Photoreduction of Methylviologen in Saponite Clay: Effect of Methylviologen Adsorption Density on the Reaction Efficiency

Abstract

To identify the mechanisms for and to estimate the photochemical reaction efficiency of molecules in solid-state host materials is difficult. The objective of the present research was to measure the photogeneration efficiency of the methylviologen cation radical (MV⁺•) hosted in a semi-transparent hybrid film composed of MV²⁺ and saponite, a 2:1 clay mineral. MV⁺• is the one-electron reduced species of MV²⁺. MV⁺• was generated by UV irradiation of these films. The fluorescence intensity of MV²⁺ and the photogeneration efficiency of MV⁺• depended on the loading level of MV²⁺. When the loading level of MV²⁺ was high (75% of the cation exchange capacity (abbreviated as % CEC) of saponite), its fluorescence was reduced considerably because of the self-fluorescence quenching reaction, and the photogeneration efficiency of MV⁺• was relatively high (quantum yield φ = 3.5×10^–2) compared to that of films with low adsorption density (10% CEC, φ = 1.1×10^–2). Furthermore, when the loading level of MV²⁺ was very low (0.13% CEC), a self-fluorescence quenching reaction was not observed and MV⁺• was not generated. From these observations, one of the principal mechanisms of the self-quenching reaction and MV⁺• formation in saponite is the electron transfer reaction between excited MV²⁺ and adjacent MV²⁺ molecules in the ground state.

Introduction

Over the last several decades, photoinduced hydrogen generation has been studied for the conversion of solar energy into chemical energy (Sprick et al. 2016; Stevenson et al. 2017; Wang et al. 2017). Methylviologen (MV²⁺) has been used commonly to evolve hydrogen, because its one-electron reduced species (methylviologen cation radical, MV⁺•) generates hydrogen in the presence of a Pt catalyst (Toshima et al. 1981; Matheson et al. 1983; Bard and Fox 2002). Many researchers have studied the generation of MV⁺• and/or hydrogen from MV²⁺ with visible light irradiation in the presence of photosensitizers (Wasielewski 1992; Yonemoto et al. 1992; Inoue et al. 1994). In addition, some researchers have found that, under UV light irradiation, MV⁺• is generated in solution in the presence of solid surfaces such as silica gels, layered transition metal oxides, and zeolites (Miyata et al. 1988; Mao et al. 1995; Alvaro et al. 1997; Peon et al. 2001). Among inorganic layered materials, clay minerals are an attractive group because of their following properties: (1) flat surfaces; (2) regularly arranged anionic charges on their surfaces; (3) exfoliation and stackability of individual nanolayers depending on the medium; (4) optical transparency in the visible region in the exfoliated state when the particle size is small (<100 nm); and (5) easy formation of hybrid materials with cationic molecules by electrostatic interaction (Shichi & Takagi 2000; Egawa et al. 2011). Various researchers have reported hybrid materials of clay minerals and MV²⁺ (Raupach et al. 1979; Villemure et al. 1986, 1991; Miyata et al. 1987; Rytwo et al. 1996), and Ogawa and co-workers (Kakegawa et al. 2003) reported that MV⁺• was generated when intercalated MV²⁺ in clay minerals was irradiated by UV light. The mechanism of MV⁺• generation in clay minerals under UV light was mentioned as being due to the generation of MV⁺• via an electron-transfer reaction from the bridging Si-O-Al oxygen lone pair to the excited MV²⁺. However, in the solid hybrid systems above, estimating the efficiency of photogeneration of MV⁺• was difficult, and the reaction details were not revealed because the amount of MV⁺• generated and decreased amount of MV²⁺ could not be estimated.

Previous work revealed the interesting structure of porphyrin−clay mineral hybrids, in which the porphyrin molecules adsorb densely on the exfoliated clay nanolayers without aggregation (Takagi et al. 2002, 2013). The crucial factor for the very dense adsorption of porphyrins was the matching of the distances between cationic sites in porphyrin and the average distance between anionic sites on the clay nanolayer surfaces (called the intercharge distance matching effect or size-matching effect). In these systems, the fluorescence intensity of some dyes adsorbed on saponite surfaces was quenched at higher adsorption densities (self-fluorescence quenching reaction) (Ishida et al. 2012b; Ohtani et al. 2014). This observation suggested that collision and/or electron transfer between excited and adjacent ground-state dye molecules might be one of the reasons for the self-fluorescence quenching reaction. The detail of the self-fluorescence quenching reaction, however, was not revealed.

