Improved charge transfer performance of eosin Y-sensitized anatase TiO₂ by anchoring group modification: from theoretical design to experiment

Abstract

Theoretical design and experimental proof of photocatalytic performance of eosin Y (EY) on anatase TiO₂ with a pyridine linker were obtained for increasing the photobiocatalytic activity of water splitting using visible light. Comparative studies on the hybrid interface of anatase and EY with carboxyl and pyridine anchors were performed by using density functional theory (DFT), time-dependent density functional theory (TD-DFT) calculations and experimental photoreduction of methyl viologen (MV). The geometries, binding interactions between dyes and anatase, electronic structures and electron transfer as well as the effect of isomers (ortho, meta, and para) on the dye/anatase systems were investigated. Theoretical results indicated that EY with carboxyl and pyridine anchors had visible absorption and electron transfer from the dye to the anatase titania. Compared to carboxyl-para, which had the best optical performance among carboxyl groups, the adsorption strength of pyridine-ortho was close to that of carboxyl-para, while the oscillator strength increased significantly, which was more than 10 times higher than that of carboxyl-para. Corresponding with the theoretical estimation, EY pyridine linked TiO₂ is active to MV reduction under visible light irradiation by fast charge transfer, in particular, pyridine-ortho and para. Furthermore, high stability is also achieved for pyridine-para. An apparent quantum yield higher than 2.00% and 0.67% under 520 nm light was experimentally achieved in biocatalytic H₂ and NH₃ formation, respectively, for EY pyridine linked TiO₂, which was correctly predicted by the DFT calculations.

---

1. Introduction

Energy transition, which involves reducing fossil fuel utilization and promoting the development of clean energies, is a crucial issue to mitigate the negative effect of climate change and to pursuit a carbon-neutral society. Hydrogen is considered to be an ideal energy carrier due to its advantages of high specific energy density, purity, versatility, possibilities for renewable production, etc.^1,2 As a renewable secondary energy, hydrogen can be generated by many low-carbon-emission pathways such as electrolysis, electro-photolysis, radiolysis, thermo-chemical, etc.^3,4 In particular, solar-driven hydrogen production has gained significant attention because it provides a promising opportunity for renewable hydrogen energy, the so-called green hydrogen.

Since the photoelectrolysis method was proposed by Fujishima and Honda in 1972,⁵ researchers have steadily increased the attempts to directly harvest solar energy for efficient water electrolysis.^6,7 In photoelectrochemical water splitting systems, photocatalysts which use sunlight energy to dissociate water molecules into hydrogen and oxygen play a key role in the efficiency of hydrogen production.^8,9 Anatase has been one of the most studied photocatalysts because of its appropriate position of the conduction band (CB) edge (lower than the reduction energy of water, 0 V) and valence band (VB) edge (higher than the oxidation energy of oxygen, 1.23 V) to trigger the photoelectrochemical reaction.^8–10 However, the band gap of anatase is ca. 3.20 eV which corresponds to ultraviolet (UV) light (only 3–4% of the solar energy).¹¹ Hence, it is important to narrow the band gap of anatase and to enhance its visible light response and photocatalytic activity.

To date, many modifications have been developed to improve the optical absorption of anatase including dopant introduction and dye sensitization. The dopants could be divided into metal and non-metal elements where metals were extensively investigated to act as co-catalysts to increase solar absorption. Some metals such as Au, Pt, Pd, Rh, Ni, etc., have been found to improve the photo electron–hole separation by trapping electrons at the interfaces of metal and anatase.^10,12,13 For example, Murdoch et al.¹² found that Au-doped anatase showed a broader visible light absorption range and a high hydrogen production rate. This was because the CB of Au was lower than that of anatase; hence, the excited electrons accumulated at anatase could easily transfer to Au. The hydrogen ions trapped by Au at the interface with anatase enhanced the hydrogen molecule reduction activity. On the other hand, Au, Ni, Cu, Al, Ag, etc., were reported to promote visible light absorption due to the surface plasmon resonance (SPR) effect and localized surface plasmon resonance (LSPR) effect.^9,14–17 These effects could increase the energy intensity of photo-excited electrons under visible light irradiation. Many studies of anatase doping with metal ions, Fe, Mo, Ru, Os, etc., and non-metal anions, N, F, P, S, Si, etc., attempt to modify the inherent bandgap of anatase TiO₂.^10,18–22 Metal ion-doping could form impurity energy levels within the anatase bandgap, while anion-doping could shift the VB upwards and cause a narrowed bandgap of anatase. However, the above methods may have some disadvantages such as the VB shifting upwards could have a detrimental effect on the oxidation activity. Metal-ion dopants may become recombination centers for electrons and holes, which have a negative effect on the photocatalytic water-splitting reaction.¹⁸

