Theoretical and experimental investigations on mono-substituted and multi-substituted functional polyhedral oligomeric silsesquioxanes

Abstract

The structure and electronic properties of polyhedral oligomeric silsesquioxane (POSS) cages functionalized with mono-substituted and multi-substituted vinyl and mercaptopropyl have been studied using density functional theory calculations. Electron density distributions and frontier molecular orbitals (FMOs) have been constructed at the B3LYP/6-31+G(d) level to understand the electronic properties. ¹H, ¹³C, and ²⁹Si NMR spectra by using the gauge including atomic orbital (GIAO) method of the studied compounds were compared with experimental data obtained.

---

Introduction

Polyhedral oligomeric silsesquioxanes (POSS), organosilicates with an empirical formula [RSiO_3/2]_n, have attracted significant interest in the past few years because of their well-defined nanostructure.¹ Numerous POSS derivatives have been used as architectures for precisely defined nanostructured functional materials with improved thermal, mechanical, and optoelectronic properties.^2–8 The hydrogen atoms of POSS-based materials can be easily substituted with functional groups such as alcohols, phenols, chlorosilanes, epoxides, esters, and so forth.^9–14 Thus, there are wide variety of applications for POSS-based materials. The cubic structure of POSS has a nanosized inorganic core containing silica and eight organic functional groups on the silica surface. Through the manipulation of the functional groups on the POSS cage, as well as the size and shape of the POSS cage itself, it is possible to improve the solubility of POSS monomers in water, organic solvents, and/or polymeric materials.¹⁵

Molecular simulation appears to be a desired instrument for improving our understanding of POSS-based hybrid materials due to the difficulty and complexity of performing experiments at the nanoscale. So far there is a very large amount of theoretical studies on POSS-based hybrid materials or complexes, although these investigations have emerged much more slowly. The majority of theoretical studies have been devoted to pure cages,^3,16–21 endohedral/exohedral complexes of POSS cages,^22–28 and reaction mechanism.^29–31 The density functional theory (DFT) method provides a precise, and computationally economical way of modeling electron correlation and is now extensively used in molecular electronic structure studies, particularly in large molecular systems where traditional ab initio methods are obviously time-consuming calculation. However, the structure–activity relationship research of functional POSS, especially containing alkenes and mercapto functional POSS compounds, has received much less attention. These compounds can be used in the synthesis of versatile POSS-based functional materials through facile and high-efficiency thiol–ene click chemistry, which is a facile, high-efficiency and stereospecific method compared with the traditional hydrosilylation reaction.

Based on the above, our interest in this work is to use both experimental and DFT methods to study the spectroscopic properties and electronic structures of mono-substituted and multi-substituted vinyl and mercaptopropyl functional polyhedral oligomeric silsesquioxanes ((C₂H₃)(CH₃)₇Si₈O₁₂, (C₃H₆SH)(CH₃)₇Si₈O₁₂, (C₂H₃)₈Si₈O₁₂, and (C₃H₆SH)₈Si₈O₁₂, shown in Fig. S1†). The computational modeling of the NMR parameter is also of abiding interest, and such calculation at DFT has emerged as a promising approach for the prediction of nuclear shielding and coupling constants of NMR active nuclei.^32–36 Thus, we have computed the proton, carbon and silicon NMR chemical shifts using the gauge-independent atomic orbital (GIAO)-DFT method,^37–39 which is aimed at providing the definitive characterization of the title compounds. It is well-known that the GIAO method can yield very accurate result.^32,39,40 It is somewhat superior as compared to the individual gauge for localized orbitals (IGLO) method and similar methods. And the GIAO method is easier to implement than the IGLO and related methods.

Experimental section

Materials

3-Mercaptopropyl trimethoxysilane (MPT) and vinyl-trimethoxysilane were purchased from J&K Chemical (China). Concentrated hydrochloric acid 37% (AR), ethyl ether, acetone, dichloromethane were purchased from Sinopharm Chemical Reagent Beijing Co., Ltd, China. All the chemicals were of analytical reagent grade and used without further purification.

