Pyridyl-Anchored Type BODIPY Sensitizer-TiO 2 Photocatalyst for Enhanced Visible Light-Driven Photocatalytic Hydrogen Production

: Dye-sensitized photocatalytic hydrogen production using a boron-dipyrromethene (BODIPY) organic material having a pyridyl group at the anchor site was investigated. Phenyl, carbazole, and phenothiazine derivatives were introduced into BODIPY dyes, and their photocatalytic activities were examined. Identification was performed by nuclear magnetic resonance (NMR), infrared (IR), mass (MS) spectra, and absorption spectra, and catalyst evaluation was performed by using visible-light irradiation and photocatalytic hydrogen production and photocurrent. These dyes have strong absorption at 600–700 nm, suggesting that they are promising as photosensitizers. When the photocatalytic activity was examined, stable catalytic performance was demonstrated, and the activity of the Pt-TiO 2 photocatalyst carrying a dye having a carbazole group was 249 μmol/g cat ·h. Photocurrent measurements suggest that dye-sensitized photocatalytic activity is occurring. This result suggests that BODIPY organic materials with pyridyl groups as anchor sites are useful as novel dye-sensitized photocatalysts.


Introduction
Research on new energy carriers that can replace fossil fuels is attracting attention. As a candidate, hydrogen can be compressed, stored, and transported, and it has a sufficient carrier density [1]. A hydrogen fuel cell is a CO2-free system that only discharges water when extracting electricity. On the other hand, in the fossil fuel combustion method and methane steam reforming method, which are the mainstays of the current hydrogen production methods, the use of a carbon material source as a hydrogen source becomes a problem, and the production process produces carbon dioxide [2]. From this perspective, attention has been paid to hydrogen production by water splitting, using a hydrogen generation that is not derived from carbon source [3]. Photocatalytic water splitting is attracting attention as the most ideal hydrogen production method among water-splitting reactions [4]. In order to obtain high photocatalytic performance, research has been conducted on the construction of a high photo-electron conversion system with a small band gap, improvement of the reaction active site by changing the shape of the catalyst, and improvement of the efficiency of the water-splitting reaction by extending the life of charge separation [5]. On the other hand, research by half-reaction of only hydrogen/oxygen sites is also important research for improving each reaction activity [6]. Among such research, there is a visible-light-responsive photocatalyst, using visible light, which is a component of 50% of sunlight. However, there are not many reports of highly efficient and stable visible-light-responsive photocatalysts [7]. Dye-sensitized photocatalysts can support various dye sensitizers on a stable photocatalyst and promise to develop a stable photocatalyst system, using the desired visible light [8]. Porphyrin is one of the dyes having a strong molar extinction coefficient in the visible-light region, and many studies have been reported [9]. However, this dye has a strong Soret band at 400-500 nm, and it is desired to develop a dye that can use the longer wavelength side. One such dye is boron-dipyrromethene (BODIPY), which has half the structure of porphyrins. It is known that BODIPY has an absorption band with a strong molar extinction coefficient at 500-900 nm [10] and has an electrochemically stable structure, and a new method for synthesizing BODIPY [11] and its related compounds [12], semiconductors and OPV [13], and biotechnology [14]. In recent years, BODIPY has attracted attention as a skeleton of photocatalysts. For hydrogen production, there have been reports on the introduction of molecular catalysts [15][16][17] and MOFs [18,19] into partial skeletons, dye-sensitized oxidation [20], CO2 reduction [21], and other photocatalytic reactions [22][23][24][25][26][27]. Kubo et al. have developed a donor-acceptor dye by using π-extended BODIPY as the photodonor site and have developed it into a dye-sensitized TiO2 photocatalyst to achieve efficient photocatalytic hydrogen production [28]. An efficient photosensitization path by introducing a donor-acceptor structure into BODIPY enables the introduction of chromophores at the α, β, and meso positions [10,13]. At this time, an electron donor site can be introduced at the α and β positions, and an electron acceptor site can be introduced at the meso position [13]. The development of an efficient dye-sensitized solar cell using this strategy has been reported [29,30]. In this study, we developed a novel BODIPY dye with a pyridyl group at the α-position as an electron donor and an electron acceptor at the meso position ( Figure 1). We have reported that these are titanium oxide photocatalysts and work as efficient and highly stable photosensitizers.