In a previous study, combination of an inorganic photocatalyst with hydrogenase coupled with MV as a redox mediator was studied. The efficiency of hydrogenase genetically formed in E. coli is 100% per transferred electron and so the total apparent quantum yield (AQY) of TiO₂/MV²⁺/hydrogenase in E. coli is higher than 30% in the presence of the sacrificial agent of triethanolamine (TEOA). However, TiO₂ can be excited by only UV light (<350 nm). Although a high AQY is achieved, the solar-to-hydrogen efficiency of the photobiocatalytic system using TiO₂ as the inorganic photocatalyst was reduced.²³

Anchoring of dye-sensitizers to anatase to improve the photocatalytic performance has attracted great attention in recent years. The dye could be excited upon visible light absorption and then excited electrons could transfer from the dye to the CB of anatase. As a result, dye-sensitized anatase can be excited in the visible light region. Organic dyes are currently attracting much interest because of their active visible-light responses, flexible structures, economical and eco-friendliness. Some organic dyes were examined for their light-harvesting capabilities and electron transfer properties such as perylene, cyanine, riboflavin, eosin Y (hereafter referred to as EY) and so on.^8,24–27 Among them, EY has been found to be an effective photosensitizer.^30,31 The Raman spectrum of bare anatase and EY adsorbed TiO₂ indicated that EY could effectively improve the intensity and the range of light absorption on anatase.²⁹ The EY-sensitized anatase also exhibited a high photo energy conversion efficiency. This is mainly because of the strong ester linkage formed between the carboxyl anchoring group of EY and anatase, which allows efficient electron injection from EY to anatase. However, this linkage was not stable in water due to hydrolysis and easily desorbed from TiO₂ and deactivated.^8,30

The anchoring group has been confirmed to influence the interaction between dyes and substrates. Several experimental and theoretical studies reported that phosphonic acid is a potential alternative anchoring group.^28–30 It has a strong binding interaction with anatase, showing better long-term stability. Compared to the carboxylic acid anchors, the most prominent benefit of phosphonic acid is the greater adsorption strength, whereas the electronic properties did not improve significantly. The pyridine group was found to be an efficient anchor for the charge injection from fluorescent dyes to anatase.³¹ The coordination bond between the pyridine ring and the Lewis acid site behaved as unique intermolecular charge transfer media. Moreover, the pyridine ring exhibited high adsorption stability in water, enabling it to function effectively over extended periods in water-splitting systems. It is necessary to consider both the binding strength and the electronic properties of the anchoring group in order to design a high-performance water-splitting system. Thus, we proposed the use of a pyridine ring as an anchoring group in place of the conventional carboxyl group to connect the EY and anatase. Since the position of the anchor attached to the molecule plays an important role in electron transfer, the isomer (ortho-, meta-, and para-) effect was considered in our theoretical study. The structures of the dye EY investigated in this work are illustrated in Fig. 1. Comparative studies on the hybrid interfaces of anatase and eosin Y with the different anchoring groups were performed by DFT and TD-DFT calculations. The geometries, binding interactions between dyes and anatase, electronic structures and electron transfer as well as the isomer effect on the dye/Ti₇₈O₁₆₀H₈ clusters were investigated and discussed. In addition, electron transport calculations were conducted using non-equilibrium Green's function at the DFT level of theory (NEGF-DFT) on the pyridine-based and the carboxyl-based molecular junctions.

Fig. 1 Molecular structures of the investigated eosin Y isomers anchored with the conventional carboxyl group and modified pyridine group.