Physical measurements

FT-IR spectra are obtained by a Nicolet-6700 FT-IR spectrophotometer (Thermo Instruments Inc., USA) within a range of 400–4000 cm⁻¹. ¹H NMR spectra are recorded on an ARX-6 NMR spectrometer (Bruker) at 300 MHz with CDCl₃ as deuterated solvent, without internal reference. ¹³C NMR and ²⁹Si NMR spectra are recorded on AV II NMR spectrometer (Bruker) at 400 MHz with CDCl₃ as deuterated solvent without the internal reference tetramethylsilane.

Preparation of octamercaptopropyl-POSS ((C₃H₆SH)₈Si₈O₁₂)

POSS-SH is prepared as shortly described in the following:⁴¹ a solution of MPT (10 mL) and concentrated HCl (37%, 20 mL) in MeOH (240 mL) were placed together in a flask (500 mL) equipped with a magnetic stirrer and a nitrogen catheter. The hydrolysis and condensation reactions were carried out at 90 °C for 36 h. The crude product was washed with cold MeOH three times to remove excess MPT. The resulting viscous solution was dissolved in CH₂Cl₂ and then washed three times with H₂O. The CH₂Cl₂ phase was dried with anhydrous Na₂SO₄ and concentrated to obtain POSS-SH in 76% yield. IR (KBr; ν, cm⁻¹): 2929 (m), 2555 (w), 2849 (m), 1401 (m), 1122 (vs), 1032 (s), 562 (m) 472 (m). ¹H NMR (300 MHz, CDCl₃; δ, ppm): 0.76 (double, Si–CH₂–), 1.71 (s, –CH₂–), 2.56 (s, –CH₂–S), 1.37 (br, –SH). ¹³C NMR (300 MHz, CDCl₃; δ, ppm): 13.46 (Si–CH₂–), 29.87 (Si–CH₂–CH₂–), 30.16 (–CH₂–SH). ²⁹Si NMR (400 MHz, CDCl₃; δ, ppm): −67.20.

Preparation of octavinyl-POSS ((C₂H₃)₈Si₈O₁₂)

50 mL of acetone and 5 mL of vinyltrimethoxysilane were charged into a 100 mL flask, which was equipped with a magnetic stirrer. The mixture of 10 mL of concentrated HCl and 10 mL of deionized water was added dropwise into the reaction mixture, stirred and refluxed at 40 °C for 5 days. White solid was deposited on the wall of the flask, meanwhile the reaction mixture turned brown. The white powder can be obtained from centrifugal separation, washing with ethanol and drying at 60 °C in vacuum drying oven. The total crude product was recrystallized from the mixed solvents of dichloromethane and acetone (volume ratio 1 : 3) to afford POSS-Vi with 40% of yield. IR (KBr; ν, cm⁻¹): 3058 (m), 3019 (m), 1604 (m), 1409 (m), 1276 (m), 1114 (vs), 778 (m), 584 (m), 465 (m). ¹HNMR (300 MHz, CDCl₃; δ, ppm): 5.86–6.14 (m, H₂C=CH–). ¹³C NMR (300 MHz, CDCl₃; δ, ppm): 126.93 (–CH=CH₂), 135.13 (–CH=CH₂). ²⁹Si NMR (400 MHz, CDCl₃; δ, ppm): −80.20.

Computational details

All the theoretical calculations reported here were performed using Gaussian 03 package.⁴² Becke's three-parameter hybrid method⁴³ and the correlation functional by Lee, Yang, and Parr⁴⁴ (B3LYP) were used with the 6-31+G(d) basis sets. Stationary structures were obtained by verifying that all the harmonic frequencies are real. Default options were employed for all optimizations. No symmetry constraints were imposed during the optimizations. An empirical uniform scaling factor of 0.983 for lower than 1700 cm⁻¹ and 0.958 for greater than 1700 cm⁻¹ (ref. 45 and 46) have been used to counterbalance the systematic errors caused by basis set incompleteness and vibrational anharmonicity. The proton, carbon and silicon NMR chemical shifts were calculated with gauge including atomic orbital (GIAO) approach³⁷ by applying B3LYP/6-311+G(d,p) method of the title molecule and compared with the experimental NMR spectra. To investigate the effect of solvent on the chemical shifts, solvation effects were modeled with the polarisable continuum model (PCM),^47,48 using chloroform (CH₃Cl) as the solvent.

