Theoretical studies on oxidation-switchable second-order nonlinear optical responses of Metallosalen-Keggin polyoxometalate derivatives

Abstract

To explore the effect of one-electron oxidation on nonlinear optical (NLO) responses, the frontier molecular orbitals, second-order NLO responses and electronic transition properties of three metallosalen–polyoxometalate compounds {SiW₁₁O₃₉[O(SiR)₂M]}⁴⁻ (R = (CH₂)₃N–CH(2-OPh); M = Pd, Ni and Co) (M-salen–POM) and their one-electron oxidized species were systematically investigated by density functional theory (DFT). Fukui function analysis shows that the M-salen is the oxidized center in M-salen–POM compounds. In addition, four functionals have been taken to confirm the switching behavior of the NLO properties by the one-electron oxidation. The β_tot value of oxo-Pd-salen–POM is 98 times larger than that of Pd-salen–POM, indicating that the first hyperpolarizability is enhanced by one-electron oxidation. Time dependent (TD) DFT calculations show that the charge transfer from POM to M-salen, and intramolecular charge transfer within M-salen in oxidized species are beneficial to reduce the transition energy and improve the static first hyperpolarizabilities. The M-salen–POM may be a potential class of switchable nonlinear optical materials by one-electron oxidized processes.

---

Introduction

Design and synthesis of the second-order nonlinear optical (NLO) materials displaying large first hyperpolarizabilities (β) have been the focus of intensive investigations for several decades owing to their potential applications in the fields of storage, telecommunications, optical data processing and high speed optical switching.^1–8 In the past few years, molecular switches exhibiting changes in color,^9,10 luminescence,^11,12 optical nonlinearity,^13,14 or magnetic properties^15–17 have been reported. Among them, switchable NLO materials have received much attention due to novel applications in electro-optic technologies.^18,19 Several strategies have been proposed to achieve the effective switchable NLO materials, such as oxidation/reduction, protonation/deprotonation and photoisomerization, and so on.^20–22 In this work, the oxidation was used to switch the NLO properties.

Polyoxometalates (POMs) are a unique class of well-defined inorganic nanocluster compounds. POMs exhibit remarkable chemical and physical properties, which have been used in a variety of fields, such as medicine, catalysis, biology, analytical chemistry, materials science and so on.^23–26 Over the past years, the POM-based organic–inorganic hybrid materials containing organic or organometallic moieties have especially received much attention, which exhibit remarkably large NLO response.^27–29 POMs enable the formation of hybrid materials in which the delocalized electrons coexist in both the organic network and the inorganic clusters. Our groups have studied the geometric structures and the first hyperpolarizabilities of these kinds of hybrid materials by density functional theory (DFT).^30,31

The modification of metal centers in organometallic compounds by changing the electronic and steric properties is an ongoing theme, which is important for controlling the reactivity of organometallic complexes. In previous report, a series of metallosalen compounds attached to a lacunary Keggin-type POM through an alkylene bridging spacer were synthesized by Neumann and co-workers.³² In these metallosalen–polyoxometalate (M-salen–POM) compounds, the POM exerts a significant intramolecular electronic effect on the metallosalen moiety leading to the formation of an oxidized metallosalen moiety. Meanwhile, the charge transfer between metallosalen donor and POM acceptor has been observed. Accompanying with the oxidization of metallosalen in these M-salen–POM compounds, many interesting phenomena might be discovered.

In this paper, three pairs of M-salen–POM (M = Pd, Ni and Co) compounds, including their one-electron oxidized states, were chosen to study the electronic and NLO properties. Herein, the M-salen–POMs, {SiW₁₁O₃₉[O(SiR)₂M]}⁴⁻ (R = (CH₂)₃N–CH(2-OPh); M = Pd^II, Ni^II and Co^II) were named as Pd-salen–POM, Ni-salen–POM and Co-salen–POM, respectively, and the corresponding one-electron oxidized species (M = Pd^III, Ni^III and Co^III) were named as oxo-Pd-salen–POM, oxo-Ni-salen–POM and oxo-Co-salen–POM. The quantum chemical investigations on the structures, molecular orbitals, and first hyperpolarizabilities of the M-salen–POM compounds were performed to understand the effect of one-electron oxidation on NLO response. The results show that the oxidation of M-salen–POM could effectively reduce the transition energy and enhance the first hyperpolarizability. In other word, the NLO properties can be switched by one-electron oxidation in these compounds. It is expected that these M-salen–POM materials possessing switchable NLO properties could enrich NLO materials.

