Porphyrins bearing corannulene pincers: outstanding fullerene receptors

Abstract

Herein, we study new fullerene receptors which are constructed employing porphyrins, corannulene pincers and metallic centers. Theoretical calculations indicated that if porphyrin and corannulene pincers are merged, the strongest hosts for fullerenes can be built. We found that there is a linear increase of the interaction energy with respect to the number of pincers. The linear trend is broken when the pincers interact amongst themselves. The strongest host–C₆₀ interaction was observed for the porphyrin which has 4 corannulene pincers and Ir or Rh centers, which exhibited interaction energies significantly larger than the one computed for the dimeric porphyrins. In addition, we report that some of the porphyrin receptors can host two fullerene molecules very efficiently, given that the host can be deformed without increasing its energy too much. Finally, we observed a good sensitivity of our hosts towards C₇₀ with respect to C₆₀. Synthetic routes of the hosts designed are discussed.

---

1. Introduction

The roadmap towards the synthesis of effective fullerene receptors has a large number of milestones, the first being calixarenes^1,2 and porphyrins.^3–6 The second family of organic compounds used to trap fullerenes is that of geodesic polyarenes, a work pioneered by Scott's group.^7,8 Some years later Sygula et al.⁹ attached two corannulene pincers to a cyclooctatetraene core to produce the C₆₀H₂₈ buckycatcher.^9–14 In the same vein, Yanney et al.¹⁵ used a different approach to join two corannulene pincers to norbornadiene producing the buckycatcher II, which has a larger affinity towards C₆₀ than the original buckycatcher, and also is capable of forming dimeric complexes with C₆₀. In order to increase the affinity of the receptors towards fullerenes, it has been proposed that the use of three corannulene pincers should be extremely effective towards that end. In line with this, Yanney and Sygula¹⁶ functionalized cyclotriveratrylene core with three corannulene pincers. Yet, the association constant measured was very low. The reason behind this unexpected result was attributed to the stacking of the corannulene pincers^11,16 and the conformational flexibility of the receptor. The opposite behavior was obtained by Huerta et al.¹⁷ who observed that C₆₀ was strongly trapped when three 2-[9-(1,3-dithiol-2-ylidene)anthracen-10(9H)-ylidene]-1,3-dithiole (exTTF) pincers were bonded to cyclotriveratrylene derivative; given that the exTTF dimer has a dimerization energy similar to that found in corannulene^18–21 the effectiveness of exTTF was quite unexpected. Considering that porphyrins are considered to be the best receptors for fullerenes, closely followed by those derived from corannulene, the next promising step towards the synthesis of improved fullerene traps is the combination of these two types of organic compounds. Over the last years, density functional theory (DFT) has been employed to help in the synthesis of new hosts for fullerenes. Examples includes the works by Rodríguez-Otero et al.,^22–24 Lein et al.,^25,26 Denis et al.,^27–30 Cassella and Saielli³¹ and Armakovic et al.,³² among many others. Herein, we have employed efficient DFT calculations to design and characterize new hosts for fullerenes which are formed by a porphyrin core and present one to four corannulene pincers attached at the rim. The receptor was carefully designed to minimize: (a) the undesired entropic effect introduced by the presence of floppy pincers and (b) the effect of intramolecular stacking. Our results indicated that the new receptor designed holds C₆₀ much more strongly than the dimeric porphyrins synthesized by Aida,^5,6 and thus constitute the next synthetic target to be considered. It is our expectation that the results presented and the synthetic route described, motivates the synthesis of the receptors proposed.