Results and discussion

Structural analysis

In order to detect the reliability of the current theoretical method, octahydrosilsesquioxane (H₈Si₈O₁₂) and the higher methyl-substituted octahydrosilsesquioxane ((Me)₈Si₈O₁₂) were studied in tentative. Corresponding geometrical parameters from the current calculation, X-ray crystal structure, and previous literature are listed in Table S1.† The optimized bond lengths and angles agree well with the experimental data and calculation results by other methods, which indicate the current theoretical calculation is reliable.

Geometries are calculated at B3LYP/6-31+G(d) level by Gaussian 03 package, the four items (maximum force, RMS force, maximum displacement, and RMS displacement) are all converged. The obtained optimized geometry of mono-substituted and multi-substituted vinyl and mercaptopropyl functional polyhedral oligomeric silsesquioxanes (POSS) are shown in Fig. S1.† And corresponding vibrational frequency calculations indicate all the optimized structures appear to be true minima. The optimized structures of (C₂H₃)₈Si₈O₁₂, and (C₃H₆SH)₈Si₈O₁₂ have low symmetry, because of the various orientations of the eight vinyl and mercaptopropyl that are attached at the eight corner Si atoms of the Si₈O₁₂ cage. A comparison between the calculated geometrical parameters of POSS derivatives are provided in Table 1. Differences of bond lengths of Si–O between H₈Si₈O₁₂ and its derivatives are all less than 0.3%. For Si–C bonds, the difference is negligible. For the Si–O–Si bond angle, the difference is less than 1.4° in H₈Si₈O₁₂, and the corresponding value is up to 13.7° in (C₂H₃)₈Si₈O₁₂. While the difference between H₈Si₈O₁₂ and its derivatives are all less than 6.2%, and those for the O–Si–O or O–Si–C angles are negligibly small.

Table 1 The selected structural parameters of POSS derivatives

	Bond length (Å)				Angle (°)
	RSi–O	RSi–C(alkyl)	RSi–C(alkenyl)	RSi–Si(diagonal)	∠Si–O–Si	∠O–Si–O	∠O–Si–C
H8Si8O12	1.643	—	—	5.472	147.5–148.9	109.5–109.7	—
(C2H3)(CH3)7Si8O12	1.648	1.861	1.854	5.498	146.6–150.4	109.2–109.8	109.6
(C3H6SH)(CH3)7Si8O12	1.647	1.860	—	5.492	148.1–149.3	109.1–109.6	109.6
(C2H3)8Si8O12	1.648	1.861	1.853	5.484	142.9–156.6	108.4–109.9	110.1
(C3H6SH)8Si8O12	1.648	1.862	—	5.500	147.2–150.6	109.0–109.4	110.0

---

The special physical and chemical properties of POSS are mainly attributed to the cage-like silica core, which is commonly found to be quiet rigid. The stability of the rigid structure is exposed by comparing the distance between two silicon atoms along body diagonal in the POSS derivatives. In H₈Si₈O₁₂, the distance is 5.472 Å, which is good agreement with previous literatures. In its derivatives, the distance between the diagonally opposed silicon atoms is always larger than that in H₈Si₈O₁₂, but the differences are less than 0.51% (shown in Table 2). The current calculated results agree well with the corresponding values of the 4-cyanophenyl and 4-carbazolephenyl substituted POSS derivatives.³ The rigid structure of POSS provides a means for spatially separating conjugated organic fragments, which is useful for the realization of novel organic molecular architectures for various applications.

Table 2 Distance of diagonal silicon atoms of the cage-like silica core

	H8Si8O12	(C2H3)(CH3)7Si8O12	(C3H6SH)(CH3)7Si8O12	(C2H3)8Si8O12	(C3H6SH)8Si8O12
DSi–Si(diagonal) (Å)	5.472	5.498	5.492	5.484	5.500
Deformation (Å)		0.026	0.020	0.012	0.028
Deformation (%)		0.48	0.37	0.22	0.51

---

The infrared (IR) spectra

The calculated fundamental vibration frequencies for H₈Si₈O₁₂ and (Me)₈Si₈O₁₂ are in excellent agreement with the published experimental IR spectrum⁴⁹ and previous theoretical method,⁵⁰ as listed in Table S2.† The largest error in the theoretical frequency is 3% (44 cm⁻¹) for the Si–C stretching mode, and the corresponding error for the important Si–H and Si–O–Si stretching modes is less than 0.3% (3 cm⁻¹). The results further indicated the current theoretical calculation is reliable.

