4-biradical. Therefore, any discussion of the potential energy surfaces computed in this work must take account of the possible excited-state reaction path that leads to a singlet-triplet intersystem crossing. A schematic representation of the relationship between the intersection and the possible reaction route is given in Scheme 2 and serves as a basis for discussion.

Scheme 2 The intersystem crossing mechanism for the photocarboxylation of o-alkylphenyl ketones with CO₂.

On irradiation with light, a singlet o-methylbenzophenone (Rea-S₀) is promoted to a triplet minimum (Min-T₁) and subsequently undergoes a 1,4-H migration to form a triplet 1,4-biradical species (Int-T₁). Finally, the intersystem crossing leads to a carboxylic acid product (Pro-S₀). The model shown in Scheme 2 is used to interpret the mechanisms for the photochemical insertion reactions of eqn (1) in the following discussion.

The photoinsertion mechanism for Rea-1-S₀ with CO₂, given in eqn (1), is firstly considered. The theoretical analysis shows that there are three possible reaction routes on the triplet excited potential energy surfaces of Rea-1-S₀: path I, path II, and path III. Fig. 1 shows all of the relative energies of the various critical points with respect to the energy of the reactant, Rea-1-S₀. The geometrical structures of these points on the possible mechanistic pathways in Fig. 1 are given in Fig. 2 (path I), Fig. 3 (path II) and Fig. 4 (path III). There are several important conclusions that can be drawn from Fig. 1–4.

Fig. 1 The energy profiles for the photocarboxylation of o-methylbenzophenone (Rea-1-S₀) with CO₂ to produce carboxylic acid (Pro-1-S₀). The relative energies are calculated at the M06-L/6-311G(d,p) level of theory. All energies (in kcal mol⁻¹) are given with respect to the reactant (Rea-1-S₀). For the M06-L optimized structures for the crucial points, see Fig. 2–4 (ESI†). For more information, see the text.

Fig. 2 The M06-L/6-311G(d,p) geometries (in Å and deg) for path I of o-methylbenzophenone (Rea-1-S₀), intermediate, transition state (TS), intersystem crossing (T₁/S₀), and insertion product (Pro). The heavy arrow in TS indicates the main atomic motions in the transition state eigenvector. The gradient difference vector of T₁/S₀-1-I computed with M06-L is shown in the square bracket. Hydrogen atoms are omitted for clarity. For more details see the ESI.†

Fig. 3 The M06-L/6-311G(d,p) geometries (in Å and deg) for path II of o-methylbenzophenone (Rea-1-S₀), intermediate, transition state (TS), intersystem crossing (T₁/S₀), and insertion product (Pro). The heavy arrow in TS indicates the main atomic motions in the transition state eigenvector. The gradient difference vector of T₁/S₀-1-II computed with M06-L is shown in the square bracket. Hydrogen atoms are omitted for clarity. For more details see the ESI.†

Fig. 4 The M06-L/6-311G(d,p) geometries (in Å and deg) for path III of o-methylbenzophenone (Rea-1-S₀), intermediate, transition state (TS), intersystem crossing (T₁/S₀), and insertion product (Pro). The heavy arrow in TS indicates the main atomic motions in the transition state eigenvector. The gradient difference vector of T₁/S₀-1-III computed with M06-L is shown in the square bracket. Hydrogen atoms are omitted for clarity. For more details see the ESI.†

Upon the absorption of light, the molecule, Rea-1-S₀, firstly goes to the Frank–Condon region at the lowest-lying triplet state (FC-1-T₁), for which the vertical excitation energy (S₀ → T₁ (S₀ geom)) is calculated to be 77.1 kcal mol⁻¹. Experimentally, it was reported that the laboratory used an LED lamp with a wavelength of 365 nm (=78.3 kcal mol⁻¹) as the light source at the DMSO¹ and the results agree reasonably well with the M06-L computational data. Considering the model reactions theoretically studied in this work are in the gaseous phase and the effect of solvents is ignored, the accuracy of these computations is sufficient for the following investigations of the mechanisms of the photochemical carboxylation reactions (vide infra).

