Effects of pore surfaces on the electronic states of metal complexes formed on bipyridine periodic mesoporous organosilica

Abstract

Bipyridine periodic mesoporous organosilica (BPy-PMO) is a porous material in which the pore walls consist of silica frameworks bridged by 2,2′-bipyridine (bpy) functional groups, and is expected to be a novel platform for catalysts and photocatalysts. Metal complexes formed on the pore surfaces of BPy-PMO are under unusual environmental conditions that are entirely different from those in the bulk solids or in solution. To understand the effects of the unique structural features of the pore surfaces on the metal complexes, this study focused on the combination of BPy-PMO with Re(bpy)(CO)₃Cl (Re-BPy-PMO) and analyzed the metal-to-ligand charge transfer (MLCT) absorption band of this material. UV/vis and infrared analyses demonstrated that the Re-complex in Re-BPy-PMO interacts with water molecules adsorbed on the pore surfaces. Quantum chemical calculations suggest that the Re-complex also interacts with SiOH groups on the silica framework. Combined interactions are responsible for the unexpectedly short MLCT absorption wavelength of Re-BPy-PMO. This hypothetical mechanism is consistent with a red shift of the MLCT absorption following trimethylsilylation, which transforms SiOH groups to SiOSiMe₃ (TMS) groups and removes adsorbed water. Calculations also suggest that the TMS group is more electron-donating than the SiOH group. Therefore, the trimethylsilylation should result in a blue shift of the MLCT absorption with regard to the substituent effect, which is inconsistent with the experimental observations. Evidently, the effect of the removal of water and SiOH groups is sufficiently powerful to counteract the blue shift. Thus, the pore-surface features have a significant influence on the metal complex.

---

1 Introduction

Periodic mesoporous organosilica (PMO) is a class of porous materials with a highly-ordered porous structure, uniform pore sizes of 2–30 nm and pore walls in which organic functional groups are homogeneously and densely integrated.^1–5 PMO is synthesized by the poly-condensation of an organosilane precursor having the formula R–{Si(OR′)₃}_x (R: organic bridging group, R′: Me, Et, i-Pr, etc. and x ≥ 2) in the presence of a template surfactant.^6–8 Thus, various functionalities originating from R can be introduced on the pore walls.^9,10 By taking advantage of this feature, PMOs with organic ligands that enable metal complex formation on the pore surfaces have been synthesized.^11–13 These PMOs show promise as novel platforms for functional materials such as catalysts and photocatalysts, and the immobilization of homogeneous complexes on PMOs may allow separation from reaction solutions, for recovery and reuse.^14–21 Porous structures having high surface areas and large pores are also advantageous with regard to obtaining efficient catalytic reactions. In particular, a PMO incorporating 2,2′-bipyridine (BPy-PMO, Fig. 1),¹³ which acts as a chelating ligand for a wide variety of metal ions (such as Ru, Ir, Re, Pd and Mo), has special potential for a wide variety of applications.^22–27

Fig. 1 Schematic images of (a) mesoporous structure of BPy-PMO, (b) molecular structure of pore-wall surface of BPy-PMO and (c) Re(bpy)(CO)₃Cl formed on the BPy-PMO pore wall. The outermost layer of the pore wall is visualized in (b) and (c).

The metal complexes formed on the pore surfaces of BPy-PMO are unusual in that they are discretely fixed by an array of bipyridine (bpy) groups without fully contacting one another, in contrast to the bulk solid. These complexes also have no interactions with solvent molecules, unlike those in solution. The silica framework of BPy-PMO, which is connected to the bpy groups through Si–C bonds, behaves as a huge substituent group and exhibits substituent effects on the complexes. In addition, silanol (SiOH) groups are densely distributed over the framework and thus are in close proximity to the complexes. These specific structural features have the potential to induce anomalous electronic properties in the metal complex, although the specific effects are still insufficiently understood.

In prior work, we formed Re(bpy)(CO)₃Cl complexes on the bpy ligands embedded in the pore surfaces of BPy-PMO (Re-BPy-PMO) and investigated the physicochemical properties²⁶ and selective CO₂ reduction photocatalytic activity of the resulting materials.^28,29 This work found that the metal-to-ligand charge transfer (MLCT) absorption band (λ_MLCT) of Re-BPy-PMO in air is close to that of Re(bpy)(CO)₃Cl in toluene. It has been reported that λ_MLCT is sensitive to both solvent polarity and the substituents on the bpy ligand, such that lower solvent polarity results in a longer λ_MLCT.^30–33 Thus, the λ_MLCT of Re-BPy-PMO in air should be longer than that of Re(bpy)(CO)₃Cl in solution because air can be regarded as a solvent having an extremely low polarity (ε_r ≈ 1). With regard to the substituent effect, λ_MLCT is red-/blue-shifted by electron-withdrawing/-donating substituents on the bpy ligand, and quantum chemical calculations suggest that a silica framework with SiOH groups behaves as an electron-withdrawing substituent.³⁴ Based on both these factors, the λ_MLCT of Re-BPy-PMO in air should be longer than that of Re(bpy)(CO)₃Cl dissolved in any solution. However, actual experimental observations are inconsistent with this theoretical prediction. In fact, the environmental effects of the pore walls, which are derived from the unique structural features discussed above, most likely decide the experimentally-observed λ_MLCT. Assessing the MLCT absorption band of Re(bpy)(CO)₃Cl formed on the pore surfaces enables a detailed investigation of environmental effects on the complex. Such information will definitely be useful in the future development of BPy-PMO-based functional materials.