Physical/Theoretical Properties of BODIPY Dyes B, B-Ph, B-Car, and B-Pheno
Absorption spectra of synthesized BODIPY B, B-Ph, B-Car, and B-Pheno were measured in THF solutions ( Figure 2a). The lowest absorption peak of B shows at 503 nm with high molar extinction coefficient (ε = 52300). On the other hand, π-expanded BODIPY system B-Ph, B-Car, and B-Pheno show bathochromic shift and a change in molar express factor, B-Ph was 650 nm (ε = 141000), B-Car was 676 nm (ε = 36000), and B-Pheno was 699 nm (ε = 21000). Among the π-extended molecules, the phenyl group-introduced molecule had the smallest bathochromic shift but a high molar extinction coefficient, while the phenothiazine-group-introduced molecule had a large bathochromic shift but a small molar extinction coefficient. The highest occupied molecular orbitals (HOMOs) and lowest occupied molecular orbitals (   The dihedral angle between the BODIPY core and the pyridyl group of B is 90 degrees, which means that the conjugation between these two parts is very weak. In B-Ph, B-Car, and B-Pheno, the dihedral angle between the BODIPY core and the pyridyl group are 89.5-90 degrees, as in B. On the other hand, the dihedral angle between the BODIPY core and the π-conjugated aromatic component is 3.9 degrees for B-Ph, 9.4 degrees for B-Car, and 7.9 degrees for B-Pheno. Thus, the dihedral angle between BODIPY and aromatic at B-Ph, B-Car, and B-Pheno is very small, and an efficient π-conjugate interaction is expected. Figure 3 (right) shows HOMO and LUMO orbitals of B, B-Ph, B-Car, and B-Pheno. BODIPY B of HOMO and LUMO are delocalized on dipyrrolomethane moiety, while no HOMO and LUMO orbitals on pyridyl group due to weak conjugation between BODIPY and pyridyl groups. In B-Ph, B-Car, and B-Pheno, HOMO and LUMO orbitals were delocalized to the extended aromatic moieties in addition to the BODIPY core, supporting the absorption spectra of bathochromic shifts and electrochemical measurements of lower HOMO-LUMO gaps. From the overlap of HOMO and LUMO on the BODIPY and aromatic core in B, B-Ph, B-Car, and B-Pheno, assigned to the π-π* transition type interaction. B, B-Ph, B-Car, and B-Pheno were loaded to the catalyst by suspending 0.75wt% Pt-TiO2 in THF solutions of the dyes. Figure 4 illustrates the reflectance spectra of B, B-Ph, B-Car, and B-Pheno/Pt-TiO2. The loading amounts of the sensitizers on the catalysts were subsequently assessed. The amount of loaded dye was estimated based on the remaining amount of the THF solution, which was separated from the photocatalyst by centrifugation. Accordingly, the loading of the dye in the B/Pt-TiO2 system was determined at 17.8 μmol/g, while in B-Ph/Pt-TiO2, it was 10.0 μmol/g. Moreover, the sensitizer loading in B-Car/Pt-TiO2 was established at 7.5 μmol/g, whereas, in the B-Pheno/Pt-TiO2 system, it was estimated at 6.8 μmol/g. These values are consistent with the dye-loading amounts previously reported for a porphyrin-pyridine pendant system [31]. Notably, the loading amount of dye B on the catalyst was the largest. This is because B-Ph, B-Car, and B-Pheno contain bulky alkyl groups, leading to alkyl chain packing on TiO2. The lowest energy peaks for B, B-Ph, B-Car, and B-Pheno on Pt-TiO2 appeared at 516, 689, 709, and 692 nm, respectively. The B, B-Ph, and B-Car dyes exhibited a peak broadening and a red shift. The adsorption modes of all the dyes and TiO2 are considered to be the same Lewis acid interaction (discussed later), suggesting that there are no differences in the adsorption modes. It is speculated that the reason for the peak shift is the formation of J-aggregates on titanium oxide by the B, B-Ph, and B-Car dyes. On the other hand, owing to the presence of a more flexible phenothiazine moiety, which is involved in a weak electronic interaction between the B-Pheno dye and titanium oxide, no significant change in the absorption spectrum was observed for B-Pheno/Pt-TiO2.  