Vibrational frequencies of mono-substituted and multi-substituted vinyl and mercaptopropyl functional polyhedral oligomeric silsesquioxanes were calculated by B3LYP/6-31+G(d) method, and the vibrational frequencies and experimental IR spectrum are shown in Table 3. Obviously, the theoretical frequencies obtained by B3LYP method are in good agreement with those obtained by experiment. A comparison between theoretical results and experimental data are shown in Fig. 1.

Table 3 Vibration frequencies of POSS derivatives calculated and experimentally measured (in parentheses) in cm⁻¹

	(C2H3)(CH3)7Si8O12	(C2H3)8Si8O12	(C3H6SH)(CH3)7Si8O12	(C3H6SH)8Si8O12
Si–O–Si (ring-asym)	1115	1115 (1116)	1115	1115 (1116)
Si–C (stretch)	1314	1291 (1279)	1314	1291 (1279)
Si–C (wag)	768	780 (780)	771	791 (797)
O–Si–O (bend)	580	583 (586)	503	542 (563)
Si–O–Si (bend)	445	448 (463)	439	451 (468)
C=C (stretch)	1659	1659 (1609)	—	—
=C–H (stretch)	3088	3088 (3067)	—	—
–C–H (stretch)	2932	— (2961)	2994	2991 (2961)
S–H (stretch)	—	—	2568	2568 (2557)

---

Fig. 1 Comparison of experimental and theoretical infrared spectra of POSS derivatives.

Vinyl compounds commonly exhibit multiple weak bands above 3000 cm⁻¹ due to vinyl C–H stretching vibrations. In the present theoretical study, the FTIR bands at 3088 cm⁻¹ was assigned to vinyl C–H stretching vibration. The difference between (C₂H₃)(CH₃)₇Si₈O₁₂ and (C₂H₃)₈Si₈O₁₂, for vinyl C–H stretching vibration is only the intensity of vibration. And the similar results were obtained for C=C stretching vibration (at 1659 cm⁻¹), but the calculated result was larger than the data obtained by experimental measured (1609 cm⁻¹, 3%). For (C₃H₆SH)(CH₃)₇Si₈O₁₂ and (C₃H₆SH)₈Si₈O₁₂, a S–H stretching peak appears at 2557 cm⁻¹, while DFT calculation gives S–H stretching vibration at 2568 cm⁻¹ with different intensity of vibration. The calculated asymmetrical Si–O–Si stretching band is 1115 cm⁻¹ in the four compounds, which indicated that the asymmetrical Si–O–Si stretching mode is likely to be little affected by the nature of the substituent.⁵¹ In experiment, the C–S stretching vibration is not observed in the FT-IR spectra, but DFT calculation gives the C–S stretching vibrations at 727 cm⁻¹.

The correlation between the experimental and calculated wavenumbers was plotted correlation graphics and shown in Fig. 2. The relations between experimental and calculated wavenumbers are usually linear and described for total and one by one as infrared by the following equation:

ν_cal = 1.0152 × ν_exp − 15.5222 (R² = 0.9998)

Fig. 2 Correlation graphic of calculated and experimental infrared spectra.

The ¹H-NMR, ¹³C-NMR and ²⁹Si-NMR properties

The NMR spectroscopy is one of the most important techniques for the structural analysis of organic compounds. This technique (¹H, ¹³C, and ²⁹Si NMR) is vulnerable to exploitation due to the useful, common and to have more information about the studied compound. The calculated values of ¹H, ¹³C, and ²⁹Si chemical shifts by GIAO method in the gas phase and solution and experimental values, which are referred to as δ_exp and δ_cal, are summarized in Table 4. The calculated results obtained from gas phase and solution basically remain unchanged.