In the present study, we examined the effects of the pore surfaces on a metal complex formed on BPy-PMO, focusing on the relationship between the λ_MLCT of Re(bpy)(CO)₃Cl and the pore surface molecular structures. For this purpose, trimethylsilylated Re-BPy-PMO (Re-BPy-PMO-TMS), in which the SiOH groups have been converted to SiOSiMe₃ groups, was prepared and compared with the original Re-BPy-PMO to explore the factors affecting the λ_MLCT. Quantum chemical calculations were utilized to corroborate the mechanism proposed for the observed shifts of the λ_MLCT.

2 Experimental and computational details

2.1 Experiments

2.1.1 Materials. Re-BPy-PMO was synthesized according to a previously reported procedure.¹³ BPy-PMO (50 mg) was added to a solution of Re(CO)₅Cl (11.6 mg, 0.032 mmol) in toluene (10 mL) under an argon atmosphere. The suspension was stirred at 90 °C for 18 h, and filtered with toluene and CH₂Cl₂ to obtain Re-BPy-PMO. In this condition, the loading amount of the Re-complex should be 0.032 mmol at maximum, accounting for approximately 30% of the bpy units exposed on the pore surfaces. Re-BPy-PMO-TMS was synthesized by end-capping of the SiOH groups on the pore surfaces of Re-BPy-PMO using trimethylsilylimidazole in toluene at 60 °C.²² Hereafter, this modification is referred to as trimethylsilylation.

2.1.2 Measurements. UV/vis diffuse reflectance spectra were acquired using a JASCO V-670 spectrometer with an integrating sphere unit following dilution of samples with BaSO₄ to avoid absorption saturation. Infrared (IR) spectra were acquired using a Thermo Fisher Scientific Nicolet Avatar-360 FT-IR spectrometer with an attenuated total reflection (ATR) attachment to confirm the presence of adsorbed water. IR spectra were also obtained after heating samples at 120 °C for 20 h under a reduced pressure of 70 Pa. In situ UV/vis spectra of each sample were acquired using a JASCO V-670 spectrometer with an integrating sphere unit and a vacuum chamber (the equipment used was different from the above). First, the spectrum in air was measured at ambient temperature, after which each sample was heated at 100 °C for 1 h under vacuum (less than 10 Pa) followed by cooling to ambient temperature. After the sample was allowed to stand for 20 min while maintaining the same reduced pressure, a second spectrum was recorded.

2.2 Calculations

Quantum chemical calculations were utilized to estimate λ_MLCT values and to explore the mechanism by which these values were shifted. All calculations were carried out utilizing the Gaussian09 program³⁵ and the following conditions were adopted unless otherwise noted. The molecular geometric structures were optimized using the density functional theory (DFT) method with the B3LYP functional^36,37 and the excited states were calculated using the time-dependent DFT (TDDFT) method^38–42 with the CAM-B3LYP functional.⁴³ The def2-TZVP and def2-SVP basis sets were employed for Re and for other elements, respectively.⁴⁴ The functional and the basis sets for the TDDFT calculations were selected on the basis of our brief assessment (Table S1, ESI†). Single excitation from the second highest occupied molecular orbital (HOMO−1) to the lowest unoccupied molecular orbital (LUMO) is primarily responsible for the MLCT excitation of Re(bpy)(CO)₃Cl, giving an absorption band in the vicinity of 400 nm in the UV/vis spectrum.^30–32 The excited states corresponding to the HOMO−1 → LUMO excitation of Re(bpy)(CO)₃Cl were identified based on their dominant excitations and relevant molecular orbitals, and the calculated excitation energies were converted to wavelengths (nm) to obtain λ_MLCT values. Although spin–orbit interactions are typically significant in molecular systems with heavy atoms, it has been reported that, in the case of the λ_MLCT of Re(bpy)(CO)₃Cl, the values obtained by spin-free calculations are in good agreement with experimental results.^30–32 Thus, the spin-free TDDFT method was employed in this study.