MV²⁺ adsorbed on clay minerals has been reported to show self-fluorescence quenching behavior (Villemure et al. 1986, 1991). Judging from those studies and previous results, the principal mechanism for the generation of MV⁺• either on the surface or in the interlayer space of saponite could be an electron transfer reaction between the excited methylviologen (abbreviated as MV²⁺*) and the adjacent ground-state MV²⁺. The objective of the current study was to measure the photogeneration efficiency of MV⁺• in a semi-transparent hybrid film composed of MV²⁺ and saponite, in which the amount of MV²⁺ and MV⁺• could be estimated by UV-Vis absorption spectra, and to determine whether the principal mechanism of the self-fluorescence quenching reaction is indeed electron transfer between MV²⁺* and the adjacent ground-state MV²⁺ molecules, all with an eye to understanding whether this electron transfer reaction in the interlayer space of the clay mineral could be utilized further to construct other photochemical reaction systems.

Methods

Preparation of MV²⁺/Saponite Hybrid Film

A synthetic saponite (Sumecton SA ((Na_0.49Mg_0.14)[(Si_7.20Al_0.80)(Mg_5.97Al_0.03)O₂₀(OH)₄]), abbreviated as SSA, from Kunimine Industries Co., Ltd., Chiyoda-ku, Tokyo, Japan) was used, and its theoretical surface area and CEC were 750 m² g ⁻¹ and 1.00 meq g ⁻¹, respectively (Egawa et al. 2011). MV²⁺ dichloride was purchased from Nacalai Tesque, Inc (>98%, Chukyo-ku, Kyoto, Japan). The purity of the MV²⁺ was checked by ¹H-NMR and thin layer chromatography. All of the chemicals were used as received.

Quartz substrates were sonicated with water for 30 min and then sonicated with ethanol for another 30 min. These substrates were further treated in sulfuric acid (Kanto Chemical, Chuo-ku, Tokyo, Japan) overnight at room temperature and washed with > 150 mL of water to remove the sulfuric acid. MV²⁺/SSA hybrid films were prepared according to the literature (Kawamata et al. 2010; Takagi et al. 2004, 2010; Suzuki et al. 2011). MV²⁺ aqueous solution (2.0×10^–4 mol/L, 600 μL) was added to the saponite dispersion (2400 μL) within 2 s to prepare a MV²⁺/SSA complex dispersion in the exfoliated state (Suquet et al. 1977; Auerbach et al. 2004). The MV²⁺ loading level as a fraction of the % CEC of the clay mineral was adjusted from 10 to 80% CEC by changing the concentration of the saponite dispersion (from 1.00 to 1.25×10^–1 g/L). The complex dispersions were filtered through PTFE membrane filters (pore size = 0.1 μm, Advantec, Bunkyo-ku, Tokyo, Japan). The residue was transferred onto a quartz substrate (Agri, Oume-city, Tokyo, Japan) and dried under vacuum overnight at room temperature to form a semi-transparent MV²⁺/SSA hybrid film.

Photogeneration of MV⁺• in Hybrid Film

MV²⁺/hybrid films were placed in a cuvette and vacuumed for >1 h to remove the oxygen that quenches MV⁺•. These films were irradiated with UV light (270 nm, FWHM of wavelength was 12 nm) from a Xe lamp through a diffraction grating to examine the photoreduction of MV²⁺. The generation of MV⁺• was monitored with a UV-Vis absorption spectrometer. The cuvette was kept under vacuum during photo-irradiation and measurement of UV-Vis absorption spectra.

Analysis

The UV-Vis absorption spectra were obtained on a UV-3150 system (Shimadzu, Chukyo-ku, Kyoto, Japan). The fluorescence spectra were measured using an FP-6600 spectrofluorometer (Jasco, Hachioji-city, Tokyo, Japan), with an excitation wavelength of 272 nm. The hybrid film was set at 45° relative to the axis of the excitation beam and the emission detector. X-ray diffraction (XRD) patterns were measured using a RINT TTR III system (Rigaku, Akishima-city, Tokyo, Japan), using CuKα radiation. These measurements were carried out at room temperature.

Results and Discussion

XRD Patterns and Absorption and Fluorescence Spectra of the Hybrid Film

The hybrid film was sufficiently transparent (Fig. 1a) for transmission-mode measurements. The photograph of the hybrid film under 254 nm UV light irradiation (Fig. 1b) revealed blue or purple emission from MV²⁺*, while MV²⁺ in the aqueous medium exhibited no emission. Fluorescence was observed, however, when MV²⁺ was intercalated in the clay mineral or adsorbed on the clay surfaces, as reported by Detellier and Szabo (Villemure et al. 1986, 1991). The rotation of pyridyl groups of MV²⁺ was suppressed by adsorption on the SSA surface, and then the rate constant of non-radiative deactivation decreased. This phenomenon has been commonly recognized as surface-fixation induced emission (S-FIE) (Ishida et al. 2014; Tokieda et al. 2017).