Synthesis of modified eosin Y was performed by using bromobenzaldehyde and pyridin-4-ylboronic acid with a catalytic amount of tetrakis(triphenylphosphine)palladium. Dye-sensitized TiO₂ (dye/TiO₂) as a photocatalytic material was prepared. Photocatalytic reduction of methyl viologen was performed to determine the light-absorbing capacity of photocatalytic materials. Photobiocatalytic H₂ production of dye/TiO₂ coupled with E. coli was performed. Finally, we demonstrated the photobiocatalytic NH₃/H₂ production using dye/TiO₂ coupled with cyanobacteria.

2. Computational and experimental methods

2.1. Computational methods

Geometry optimizations of the isolated EY molecular structures shown in Fig. 1 and their complexes with a charge neutral Ti₇₈O₁₆₀H₈ cluster were performed by the DFT method with the PBE exchange–correlation functional.³² Based on the optimized structures of dye/anatase, we studied the binding energies (which was the difference between the total energy of dye/anatase, isolated dye and isolated anatase) and the optical properties of the six complexes, by performing TD-DFT calculations with the hybrid B3LYP functional.^33,34 The Karlsruhe def2-SV(P)³⁵ basis set was applied for all atoms and N-1 Hay–Wadt³⁶ effective-core potential was employed for the titanium atoms. Based on this method, the optimized lattice properties of anatase crystal were a = b = 3.78 Å and c = 9.51 Å, close to the experimental values of a = b = 3.79 Å and c = 9.51 Å.³⁷ The band gap of anatase calculated using the periodic boundary condition (B3LYP functional and pob-TZVP basis set) was estimated to be 3.28 eV, which was in good agreement with the experimental values of 3.20 eV.^38,39 All the dyes were adsorbed on the fivefold coordinated titanium atoms, which was the Lewis acid site with high reactivity.^33,40 The above calculations were implemented in the Turbomole 7.7 program.⁴¹

To elucidate the mechanism of enhanced photocatalytic activity of anatase/EY-pyridine, electron transport calculations were performed by the NEGF-DFT method,⁴² implemented in the QuantumATK package.⁴³ Local density approximation (LDA)⁴⁴ with the FHI⁴⁵ pseudopotential was employed for the calculations. A single zeta-polarized basis set was chosen for the titanium atoms and a double zeta-polarized basis set was chosen for the other atoms.⁴⁶ The electric current and transmission spectra at the carboxyl-linked molecular junction and the carboxyl-linked molecular junction, as shown in Fig. 2, were quantitatively investigated in this work. The electrode was modelled by the optimized anatase (001) – (4 × 4) surface with one O–Ti–O layer. For the central scattering region, it was composed of the investigated anchoring group (i.e., carboxyl group and pyridine group), benzene and three layers from each (i.e., three O–Ti–O layers). We used the simplest aromatic hydrocarbon benzene ring as the connection medium to investigate the electric transport through anatase and the different anchors. During the calculations, only the two top O–Ti–O layers as well as the investigated molecules in the scattering region were relaxed, while the two layers on each side in the scattering region and the two electrodes were frozen.

Fig. 2 Structures of the carboxyl-linked device (upper panel) and the pyridine-linked device (lower panel). In all systems, the symmetrical electrodes had a length of 5.55 Å, while the length of the central region in the carboxyl-linked device and the pyridine-linked device was 30.02 Å and 33.73 Å, respectively.

2.2. Synthesis of modified Eosin-Y (EY)

Synthesis of modified eosin Y was performed by using bromobenzaldehyde and pyridin-4-ylboronic acid with a catalytic amount of tetrakis(triphenylphosphine)palladium resulting in the corresponding (pyridin-4-yl)benzaldehyde in 64–94% yields. The obtained products were reacted with resorcinol in methanesulfonic acid to give fluorescein type derivatives. These obtained compounds were reacted with an excess amount of bromine in ethanol. The precipitated red colored powders were collected to give the corresponding modified eosin Y in 36–42% yields. Thus, the obtained modified products were confirmed using ¹HNMR. Details of the synthesis method are mentioned in the SI. ¹HNMR analyses showed that the obtained final product is of high purity and does not contain any by-product. ¹HNMR was performed with a Bruker AV600 (600 MHz) or a Bruker Avance 400 III HD (400 MHz) spectrometer. The ¹H NMR chemical shifts were reported to be δ values (ppm) relative to Me₄Si. The absorption spectra were recorded on a Shimadzu UV-3600. All the solvents and reagents were of reagent quality, purchased commercially, and used without further purification.