2.2.1 Model complexes. Calculations were performed for Re(bpy)(CO)₃Cl, Re[5,5′-{Si(OH)₃}₂bpy](CO)₃Cl and Re[5,5′-{Si(OH)₂(OSiMe₃)}₂bpy](CO)₃Cl to estimate the substituent effect of the silica framework on λ_MLCT. Henceforth, Re[5,5′-{Si(OH)₃}₂bpy](CO)₃Cl and Re[5,5′-{Si(OH)₂(OSiMe₃)}₂bpy](CO)₃Cl are denoted as XOH and XOTMS, respectively. XOH is the simplest model of a Re-complex on BPy-PMO for which the bpy ligands are connected to the silica framework through Si–C bonds (Fig. 1(c)), and variation of the substituent effect following trimethylsilylation was examined by comparing XOH to XOTMS. In both cases, TDDFT calculations using the polarizable continuum model (PCM)^45–47 were also performed so as to examine the solvent effects of toluene, dichloromethane (CH₂Cl₂), acetonitrile (CH₃CN) and water (in ascending order of polarity). Re(bpy)(CO)₃Cl–(H₂O)_n (n = 1, 2) clusters were additionally examined to investigate the interactions between Re(bpy)(CO)₃Cl and water molecules in more detail.

2.2.2 Pore-surface cluster models. Pore-surface cluster models were prepared utilizing first principles calculations. The pore walls of BPy-PMO were modeled as a periodic system with the unit cell consisting of SiO₃-bpy-Si₂O₃-bpy-Si (C₂₀H₁₂N₄O₆Si₄). The atomic positions, cell shape and volume were optimized using the DFT method with the optPBE-vdW functional.^48–52 The cutoff energy for the plane wave was set to 400 eV, a 3 × 6 × 6 gamma-centered Monkhorst–Pack k-point mesh⁵³ was used and Gaussian smearing with a width of 0.2 eV was applied. The convergence criterion for the optimization was set to 0.02 eV Å⁻¹ of the forces on the nuclei. The calculations were performed using the VASP 5.4 program.^54,55 A supercell was created by repeating the optimized unit cell five times along the a and c axes, while two times along the b axis. Three bpy groups within the silica framework were cut out from the super cell along the c axis to prepare a BPy-PMO pore surface model, and the dangling bonds were saturated with OH groups. ReBP, the Re-BPy-PMO cluster model (Fig. 2(a)), was obtained by mounting Re(CO)₃Cl on the central bpy group, while ReBPT, a cluster model of Re-BPy-PMO-TMS (Fig. 2(b)), was prepared by converting six SiOH groups of ReBP (circled in Fig. 2(a)) to SiOSiMe₃ groups.

Fig. 2 Molecular structures of (a) ReBP and (b) ReBPT. ReBPT was prepared by converting circled OH groups of ReBP to OSiMe₃ groups.

The molecular structures of ReBP and ReBPT were optimized and their excited states were calculated. The silica framework of the PMO forms a three-dimensional network structure, and both T² (R–Si(OH)(OSi)₂) and T³ (R–Si(OSi)₃) species are present in large amounts in BPy-PMO synthesized under basic conditions,¹³ suggesting that the silica framework of BPy-PMO is inflexible. Therefore, the positions of Si atoms were fixed for the duration of the optimizations of the cluster models, except for those in the OSiMe₃ groups of ReBPT.

3 Results and discussion

3.1 Experimental results

The UV/vis spectra acquired in this work are shown in Fig. 3. The λ_MLCT of Re-BPy-PMO had an absorption maximum at 403 nm, which is close to that of Re(bpy)(CO)₃Cl in toluene.²⁶ However, the λ_MLCT of Re-BPy-PMO-TMS was 416 nm, meaning that the trimethylsilylation of Re-BPy-PMO resulted in a red shift of the MLCT absorption.

Fig. 3 UV/vis spectra of Re-BPy-PMO (black) and Re-BPy-PMO-TMS (gray).

The IR spectra are presented in Fig. 4. The Re-BPy-PMO spectrum exhibits an absorption peak around 1600 cm⁻¹ that is assigned to the bending vibration of water.⁵⁶ This peak almost disappears upon heating under reduced pressure, indicating that the adsorbed water was almost completely removed. In the case of the Re-BPy-PMO-TMS, only a small peak was observed in this region and the spectral change upon heating under reduced pressure was insignificant, suggesting that only a small amount of water was adsorbed on the pore surfaces. The lack of water is consistent with an increase in the hydrophobicity of the pore surfaces due to trimethylsilylation.

Fig. 4 IR spectra of (a) Re-BPy-PMO before (black) and after (blue) heating, and (b) Re-BPy-PMO-TMS before (gray) and after (red) heating.

In the in situ UV/vis spectra, the MLCT absorption band of Re-BPy-PMO is red-shifted following the heating under vacuum and appears close to that generated by Re-BPy-PMO-TMS (Fig. 5). In contrast, the shift in the Re-BPy-PMO-TMS peak following the same treatment was negligible.

Fig. 5 In situ UV/vis spectra of Re-BPy-PMO before (black) and after (blue) vacuum heating; and Re-BPy-PMO-TMS before (gray) and after (red) vacuum heating. Detailed procedures are described in Section 2.1.2.

3.2 Calculation results

Vibrational analyses were carried out for the B3LYP-optimized geometries to confirm the absence of imaginary number frequencies. Because the second singlet excited state (S₂) of each model exhibited the largest oscillator strength in the wavelength region above 300 nm (Table S2, ESI†), this excited state was regarded as the origin of the visible MLCT absorption band. The HOMO−1 → LUMO single excitation was dominant in the S₂ states, while contributions from other excitations were negligible (Table S2, ESI†). Thus, these orbitals were carefully analyzed (Fig. S1, ESI†).