Computational details

The density functional theory (DFT) with Becke's three parameter exchange functional combined with the Lee–Yang–Parr correlation functional (B3LYP)^33–35 has been widely used to optimize the molecular geometries. In fact, the B3LYP functional can well describe the geometric structures of POMs-based inorganic–organic hybrid compounds. The geometric parameters obtained by the B3LYP functional are in good agreement with the experimental values.^36–38 Herein, the geometries of all compounds were optimized and characterized as energy minima by B3LYP functional. The solvent effect of N,N-dimethylformamide (DMF) was considered by the polarizable continuum model (PCM).³⁹ The LANL2DZ basis set with Los Alamos relativistic effective core potentials (ECPs)⁴⁰ was employed for metal atoms (W, Co, Ni, Pd), while the 6-31G* basis set was used for C, H, O, N and Si elements.

The static first hyperpolarizability is termed as the zero frequency first hyperpolarizability, which is an estimate of the intrinsic molecular first hyperpolarizability in the absence of resonance effect. The first hyperpolarizability, β_tot, for all compounds were calculated by using the following equation (eqn (1)):⁴¹

here, β_i is defined by (eqn (2)):

The functionals for evaluating the first hyperpolarizabilities are usually general gradient approximation (GGA) in recent years. However, the Hartree–Fock (HF) exchange function and the long-range function must be included to obtain more reasonable results. Thus, several types of density functionals were selected to examine the reliability of the calculated results. The meta-GGA M06-2X functional⁴² includes a high percentage of HF exchange has been proved to be a good method to calculate the first hyperpolarizabilities.^21,43 Also, the CAM-B3LYP⁴⁴ and LC-BLYP⁴⁵ including long range corrections were employed because these DFT functionals are excellent for computing the first hyperpolarizabilities of transition metal complexes.^{21,38,43,46,47}

Moreover, in order to obtain a more intuitive description on NLO behaviors of compounds, time dependent (TD) DFT method was applied to determine the excitation energies of excited states at the PBE0 (ref. 48)/6-31G** level (LANL2DZ basis set for W, Co, Ni and Pd atoms) due to its efficiency and accuracy,^37,49,50 and PCM was employed in TDDFT calculations. Furthermore, in order to better describe the connection of the first hyperpolarizabilities and the electronic transition of excited states, the PBE0 functional were selected to calculate the first hyperpolarizabilities. The open-shell calculations were performed using unrestricted methods. All calculations in this work were carried out by using the GAUSSIAN 09 program package.⁵¹

Results and discussion

Geometric and electronic structures

In this paper, the electronic configuration of ground state of systems Pd-salen–POM, oxo-Pd-salen–POM, Ni-salen–POM, oxo-Ni-salen–POM, Co-salen–POM and oxo-Co-salen–POM are singlet, doublet, singlet, doublet, quartet and triplet, respectively. The energies on different electronic configuration for the systems are listed in Table S1.† The geometrical structures of M-salen–POM compounds are similar and are sketched in Fig. 1. The comparison on structural parameters between computed results and experimental ones cannot be carried out as crystal structural parameters are not available. Furthermore, to analyze the effect of oxidation processes on the geometric structure, the selected bond lengths for three pairs of compounds obtained at B3LYP/6-31G* (LANL2DZ basis set for W, Co, Ni and Pd atoms) level are shown in Table 1. It can be seen that the differences on the calculated bond lengths between M-salen–POM compounds and their one-electron oxidized species mainly concentrate on the M-salen part. It suggests the oxidation may take place on M-salen. There is no doubt that the redox property of compound is closely relative to the nature of the frontier molecular orbitals. And the highest occupied molecular orbital (HOMO) reflects the nature of the oxidation of compounds. So the HOMOs of Pd-salen–POM, Ni-salen–POM and Co-salen–POM are analyzed. As shown in Fig. S1,† the HOMOs of studied M-salen–POMs are similar and entirely locate on the M-salen unit. It proposes that the M-salen unit will be the oxidized center in one-electron oxidized process. Combined with the prediction from the geometrical structures, it is reasonable that the oxidation process occurs on the M-salen part.

Fig. 1 The structure of calculation model.