The theoretical findings shown in Fig. 1 suggest that the photocarbonylation of o-methylbenzophenone with CO₂ proceeds along three types of reaction pathways, as shown below:

Path I:

Rea-1-S₀ + CO₂ + hν → FC-1-T₁ + CO₂ → Min-1-T₁ + CO₂ → TS1-1-T₁ → T₁/S₀-1-I → Pro-1-S₀.

Path II:

Rea-1-S₀ + CO₂ + hν → FC-1-T₁ + CO₂ → Min-1-T₁ + CO₂ → TS2-1-T₁ + CO₂ → Int-1-T₁ + CO₂ → TS3-1-T₁ → T₁/S₀-1-II → Pro-1-S₀.

Path III:

Rea-1-S₀ + CO₂ + hν → FC-1-T₁ + CO₂ → Min-1-T₁ + CO₂ → TS2-1-T₁ + CO₂ → T₁/S₀-1-III + CO₂ → Int-1-S₀ + CO₂ → TS1-1-S₀ → Pro-1-S₀.

On path I, after the Frank–Condon point (FC-1-T₁), Rea-1-S₀ relaxes to a local minimum of the triplet surface near to the S₀ geometry, which is denoted Min-1-T₁. The M06-L calculations demonstrate that Min-1-T₁ is about 62 kcal mol⁻¹ above the ground state, Rea-1-S₀, as shown in Fig. 1. The optimized geometrical parameters are given in Fig. 2. It is seen that the bond of the carbonyl group of Min-1-T₁ (1.305 Å) is longer than that in the corresponding Rea-1-S₀ conformation (1.219 Å). This triplet minimum then interacts with CO₂ to give a triplet transition state (TS1-1-T₁). As seen in Fig. 2, following its transition state vector, which corresponds to the insertion of CO₂ into the C–H bond of the methyl group, it arrives at an intersystem crossing, T₁/S₀-1-I. This species then finally relaxes to the photoproduct, Pro-1-S₀. However, the M06-L computational results in Fig. 1 indicate that the TS1-1-T₁ transition state lies above the Rea-1-S₀ point by 146 kcal mol⁻¹, which is much higher than its vertical excitation energy (77 kcal mol⁻¹). Therefore, there is strong theoretical evidence that o-methylbenzophenone is unlikely to follow the mechanism of path I.

In path II, the photo-excited triplet minimum, Min-1-T₁, experiences the [1,5] sigmatropic migration of the H atom from the methyl group to the carbonyl group, which gives the transition state on the triplet surface (TS2-1-T₁), with an activation energy of about 10 kcal mol⁻¹. From TS2-1-T₁, a triplet 1,4-biradical intermediate, Int-1-T₁, is then obtained, as shown in Fig. 3. It is worthy of note that the twisted conformation of this triplet biradical species, which determines the stereochemistry of the final cyclo-butenol product, was discussed about 25 years ago by Wagner's group.¹¹ Subsequently, this triplet Int-1-T₁ reacts with CO₂ to reach a triplet transition state, TS3-1-T₁, followed by an intersystem crossing, T₁/S₀-1-II, and the final photoproduct, Pro-1-S₀, is then yielded. However, at the M06-L level of theory, the relative energy of TS3-1-T₁ is calculated to be 84 kcal mol⁻¹, with respect to the starting material, Rea-1-S₀. This value is larger than the vertical excitation energy (77 kcal mol⁻¹) of Rea-1-S₀, so when the triplet minimum (Min-1-T₁) is produced, the reactant cannot produce the photo-insertion product, Pro-1-S₀, via the process for path II.