Table 4 Experimental δ_exp and calculated δ_cal ¹H, ¹³C, and ²⁹Si chemical shifts (ppm) of the title compounds by B3LYP method, the values considering the effect of solvent are shown in brackets

	(C2H3)(CH3)7Si8O12	(C2H3)8Si8O12		(C3H6SH)(CH3)7Si8O12	(C3H6SH)8Si8O12
	δcal	δcal	δexp	δcal	δcal	δexp
H(a)	6.13 (6.37)	6.15 (6.40)	5.90–6.15	—	—	—
H(b)	6.30 (6.42)	6.36 (6.48)		—	—	—
H(c)	0.11 (0.15)	—	—	0.13 (0.15)	—	—
H(d)	—	—	—	0.60 (0.70)	0.71 (0.76)	0.76
H(e)	—	—	—	1.63 (1.58)	1.65 (1.67)	1.71
H(f)	—	—	—	2.53 (2.64)	2.56 (2.62)	2.56
H(g)	—	—	—	1.82 (1.97)	1.87 (2.01)	1.37
C(a)	138.28 (138.70)	138.31 (138.02)	126.93	—	—	—
C(b)	141.98 (143.12)	142.64 (143.42)	135.13	—	—	—
C(c)	−2.83 (−2.81)	—	—	−2.93 (−2.88)	—	—
C(d)	—	—	—	18.40 (18.12)	18.18 (18.30)	13.46
C(e)	—	—	—	28.28 (29.25)	28.24 (28.97)	29.87
C(f)	—	—	—	37.32 (37.25)	36.84 (37.31)	30.16
Si(a)	−85.42 (−85.22)	−83.84 (−83.25)	−80.20	−71.32 (−71.16)	−70.52 (−70.48)	−67.04
Si(b)	−68.56 (−68.49)	—	—	−68.56 (−68.40)	—	—

---

There are two typological hydrogen atoms in (C₂H₃)₈Si₈O₁₂ molecule. The chemical shifts of olefinic protons of organic molecules are usually observed in the region 4.5–6.5 ppm. Without doubt, these chemical shifts change greatly with the electronic environment of the proton. Electron-withdrawing atom or group (hydrogen attached or nearby) can decrease the shielding and move the resonance of attached proton towards a higher frequency, while electron-donating atom or group increases the shielding and moves the resonance to a lower frequency.⁵² The chemical shifts of these protons are calculated at 6.15 and 6.36 ppm, and the corresponding values were observed at 5.90–6.15 ppm in CDCl₃ by experiment. In (C₃H₆SH)₈Si₈O₁₂ molecule, the calculated chemical shifts of methylene hydrogens are at 0.71, 1.65, and 2.56 ppm, and the corresponding experimental values are 0.76, 1.71, and 2.56 ppm. Thus, these results indicated that the calculated chemical shifts of protons in high magnetic field (methylene hydrogens) are closer to the experimental values, compared with those in low magnetic field (olefinic protons). Notably, the calculated chemical shift of sulfhydryl proton (1.87 ppm) is much higher than the experimental value (1.37 ppm) because it is active hydrogen.

The calculated and experimental ¹³C NMR chemical shifts of multi-substituted POSS ((C₂H₃)₈Si₈O₁₂ and (C₃H₆SH)₈Si₈O₁₂) were showed same parallel with reasonable range and good coherent with each other. The chemical shifts of the carbon atoms of (C₂H₃)₈Si₈O₁₂ are calculated at 138.31 and 142.64 ppm, and corresponding values of (C₃H₆SH)₈Si₈O₁₂ molecule are 18.18, 28.24, and 36.84 ppm. Both of them are slightly larger than the results obtained by experiment. Also, the calculated chemical shifts of the silicon atoms are slightly lower than the experimental values, the differences of chemical shifts are all less than 4.5% (3.64 ppm).

Fig. 3 gives the relationship between δ_exp and δ_cal of ¹H and ¹³C chemical shifts calculated by B3LYP/6-311+G(d,p) method. The function relationship for ¹H chemical shifts is δ_cal = 1.0588 × δ_exp − 0.1307, with the linear correlation coefficient R is 0.9998. This shows that the correlation between theory and experiment for the studied compounds is good. For ¹³C chemical shifts, the function relationship is δ_cal = 1.0544 × δ + 2.0833 (R = 0.9957). It indicated that the calculated values are also agree with the experimental values.