Table 1 The selected bond lengths (Å) of compounds

Compounds	Pd-salen–POM	Oxo-Pd-salen–POM	Ni-salen–POM	Oxo-Ni-salen–POM	Co-salen–POM	Oxo-Co-salen–POM
M–O	2.031	2.025	1.874	1.878	1.909	1.841
M–N	2.118	2.102	1.991	1.988	2.036	1.987
O–C1	1.308	1.293	1.307	1.296	1.304	1.293
N–C2	1.296	1.297	1.299	1.297	1.299	1.311

---

In addition, the Fukui function has been considered as a common descriptor of reaction activity in the corresponding system.⁵² Fukui function f_(r) is defined as

where ρ_(r) is the total electron density of the molecule. Thus, the condensed Fukui function of the atom A in a molecule with N electrons can be defined as

f_A⁺ = q_A^N+1 − q_A^N (for nucleophilic attack)

f_A⁻ = q_A^N − q_A^N−1 (for electrophilic attack)

where q_A is the electronic population of atom A in a molecule.

In order to further verify the oxidized center by an intuitive expression, the Fukui function was visualized with Multiwfn3.2.1 (ref. 53) and shown in Fig. 2. The oxidation process is discussed in this work, so the f⁻ was analyzed here. The result is well agreement with the experimental phenomenon³² that the M-salen is the oxidized center in M-salen–POM (M = Pd, Ni and Co). Accompanying with the oxidized process, the electronic structure might be affected, which further affects the static first hyperpolarizabilities. The influences of oxidized process on the NLO response of compounds are investigated in the following sections.

Fig. 2 The isosurface of Pd-salen–POM, Ni-salen–POM and Co-salen–POM.

Switchable second-order NLO properties

The static first hyperpolarizability (β_tot) have been calculated by M06-2X, CAM-B3LYP, LC-BLYP and PBE0 functionals with the 6-31G** basis set (LANL2DZ basis set for metal atoms) in this work and the β_tot values are shown in Table 2. The β_tot values reveal a dependence on the functional. For example, the β_tot value of oxo-Pd-salen–POM by LC-BLYP is twenty-one and fifteen percent of those by M06-2X and CAM-B3LYP, respectively. And the β_tot value by PBE0 is larger than those of other three functionals. It seems that the LC-BLYP functional underestimates and the PBE0 overestimates the first hyperpolarizabilities for studied oxidized compounds in this work. Although the β_tot values by the hybrid functionals M06-2X and PBE0 and the long range correction functionals CAM-B3LYP and LC-BLYP are different, the four functionals yield the same trend on β_tot values. As can be seen in Table 2, the β_tot values of Pd-salen–POM, Ni-salen–POM and Co-salen–POM are 3.59 × 10⁻³⁰, 3.82 × 10⁻³⁰ and 4.62 × 10⁻³⁰ esu by M06-2X functional, and the difference on β_tot value among three compounds is subtle. But the β_tot values of their one-electron oxidized species are obviously increased. In order to investigate the influence of one-electron oxidation on the switching behavior of first hyperpolarizability, the β_tot values of M-salen–POM and their corresponding one-electron oxidized species are compared. From the results in Table 2, the β_tot value of oxo-Pd-salen–POM is 377.27 × 10⁻³⁰ esu by the CAM-B3LYP functional, which is 98 times larger than that of Pd-salen–POM (3.85 × 10⁻³⁰ esu). The other two pairs of compounds also show good switching properties on second-order NLO coefficients, and the multiplying factor is about 35 (β_{(oxo-Ni-salen–POM)}/β_{(Ni-salen–POM)}) and 218 (β_{(oxo-Co-salen–POM)}/β_{(Co-salen–POM)}) times by CAM-B3LYP functional, respectively. It proposes that the oxidized process tremendously enhances the static first hyperpolarizability. So these kinds of compounds might be promising candidates for switchable NLO materials.

Table 2 The static first hyperpolarizability β_tot (10⁻³⁰ esu) for all compounds obtained by M06-2X, CAM-B3LYP, LC-BLYP and PBE0 functionals

Compounds	M06-2X	CAM-B3LYP	LC-BLYP	PBE0
Pd-salen–POM	3.59	3.85	5.84	8.12
Oxo-Pd-salen–POM	269.42	377.27	57.13	2293.41
Ni-salen–POM	3.82	4.02	5.73	8.98
Oxo-Ni-salen–POM	106.21	140.36	25.82	581.76
Co-salen–POM	4.62	10.10	7.63	6.30
Oxo-Co-salen–POM	535.64	2178.81	752.80	8495.04