In path III, similarly to path II, starting from the FC-1-T₁ point, the triplet Min-1-T₁ reactant undergoes the triplet-state γ-hydrogen abstraction through the TS2-1-T₁ transition state to generate an intersystem crossing, T₁/S₀-1-III. This species, in turn, decays to a singlet intermediate, o-quinodimethane (Int-1-S₀). As shown in Fig. 1, the M06-L computations predict that the triplet Int-1-T₁ point relaxes into the singlet Int-1-S₀ point via a short release of energy. The optimized geometrical structures for the intermediates at both the triplet (Int-1-T₁) and singlet (Int-1-S₀) states are shown in Fig. 3 and 4, respectively. The M06-L calculations show that the energy of Int-1-T₁ is greater than that of Int-1-S₀ by about 20 kcal mol⁻¹. The reason for this large energy difference is ascribed to the fact that the phenyl rings increase the effect of resonance, which significantly stabilizes the singlet intermediate (Int-1-S₀). The supporting evidence comes from the fact that the singlet Int-1-S0 species has a C₁=C₂–C₃=C₄ pattern (Fig. 4), but the triplet Int-1-T1 intermediate has a C₁–C₂=C₃–C₄ skeleton (Fig. 3). The theoretical findings suggest that two reaction pathways to the singlet Int-1-S₀ intermediate are possible. Fig. 1 shows that one pathway originates at the triplet Int-1-T₁ intermediate and directly relaxes to the Int-1-S₀ species after a release of energy. The other path originates at the T₁/S₀-1-III crossing point and subsequently goes to the Int-1-S₀ point. Therefore, all of these routes produce a large amount of the Int-1-S₀ species during light-assisted carboxylation reactions. This may explain why the triplet o-alkylphenyl ketone that is produced by UV irradiation is an essential precursor for photocarboxylation with carbon dioxide (see below).¹

From the Int-1-S₀ point, o-methylbenzophenone can then either interact with CO₂ to produce (2-benzoylphenyl)acetic acid (Pro-1-S₀) through the TS1-1-S₀ state or revert to the original reactant, Rea-1-S₀, via TS2-1-S₀ point, as shown in Fig. 1. The M06-L computational data indicates that the energy of TS1-1-S₀ that connect Int-1-S₀ and Pro-1-S₀ on the singlet potential energy surface lies about 51 kcal mol⁻¹ above that of the starting material, Rea-1-S₀. However, for the Int-1-S₀ → TS2-1-S₀ → Rea-1-S₀ process, TS2-1-S0 is computed to be about 45 kcal mol⁻¹ above the corresponding reactant, Rea-1-S₀. It is emphasized that o-methylbenzophenone maintains an excess energy of about 62 kcal mol⁻¹, which arises from the relaxation of the Frank–Condon point (FC-1-T₁) to the triplet minimum (Min-1-T₁). This energy is much greater than the energy difference between Int-1-S₀ and TS1-1-S₀ (16 kcal mol⁻¹) or between Int-1-S₀ and TS2-1-S₀ (9.0 kcal mol⁻¹), as seen in Fig. 1. These theoretical observations show that this photocarboxylation reaction has sufficient internal energy to overcome the energy barriers. Therefore, Int-1-S₀ is a branching point on the singlet profile, where there is either a return to the initial reactant or a reaction with CO₂ to produce the insertion reaction and generate the final product, Pro-1-S₀. The computational evidence is in good agreement with the experimental observations,¹ in that upon photo-irradiation with UV light, o-methylbenzophenone either reacts with carbon dioxide to produce carboxylic acid (Pro-1-S₀) or it regenerates to the singlet reactant (Rea-1-S₀) without any difficulty.

The photocarboxylation reaction for o-methylacetophenone (Rea-2-S₀) shown in eqn (1) was also studied. Similarly to the case for Rea-1-S₀, there are three reaction routes. In path I, a triplet reactant (Min-2-T₁) reacts with CO₂ and there is then an intersystem crossing to produce a final product. In path II, Min-2-T₁ firstly undergoes the 1,5-H migration and subsequently interacts with CO₂. This precedes an intersystem crossing to yield the same product. In path III, Min-2-T₁ begins an intersystem crossing to produce a singlet intermediate (Int-2-S₀). This then either interacts with CO₂ to produce the final insertion product or returns to the original reactants. The computed energy profiles using the M06-L/6-311G(d,p) level of theory are summarized in Fig. SB† and the optimized geometries for the corresponding stationary points are given in Fig. SA–SD.† These model computations for the potential energy surfaces of Rea-2-S0 shown in the Fig. SA† are quite similar to those for Rea-1-S₀, which are shown in Fig. 1. The theoretical study also demonstrates that path III is the preferred reaction route. All of the theoretical findings that are presented in this work agree well with experimental observations.¹