2. Methods

Because of the large size of the systems studied, the gold standard of Quantum Chemistry, Coupled Cluster Theory, cannot be employed. For this reason we adopted the approach that we have been employing in our previous investigations of fullerene receptors,^{11,12,18,19,27–30} which is similar to that selected by Lein and coworkers.^25,26 All complexes were studied employing the M06-2X functional developed by Zhao and Truhlar.^33,34 Two basis sets were selected: 6-31G and 6-31G*.³⁵ In those cases where metal atoms were utilized to construct the porphyrins the LANL2DZ basis set was utilized with the accompanying pseudopotentials.^36–38 All complexes were optimized at the M06-2X/6-31G and M06-2X/6-31G* levels of theory. It has been shown by us that the M06-2X functional adequately describes the stacking interactions which involve the corannulene dimer¹⁹ and the C₆₀@buckycatcher complex.¹¹ For example, the interaction energy of the corannulene dimer is −15.5 kcal mol⁻¹ at the QCISD(T)/aug-cc-pVTZ+BSSE level of theory,³⁹ while we have determined that it is −15.1, −15.2 and −14.1 kcal mol⁻¹ at the M06-2X/6-31G, M06-2X/6-31G* and M06-2X/6-311G+BSSE levels of theory, respectively. In the case of the buckycatcher complex C₆₀@C₆₀H₂₈, the M06-2X/6-31G* interaction energy is −31.7 kcal mol⁻¹, only 3.7 kcal mol⁻¹ outside the error bars of the value computed using Diffusion Montecarlo techniques.¹⁰ Therefore, the performance of the M06-2X/6-31G* method is good for the purposes of this work, given that we are not interested in computing very accurate interaction energies but in designing better fullerene receptors. Notwithstanding the fact that M06-2X does not have the correct asymptotic behavior at long distances, it has been pointed out by Ristauss and Grimme⁴⁰ that this functional performs well at equilibrium and when three body corrections are not included. Moreover, we have shown that when combined with the 6-31G/6-31G* basis set and excluding basis set superposition error (BSSE) this method good interaction energies as mentioned above for the corannulene dimer¹⁹ and the C₆₀@buckycatcher complex.¹¹ Therefore, BSSE was not included, because by not considering it we can compensate for the underestimation of dispersion interactions that affects the M06-2X functional. For comparative purposes, we performed PBE-D3BJ^41,42 calculations which include D3 Grimme⁴¹ corrections to dispersion and the Becke–Johnson damping. The basis set employed for the PBE-D3BJ calculations was the 6-31G*, while for metals the LANL2DZ basis set was utilized with the accompanying pseudopotentials. We note that all the conclusions obtained at the M06-2X/6-31G* level were supported by PBE-D3BJ/6-31G* calculations. For all calculations the ultrafine (99 950 point) grid was used and the structures were optimized for both basis sets. Finally, the effect of the solvent was considered employing the model developed by Marenich et al.⁴³ The solvent selected was toluene, at the M06-2X/6-31G* level of theory. Vibrational frequencies were calculated at the M06-2X/6-31G and M06-2X/6-31G* levels and free energy changes were computed at 1 atm and 298 K. All of the DFT calculations were carried out using Gaussian 09.⁴⁴