3.2.1 Model complexes and cluster models. The λ_MLCT wavelengths of model complexes calculated without PCM were in the order of XOH > XOTMS > Re(bpy)(CO)₃Cl and well correlated with the gaps between the HOMO−1 and LUMO energy levels, Δ_H−1/L (Table 1). Since the LUMO of Re(bpy)(CO)₃Cl is distributed over the bpy ligand (Fig. S1, ESI†), its energy level is greatly affected by the substituents on the ligand. In contrast, the HOMO−1 of Re(bpy)(CO)₃Cl is distributed over the Re(CO)₃Cl moiety and therefore only slightly affected by the ligand substituents. The LUMO level of XOH was lower than that of Re(bpy)(CO)₃Cl. The λ_MLCT wavelengths and the orbital levels confirm that Si(OH)₃ groups behave as electron-withdrawing groups, and so with silica frameworks of BPy-PMO. Meanwhile, the HOMO−1 and LUMO levels of XOTMS are higher than XOH, indicating that the Si(OH)₂(OSiMe₃) group is more electron-donating than the Si(OH)₃ group. The LUMO rises more largely than HOMO−1, resulting in the larger Δ_H−1/L and the shorter λ_MLCT wavelength than XOH. The results predict that Re-BPy-PMO-TMS should have a shorter λ_MLCT wavelength than Re-BPy-PMO based on substituent effects, although the prediction is inconsistent with the experimental observations in which the λ_MLCT of Re-BPy-PMO-TMS is longer than that of Re-BPy-PMO.

Table 1 MLCT absorption wavelength (λ_MLCT), energy levels of HOMO−1 and LUMO, and their gaps (Δ_H−1/L) calculated without PCM

Molecule	λ MLCT (nm)	Orbital energy level (eV)
		HOMO−1	LUMO	Δ H−1/L
a Re[5,5′-{Si(OH)3}2bpy](CO)3Cl. b Re[5,5′-{Si(OH)2(OSiMe3)}2bpy](CO)3Cl.
Re(bpy)(CO)3Cl	460.7	−6.85	−1.76	5.09
XOH^a	471.0	−6.83	−1.85	4.98
XOTMS^b	465.1	−6.80	−1.79	5.01
ReBP	404.7	−7.28	−1.82	5.46
ReBPT	432.8	−6.68	−1.46	5.22

---

Fig. 6 and Fig. S2 show the 5 × 2 × 5 supercell created by replicating the optimized unit cell. The optimized cell parameters and the Cartesian coordinates of the atoms are provided in Table S3 (ESI†). The calculated length of the c axis (4.75 Å) is close to the intermolecular distance between the organic groups (4.4 Å), which was previously estimated by molecular mechanics calculations.⁵ A cluster including three organic groups within the silica framework was cut out from this supercell along the c axis and was utilized to prepare the ReBP and ReBPT models (see Section 2.2.1).

Fig. 6 Molecular structure of BPy-PMO pore-wall model optimized by utilizing first principle calculation. The bc plane of 5 × 2 × 5 supercell is presented. A cluster including three bipyridine groups and silica framework in a square were cut out and used for preparing ReBP and ReBPT (Fig. 2).

The calculated λ_MLCT values of the model clusters are also collected in Table 1. The ReBP and ReBPT values were 404.7 and 432.8 nm, respectively, and thus were in good agreement with the experimental results, where the λ_MLCT values were in the order of Re-BPy-PMO < Re-BPy-PMO-TMS (Fig. 3). The HOMO−1 and LUMO levels of ReBPT were shifted upward by 0.60 and 0.36 eV from those of ReBP, leading to a smaller Δ_H−1/L and a longer λ_MLCT. In this case, in contrast to the changes from XOH to XOTMS, the HOMO−1 level rises more significantly than the LUMO level. The optimized structures for ReBP and ReBPT are shown in Fig. 7. In ReBP, the OH bonds of SiOH groups are oriented relative to the CO and Cl ligands of the Re-complex moiety, suggesting hydrogen-bond-like interactions (Fig. 7(a)). These interactions attract the CO and Cl ligands to the SiOH groups and induce a slight distortion of the Re-complex. Such interactions and distortion of the complex are not observed in the case of the ReBPT, for which there are no SiOH groups around the complex.

Fig. 7 Optimized structures of (a) ReBP and (b) ReBPT. OH groups interacting with the Re-complex moiety in ReBP are circled.