---

TDDFT calculations

According to an approximately two-level expression proposed by Oudar,⁵⁴ the β_tot value is proportional to the oscillator strength (f) and the difference of dipole moment between ground state and crucial excited state (Δμ) and inversely proportional to the third power of transition energy (ΔE³). To provide the understanding about the oxidation effect on the first hyperpolarizabilities of M-salen–POM, we focus on the crucial excited states of these compounds. The crucial excited state is the lowest excited state with relatively large oscillator strength among all excited states. In this work, the ΔE, f, Δμ and λ_max (maximum absorption wavelengths of the UV-vis spectra) were computed by PBE0 functional, and the results are listed in Tables S2† and 3. Firstly, the calculated λ_max and the experimental one (Table S2†) are compared, and the difference is less than 30 nm, confirming that the PBE0 functional is exact for simulating the absorption spectra of these compounds. Furthermore, the ΔE value of oxo-Pd-salen–POM is 1.539 (2.655) eV, which is smaller than that of Pd-salen–POM (3.518 eV). Similarly, the ΔE values of oxo-Ni-salen–POM (2.010 eV) and oxo-Co-salen–POM (1.603 eV) are also smaller than those of Ni-salen–POM (3.641 eV) and Co-salen–POM (3.431 eV), respectively. The results show that the oxidization greatly reduce the ΔE value. Obviously, the order of ΔE³ is inversely proportional to that of the β_tot values from the two-level expression. The Δμf/ΔE³ values of oxidized species are larger than their non-oxidized species, which result in the larger β_tot values for oxidized species. The result show that the order of Δμf/ΔE³ values is in agreement with β_tot values calculated with the PBE0 functional that the β_tot values of oxidized species are larger than their non-oxidized species. Significantly, the transition energy is the decisive factor on determining the β_tot values of these compounds.

Table 3 The transition energy ΔE (eV), the oscillator strength f, and the difference of dipole moment between the ground state and crucial excited state (Δμ) obtained by PBE0 functional

Compounds	Pd-salen–POM	Ni-salen–POM	Co-salen–POM	Oxo-Pd-salen–POM	Oxo-Ni-salen–POM	Oxo-Co-salen–POM
Excited state	S24	S33	S60	S4/S19	S15	S10
ΔE (eV)	3.518	3.641	3.431	1.539/2.655	2.010	1.603
f	0.148	0.200	0.027	0.027/0.035	0.049	0.106
Δμ	1.311	1.498	0.570	0.847/0.735	1.005	1.646
Δμf/ΔE^3 (10^−3)	4.456	6.207	0.381	6.274/1.375	6.188	42.358

---

Furthermore, we focused on analyzing the molecular orbitals of the crucial excited states for the three pairs of compounds to understand the origin of β_tot values. As shown in Fig. 3 and Table S2,† the crucial excited states of Pd-salen–POM are mainly made up of H(HOMO) → L(LUMO)+21 (34%), H−1 → L+19 (31%) and H → L+7 (8%) electron transitions. The occupied molecular orbitals (H and H−1) entirely localize on the M-salen (the p orbitals of C, O and N atoms and the d orbitals of Pd atoms), and the unoccupied molecular orbitals (L+7, L+19 and L+21) mainly localize on the POM (the d orbitals of W atoms and the p orbitals of O atoms) and M-salen. Therefore, these electron transitions generate the mixing charge transfer from M-salen to POM and the intramolecular π → π* transition within M-salen, which is classified as type I charge transfer. The crucial excited states of the one-electron oxidized species oxo-Pd-salen–POM contain the βH−38 → βL (41%), αH−1 → αL+4 (12%) and αH−39 → αL+4 (8%) transitions in excited state S19 and βH−1 → βL (98%) transitions in excited state S3. Contrast to the Pd-salen–POM, the occupied molecular orbitals of oxo-Pd-salen–POM mainly localize on the M-salen and POM, and the unoccupied molecular orbitals entirely localize on M-salen, which is assigned the charge transfer from POM to the M-salen and the intramolecular charge transfer (π → π*) in M-salen, which is defined as type II charge transfer. It clearly shows that POM acts as the electron acceptor in type I charge transfer (Pd-salen–POM), while it acts as electron donor in type II charge transfer (oxo-Pd-salen–POM). As mentioned above, the β_tot value of oxo-Pd-salen–POM is much larger than that of Pd-salen–POM. So the mixing transitions for the type II charge transfer might be helpful to enhance the first hyperpolarizability of oxo-Pd-salen–POM. Furthermore, the molecular orbitals of the crucial excited states for other two pairs of compounds are list in Fig. S2 and S3.† The results show that the charge transfers in Ni-salen–POM and Co-salen–POM are similar to that of Pd-salen–POM, which exhibits the type I charge transfer. And the charge transfers in oxo-Ni-salen–POM and oxo-Ni-salen–POM resemble to that of oxo-Pd-salen–POM, which belongs to the type II charge transfer.