3. Results and discussion

In first place we investigated the interaction between C₆₀ and the 2H-porphyrin shown in Fig. 1. The latter has no metal atoms and no alkyl chains are attached at the rim, so a weak interaction with fullerene is expected. All the energetic data is gathered in Table 1, and the following discussion will be based on the M06-2X/6-31G* results. In effect, we found that the interaction energy (IE) is −17.9 kcal mol⁻¹, which is translated to a ΔG^°₂₉₈ = −4.3 kcal mol⁻¹, at the M06-2X/6-31G* level. The closest contacts between the fullerene and the porphyrin occur between the H atoms of the NH groups and the carbon atoms of the C=C bond shared by two hexagons, which is positioned above. The C⋯H distances are not equal, 2.69 and 2.76 Å, respectively. These values are in line with previous experimental investigations,³ which indicated that the bonding between porphyrins and fullerenes is stronger than regular pi stacking interactions. In order to increase the interaction between the porphyrin and C₆₀ we considered the possibility of joining corannulene pincers to the rim of the host. To that end, we have employed the norbornadiene units that have been recently employed to create buckycatcher II.¹⁵ In Fig. 2 and 3 we present the structures of the 4C-porphyrin and the complex formed with C₆₀, respectively. The distances between the atoms of C₆₀ and those of the pincers lie between 3.3 and 3.6 Å, which is very good to obtain strong ball–socket interactions. The outcome of the M06-2X/6-31G* calculations supports this statement, as the IE is −62.8 kcal mol⁻¹ and the ΔG^°₂₉₈ = −46.0 kcal mol⁻¹. This host–guest complex is the strongest among all the complexes that we have assayed over the last years. For example, C₆₀@buckycatcher has a IE of −31.7 kcal mol⁻¹, while for buckycatcher II the IE is −32.9 kcal mol⁻¹, both computed at the M06-2X/6-31G* level of theory. Can we expect such a large IE between C₆₀ and 4C-porphyrin? The answer to that question can be obtained by following raw estimation; the IE between C₆₀ of corannulene is −15.0 kcal mol⁻¹, at the M06-2X/6-31G* level of theory, while that of the porphyrin is −17.9 kcal mol⁻¹ as mentioned above. In a perfect situation the four corannulene pincers will contribute with −15 × 4 = −60 kcal mol⁻¹, which when summed with the IE of the porphyrin give an “ideal” IE of −77.9 kcal mol⁻¹. This value is 15.1 kcal mol⁻¹ larger than the IE computed by us. This difference is small if we consider that for each pincer it is very difficult to obtain an optimal interaction with the fullerene in order to maximize the stacking. Also, the deformation experienced by the host is an important factor given that 15.6 kcal mol⁻¹ are required to bring the receptor from its isolated configuration, to that observed in the complex. For comparative purposes we computed the IE of C₆₀@4C-porphyrin at the B3LYP/6-31G* level of theory.^45,46 As expected, because B3LYP does not treat dispersion interactions properly, we found that the IE is +1.8 kcal mol⁻¹. This value is an indicator of the importance of non-bonded interactions for the complexes studied. In addition to this, we determined the IE at the M06-2X/6-311G*+BSSE level of theory. Interestingly, we found that it is 57.1 kcal mol⁻¹, very close to the value obtained using the smaller basis set and without including BSSE. We also investigated the C₇₀@4C-porphyrin complex. In this case, two structures were found: (a) complex 1, in which the long axis of C₇₀ has an almost upright configuration; and (b) complex 2 where the long axis of C₇₀ is parallel with the plane of the porphyrin. The computed IE are −69.3 and −68.4 kcal mol⁻¹, at the M06-2X/6-31G* level, for complexes 1 and 2, respectively, whereas the ΔG^°₂₉₈ are −53.7 and −53.0 kcal mol⁻¹, respectively. Thus, in terms of ΔG^°₂₉₈, the value computed for C₇₀ is 7.7 kcal mol⁻¹ larger than the one determined for C₆₀ and for this reason we expect that the 4C-porphyrin host may display a strong sensitivity towards C₇₀.

Fig. 1 Optimized structure for the isolated 2H-porphyrin (left) and its complex with C₆₀ (right), at the M06-2X/6-31G* level of theory.

Table 1 Interaction energies and free energies (kcal mol⁻¹ determined for the complexes formed by C₆₀ and C₇₀ with different hosts