3.2.2 Models with PCM and Re(bpy)(CO)₃Cl–(H₂O)_n. The calculated results for Re(bpy)(CO)₃Cl, XOH and XOTMS using the PCM approach are summarized in Table 2. A higher solvent polarity evidently resulted in a shorter λ_MLCT, and the calculated λ_MLCT values for Re(bpy)(CO)₃Cl are in good agreement with corresponding experimental values.³² Although Re(bpy)(CO)₃Cl is poorly soluble in water, the estimated λ_MLCT in water was close to that in acetonitrile. Re(bpy)(CO)₃Cl in the ground state is polarized as Re(CO)₃Cl^δ−–(bpy)^δ+ because of the electron-withdrawing properties of the CO and Cl ligands, and the MLCT excitation results in reduced polarization.⁵⁷ For this reason, the ground state is stabilized by the solvent to a greater extent than the MLCT excited state and this effect is significant in a polar solvent. Calculated molecular orbital energy levels (Table 2), populations and dipole moments (Table S4, ESI†) are consistent with this mechanism. The same mechanism is also applicable in the cases of the XOH and XOTMS, as suggested by the results in Table 2. The λ_MLCT values decreased in the order of XOH > XOTMS > Re(bpy)(CO)₃Cl regardless of the solvent, indicating that this order reflects substituent effects, so long as the solvent effects are equal.

Table 2 MLCT absorption wavelength (λ_MLCT), energy levels of HOMO−1 and LUMO, and their gaps (Δ_H−1/L) calculated with PCM. Available experimental values are also presented

Molecule	Solvent	λ MLCT (nm)		Orbital energy level (eV)
		Calc.	Exptl.^a	HOMO−1	LUMO	Δ H−1/L
a Ref. 32. b Re[5,5′-{Si(OH)3}2bpy](CO)3Cl. c Re[5,5′-{Si(OH)2(OSiMe3)}2bpy](CO)3Cl.
Re(bpy)(CO)3Cl	Toluene	408.9	403	−7.06	−1.59	5.47
	CH2Cl2	374.8	387	−7.24	−1.48	5.76
	CH3CN	364.7	371	−7.31	−1.45	5.86
	Water	362.9		−7.32	−1.45	5.87
XOH^b	Toluene	421.4		−7.08	−1.76	5.32
	CH2Cl2	389.1		−7.26	−1.69	5.57
	CH3CN	379.5		−7.32	−1.67	5.66
	Water	377.7		−7.33	−1.66	5.67
XOTMS^c	Toluene	417.8		−7.07	−1.72	5.35
	CH2Cl2	386.4		−7.26	−1.66	5.60
	CH3CN	377.0		−7.32	−1.64	5.68
	Water	375.3		−7.33	−1.63	5.70

---

The results for Re(bpy)(CO)₃Cl–(H₂O)_n are collected in Table 3. The λ_MLCT was shortened with increases in n and the values were in the order of 1BPy > 1CO ≈ 1CL for n = 1 and 2BPy > 2CO ≈ 2CL for n = 2. These data demonstrate that the blue shift induced by the solvent primarily originates from the interactions of the solvent molecules with the CO and Cl ligands. The HOMO−1 levels of 1CO, 1CL, 2CO and 2CL were all shifted downward from those of Re(bpy)(CO)₃Cl, while the LUMO levels varied over a narrow range. The Δ_H−1/L value was therefore increased with increases in n and correlated with λ_MLCT. The optimized structures for Re(bpy)(CO)₃Cl–(H₂O)_n are shown in Fig. 8. The O–H bonds of H₂O, which are polarized as O^δ−–H^δ+, are determined to be oriented relative to the CO and Cl ligands in 1CO, 1CL, 2CO and 2CL. The HOMO−1 of Re(bpy)(CO)₃Cl is stabilized when H^δ+ approaches the CO and/or Cl ligands, since the orbital is distributed over the Re(CO)₃Cl moiety (Fig. S1, ESI†). Although the Re(bpy)(CO)₃Cl complex formed on BPy-PMO is half embedded in the pore wall, the CO and Cl ligands stick out from the pore surface and the space around these ligands is expected to be vacant. Therefore, the adsorbed water molecules can readily approach these ligands to induce a blue shift in the MLCT adsorption.

Table 3 Calculated MLCT absorption wavelengths (λ_MLCT), energy levels of HOMO−1 and LUMO, and their gaps (Δ_H−1/L) in Re(bpy)(CO)₃Cl–(H₂O)_n. Optimized structures of the models are shown in Fig. 8

Model	λ MLCT (nm)	Orbital energy level (eV)
		HOMO−1	LUMO	Δ H−1/L
Re(bpy)(CO)3Cl	460.7	−6.85	−1.76	5.09
1CO	423.9	−7.03	−1.69	5.34
1CL	421.3	−7.11	−1.77	5.34
1BPy	448.1	−6.69	−1.52	5.17
2CO	408.7	−7.05	−1.59	5.46
2CL	393.0	−7.35	−1.76	5.59
2BPy	434.0	−6.55	−1.29	5.26

---

Fig. 8 Optimized structures of Re(bpy)(CO)₃Cl–(H₂O)_n: (a) 1CO, (b) 1CL, (c) 1BPy, (d) 2CO, (e) 2CL and (f) 2BPy.