Fig. 3 The relationship between δ_cal and δ_exp for ¹H(a) and ¹³C(b) chemical shifts.

Electronic structure

The spatial arrangement of the frontier orbitals was determined for the ground state optimized geometries corresponding to the singlet spin state for POSS derivatives. The calculated HOMO–LUMO energies of those molecules are shown in Table 5.

Table 5 The calculated HOMO–LUMO energies of POSS molecules

	HOMO	LUMO	Egap
(CH3)8Si8O12	−7.75	0.82	8.57
(C2H3)(CH3)7Si8O12	−7.40	−0.11	7.29
(C2H3)8Si8O12	−7.31	−0.30	7.02
(C3H6SH)(CH3)7Si8O12	−6.28	0.71	6.99
(C3H6SH)8Si8O12	−6.34	0.30	6.64

---

The HOMO energy is −7.75 eV and LUMO energy is 0.82 eV for (CH₃)₈Si₈O₁₂. The HOMO and LUMO gap is 8.57 eV, which is larger than the value of substituted POSS. The energy gaps for the mono-substituted and multi-substituted vinyl and mercaptopropyl functional polyhedral oligomeric silsesquioxanes ((C₂H₃)(CH₃)₇Si₈O₁₂, (C₃H₆SH)(CH₃)₇Si₈O₁₂, (C₂H₃)₈Si₈O₁₂, and (C₃H₆SH)₈Si₈O₁₂) have a little change, but it presents a regular decrease. For (C₃H₆SH)₈Si₈O₁₂, the decrease of energy gap is up to 1.93 eV. A comparison of (C₂H₃)(CH₃)₇Si₈O₁₂ with (C₂H₃)₈Si₈O₁₂, adding seven more vinyl along the side chain, shows that the LUMO–HOMO gap decreases: that is, the wavelength is red-shifted. And mercaptopropyl functional polyhedral oligomeric silsesquioxanes have analogous results. This observation is similar to that from monosubstituted and higher phenyl-substituted octahydrosilsesquioxanes,⁵³ but the values of LUMO–HOMO gap (6.64 eV to 7.29 eV) are larger than those of phenyl-substituted octahydrosilsesquioxanes (4.13 eV to 4.80 eV). From Fig. 4, we can see that all these HOMO are localized in the substituted groups, and share a similar electron density surface. However, the LUMO of vinyl substituted POSS are localized in the partial substituted groups, and the LUMO of mercaptopropyl substituted POSS are mainly localized in the oxygen atoms of POSS skeleton. In the vinyl substituted POSS, the HOMO and LUMO are the π and π* orbitals of the vinyl double bond. While the HOMO is the lone pair of sulphur atom and the LUMO remaining essentially the same of the unsubstituted POSS in the mercapto substituted POSS. This slightly differs from the 4-cyanophenyl and 4-carbazolephenyl substituted POSS derivatives. In the 4-cyanophenyl and 4-carbazolephenyl substituted POSS, the HOMO and LUMO are the π and π* orbitals of 4-cyanophenyl and 4-carbazolephenyl groups similar to vinyl substituted POSS.³

Fig. 4 Molecular orbitals (top (HOMO), bottom (LUMO)) of the compounds. From left to right are (C₂H₃)(CH₃)₇Si₈O₁₂, (C₂H₃)₈Si₈O₁₂, (C₃H₆SH)(CH₃)₇Si₈O₁₂, and (C₃H₆SH)₈Si₈O₁₂, respectively.

Conclusions

In the present study, we have carried out the structures and spectroscopic analysis of mono-substituted and multi-substituted POSS, using FT-IR, and NMR techniques and tools derived from the DFT. In general, a good agreement between experimental and calculated results has been observed. The inorganic core of the POSS is quite rigid. The deformation of the POSS core upon functionalizing the corners of the cube with organic groups is very small. The rigidity of the POSS cube may help to prevent the aggregation of planar organic conjugated fragments, which is important in some applications such as organic electronics devices. The calculated results indicate that the energy gap (HOMO–LUMO) can be tuned through functionalization, which provides the flexibility to design molecules with desired properties and for specific applications. The POSS derivatives are conducive to the development of highly efficient emitters in OLEDs. Although the current investigation focuses on a specific POSS system, the calculation results on the designed molecular structures are expected to be applicable in a broader context for organic molecular materials and organic–inorganic hybrid structures that are useful for organic electronics or further functionalization reaction.