	ΔEgas	ΔH^°gas 298	ΔG^°gas 298	ΔEgas	ΔH^°gas 298	ΔG^°gas 298	ΔEtoluene	ΔG^°toluene 298
a The binding energy per C60 molecule is 42.6 kcal mol^−1.
M06-2X	6-31G	6-31G	6-31G	6-31G*	6-31G*	6-31G*	6-31G*	6-31G*
C60@2H-porphyrin	−17.4	−16.3	−4.1	−17.9	−16.8	−4.3	−10.4	3.2
C60@1C-porphyrin	−33.0	−32.0	−15.6	−34.2	−33.0	−16.5	−19.0	−1.3
C60@2C-porphyrin	−48.7	−47.9	−28.2	−50.0	−49.2	−29.2	−27.9	−7.1
C60@3C-porphyrin	−53.0	−52.3	−36.7	−53.8	−53.2	−36.6	−30.3	−13.1
C60@4C-porphyrin	−62.1	−61.6	−45.2	−62.8	−62.5	−46.0	−34.5	−17.7
C60@4C-porphyrin-Ir				−67.8	−66.9	−49.7	−36.8	−18.7
C60@4C-porphyrin-Rh				−67.0	−65.8	−48.9	−37.1	−19.0
C60@porphyrin-Ir-Aida				−41.4	−39.2	−16.5	−11.3	13.6
2C60@4C-porphyrin-2up-2dn	−83.7	−81.8	−53.5	−85.2^a	−83.1	−54.5	−50.0	−19.3
C70@4C-porphyrin-1	−68.3	−68.1	−52.3	−69.3	−69.2	−53.4	−38.4	−22.4
C70@4C-porphyrin-2	−67.8	−67.5	−52.3	−68.4	−68.2	−52.9	−38.8	−23.3

---

Fig. 2 Optimized structure for the 4C-porphyrin, at the M06-2X/6-31G* level of theory. The H atoms of the corannulene pincers were omitted (for clarity) and the pincer which bends forward to interact with the other corannulene molecules is shown in a different color.

Fig. 3 Optimized structure for the C₆₀@4C-porphyrin complex, at the M06-2X/6-31G* level of theory. The H atoms of the corannulene pincers were omitted (for clarity) and C₆₀ is shown in a different color.

The 4C-porphyrin has two structural isomers depending on the position of the pincers. They are shown in Fig. 4. For 4C-porphyrin-3up-1dn we found that it is 8.2 kcal mol⁻¹ less stable than 4C-porphyirin. In the case of 4C-porphyin-2up-2dn we found that it is −10.9 kcal mol⁻¹ less stable than the one with the four pincers up. Although it is less stable, 4C-porphyin-2up-2dn can trap 2C₆₀ molecules. We found that the IE for 2C₆₀@4C-porphyin-2up-2dn is −85.2 kcal mol⁻¹, about −42.6 kcal mol⁻¹ per C₆₀ molecule, at the M06-2X/6-31G* level. Therefore, the reduction from 4 to 2 pincers which interact with C₆₀ decreases the IE by 20.2 kcal mol⁻¹, a value which is smaller than what we expected given that the IE of C₆₀@corannulene is 15.0 kcal mol⁻¹. We also computed the deformation of the host, finding that 3.4 kcal mol⁻¹ are needed to reorganize the structure of the host. For comparative purposes, we also determined the IE of the complex C₆₀@4C-porphyin-2up-2dn, which is −40.1 kcal mol⁻¹, (ΔG^°₂₉₈ = −30.5 kcal mol⁻¹). This IE is 2.5 kcal mol⁻¹ above the IE obtained when two C₆₀ molecules are hosted and is another indicator about the good flexibility observed above for the norbornadiene pincers, which can be altered without increasing the energy of the host too much.

Fig. 4 Optimized structures for the 4C-porphyrin-2up-2dn (top) and 4C-porphyrin-3up-1dn (bottom) structural isomers of 4C-porphyrin, at the M06-2X/6-31G* level of theory.

Receptors with several pincers can be prone to aggregation, either with other receptors or itself. As regards the aggregation with other receptor molecules, we studied two dimers: (a) two porphyrin-4C pincers stack via the porphyrin planes; (b) the pincers of the receptors interact. In both cases we found that the interaction energies are 18–20 kcal mol⁻¹, much smaller than that computed for the C₆₀@porphyrin-4C complex.