In the optimized structures of 1BPy and 2BPy, water molecules interact with C–H bonds of the bpy ligand (Fig. 8(c and f)). In actual systems, the O–H bonds of water molecules may interact with π orbitals of the bpy ligand.⁵⁸ In our calculations of Re(bpy)(CO)₃Cl–(H₂O)_n, however, such structures were not obtained. In the calculation employing an initial structure of Re(bpy)(CO)₃Cl–(H₂O) in which a water molecule is placed above the bpy-ligand molecular plane, the water molecule departed from the bpy ligand and moved toward the CO or Cl ligands as a result of the geometry optimizations. The result suggests that the interactions of water molecules with the CO and Cl ligands are stronger than those with the bpy ligand. The O–H/π interactions may be predicted by incorporating the cluster of water molecules with hydrogen bond networks.

3.3 Discussion

The IR spectra (Fig. 4) suggest that water is adsorbed on the pore surfaces of Re-BPy-PMO in air, but is almost absent on Re-BPy-PMO-TMS. The interactions between the Re(bpy)(CO)₃Cl complex and water molecules induces a blue shift in the MLCT absorption band. This mechanism is supported by the red shift in the MLCT band of Re-BPy-PMO following heating under vacuum (Fig. 5) and the results of calculations incorporating solvent effects (Tables 2 and 3). Thus, it is likely that the blue shift due to adsorbed water dictates the experimentally-obtained ordering of the λ_MLCT values (Re-BPy-PMO < Re-BPy-PMO-TMS).

The optimized structure of ReBP shown in Fig. 7(a) suggests that the Re(bpy)(CO)₃Cl moieties of Re-BPy-PMO also interact with SiOH groups on the silica framework. The OH bonds of these SiOH groups are oriented relative to the CO and Cl ligands as well as the OH bonds of water molecules in Re(bpy)(CO)₃Cl–(H₂O)_n (Fig. 8). The optimized structures suggest that the SiOH groups have similar effects to the water molecules and also contribute to a blue shift of the MLCT absorption band through the downward shift of the HOMO−1 level. To examine this hypothetical mechanism, ReBPH was prepared by substituting H atoms for the two OH groups that interact with the Re-complex moiety (Fig. S3, ESI†). The positions of these two H atoms were re-optimized and the λ_MLCT was calculated adopting the conditions described in Section 2.2. The resulting λ_MLCT of 446.1 nm was significantly longer than that for ReBP (404.7 nm). The HOMO−1 and LUMO levels of ReBPH were −6.79 and −1.64 eV, respectively, while the Δ_H−1/L was 5.15 eV (Table S2, ESI†). The reduced Δ_H−1/L is primarily attributed to the increase in the HOMO−1 level and is consistent with the change in λ_MLCT. The calculated λ_MLCT for ReBPH is longer even than that for ReBPT, which is consistent with the substituent effects of SiOH and SiOSiMe₃ groups estimated from the results for XOH and XOTMS (Table 1). As the predicted interactions between SiOH groups and Re(bpy)(CO)₃Cl moieties are originated from the local molecular structures around the Re-complex moieties, the calculation results are expected to be independent of the ReBP cluster size.

These results indicate that both adsorbed water and SiOH groups in Re-BPy-PMO induce a blue shift in the MLCT absorption band, resulting in the λ_MLCT values in the order of Re-BPy-PMO < Re-BPy-PMO-TMS. The combined effects of the adsorbed water and SiOH groups are responsible for the unexpectedly short MLCT absorption wavelength of Re-BPy-PMO which is close to that of Re(bpy)(CO)₃Cl in toluene.²⁶ The MLCT band of Re-BPy-PMO was red-shifted following heating in a vacuum, and its λ_MLCT approached that of Re-BPy-PMO-TMS (Fig. 5). The red shift can be attributed to the removal of adsorbed water, while the interactions between Re(bpy)(CO)₃Cl moieties and SiOH groups remain even after the vacuum heating. The blue shift resulting from SiOH groups is responsible for the λ_MLCT of Re-BPy-PMO which is shorter than that of Re-BPy-PMO-TMS after the vacuum heating.

In the case of ReBP, the Re-complex moiety was distorted by the attractive interactions with SiOH groups (Fig. 7(a)). To examine the effect of this distortion on λ_MLCT, the complex moiety of ReBP was cut out from the optimized structure by breaking the Si–C bonds. The dangling bonds were terminated by H atoms and the positions of the two H atoms were re-optimized. The λ_MLCT was then calculated utilizing the same procedure and conditions discussed in Section 2.2. The calculated λ_MLCT was 461.4 nm (Table S2, ESI†) and thus similar to that of Re(bpy)(CO)₃Cl having an optimized structure (460.7 nm, Table 1), indicating that the influence of the distortion was negligible.