To confirm the adsorption of the dyes on TiO2, we measured the IR absorption spectra ( Figure  5). The bulk powders of the B, B-Ph, B-Car, and B-Pheno dyes were analyzed, and vibration modes corresponding to the pyridyl moiety appeared at 1600 cm -1 for B, 1608 cm -1 for B-Ph, 1612 cm -1 for B-Car, and 1596 cm -1 for B-Pheno. Pt-TiO2 exhibited an absorption at 1647 cm -1 , which belonged to the vibration mode of adsorbed water. For B, B-Ph, B-Car, and B-Pheno/Pt-TiO2, the peak at 1647cm -1 shifted, and the appearance of peaks at 1631 and 1442 cm -1 (B/Pt-TiO2), 1635 and 1438 cm -1 (B-Ph/Pt-TiO2), 1631 and 1434 cm -1 (B-Car/Pt-TiO2), and at 1635 and 1446 cm -1 (B-Pheno/Pt-TiO2) was observed. When the pyridyl anchor moiety was changed to a phenyl group, the Pt-TiO2 loading test demonstrated that compound B-ref did not show this characteristic peak in the reflectance spectra (Supplementary Figure S1a,b). However, a peak at 1542 cm -1 was observed in the IR spectrum of the B-ref. Following the Pt-TiO2 loading test, the peak at 1644 nm -1 did not shift, indicating that the dye was not adsorbed on the TiO2 surface in the absence of a pyridyl moiety (Supplementary Figure S1c). The pyridyl group is known as an anchor site to titanium oxide [32]. The observed peaks were attributed to adsorption involving a Lewis-acid-type coordination bond to titanium oxide [33]. This implies that the pyridyl moiety is adsorbed on the titanium oxide surface via a N-Ti coordinate-type bond between the nitrogen atom of the pyridyl group and metal in the TiO2. The physical and theoretical properties of the BODIPY dyes, i.e., B, B-Ph, B-Car, and B-Pheno, are summarized in Table  1.

Photocatalytic Reaction
The visible-light-driven photocatalytic hydrogen production in water medium using B, B-Ph, B-Car, and B-Pheno/0.75wt%Pt-TiO2 nanoparticles was investigated. The dye-loaded 0.75 wt% Pt-TiO2 catalyst powder was suspended in water, with ascorbic acid (pH 4) acting as a sacrificial reagent. This suspension was purged with argon, and a Xe lamp (>420 nm filter) was used to irradiate the system. Further, the amount of hydrogen gas output was measured by using the thermal conductivity detector-gas chromatography (Figure 6a). When the hydrogen production reaction of B, B-Ph, B-Car, and B-Pheno/0.75wt%Pt-TiO2 were tested in dark condition, no hydrogen production was confirmed (Supplementary Figure S2). In this experiment, 0.75wt% Pt-TiO2, which does not carry a dye, hardly generated hydrogen (8.0 μmol/gcat·h) because it cut off ultraviolet light (> 420 nm). On the other hand, in the case of dye B, B-Ph, B-Car, and B-Pheno/0.75wt% Pt-TiO2, the generation of hydrogen under visible-light irradiation increased. These results suggest that the dye acted as a sensitizer under visible-light irradiation and improved the photocatalytic ability. The hydrogen production rates were in the order of B-Car/0.75wt% Pt-TiO2 (249 μmol/gcat·h) > B-Pheno/0.75wt% Pt-TiO2 (235 μmol/gcat·h) > B/0.75wt% Pt-TiO2 (187 μmol/gcat·h) > B-Ph/0.75wt% Pt-TiO2 (97 μmol/gcat·h). Figure 6b demonstrates the hydrogen production efficiency as a turnover number (TON), which was used to evaluate the catalytic reactivity of dye-sensitizer B, B-Ph, B-Car, and B Figure S3b). We considered that the differences in the orientations of B-Car/Pt-TiO2 and B-Pheno/Pt-TiO2 before and after the reaction were small, because both B-Car and B-Pheno exhibit bulkier structures, enabling the formation of optimal packing structures on titanium oxide [34]. To further confirm the stability of the photocatalyst, we conducted a B-Car/Pt-TiO2 recycle test (Figure 7). The photocatalytic reaction involving B-Car/Pt-TiO2 was repeated five times. After five sequences lasting 24 h each (120 h in total), the dye remained active, suggesting that the catalyst was stable and reusable.  To elucidate the hydrogen production mechanism, the photocatalytic reaction of B-Car/TiO2 with and without a UV cut filter (>420 nm) was investigated. No hydrogen was detected under dark conditions (Supplementary Figure S2). On the other hand, the production of hydrogen was evident under visible light conditions ( Figure 6 and Supplementary Figure S4). It is noteworthy that the photocatalytic reaction without a >420 nm filter enhanced the photocatalytic activity. This outcome implied that the developed photocatalyst was effective in the visible-light-driven photocatalytic reactions. Conversely, the photocatalytic reaction with B-Car/TiO2 in the absence of Pt did not produce hydrogen under visible light, demonstrating the critical role of the Pt cocatalyst in this photocatalytic system (Supplementary Figure S4).