Fig. 3 Molecular orbitals of Pd-salen–POM and oxo-Pd-salen–POM involved in the crucial excited states.

Conclusions

In summary, three pairs of compounds were systematically investigated with DFT and TDDFT methods to explore the one-electron oxidation effect on the static first hyperpolarizability. Fukui function analysis shows that the M-salen is the oxidized center in M-salen–POM compounds. The one-electron oxidized process obviously enhances the static first hyperpolarizability. The TDDFT calculations exhibit that the type II charge transfer in one-electron oxidized species that POM acts as electron donor might play a key role for effectively reducing the transition energy and improving the static first hyperpolarizability. The β_tot value of oxo-Pd-salen–POM is 377.27 × 10⁻³⁰ esu, which is 98 times larger than that of Pd-salen–POM by CAM-B3LYP functional. So this kind of M-salen–POM compounds might be excellent switchable NLO material. The analysis of the effect of one-electron oxidation on their optical responses might provide important information of these compounds for switchable NLO applications.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No. 21571031 and 21131001).

Notes and references

1. J. Zyss, Molecular Nonlinear Optics: Materials, Physics and Devices, Academic Press, Boston, 1994.
2. H. S. Nalwa and S. Miyata, Nonlinear Optics of Organic Molecules and Polymers, CRC Press, Boca Raton, Florida, 1997.
3. S. Di Bella, I. Fragalà, I. Ledoux, M. A. Diaz-Garcia and T. J. Marks, J. Am. Chem. Soc., 1997, 119, 9550.
4. J. P. Costes, J. F. Lamère, C. Lepetit, P. G. Lacroix and F. Dahan, Inorg. Chem., 2005, 44, 1973.
5. E. Cariati, M. Pizzotti, D. Roberto, F. Tessore and R. Ugo, Coord. Chem. Rev., 2006, 250, 1210.
6. R. L. Zhong, J. Zhang, S. Muhammad, Y. Y. Hu, H. L. Xu and Z. M. Su, Chem.–Eur. J., 2011, 17, 11773.
7. H. L. Xu, Z. R. Li, D. Wu, B. Q. Wang, Y. Li, F. L. Gu and Y. Aoki, J. Am. Chem. Soc., 2007, 129, 2967.
8. R. L. Zhong, H. L. Xu, S. Muhammad, J. Zhang and Z. M. Su, J. Mater. Chem., 2012, 22, 2196.
9. M. Irie, Chem. Rev., 2000, 100, 1683.
10. V. De Waele, U. Schmidhammer, T. Mrozek, J. Daub and E. Riedle, J. Am. Chem. Soc., 2002, 124, 2438.
11. A. P. de Silva, H. Q. Nimal Gunaratne, T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy, J. T. Rademacher and T. E. Rice, Chem. Rev., 1997, 97, 1515.
12. S. J. Yoon, J. W. Chung, J. Gierschner, K. S. Kim, M. G. Choi, D. Kim and S. Y. Park, J. Am. Chem. Soc., 2010, 132, 13675.
13. K. A. Green, M. P. Cifuentes, M. Samoc and M. G. Humphrey, Coord. Chem. Rev., 2011, 255, 2530.
14. I. Asselberghs, K. Clays, A. Persoons, M. D. Ward and J. Mc Cleverty, Mater. Chem., 2004, 14, 2831.
15. P. Gutlich, Y. Garcia and T. Woike, Coord. Chem. Rev., 2001, 219, 839.
16. Z. P. Ni, X. M. Ren, J. Ma, J. L. Xie, C. L. Ni, Z. D. Chen and Q. J. Meng, J. Am. Chem. Soc., 2005, 127, 14330.
17. M. Z. Zhu, C. M. Aikens, M. P. Hendrich, R. Gupta, H. F. Qian, G. C. Schatz and R. C. Jin, J. Am. Chem. Soc., 2009, 131, 2490.
18. B. J. Coe, S. Houbrechts, I. Asselberghs and A. Persoons, Angew. Chem., Int. Ed., 1999, 38, 366.