Acknowledgements

The authors wish to acknowledge the financial supports from Jiangsu Province Transformation of Scientific and Technological Achievements Program (BA 2014123), National major scientific instruments and equipment development projects (2014YQ060773), and the Fundamental Research Funds for the Central Universities (CXLX13_106).

References

1. R. H. Baney, M. Itoh, A. Sakakibara and T. Suzuki, Chem. Rev., 1995, 95, 1409–1430.
2. C. B. He, Y. Xiao, J. C. Huang, T. T. Lin, K. Y. Mya and X. H. Zhang, J. Am. Chem. Soc., 2004, 126, 7792–7793.
3. C. G. Zhen, U. Becker and J. Kieffer, J. Phys. Chem. A, 2009, 113, 9707–9714.
4. S. M. Ramirez, Y. J. Diaz, R. Campos, R. L. Stone, T. S. Haddad and J. M. Mabry, J. Am. Chem. Soc., 2011, 133, 20084–20087.
5. K. Tanaka and Y. Chujo, J. Mater. Chem., 2012, 22, 1733–1746.
6. M. Carraro and S. Gross, Materials, 2014, 7, 3956–3989.
7. K. Wu, M. J. Huang, K. Yue, C. Liu, Z. W. Lin, H. Liu, W. Zhang, C. H. Hsu, A. C. Shi, W. B. Zhang and S. Z. D. Cheng, Macromolecules, 2014, 47, 4622–4633.
8. K. Q. Zhang, B. Li, Y. H. Zhao, H. Li and X. Y. Yuan, Prog. Chem., 2014, 26, 394–402.
9. E. G. Shockey, A. G. Bolf, P. F. Jones, J. J. Schwab, K. P. Chaffee, T. S. Haddad and J. D. Lichtenhan, Appl. Organomet. Chem., 1999, 13, 311–327.
10. L. Liu, S. J. Lee, M. E. Lee, P. Kang and M.-G. Choi, Tetrahedron Lett., 2015, 56, 1562–1565.
11. H. Araki and K. Naka, Macromolecules, 2011, 44, 6039–6045.
12. E. L. Heeley, D. J. Hughes, Y. El Aziz, P. G. Taylor and A. R. Bassindale, Macromolecules, 2013, 46, 4944–4954.
13. Y. Gao, W. Xu, D. Zhu, L. Chen, Y. Fu, Q. He, H. Cao and J. Cheng, J. Mater. Chem. A, 2015, 3, 4820–4826.
14. L. Wang, Y. Ishida, R. Maeda, M. Tokita and T. Hayakawa, RSC Adv., 2014, 4, 34981–34986.
15. K. Y. Mya, X. Li, L. Chen, X. P. Ni, J. Li and C. B. He, J. Phys. Chem. B, 2005, 109, 9455–9462.
16. A. Pasquarello, M. S. Hybertsen and R. Car, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, R2339–R2342.
17. K. H. Xiang, R. Pandey, U. C. Pernisz and C. Freeman, J. Phys. Chem. B, 1998, 102, 8704–8711.
18. W. D. Cheng, K. H. Xiang, R. Pandey and U. C. Pernisz, J. Phys. Chem. B, 2000, 104, 6737–6742.
19. R. Franco, A. K. Kandalam, R. Pandey and U. C. Pernisz, J. Phys. Chem. B, 2002, 106, 1709–1713.
20. D. A. Wann, R. J. Less, F. Rataboul, P. D. McCaffrey, A. M. Reilly, H. E. Robertson, P. D. Lickiss and D. W. H. Rankin, Organometallics, 2008, 27, 4183–4187.
21. H. Phillips, S. H. Zheng, A. Hyla, R. Laine, T. Goodson, E. Geva and B. D. Dunietz, J. Phys. Chem. A, 2012, 116, 1137–1145.
22. E. C. Allen and K. J. Beers, Polymer, 2005, 46, 569–573.
23. D. Hossain, C. U. Pittman, S. Saebo and F. Hagelberg, J. Phys. Chem. C, 2007, 111, 6199–6206.