Fig. 1. Photographs of MV²⁺/SSA hybrid film: (a) light transmission and (b) under UV irradiation

From the XRD patterns of MV²⁺/SSA hybrid films at various MV²⁺ loading levels (Fig. 2), the interlayer spaces of hybrid films were estimated to be 0.30–0.32 nm, which is almost the same as that of montmorillonite and other clay minerals where MV²⁺ was intercalated (Raupach et al. 1979; Rytwo et al. 1996; Kakegawa et al. 2003). Judging from the thickness of pyridyl groups and the basal space of SSA in the hybrid film, the orientation of MV²⁺ aromatic rings in the interlayer of SSA was parallel to the SSA layer. The XRD pattern of MV²⁺/SSA shifted slightly to a higher diffraction angle, indicating the decrease in interlayer distance, when the loading level of MV²⁺ was increased. This may indicate that the interlayer space of MV²⁺/SSA hybrid film became hydrophobic with the increased MV²⁺, and then the interlayer water was excluded.

Fig. 2. XRD patterns of MV²⁺/SSA hybrid films, with MV²⁺ loading levels of 20%, 40%, 60%, and 80% CEC

The UV-Vis absorption spectra of hybrid films varied with different loading levels of MV²⁺ (Fig. 3). The film with no MV²⁺ loading displayed no absorption band in the 200–600 nm region (not shown). The background of each film was different because of the scattering by SSA; however, the absorption band ascribed to MV²⁺ was clearly observed in the UV region in all cases. Note that when the loading level of MV²⁺ was low, a correspondingly greater amount of SSA was in the film, thus causing greater scattering. The maximum absorption wavelength (λ_max) of MV²⁺ was shifted to longer wavelength (272 nm) compared to that in the aqueous solution (257 nm, Fig. S1). A similar spectral shift was observed when MV²⁺ was adsorbed on the surfaces of SSA or other clay minerals (Villemure et al. 1991). This result indicated that the π-conjugated system of MV²⁺ was extended by the co-planarization of its two pyridinium groups (Fig. S2) because of the steric effect of SSA (Kuykendall & Thomas 1990; Chernia & Gill 1999; Ishida et al. 2012a). The λ_max of MV²⁺ was independent of the MV²⁺ loading level, and the spectral shape was also almost the same. Hence, the MV²⁺ intercalated in the interlayer of SSA did not aggregate in spite of its dense intercalation. When the MV²⁺ loading level in the hybrid film was set as high as 80% CEC, the absorbance at 272 nm was relatively small compared to other films, and MV²⁺ was observed in the filtrate, according to the UV-Vis absorption spectra. This indicated that a small amount of MV²⁺ was not adsorbed on SSA, because its saturation adsorption amount on the SSA surface was <80% CEC. Judging from the optical absorbance of excess MV²⁺ in the filtrate solution and the extinction coefficient of MV²⁺ at 257 nm (ε = 20700 mol·L^–1·cm^–1), the maximum adsorption of MV²⁺ in the above hybrid film was calculated more accurately to be 75% CEC (Watanabe & Honda 1982). In the fluorescence spectra of MV²⁺/SSA hybrid films (Fig. 4), fluorescence of MV²⁺ was observed at the maximum emission wavelength of 344 nm.

Fig. 3. UV-Vis absorption spectra of MV²⁺/SSA hybrid films, with MV²⁺ loading levels of 10% (black solid line), 20% (black broken line), 40% (black dotted line), 60% (gray solid line), and 80% (gray broken line) CEC

Fig. 4. Fluorescence spectra of MV²⁺/SSA hybrid films at various MV²⁺ loadings, with MV²⁺ loading levels of 10% (black solid line), 20% (black broken line), 40% (black dotted line), 60% (gray solid line), and 80% (gray broken line) CEC. The excitation wavelength was set at 272 nm

The fluorescence of MV²⁺ is known to occur when MV²⁺ is adsorbed on/intercalated in clay minerals, although aqueous solutions of MV²⁺ do not show the fluorescence (Villemure et al. 1986, 1991). This phenomenon is explained as being due to the suppression of rotation of the pyridinium groups of MV²⁺ by the steric effect of the clay minerals, and the rate constant of internal conversion is decreased, as described above.