2.3. Photocatalytic activity measurement

Photocatalytic reduction of methyl viologen aims to determine the light-absorbing capacity of photocatalytic materials. In principle, photocatalysts can absorb light energy and transfer photoexcited electrons to methyl viologen (MV) as an electron acceptor, facilitating the reduction of oxidized MV²⁺ to reduced MV˙⁺. Therefore, measuring the amount of reduced MV˙⁺ formed in the reaction directly indicates the capacity of the photocatalysts.

In this work, dye-sensitized TiO₂ (dye/TiO₂) as a photocatalytic material was prepared as follows. The hydrothermal grafting of dye onto the surface of TiO₂ nanoparticles was conducted without any coupling reagents.⁴⁷ First, 12.5 mmol of TiO₂ powder (1 g) was dispersed in 60 mL of H₂O using ultrasonication for 5 min, followed by stirring for 30 min to form a homogeneous slurry. Next, 7.5 μmol of dye (approximately 5 mg) was added, and the mixture was stirred for 2 h in the dark to create a dye/TiO₂ suspension. The resulting suspension was then transferred to a Teflon-lined autoclave and heat-treated at 160 °C for 8 h. After cooling to room temperature, the product was collected by centrifugation at 10 000×g for 5 min, thoroughly washed with deionized distilled water to remove any unabsorbed dye molecules and dried at 100 °C under vacuum to obtain dye/TiO₂ nanoparticles.

To determine the visible light-absorbing capacity of dye/TiO₂ nanoparticles, the reaction mixture comprised 100 mM L-cysteine pH 6–8, 1 mg per mL photocatalyst, and 5 mM methyl viologen (MV²⁺) solution, with a total reaction volume of 3 mL in a 10 mm × 10 mm quartz cuvette with a screw cap and a polytetrafluoroethylene/silicone septum. Oxygen was displaced by bubbling nitrogen gas through the solution for 2 min. The reaction was initiated by irradiation under visible light using a 300-W Xenon lamp (MAX-303, Asahi Spectra, Japan) equipped with a 520 nm bandpass filter (XBPA520, Asahi Spectra, Japan). At each time point during the reaction, the cuvette was centrifuged at 1000×g for 1 min to precipitate the particles that might interfere with absorbance measurements. The reduced MV˙⁺ formed during the photocatalytic reaction of dye/TiO₂ was monitored using a UV-vis spectrophotometer (SH-1000Lab, Corona Electric Co., Ltd, Japan) and quantified using a molar absorption coefficient (ε) of 1.3 × 10⁴ M⁻¹ cm⁻¹, based on the absorbance value at 605 nm.^48,49 The AQY for photocatalytic MV˙⁺ reduction is defined as shown below:⁵⁰

where light intensity at 520 nm is 0.013 W cm⁻², irradiation area is 1 cm², Avogadro's number is 6.022 × 10²³ mol⁻¹, and the rate of incident photons is 0.0562 × 10⁻⁶ mol s⁻¹.

2.4. Photobiocatalytic H₂ production of dye/TiO₂ coupled with E. coli

The combination of photocatalysts and biocatalysts was carried out according to our previous report⁴⁸ with slight modifications for H₂ production. The reaction solution consisted of 100 mM L-cysteine pH 8, 3 mg per mL dye/TiO₂ composite, 5 mM MV²⁺ and 5 mg mL⁻¹ recombinant E. coli in a quartz reactor with a total volume of 30 mL under anaerobic conditions. After connecting the reactor to a closed gas circulation system linked to a gas chromatograph, the air in the reactor was evacuated and replaced with argon. The reaction was then initiated by irradiation with visible light through a 300-W Xenon lamp equipped with a 520 nm bandpass filter. At each time point, H₂ produced was sampled directly using argon as a carrier gas and monitored by using a gas chromatograph (GC-8A, Shimadzu Corp., Japan) equipped with a molecular sieve 5 A column (GL Sciences Inc., Japan), a thermal conductivity detector at an oven temperature of 50 °C and an integrator (C-R6A, Shimadzu Corp.). The AQY for H₂ production is calculated as shown below:
where light intensity at 520 nm is 0.013 W cm⁻², irradiation area is 1 cm², Avogadro's number is 6.022 × 10²³ mol⁻¹, and the rate of incident photons is 0.0562 × 10⁻⁶ mol s⁻¹.