The red shift of λ_MLCT following trimethylsilylation was inconsistent with the substituent effects suggested by the λ_MLCT values in the order of XOH > XOTMS (Table 1). The results demonstrate that the red shift due to the removal of the interactions of Re(bpy)(CO)₃Cl moieties with the SiOH groups and adsorbed water counteracts the blue shift due to the change in the substituent effect. In this case, the structural properties of the pore surfaces have a significant influence on the electronic properties of the metal complex formed on the BPy-PMO. Thus, modification of the pore surfaces is a potential means of controlling the properties and functionalities of the metal complex, in addition to selecting specific metal ions and ligands. The effects of the modification of the pore surfaces on catalytic activity, selectivity and stability, as implied in our recent study,²⁷ are currently being investigated in our laboratory.

4 Conclusions

Metal complexes formed on the pore surfaces of BPy-PMO exist under unusual environmental conditions that are entirely different from those in the bulk solids or in solution. In this study, we examined the MLCT absorption band of Re(bpy)(CO)₃Cl to understand the effects of the pore surfaces on the electronic properties of the complex. The experimental and calculation results indicate that the Re(bpy)(CO)₃Cl moieties of Re-BPy-PMO interact with both adsorbed water and SiOH groups on the silica framework. These interactions contribute to a blue shift of the MLCT absorption band and are responsible for the significantly shorter λ_MLCT wavelength of Re-BPy-PMO which is close to that of Re(bpy)(CO)₃Cl in toluene. These interactions are largely absent in Re-BPy-PMO-TMS, resulting in λ_MLCT values that increase in the order of Re-BPy-PMO < Re-BPy-PMO-TMS. The calculation results suggest that the SiOSiMe₃ group is more electron-donating than the SiOH group, and therefore trimethylsilylation contributes to a blue shift of the MLCT absorption with regard to the substituent effect. However, a red shift was experimentally observed following trimethylsilylation as a result of the removal of water molecules and SiOH groups, which was more powerful than the blue shift. Thus, the pore surface has dominant effects on the electronic properties of the metal complex.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was partially supported by the Advanced Catalytic Transformation Program for Carbon Utilization (ACT-C) of the Japan Science and Technology Agency (JST) (Grant number: JPMJCR12Y1). Prof. Atsushi Satsuma at Nagoya University is acknowledged for his technical support while acquiring in situ UV/vis spectra. The authors are also grateful to Drs Ryoji Asahi and Nobuko Ohba at Toyota Central R&D Labs., Inc. for their suggestions, supports and encouragement.