To gain deeper insight into the influence of the pincers, we studied three more porphyrin based receptors which bears one (1C-porphyrin), two (2C-porphyrin) and three corannulenes (3C-porphyrin), as shown in Fig. 5. The IE computed are −34.2, −50.0 and −53.8 kcal mol⁻¹, for one, two and three pincers, respectively, at the M06-2X/6-31G* level. In Fig. 6 we plot the interaction energy against the number of pincers. We can appreciate that there is a linear decrease of the IE vs. the number of pincers up to n = 2. For three pincers the IE is slightly above the value expected from a linear trend. The reason for such deviation is that for three pincers, one of the corannulene molecules bends forward and the receptor folds. For this reason, the energy required to alter the equilibrium configuration of the free receptor becomes larger, and concomitantly the IE becomes more positive. We would like to stress that pincer–pincer contacts can be a problem when they are attached to flexible arms,¹¹ as we have shown for cyclotriveratrylene based receptors.¹¹ For this reason, we have specifically designed a receptor with 4 pincers which does not suffer from aggregation problems. The key to solve this problem is the use of norbornadiene tethers which do not permit the aggregation of the four pincers. In the worst scenario found, for the porphyrin-4C and porphyrin-3C receptors, only two pincers interact and for this reason the linear trend is broken in Fig. 6. As explained above, the energy lost due to aggregation is only 15.6 kcal mol⁻¹, a value which nicely matches the interaction energy of the corannulene dimer, i.e. two pincers stacked.

Fig. 5 Optimized structure for the C₆₀@1C-porphyrin, C₆₀@2C-porphyrin and C₆₀@3C-porphyrin complexes, at the M06-2X/6-31G* level of theory. In the case of the 3C-porphyin C₆₀ is shown in a different color for clarity.

Fig. 6 Plot of the number of corannulene pincers attached to the 2H-porphyrin vs. the interaction energy computed for the NC-porphyrins N = 1, 2, 3 and 4, at the M06-2X/6-31G* level of theory.

It is interesting to compare the performance of the 4C-porphyrin receptor with respect to the Ir-based porphyrin receptors developed by Aida (dubbed as porphyrin-Ir-Aida). We determined that the IE of C₆₀@porphyrin-Ir-Aida is −41.4 kcal mol⁻¹, about 21.4 kcal mol⁻¹ below the IE computed for the 4C-porphyrin receptor. However, when free energy is considered, we found that the ΔG^°₂₉₈ is much lower than the ΔG^°₂₉₈ determined for C₆₀@4C-porphyrin. The reason behind that larger difference is the presence of 16 ethyl groups attached to the rim of the Ir-porphyrin, which introduce a large entropic contribution due to their free motion. To understand the effect of the metal center, we replaced the 2H atoms of 4C-porphyin by Ir–CH₃ or Rh–CH₃ groups. The IE determined are −67.8 and −67.0 kcal mol⁻¹ for the Ir–CH₃ or Rh–CH₃ 4C-porphyrins, respectively. Thus, the introduction of an Ir/Rh metal increases the IE by 5.0 and 4.2 kcal mol⁻¹, respectively, providing the best host for C₆₀ proposed in this work. With regards to solvation effects, the results listed in Table 1 indicated that the ΔG^°_{toluene 298} values follow the same trend as the ΔE and ΔG^°_{gas 298} values, i.e. the combination of porphyrins/corannulenes is more effective than a dimeric porphyrin to host fullerenes and the selectivity between C₇₀ and C₆₀ is maintained. Finally, for comparative purposes we performed PBE-D3BJ calculations for all complexes studied. The results are gathered in Table 2. Comparison of the M06-2X and PBE-D3BJ reveals that the latter method supports the results obtained with M06-2X. In effect, four corannulene pincers trap C₆₀ much stronger than a dimeric porphyrin and the combination of 4 corannulene pincers with an Ir containing porphyrin provides the best receptor for C₆₀.