2.5. Photobiocatalytic NH₃/H₂ production using dye/TiO₂ coupled with cyanobacteria

The combination of photocatalysts and biocatalysts was performed based on our previous report,⁴⁹ with slight modifications. Cyanobacterium Anabaena variabilis ATCC 29413 was utilized in this study. Cyanobacterial cells were aerobically cultivated in Allen & Arnon medium supplemented with 20 μM Na₂MoO₄ to promote heterocyst differentiation under the light intensity of 5000 Lux (68 μmol m⁻² s⁻¹) at 26 °C and in a 140 rpm shaking incubator. At a cell density of 1.0 at A₆₈₃ and a chlorophyll a (Chl a) content of ∼12 μg mL⁻¹, the cells were harvested by centrifugation at 10 000×g for 5 min for experiments. The reaction solution consisted of 100 mM L-cysteine pH 7, 2 mg mL⁻¹ of dye/TiO₂ composite, 10 mM MV²⁺ and 50 mg mL⁻¹ of cyanobacterial cell mass in a quartz reactor with a total volume of 100 mL under anaerobic conditions. After connecting the reactor to a closed gas circulation system, the reaction was initiated by irradiating it with visible light from a 300-W xenon lamp equipped with a 520 nm bandpass filter.

At each time point, H₂ production and N₂ consumption were monitored using a gas chromatograph (GC-8A, Shimadzu Corp., Japan) equipped with a thermal conductivity detector and an integrator C-R6A (Shimadzu Corp.), with the oven temperature set at 50 °C. The reaction gas was passed through a molecular sieve 5A column (GL Sciences Inc., Japan) using argon as a carrier gas.⁴⁸ For NH₃ determination, samples were collected and filtered using Amicon^® Ultra-15 Centrifugal Filters-10 K (Merck Millipore) at 5000×g for 1 h to remove insoluble components. A 1 mL filtrate was injected into a cation exchange chromatograph (Dionex, USA) equipped with an IonPac™ CG12A (4 × 50 mm) guard column and an IonPac^® CS12A (4 × 250 mm) analytical column.^49,51 The concentration of NH₄⁺ was calculated using a calibration curve prepared with standard NH₄Cl. The AQY for NH₃ production is calculated as shown below:

where light intensity at 520 nm is 0.013 W cm⁻², irradiation area is 1 cm², Avogadro's number is 6.022 × 10²³ mol⁻¹, and the rate of incident photons is 0.0562 × 10⁻⁶ mol s⁻¹.

3. Results and discussion

3.1. Computational design of eosin Y (EY)

Optimized geometries of the EY isomers with the carboxyl and pyridine anchors shown in Fig. 1 adsorbed on the anatase nanoparticle are shown in Fig. 3. Here, we show the magnified views of the adsorption points of the molecules on anatase. Regarding the full structures of the investigated dyes on anatase, the Cartesian coordinates of the complexes can be found in Tables S1–S6, in the SI. EY isomers with the carboxyl anchoring group on the anatase particle were in the H-bonded monodentate adsorption mode with the H atom of the carboxyl group connected to the two-coordinated O atom on the anatase substrate, whereas the pyridine ring adsorbed on the anatase by a sole Ti–N bond. The length of the main electron transport path, Ti–O bond in the anatase/EY-carboxyl, was around 2.15 Å, which was close to that reported previously between anatase and the dyes with the carboxylic acid group, indicating a strong binding interaction between the carboxyl group and anatase.^52,53 This distance was not far from the Ti–N bond connecting the pyridine ring and anatase (in the region of 2.26 Å to 2.28 Å). As can be seen from the calculated binding energies in Fig. 3, EY isomers anchored by the carboxyl group bind stronger than the corresponding isomers anchored by pyridine to the anatase nanoparticle. The total energies of the isolate dyes, isolate anatase and their combined systems are given in Table S7. On the other hand, the dihedral angles (from 82.05° to 84.93°) formed by the O and Ti atoms of EY-carboxyl/anatase were smaller than those of EY-pyridine/anatase. The smaller dihedral angle could maximize the interaction between the adsorbate and substrate, thus contributing to the high binding energy.⁵⁴ Though the EY-pyridine isomers had weaker binding, they show the advantage in stabilization in water compared to the unstable carboxyl group in water due to hydrolysis.⁸

Fig. 3 Optimized geometries and the binding energies of the eosin Y isomers anchored by the carboxyl and pyridine groups on anatase. (a) EY, (b) MEY, (c) PEY, (d) OPyE, (e)MPyE, and (f) PPyE.