To further evaluate the visible light-driven photo-induced catalytic reactions involving the sensitizers, fluorescence decay spectra of B, B-Ph, B-Car, and B-Pheno were measured.    [35]. However, all dyes showed an unstable character after the photocurrent spike; the decreasing of the photocurrent rapidly under photo illumination indicated that a fast charge-recombination process occurred between dyes and TiO2 surfaces [36]. The order of photocurrent follows the same tendency as the hydrogen production rate or TOF, as the suggested number of excited photons played a critical role in the production of hydrogen in the B-, B-Ph-, B-Car-, and B-Pheno-loaded photocatalyst system. As a result of DFT calculation, the orbits of all the dyes were delocalized in the BODIPY skeleton for both HOMO-LUMO, and the pyridyl group as the anchor site did not have orbitals. From the result of IR spectrum, the dye was adsorbed on titanium oxide via the pyridyl group. Furthermore, the fluorescence lifetime spectrum suggested that electron transfer occurred from the dyes to titanium oxide. The fact that all dyes produce higher photocurrents than titanium oxide alone indicates that efficient through-space-type charge transfer from the sensitizer to titanium oxide occurs under visible light, resulting in hydrogen production. The LUMO potential is more negative than TiO2 CB (−0.50 V vs. NHE) and provides sufficient driving force for electron injection from excited state B, B-Ph, B-Car, and B-Pheno into the conduction band of TiO2. The injected electrons move to the cocatalyst of platinum (Pt) and react with protons (H + ) to generate hydrogen. In this reaction, the BODIPY dye becomes an oxidized dye after injecting electrons into titanium oxide in the excited state. As an electron donor, ascorbic acid (AA) (Figure 10b, + 0.46 V vs. NHE) transfers electrons to the oxidized dyes and converted to dehydroascorbic acid (DHA) (Figure 10b) [37]. Correspondingly, oxidized dyes are reduced by accepting electrons from AA, affording BODIPY dyes, which exhibit catalytic activity. During a photocatalytic reaction, the photo-generated chemical species include the oxidized dye, hydrogen, and DHA. Thus, if the oxidized dyes react with the BODIPY dyes prior to receiving electrons from AA, the photocatalytic activity should be suppressed. Nonetheless, the recycle test shown in Figure  7 demonstrated that the hydrogen production rate does not change, and the catalyst retains its stability. The same photocatalytic activity was determined, meaning that the oxidized dye rapidly received electrons from AA and returned to the BODIPY form. To confirm the reactivity of these photo-generated species, we carried out a 1 H nuclear magnetic resonance (NMR) analysis of B-Car in the presence of hydrogen gas and the aqueous solution after the photoreaction (i.e., the DHA solution [37]). The obtained 1 H NMR results showed no changes in the conducted experiment (Supplementary Figure S6). Hence, the dye was determined to be stable to photogenerated hydrogen and DHA. It was also confirmed that, as an electron donor, the oxidized dye transferred electrons to titanium oxide in the photoexcited state and reacted with AA, resulting in a stable catalytic cycle. This phenomenon leads to charge separation and photocatalytic hydrogen generation observed in the dye/Pt-TiO2-AA system.