The shape of the emission spectrum was independent of the loading level of MV²⁺. This suggests the absence of new luminescent species, such as J-type aggregates (head-to-tail type aggregate), and of excimers that could be formed in the interlayer space of SSA nanolayers, in spite of the large MV²⁺ loadings. The emission intensities of MV²⁺/SSA hybrid films decreased when the loading level of MV²⁺ was increased, in spite of the same amount of MV²⁺. The formation of H-type aggregates, which is one cause of fluorescence quenching, was not observed judging from the above-mentioned XRD and absorption spectra analyses. These results indicate that the fluorescence of MV²⁺ was quenched via a self-fluorescence quenching reaction. The occupied surface area of an anion on SSA was calculated as 1.25 nm², based on the theoretical surface area and the cation exchange capacity. On the other hand, assuming that MV²⁺ is rectangular (dihedral angle of pyridinium substrate is 0°), the cross-sectional footprint of MV²⁺ would be 0.844 nm²/molecule (viewed from the normal direction of the pyridinium group (Raupach et al. 1979). Calculating from these values, the occupancy ratio of the SSA surface by MV²⁺ was 50% (at 75% CEC). Judging from the result of this calculation, MV²⁺ should be intercalated into SSA without aggregation.

Generation of MV⁺• with UV Irradiation in the Interlayer of Saponite

The UV-Vis absorption spectra of MV²⁺/SSA hybrid films (75% CEC) changed under UV irradiation at 270 nm (Fig. 5); the absorbance of MV²⁺ decreased quickly under UV irradiation, and new absorption bands appeared at ~400 and ~600 nm. The newly appeared absorption bands at ~350–420 nm, the broadened band at ~450–750 nm, and their characteristic vibrational structures were almost the same as in the reported absorption spectrum of MV⁺•, which is the one-electron reduced species of MV²⁺ (Watanabe & Honda 1982; Bockman & Kochi 1990). In the photograph of the MV²⁺/SSA hybrid film after UV irradiation (Fig. 6), the color of the film changed from colorless to blue, the latter being the typical color of MV⁺• (Palenzuela et al. 2014). In addition, this absorption peak disappeared upon exposure to air, and then the absorption band of MV²⁺ increased. This result also supports the conclusion that the newly generated species was MV⁺• because it reacted with oxygen in air to produce MV²⁺. The peak absorption wavelengths of MV⁺• in the interlayer space of SSA were 398 and 608 nm. Interestingly, λ_max of MV⁺• was shifted slightly to a longer wavelength (by 2 nm) compared to that in water, although the λ_max of MV²⁺ was shifted to an even longer wavelength (by >10 nm) by intercalation into the interlayer space of SSA. These observations could be ascribed to the different stable structures of MV²⁺ and MV⁺• in water. In the case of MV²⁺, the pyridinium groups are not co-planar in water because of the steric effect, while the stable structure of MV⁺• is approximately co-planar (Bockman & Kochi 1990; Kubota 1999; Porter & Vaid 2005). Thus, the λ_max of intercalated MV⁺• in SSA was almost the same as in water, because the effect of co-planarization by intercalation is relatively weak compared to that for MV²⁺. MV⁺• was not generated without UV light irradiation, indicating that the MV²⁺ intercalated in SSA was reduced by UV irradiation.

Fig. 5. UV-Vis absorption spectra of MV²⁺/SSA hybrid film during UV-irradiation (concentration of MV²⁺: 75% CEC) at 0 min (gray solid line), 1 min (black dotted line), 3 min (black broken line), and 5 min (black solid line)