The frontier molecular orbitals of the EY anchored by the carboxyl group and pyridine group are displayed in Fig. 4 and Fig. 5, respectively. The highest occupied molecular orbitals (HOMOs) of all the dyes were delocalized over the three benzene rings and mainly consisted of 2p contributions of the C, Br, and O atoms of the EY.⁵⁴ The similar HOMO distributions imply that the modification of the anchoring group for the EY dye retained its molecular character. The lowest unoccupied molecular orbital (LUMO) contributions mainly came from the Ti atoms in the center of anatase nanoparticles. Furthermore, the LUMOs were spatially separated from the HOMOs, which could reduce the probability of electron–hole recombination upon photoexcitation.⁵⁵ Similar to the adsorption energy, the HOMO–LUMO gaps of the EY isomers anchored by the carboxyl group were smaller than the corresponding isomers anchored by the pyridine ring. All the linkage types were characterized by a narrow band gap (<2.5 eV), in visible light absorption ranges.

Fig. 4 Frontier orbitals and energy levels of the HOMO and LUMO for the eosin Y isomers anchored by the carboxyl group.

Fig. 5 Frontier orbitals and energy levels of the HOMO and LUMO for the eosin Y isomers anchored by the pyridine group.

Though all six forms could have visible absorption and electron transfer from the dye to the particle, their UV-vis absorption spectra differed significantly. The wavelengths, oscillator strengths and assignments for the two most probable singlet excitation states are reported in Table 1. The main difference is in the oscillator strength for singlet excitations of the composites, with those linked by the pyridine ring showing the higher photoactivity, compared to those linked by the carboxyl group. The oscillator strength of the anatase sensitized by EY anchored by OPyE was the highest (4.76 × 10⁻²) among the investigated complexes, which was 10 times higher than that anchored by PEY (4.44 × 10⁻³). The higher transition probability indicates the stronger optical interaction of the designed EY with the anatase particle, as well as the intense transition corresponding to the HOMO → LUMO+1 excitation. At the same time, anatase/OPyE showed a significant advantage in green light absorption, ca. 520 nm, whereas a red-shift of the maximum absorption over the range of 606 nm to 655 nm was observed for anatase/PEY. It is worth noting that the difference between the binding energies of anatase/PEY and anatase/OPyE was only 1.78 kcal mol⁻¹. This slight destabilization caused by anchor modification could be masked by the non-negligible optically excited transition. In other words, anatase/OPyE presented unique features in binding interaction, stability in water, electron transfer and light absorption. In addition, there were two absorption peaks that appeared at 571 nm and 615 nm in the combined system of anatase and PPyE, which were also stronger than those calculated for anatase/PEY. The intense absorption of visible light observed for the anatase/EY-pyridine will have a marked effect on its application in harvesting solar energy photocatalytic systems.

Table 1 Computed excitation properties of the combined systems of anatase and the eosin Y isomers anchored by the carboxyl and pyridine groups. The two most intense singlet excitation states are reported