Table 2 Comparison of the interaction energies (kcal mol⁻¹) determined for the complexes formed by C₆₀ and C₇₀ with different hosts at the PBE-D3BJ/6-31G* and M06-2X/6-31G* levels of theory

	M06-2X 6-/31G*	PBE-D3BJ/6-31G*
C60@2H-porphyrin	−17.9	−21.7
C60@1C-porphyrin	−34.2	−40.5
C60@2C-porphyrin	−50.0	−59.0
C60@3C-porphyrin	−53.8	−63.0
C60@4C-porphyrin	−62.8	−69.6
C60@4C-porphyrin-Ir	−67.8	−78.4
C60@4C-porphyrin-Rh	−67.0	−76.0
C60@porphyrin-Ir-Aida	−41.4	−54.4
2C60@4C-porphyrin-2up-2dn	−85.2	−119.2
C70@4C-porphyrin-1	−69.3	−75.0
C70@4C-porphyrin-2	−68.4	−72.5

---

Although theoretical calculations are a valuable tool to design receptors, we comment on possible synthetic routes for the receptors proposed. Molecular receptor 4C-porphyrin (4 in Scheme 1) composed of a porphyrin core and four corannulene pincers could be synthesized by combining the synthetic procedures utilized by Yanney et al.¹⁵ and Smith et al.⁴⁷ The first step, which is a known synthetic procedure employed by Yanney et al.¹⁵ in the synthesis of buckycatcher II is the reaction between isocorannuleno furan⁴⁸ 5 and 2,5-norbornadiene 6 to afford compound 7. Dehydration of 7 using p-toluene sulfonic acid should lead to the formation of 8 which could then be reacted with tosylmethylisocyanide (TOSMIC) in a similar scheme demonstrated by Smith et al.⁴⁷ to afford the pyrrole derivative 9. Finally, cyclocondensation of 9 using formaldehyde in a suitable solvent should lead to the desired porphyrin–corannulene receptor 4.

Scheme 1 Proposed synthetic route for the 4C-porphyrin receptor (4).

4. Conclusions

The M06-2X density functional was employed to design and characterize fullerene receptors which we link a porphyrin core with corannulene pincers. The receptors studied exhibited large interaction energies. In particular, the Ir based porphyrins with 4 corannulene pincers showed the largest affinity towards fullerenes and are very sensitive to distinguish C₆₀ and C₇₀. The strength of the supramolecular complexes formed was confirmed by comparing with the complexation energies computed for the dimeric iridium containing porphyrin prepared by Aida. Finally, we observed that one of the porphyrin receptors studied can trap two fullerene molecules with great affinity.

Acknowledgements

The author thanks PEDECIBA Quimica, CSIC and ANII Uruguayan institutions for financial support.