24. D. Hossain, S. R. Gwaltney, C. U. Pittman and S. Saebo, Chem. Phys. Lett., 2009, 467, 348–353.
25. N. L. Dias, F. C. M. Portugal, J. M. F. Nogueira, P. Brandao, V. Felix, P. D. Vaz, C. D. Nunes, L. F. Veiros, M. J. V. de Brito and M. J. Calhorda, Organometallics, 2012, 31, 4495–4503.
26. S. G. Semenov and M. E. Bedrina, J. Struct. Chem., 2013, 54, 159–163.
27. A. A. Skelton and J. R. Fried, Phys. Chem. Chem. Phys., 2013, 15, 4341–4354.
28. X. Q. Wang, J. Corn and F. Hagelberg, Chem. Phys., 2013, 426, 48–53.
29. T. Kudo and M. S. Gordon, J. Am. Chem. Soc., 1998, 120, 11432–11438.
30. M. Cypryk, J. Phys. Chem. A, 2005, 109, 12020–12026.
31. X. Solans-Monfort, J. S. Filhol, C. Coperet and O. Eisenstein, New J. Chem., 2006, 30, 842–850.
32. K. Wolinski, J. F. Hinton and P. Pulay, J. Am. Chem. Soc., 1990, 112, 8251–8260.
33. F. Cloran, I. Carmichael and A. S. Serianni, J. Am. Chem. Soc., 2001, 123, 4781–4791.
34. P. Tahtinen, A. Bagno, K. D. Klika and K. Pihlaja, J. Am. Chem. Soc., 2003, 125, 4609–4618.
35. S. Basak, D. Chopra and K. K. Rajak, J. Organomet. Chem., 2008, 693, 2649–2656.
36. P. Mondal, A. Hens, S. Basak and K. K. Rajak, Dalton Trans., 2013, 1536–1549.
37. K. Wolinski, R. Haacke, J. F. Hinton and P. Pulay, J. Comput. Chem., 1997, 18, 816–825.
38. K. Wolinski, J. F. Hinton and P. Pulay, J. Am. Chem. Soc., 1990, 112, 8251–8260.
39. G. Schreckenbach and T. Ziegler, J. Phys. Chem., 1995, 99, 606–611.
40. S. Pathak, R. Bast and K. Ruud, J. Chem. Theory Comput., 2013, 9, 2189–2198.
41. H. Lin, X. Wan, X. S. Jiang, Q. K. Wang and J. Yin, Adv. Funct. Mater., 2011, 21, 2960–2967.
42. M. J. Frisch, G. W. Trucks and H. B. Schlegel, et al., Gaussian 2003W Revision B. 05, Gaussian Inc, Pittsburgh, PA, 2003.
43. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
44. C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789.
45. N. Sundaraganesan, S. Ilakiamani, H. Saleem, P. M. Wojciechowski and D. Michalska, Spectrochim. Acta, Part A, 2005, 61, 2995–3001.
46. M. Karabacak, M. Kurt, M. Cinar and A. Coruh, Mol. Phys., 2009, 107, 253–264.
47. S. Miertus, E. Scrocco and J. Tomasi, Chem. Phys., 1981, 55, 117–129.
48. S. Miertus and J. Tomasi, Chem. Phys., 1982, 65, 239–245.
49. M. Bartsch, P. Bornhauser, G. Calzaferri and R. Imhof, J. Phys. Chem., 1994, 98, 2817–2831.
50. E. S. Park, H. W. Ro, C. V. Nguyen, R. L. Jaffe and D. Y. Yoon, Chem. Mater., 2008, 20, 1548–1554.
51. J. F. Brown, J. Am. Chem. Soc., 1965, 87, 4317–4324.
52. N. Subramanian, N. Sundaraganesan and J. Jayabharathi, Spectrochim. Acta, Part A, 2010, 76, 259–269.
53. T. T. Lin, C. B. He and Y. Xiao, J. Phys. Chem. B, 2003, 107, 13788–13792.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra14298a

---