Materials and Methods
TiO2, (Sigma-Aldrich, Japan, anatase nanopowder, <25 nm particle size, 99.7% trace metals basis, surface area 45-55 m 2 /g) and chloroplatinic acid hexahydrate (Sigma-Aldrich, Japan, 8 wt.% in H2O) were used as the starting materials. Photodeposition in methanol was used to obtain 0.75wt% Pt-TiO2. 0.75wt% Pt-TiO2 (100 mg) was suspended in a THF solution of B, B-Ph, B-Car, and B-Pheno and allowed to stand at room temperature for 26 h, in the dark. The amount of loaded dye was estimated based on the remaining amount of the THF solution, which was separated from the photocatalyst by centrifugation. The photocatalytic hydrogen production reaction was performed by utilizing a quartz-closed reactor circulating system equipped with a thermal conductivity detector (Ar-carrier, column: molecular sieve 4Å, Shimadzu, Japan, GC-8A). The aqueous reaction solution contained the catalyst powder (20 mg), DI water, and 0.56 M AA as the sacrificial reagent (18 mL, pH 4 adjusted by NaOH). The photocatalytic reaction was performed by using a solar simulator (0.23 mW/cm 2 , HAL-302, ASAHI spectra, Japan, 0.15 W/cm 2 after a >420 nm filter). The absorption/reflectance spectra were obtained by using a Shimadzu UV-3600 spectrophotometer. Cyclic voltammetry and photocurrent spectra were recorded on an electrochemical analyzer (ALS Co., Ltd, Japan, BAS ALS-1200C). IR spectra were tested by ATR-IR (Thermo Scientific, Japan, Nicolet iN10 MX). The fluorescence decay spectra were measured by employing a fluorescence lifetime spectrofluorometer (HORIBA, Japan, TemPro). The utilized excitation laser was 494 nm nanoLED for

Conclusions
In conclusion, new absorption was observed in the range from 400 to 700 nm by modifying B, B-Ph, B-Car, and B-Pheno to TiO2. The lowest absorption peak of B shows at 503 nm with a high molar extinction coefficient (ε = 52300). On the other hand, π-expanded BODIPY system B-Ph, B-Car, and B-Pheno show bathochromic shift and a change in the molar extinction coefficient, B-Ph was 650 nm (ε = 141000), B-Car was 676 nm (ε = 36000), and B-Pheno was 699 nm (ε = 21000). From the results of CV and DFT, the dye HOMO-LUMO was localized in BODIPY and the extended aromatic compound, and the pyridyl group had no charge. This suggests that the light injection from the dye into the titanium oxide is a through-space-type injection. When conducting a photocatalytic reaction with B-Ph, B-Car, and B-Pheno/0.75wt% Pt-TiO2, more photocatalytic activity than 0.75wt% Pt-TiO2 was confirmed in all modified organic dye B-Ph, B-Car, and B-Pheno. The hydrogen production rates of orders were B-Car/0.75wt% Pt-TiO2 (249 μmol/gcat·h) > B-Pheno/0.75wt% Pt-TiO2 (235 μmol/gcat·h) > B/0.75wt% Pt-TiO2 (187 μmol/gcat·h) > B-Ph/0.75wt% Pt-TiO2 (97 μmol/gcat·h). The photocurrent of intensity order showed B-Car/0.75wt% Pt-TiO2 > B-Pheno/0.75wt% Pt-TiO2 > B/0.75wt% Pt-TiO2 > B-Ph/0.75wt% Pt-TiO2 > 0.75wt% Pt-TiO2. This order shows a similar tendency with hydrogen production rate, as the suggested number of excited photons played a critical role in the production of hydrogen in the B-Ph-, B-Car-, and B-Pheno-loaded photocatalyst system. When the photocatalytic activity was examined, stable catalytic performance was demonstrated, and the activity of the Pt-TiO2 photocatalyst carrying a dye having a phenyl group was 400 μmol/h. Photocurrent measurements suggest that dye-sensitized photocatalytic activity is occurring. This result suggests that BODIPY organic materials with pyridyl groups as anchor sites are useful as novel dye-sensitized photocatalysts.