M06-L DFT studies are performed in order to gain a deeper understanding of the mechanism for the photocarboxylation reaction of o-alkylphenyl ketones with carbon dioxide. Three reaction pathways (path I, path II and path III) are theoretically studied. These are characterized on the basis of the computed potential energy surface profiles and the surface intersections, as schematically illustrated in Fig. 1 (and Fig. SA in ESI†). The computational evidence shows that path III is the most favorable pathway, from both an energetic and a kinetic viewpoint. Therefore, when o-alkylphenyl ketone populates on the triplet surface (T₁), its intersystem crossing to the ground singlet (S₀) surface is energetically allowed. After relaxation to the S₀ surface, o-alkylphenyl ketone retains sufficient internal energy. It then either interacts with carbon dioxide to undergo the insertion reaction, which produces carboxylic acid, or surmounts the energy barrier of the 1,5 H migration to go back to the initial reactants. The model investigations demonstrate that the photocarboxylation reaction for o-alkylphenyl ketones with CO₂ should occur on the ground singlet (S₀) surface, rather than on the excited triplet surface (T₁). This theoretical analysis provides a good explanation for the available experimental evidence and agrees well with the proposed mechanisms that have been proffered by Murakami and co-workers.¹

Acknowledgements

The authors are grateful to the National Center for High-Performance Computing of Taiwan for generous amounts of computing time, and the Ministry of Science and Technology of Taiwan for the financial support. In particular, one of the authors (M.-D. Su) also wishes to thank Professor Michael A. Robb, Dr S. Wilsey, Dr Michael J. Bearpark, (University of London, UK) and Professor Massimo Olivucci (Universita degli Studi di Siena, Italy), for their encouragement and support during his stay in London. Special thanks are also due to reviewers 1 and 2 for very help suggestions and comments.