References

1. S. Inagaki, S. Guan, Y. Fukushima, T. Ohsuna and O. Terasaki, J. Am. Chem. Soc., 1999, 121, 9611–9614.
2. T. Asefa, M. J. MacLachlan, N. Coombs and G. A. Ozin, Nature, 1999, 402, 867–871.
3. B. J. Melde, B. T. Holland, C. F. Blanford and A. Stein, Chem. Mater., 1999, 11, 3302–3308.
4. C. Yoshina-Ishii, T. Asefa, N. Coombs, M. J. MacLachlan and G. A. Ozin, Chem. Commun., 1999, 2539–2540.
5. S. Inagaki, S. Guan, T. Ohsuna and O. Terasaki, Nature, 2002, 416, 304–307.
6. K. J. Shea, D. A. Loy and O. Webster, J. Am. Chem. Soc., 1992, 114, 6700–6710.
7. R. J. P. Corriu, Angew. Chem., Int. Ed., 2000, 39, 1376–1398.
8. D. A. Loy and K. J. Shea, Chem. Rev., 1995, 95, 1431–1442.
9. S. Fujita and S. Inagaki, Chem. Mater., 2008, 20, 891–908.
10. N. Mizoshita, T. Tani and S. Inagaki, Chem. Soc. Rev., 2011, 40, 789–800.
11. M. Waki, N. Mizoshita, T. Ohsuna, T. Tani and S. Inagaki, Chem. Commun., 2010, 46, 8163–8165.
12. M. Waki, N. Mizoshita, T. Tani and S. Inagaki, Angew. Chem., Int. Ed., 2011, 50, 11667–11671.
13. M. Waki, Y. Maegawa, K. Hara, Y. Goto, S. Shirai, Y. Yamada, N. Mizoshita, T. Tani, W. J. Chun, S. Muratsugu, M. Tada, A. Fukuoka and S. Inagaki, J. Am. Chem. Soc., 2014, 136, 4003–4011.
14. Z. M. Michalska and D. E. Webster, Platinum Met. Rev., 1984, 18, 65–73.
15. X. S. Zhao, X. Y. Bao, W. P. Guo and F. Y. Lee, Mater. Today, 2006, 9, 32–39.
16. C. E. Song and S.-g. Lee, Chem. Rev., 2002, 102, 3495–3524.
17. D. E. De Vos, M. Dams, B. F. Sels and P. A. Jacobs, Chem. Rev., 2002, 102, 3615–3640.
18. P. McMorn and G. J. Hutchings, Chem. Soc. Rev., 2004, 33, 108–122.
19. C. Li, Catal. Rev.: Sci. Eng., 2004, 46, 419–492.
20. A. E. C. Collis and I. T. Horváth, Catal. Sci. Technol., 2011, 1, 912–919.
21. C. Perego and R. Millini, Chem. Soc. Rev., 2013, 42, 3956–3976.
22. Y. Maegawa and S. Inagaki, Dalton Trans., 2015, 44, 13007–13016.
23. X. Liu, Y. Maegawa, Y. Goto, K. Hara and S. Inagaki, Angew. Chem., Int. Ed., 2016, 55, 7943–7947.
24. N. Ishito, H. Kobayashi, K. Nakajima, Y. Maegawa, S. Inagaki, K. Hara and A. Fukuoka, Chem. – Eur. J., 2015, 21, 15564–15569.
25. Y. Kuramochi, M. Sekine, K. Kitamura, Y. Maegawa, Y. Goto, S. Shirai, S. Inagaki and H. Ishida, Chem. – Eur. J., 2017, 23, 10301–10309.
26. M. Waki, K.-i. Yamanaka, S. Shirai, Y. Maegawa, Y. Goto, Y. Yamada and S. Inagaki, Chem. – Eur. J., 2018, 24, 3846–3853.
27. S. Ishikawa, Y. Maegawa, M. Waki and S. Inagaki, ACS Catal., 2018, 8, 4160–4169.
28. J. Hawecker, J. M. Lehn and R. Ziessel, J. Chem. Soc., Chem. Commun., 1983, 536–538.
29. J. Hawecker, J. M. Lehn and R. Ziessel, Helv. Chim. Acta, 1986, 69, 1990–2012.
30. L. Yang, A. M. Ren, J. K. Feng, X. J. Liu, Y. G. Ma, M. Zhang, X. D. Liu, J. C. Shen and H. X. Zhang, J. Phys. Chem. A, 2004, 108, 6797–6808.
31. A. Vlček and S. Zális, J. Phys. Chem. A, 2005, 109, 2991–2992.
32. R. Heydová, E. Gindensperger, R. Romano, J. Sýkora, A. Vlček, S. Zális and C. Daniel, J. Phys. Chem. A, 2012, 116, 11319–11329.
33. L. A. Worl, R. Duesing, P. Y. Chen, L. Dellaciana and T. J. Meyer, J. Chem. Soc., Dalton Trans., 1991, 849–858.
34. S. Shirai, S. Iwata, Y. Maegawa, T. Tani and S. Inagaki, J. Phys. Chem. A, 2012, 116, 10194–10202.
35. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox, Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford CT, 2016.
36. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
37. P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch, J. Phys. Chem., 1994, 98, 11623–11627.
38. E. Runge and E. K. U. Gross, Phys. Rev. Lett., 1984, 52, 997–1000.
39. E. K. U. Gross and W. Kohn, Adv. Quantum Chem., 1990, 21, 255–291.
40. C. Jamorski, M. E. Casida and D. R. Salahub, J. Chem. Phys., 1996, 104, 5134–5147.
41. M. Petersilka, U. J. Gossmann and E. K. U. Gross, Phys. Rev. Lett., 1996, 76, 1212–1215.
42. M. E. Casida, C. Jamorski, K. C. Casida and D. R. Salahub, J. Chem. Phys., 1998, 108, 4439–4449.
43. T. Yanai, D. P. Tew and N. C. Handy, Chem. Phys. Lett., 2004, 393, 51–57.
44. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
45. E. Cancès, B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 107, 3032–3041.
46. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3093.
47. G. Scalmani and M. J. Frisch, J. Chem. Phys., 2010, 132, 114110.
48. J. Klimeš, D. R. Bowler and A. Michaelides, J. Phys.: Condens. Matter, 2010, 22, 022201.
49. M. Dion, H. Rydberg, E. Schröder, D. C. Langreth and B. I. Lundqvist, Phys. Rev. Lett., 2004, 92, 246401.
50. M. Dion, H. Rydberg, E. Schröder, D. C. Langreth and B. I. Lundqvist, Phys. Rev. Lett., 2005, 95, 109902.
51. G. Román-Pérez and J. M. Soler, Phys. Rev. Lett., 2009, 103, 096102.
52. J. Klimeš, D. R. Bowler and A. Michaelides, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 195131.
53. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188–5192.
54. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
55. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
56. D. B. Asay and S. H. Kim, J. Phys. Chem. B, 2005, 109, 16760–16763.
57. K. A. Walters, Y. J. Kim and J. T. Hupp, Inorg. Chem., 2002, 41, 2909–2919.
58. L. V. Slipchenko and M. S. Gordon, J. Phys. Chem. A, 2009, 113, 2092–2102.

---

Footnote

† Electronic supplementary information (ESI) available: Assessment on functional and basis set dependence in TDDFT calculation (Table S1); detailed results of the excited state calculations (Table S2); molecular orbitals of the models (Fig. S1); molecular structure of the BPy-PMO-pore-wall model (Fig. S2); cell parameters and Cartesian coordinates of the BPy-PMO-pore-wall model (Table S3); results of population analyses of the ground and MLCT excited states of Re(bpy)(CO)₃Cl (Table S4); molecular structure of ReBPH (Fig. S3). See DOI: 10.1039/c8nj06277c

---