References

1. S. K. Nakashima and S. Shinkai, Chem. Lett., 1994, 23, 699–702.
2. J. L. Atwood, G. A. Koutsantonis and C. L. Raston, Nature, 1994, 368, 229.
3. P. D. W. Boyd, M. C. Hodgson, C. E. F. Rickard, A. G. Oliver, L. Chaker, P. J. Brothers, R. D. Bolskar, F. S. Tham and C. A. Reed, J. Am. Chem. Soc., 1999, 121, 10487–10495.
4. M. M. Olmstead, D. A. Costa, K. Maitra, B. C. Noll, S. L. Phillips, P. M. Van Calcar and A. L. Balch, J. Am. Chem. Soc., 1999, 121, 7090–7097.
5. J.-Y. Zheng, K. Tashiro, Y. Hirabayashi, K. Kinbara, K. Saigo, T. Aida, S. Sakamoto and K. Yamaguchi, Angew. Chem., Int. Ed., 2001, 40, 1857.
6. M. Yanagisawa, K. Tashiro, M. Yamasaki and T. Aida, J. Am. Chem. Soc., 2007, 129, 11912.
7. S. Mizyed, P. E. Georghiou, M. Bancu, B. Cuadra, A. K. Rai, P. Cheng and L. T. Scott, J. Am. Chem. Soc., 2001, 123, 12770–12774.
8. P. E. Georghiou, A.-H. Tran, S. Mizyed, M. Bancu and L. T. Scott, J. Org. Chem., 2005, 70, 6158–6163.
9. A. Sygula, F. R. Fronczek, R. Sygula, P. W. Rabideau and M. M. Olmstead, J. Am. Chem. Soc., 2007, 129, 3842.
10. A. Tkatchenko, D. Alfè and K. S. Kim, J. Chem. Theory Comput., 2012, 8, 4317–4322.
11. P. A. Denis, RSC Adv., 2013, 3, 25296.
12. P. A. Denis, Chem. Phys. Lett., 2014, 591, 323.
13. V. E. Le, M. Yanney, M. McGuire, A. Sygula and E. A. Lewis, J. Phys. Chem. B, 2014, 118, 11956–11964.
14. S. Grimme, Chem.–Eur. J., 2012, 18, 9955.
15. M. Yanney, F. R. Fronczek and A. Sygula, Angew. Chem., Int. Ed., 2015, 54, 1153–11156.
16. M. Yanney and A. Sygula, Tetrahedron Lett., 2013, 54, 2604.
17. E. Huerta, H. Isla, E. M. Perez, C. Bo, N. Martın and J. de Mendoza, J. Am. Chem. Soc., 2010, 132, 5351–5353.
18. P. A. Denis, Struct. Chem., 2015, 26, 171–176.
19. P. A. Denis, Chem. Phys. Lett., 2011, 516, 82.
20. T. Janowski, P. Pulay, A. A. S. Karunarathna, A. Sygula and S. Saebo, Chem. Phys. Lett., 2011, 512, 155.
21. D. Vijay, H. Sakurai and G. N. Sastry, Int. J. Quantum Chem., 2011, 111, 1893.
22. D. Josa, J. Rodríguez-Otero, E. M. Cabaleiro-Lago, L. A. Santos and T. C. Ramalho, J. Phys. Chem. A, 2014, 118, 9521–9528.
23. D. Josa, L. A. dos Santos, I. Gonzalez-Veloso, J. Rodríguez-Otero, E. M. Cabaleiro-Lago and T. C. Ramalho, RSC Adv., 2014, 4, 29826–29833.
24. D. Josa, J. Rodríguez-Otero and E. M. Cabaleiro-Lago, Phys. Chem. Chem. Phys., 2015, 17, 13206–13214.
25. R. Kleingeld, J. Wang and M. Lein, Polyhedron, 2016 DOI:10.1016/j.poly.2015.11.016.
26. I. Welsh and M. Lein, J. Comput. Chem., 2014, 35, 181–191.
27. P. A. Denis, New J. Chem., 2014, 38, 5608.
28. P. A. Denis, J. Phys. Org. Chem., 2014, 27, 918–925.
29. P. A. Denis and M. Yanney, New J. Chem., 2016, 40, 202.
30. P. A. Denis and M. Yanney, Chem. Phys. Lett., 2015, 640, 140–146.
31. G. Casella and G. Saielli, New J. Chem., 2011, 35, 1453–1459.
32. S. Armakovic, S. J. Armakovic, J. P. Setrajcic and I. J. Setrajcic, Chem. Phys. Lett., 2013, 578, 156–161.
33. Y. Zhao and D. G. Truhlar, J. Chem. Phys., 2006, 125, 194101.
34. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215.
35. R. Ditchfeld, W. J. Hehre and J. A. Pople, J. Chem. Phys., 1971, 54, 724.
36. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 270.
37. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 284.
38. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299.
39. T. Janowski, P. Pulay, A. A. S. Karunarathna, A. Sygula and S. Saebo, Chem. Phys. Lett., 2011, 512, 155.
40. T. Risthaus and S. Grimme, J. Chem. Theory Comput., 2013, 9, 1580–1591.
41. S. Grimme, S. Ehrlich and L. Goerigk, J. Chem. Phys., 2010, 132, 154104.
42. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
43. A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B, 2009, 113, 6378–6396.
44. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
45. A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
46. C. Lee, W. Yang and R. B. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785.
47. N. D. Smith, D. Huang and N. D. P. Cosford, Org. Lett., 2002, 4, 3537–3539.
48. A. Sygula, R. Sygula and P. W. Rabideau, Org. Lett., 2006, 8, 5909.

---