19. B. J. Coe, Acc. Chem. Res., 2006, 39, 383.
20. N. N. Ma, S. L. Sun, C. G. Liu, X. X. Sun and Y. Q. Qiu, J. Phys. Chem. A, 2011, 115, 13564.
21. T. Y. Ma, N. N. Ma, L. K. Yan, T. Zhang and Z. M. Su, J. Phys. Chem. A, 2013, 117, 10783.
22. V. Aubert, V. Guerchais, E. Ishow, K. Hoang-Thi, I. Ledoux, K. Nakatani and H. Le Bozec, Angew. Chem., Int. Ed., 2008, 47, 577.
23. M. T. Pope and A. Müüller, Angew. Chem., Int. Ed. Engl., 1991, 30, 34.
24. L. P. Kazansky and B. R. Mc Garvey, Coord. Chem. Rev., 1999, 188, 157.
25. N. Mizuno, K. Yamaguchi and K. Kamata, Coord. Chem. Rev., 2005, 249, 1944.
26. D. L. Long, E. Burkholder and L. Cronin, Chem. Soc. Rev., 2007, 36, 105.
27. P. Gouzerh and A. Proust, Chem. Rev., 1998, 98, 77.
28. A. Proust, R. Thouvenot and P. Gouzerh, Chem. Commun., 2008, 1837.
29. Y. G. Wei, B. B. Xu, C. L. Barnes and Z. H. Peng, J. Am. Chem. Soc., 2001, 123, 4083.
30. W. Guan, G. C. Yang, L. K. Yan and Z. M. Su, Inorg. Chem., 2006, 45, 7864.
31. G. C. Yang, W. Guan, L. K. Yan, Z. M. Su, L. Xu and E. B. Wang, J. Phys. Chem. B, 2006, 110, 23092.
32. I. Bar-Nahum, H. Cohen and R. Neumann, Inorg. Chem., 2003, 42, 3677.
33. A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
34. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785.
35. P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frishch, J. Phys. Chem., 1994, 98, 11623.
36. T. Zhang, W. Guan, L. K. Yan, T. Y. Ma, J. Wang and Z. M. Su, Phys. Chem. Chem. Phys., 2015, 17, 5459.
37. T. Zhang, W. Guan, S. Z. Wen, T. Y. Ma, L. K. Yan and Z. M. Su, J. Phys. Chem. C, 2014, 118, 29623.
38. T. Zhang, N. N. Ma, L. K. Yan, T. Y. Ma and Z. M. Su, Dyes Pigm., 2014, 106, 105.
39. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999.
40. P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 270.
41. D. R. Kanis, M. A. Ratner and T. J. Marks, Chem. Rev., 1994, 94, 195.
42. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215.
43. N. N. Ma, S. J. Li, L. K. Yan, Y. Q. Qiu and Z. M. Su, Dalton Trans., 2014, 43, 5069.
44. T. Yanai, D. Tew and N. Handy, Chem. Phys. Lett., 2004, 393, 51.
45. H. Iikura, T. Tsuneda, T. Yanai and K. Hirao, J. Chem. Phys., 2001, 115, 3540.
46. N. N. Ma, C. G. Liu, Y. Q. Qiu, S. L. Sun and Z. M. Su, J. Comput. Chem., 2012, 33, 211.
47. T. Zhang, W. Guan, T. Y. Ma, L. Yan, L. K. Yan and Z. M. Su, Inorg. Chem. Front., 2015, 2, 544.
48. C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158.
49. D. Ravelli, D. Dondi, M. Fagnoni, A. Albini and A. Bagno, Phys. Chem. Chem. Phys., 2013, 15, 2890.
50. D. Ravelli, D. Dondi, M. Fagnoni, A. Albini and A. Bagno, J. Comput. Chem., 2011, 32, 2983.
51. 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 09W, revision A.02, Gaussian, Inc, Wallingford, CT, 2009.
52. D. D. Qi, L. J. Zhang, L. Wan, Y. X. Zhang, Y. Z. Bian and J. Z. Jiang, Phys. Chem. Chem. Phys., 2011, 13, 13277.
53. T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580.
54. J. L. Oudar and D. S. Chemla, J. Chem. Phys., 1977, 66, 2664.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00146g

---