References

1. Y. Masuda, N. Ishida and M. Murakami, J. Am. Chem. Soc., 2015, 137, 14063.
2. For transition metal catalysts, see: (a) M. Shi and K. M. Nicholas, J. Am. Chem. Soc., 1997, 119, 5057; (b) M. Takimoto, J. Am. Chem. Soc., 2002, 124, 10008; (c) K. Ukai, M. Aoki, J. Takaya and N. Iwasawa, J. Am. Chem. Soc., 2006, 128, 8706; (d) C. S. Yeung and V. M. Dong, J. Am. Chem. Soc., 2008, 130, 7826; (e) C. M. Williams, J. B. Johnson and T. Rovis, J. Am. Chem. Soc., 2008, 130, 14936; (f) J. Takaya and N. Iwasawa, J. Am. Chem. Soc., 2008, 130, 15254; (g) A. Correa and R. Martin, J. Am. Chem. Soc., 2009, 131, 15974; (h) I. I. F. Boogaerts and S. P. Nolan, J. Am. Chem. Soc., 2010, 132, 8858; (i) L. Zhang, J. Cheng, T. Ohishi and Z. Hou, Angew. Chem., Int. Ed., 2010, 49, 8670; (j) T. Fujihara, T. Xu, K. Semba, J. Terao and Y. Tsuji, Angew. Chem., Int. Ed., 2011, 50, 523; (k) H. Mizuno, J. Takaya and N. Iwasawa, J. Am. Chem. Soc., 2011, 133, 1251; (l) S. Li, W. Yuan and S. Ma, Angew. Chem., Int. Ed., 2011, 50, 2578; (m) D. P. Hruszkewycz, J. Wu, N. Hazari and C. D. Incarvito, J. Am. Chem. Soc., 2011, 133, 3280; (n) T. Fujihara, K. Nogi, T. Xu, J. Terao and Y. Tsuji, J. Am. Chem. Soc., 2012, 134, 9106; (o) K. Sasano, J. Takaya and N. Iwasawa, J. Am. Chem. Soc., 2013, 135, 10954; (p) T. G. Ostapowicz, M. Schmitz, M. Krystof, J. Klankermayer and W. Leitner, Angew. Chem., Int. Ed., 2013, 52, 12119; (q) X. Wang, M. Nakajima and R. Martin, J. Am. Chem. Soc., 2015, 137, 8924.
3. For base-promoted carboxylation reactions, see: (a) O. Vechorkin, N. Hirt and X. Hu, Org. Lett., 2010, 12, 3567; (b) K. Inamoto, N. Asano, Y. Nakamura, M. Yonemoto and Y. Kondo, Org. Lett., 2012, 14, 2622; (c) W.-J. Yoo, M. Guiteras, C. Capdevila, X. Du and S. Kobayashi, Org. Lett., 2012, 14, 5326; (d) Y. Suzuki, T. Hattori, T. Okuzawa and S. Miyano, Chem. Lett., 2002, 31, 102; (e) G. A. Olah, B. Torok, J. P. Joschek, I. Bucsi, P. M. Estevens, G. Rasul and G. K. S. Prakash, J. Am. Chem. Soc., 2002, 124, 11379.
4. For instance, see: (a) P. Braunstein, D. Matt and D. Nobel, Chem. Rev., 1988, 88, 747; (b) Carbon Dioxide as Chemical Feedstock, ed. M. Aresta, Wiley-VCH, Weinheim, 2010; (c) R. Martin and A. W. Kleij, ChemSusChem, 2011, 4, 1259; (d) K. Huang, C.-L. Sun and Z.-J. Shi, Chem. Soc. Rev., 2011, 40, 2435; (e) L. Zhang and Z. Hou, Chem. Sci., 2013, 4, 3395.
5. For instance, see: (a) A. S. Lindsey and H. Jeskey, Chem. Rev., 1957, 57, 583; (b) X. Yin and J. R. Moss, Coord. Chem. Rev., 1999, 181, 27; (c) T. Sakakura, J.-C. Choi and H. Tasuda, Chem. Rev., 2007, 107, 2365; (d) S. N. Riduan and Y. Zhang, Dalton Trans., 2010, 39, 3347; (e) I. I. F. Boogaerts and S. P. Nolan, Chem. Commun., 2011, 47, 3021; (f) M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann and F. E. Kuhn, Angew. Chem., Int. Ed., 2011, 50, 8510; (g) Y. Tsuji and T. Fujihara, Chem. Commun., 2012, 48, 9956.
6. (a) D. B. Dell'Amico, F. Calderazzo, L. Labella, F. Marchetti and G. Pampaloni, Chem. Rev., 2003, 103, 3857; (b) T. Sakakura, J.-C. Choi. and H. Yasuda, Chem. Rev., 2007, 107, 2365; (c) D. J. Darensbourg, Chem. Rev., 2007, 107, 2388; (d) M. S. Aresta, Carbon Dioxide as Chemical Feedstock, Wiley-VCH, Weinheim, 2010; (e) S. N. Riduan and Y. Zhang, Dalton Trans., 2010, 39, 3347; (f) A. Behr and G. Henze, Green Chem., 2011, 13, 25; (g) M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann and F. E. Kühn, Angew.Chem., 2011, 123, 8662 (Angew. Chem. Int. Ed., 2011, 50, 8510); (h) I. Omae, Coord. Chem. Rev., 2012, 256, 1384.
7. Y. Zhao and D. G. Truhlar, Acc. Chem. Res., 2008, 41, 157.
8. (a) J. N. Harvey, M. Aschi, H. Schwarz and W. Koch, Theor. Chem. Acc., 1998, 99, 95; (b) J. N. Harvey and M. Aschi, Phys. Chem. Chem. Phys., 1999, 1, 5555.
9. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci and G. A. Petersson, et al., Gaussian, Gaussian, Inc., Wallingford, CT, 2013.
10. (a) R. Haag, J. Wirz and P. J. Wagner, Helv. Chim. Acta, 1977, 60, 2595; (b) P. K. Das, M. V. Encinas, R. D. Small Jr and J. C. Scaiano, J. Am. Chem. Soc., 1979, 101, 6965; (c) P. J. Wagner and C.-P. Chen, J. Am. Chem. Soc., 1976, 98, 239.
11. (a) P. J. Wagner, D. Subrahmanyam and B.-S. Park, J. Am. Chem. Soc., 1991, 113, 709; (b) P. J. Wagner, M. Sobczak and B.-S. Park, J. Am. Chem. Soc., 1998, 120, 2488.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra05890f

---