Fig. 6. Photograph of MV²⁺/SSA hybrid film after UV light irradiation

Several reports have discussed the photoreduction of MV²⁺ with UV irradiation in several media, such as water, organic solvents, and solid surfaces, as mentioned above. The electron donor to reduce the MV²⁺ via a photoinduced electron transfer reaction was discussed in those reports. Although some reports indicated that the chloride ion is the electron donor to reduce MV²⁺* (Ebbesen et al. 1984; Peon et al. 2001), MV²⁺ adsorbs on SSA by electrostatic interaction and, thus, few chloride ions should be present in the hybrid film. Kohler and co-workers (Peon et al. 2001) reported that gas-phase water has a high ionization potential (12.62 eV), and they did not observe the quenching of MV²⁺* via electron transfer. Considering the previous reports, the residual interlayer water should not be the electron donor. In aluminosilicate media (clays and zeolites), some reports (Alvaro et al. 1997; Kakegawa et al. 2003) suggested that the bridging Si-O-Al oxygen lone pair can donate an electron to MV²⁺*. To confirm this electron transfer pathway, MV²⁺/SSA hybrid films with very low adsorption density of MV²⁺ (0.13% and 0.06% CEC) were prepared to avoid the self-fluorescence quenching reaction. The fluorescence intensities of these films were approximately the same in spite of their different loading levels of MV²⁺. Thus, the self-quenching reaction was absent in these films (Fig. S3). The UV-Vis absorption spectra of MV²⁺/SSA hybrid film (MV²⁺ loading level of 0.13% CEC) varied during UV irradiation (Fig. S4). The absorbance ascribed to MV²⁺ decreased with UV irradiation because of decomposition after prolonged UV exposure, and a small absorption band was observed at ~400 nm. The absorption peak at ~600 nm attributed to MV⁺•, however, was not observed, and the existing peak did not disappear in the presence of oxygen. These observations suggested that MV⁺• was not generated because it would be quenched by oxygen molecules. This absorption band was similar to that of the decomposition product of MV²⁺ (Solar et al. 1982; Bahnemann et al. 1987). Hence, the new species was not MV⁺• but a decomposition product. All these results, taken together, indicate that the electron transfer from the bridging Si-O-Al oxygen lone pair to MV²⁺* did not take place. Considering these results and the dependency of MV⁺• photogeneration on the adsorption density of MV²⁺ (as described below), the photoinduced electron transfer reaction between excited MV²⁺ and adjacent ground-state MV²⁺ would be one of the principal pathways of the self-quenching reaction. An electron transfer reaction between electron donor and acceptor depends on their redox potentials and re-organization energy, requiring a difference between the reduction potential of the acceptor and the oxidation potential of the donor. In this case, the redox potential of MV²⁺ would be affected by the negative charges of SSA and co-planarization of the dye molecules, both of which will depend on the distance between the anionic site on SSA and the cationic site of MV²⁺. Thus, a difference in the adsorption site might change the redox potentials slightly, thereby allowing the photoinduced electron transfer reaction to happen.

Although a back electron transfer reaction between MV⁺• and oxidized MV²⁺ was expected, MV⁺• could be observed by the typical steady-state UV-Vis spectrophotometer. The reason is that the one-electron oxidized MV²⁺ will be unstable and, thus, undergo a decomposition reaction, and then the back electron transfer may be suppressed.

The generation of MV⁺• with UV irradiation in MV²⁺/SSA hybrid films at various adsorption densities of MV²⁺ is summarized in Fig. 7. The amount of MV⁺• generated was calculated from Δabs. = abs_t – abs₀ and the extinction coefficient of MV⁺• in water (Watanabe & Honda 1982). The amount of MV⁺• generated depended on the adsorption density of MV²⁺, and reached a plateau in all hybrid films. The maximum conversion ratio from MV²⁺ to MV⁺• (generated MV⁺• / initial MV²⁺) was 0.144 (75% CEC), although the maximum value of the conversion ratio will be 0.5. This is because MV²⁺* may be quenched by the MV⁺• generated and decomposition products. According to the quantum yield (φ) at 1 min of the photogeneration of MV⁺• (Table 1), the photogeneration efficiency of MV⁺• increased with increasing adsorption density of MV²⁺, indicating that the self-fluorescence quenching reaction is important for generating MV⁺•. φ drastically increased at large dye loadings (>60% CEC), suggesting that the roaming range of MV²⁺* within the excited lifetime will be limited, because MV²⁺ molecules have to approach each other to some extent.

Fig. 7. Generation of MV⁺• with UV irradiation at different MV²⁺ loading levels: 75% (open circle), 60% (solid circle), 40% (open square), 20% (solid square), and 10% (solid triangle) CEC

Table 1. Quantum yields (φ) of the MV⁺• photogeneration in hybrid films

Conclusions

The generation of MV⁺• in semi-transparent MV²⁺/SSA hybrid films under UV irradiation was observed clearly by transmission spectra. The efficiency of MV⁺• generation depended on the adsorption density of MV²⁺. When the adsorption density of MV²⁺ was low, where a self-quenching reaction was not observed, MV⁺• was not generated. This indicates that MV⁺• was produced via electron transfer between MV²⁺* and MV²⁺ on the saponite surfaces.

Electronic supplementary material

The online version of this article (https://doi.org/10.1007/s42860-019-00047-8) contains supplementary material, which is available to authorized users.