	Excited state	Wavelength (nm)	Oscillator strength	Transition assignment^a
a Contributions over 15% are displayed. H and L refer to the HOMO and LUMO, respectively.
Carboxyl-ortho (EY)	S0 → S1	635.72	3.20 × 10^−5	H → L (74%), H → L+2 (15%)
	S0 → S5	528.12	1.62 × 10^−5	H → L+8 (33%), H → L+6 (18%)
Carboxyl-meta (MEY)	S0 → S4	626.24	4.28 × 10^−4	H → L+3 (81%)
	S0 → S1	736.38	1.78 × 10^−4	H → L (95%)
Carboxyl-para (PEY)	S0 → S4	605.68	4.44 × 10^−3	H → L+1 (96%)
	S0 → S2	655.19	1.16 × 10^−3	H → L+3 (92%)
Pyridine-ortho (OPyE)	S0 → S3	523.67	4.76 × 10^−2	H → L+1 (42%), H → L+3 (40%)
	S0 → S2	530.39	4.58 × 10^−2	H → L+1 (50%), H → L+3 (28%)
Pyridine-meta (MPyE)	S0 → S4	574.62	4.77 × 10^−4	H → L+3 (95%)
	S0 → S2	619.14	1.24 × 10^−4	H → L+1 (97%)
Pyridine-para (PPyE)	S0 → S4	571.11	8.05 × 10^−3	H → L+3 (95%)
	S0 → S2	614.73	1.89 × 10^−3	H → L+1 (98%)

---

The above results elucidate that modifying the anchoring group to the pyridine ring could significantly improve the photocatalytic performance of anatase. To further understand the electron transfer by the pyridine ring, we examined the electric current in the carboxyl-linked device and the pyridine-linked device, respectively. In the plotted I–V curves in Fig. 6a and b, the current increased with the increasing bias voltage for the two devices. In particular, the current increased sharply at the voltage of 1.6 V, where the current in the carboxyl-linked molecular junction was over three orders of magnitude lower than that in the pyridine-linked one. The transmission spectra at 1.6 V bias for the two investigated devices are shown in Fig. 6c and d. Both transmission spectra were characterized by the sharp peak above the Femi level, which originated from the LUMO of the molecules. In the case of the pyridine-based device, two peaks appeared among the bias window of [−0.8 : 0.8] eV at 1.6 V. This indicated that the LUMO served as a conducting channel when the pyridine ring connected to the anatase. Meanwhile, the calculated energy level of the LUMOs for the combined systems of anatase and EY anchored by pyridine (PyE) in Fig. 5 was closer to the Fermi energy compared to the HOMO energy. Thus, it further confirmed that the LUMO dominated the electron transport through the pyridine-connected junction.⁵⁶ This was in line with the lower energy positions of LUMOs for the pyridine-linked anatase and EY compared to anatase/EY-carboxyl. The intense resonance tunneling at high applied biases through the LUMO channel contributed significantly to the high current of the pyridine-based device. The occurrence of a similar electron injection mechanism was found in other pyridine-linked molecular junctions, such as pyridine bridged between two Ag, Au and graphene nanoribbon,^57–59 based on experimental and theoretical studies. The implications of the photocatalytic performance for the pyridine-anchored EY-sensitized anatase are under continuous experimental investigation. Our work provides a design of EY with a pyridine ring to maximize the electron injection. We performed Mulliken population analysis of representative EY/TiO₂ complexes to quantify interfacial charge redistribution. The results are consistent with our conclusions on electron distribution and interface charge dynamics. In the case of MPyE, it donates ∼0.29e⁻ to the Ti₇₈O₁₆₀H₈ cluster, whereas OPyE donates ∼0.47e⁻, leading to the higher charge transfer ability of OPyE.

Fig. 6 I–V curves of (a) the carboxyl-based device as well as (b) the pyridine-based device, and the transmission spectra (c) of the carboxyl-based device as well as (d) the pyridine-based device at 1.6 V bias, where the insets show the magnified spectra around the bias window.

3.2. Photospectroscopic characterization of Eosin Y (EY) on TiO₂

To validate the theoretical calculation suggestions, experimental measurements for the charge transfer from surface modified Eosin Y (EY) to TiO₂ were conducted. Fig. 7 shows the UV-vis absorption spectra of organic dye/TiO₂. A strong absorption peak was observed around 550 nm for all EY modified TiO₂ and this absorption peak is a typical peak of EY. On the other hand, modification of EY shows peaks shift depending on the modified functional group and position. Compared to the pristine EY, carboxyl modified EY shows a higher wavenumber of absorption peak suggesting strong electronic interaction with TiO₂, which is also suggested by DFT calculations. The absorbance of the peak around 520 nm was weak for QPyE, MPyE, and PPyE modified TiO₂ compared with carboxyl EY. This suggests that the molecular photoexcitation coefficient seems to be small for pyridine modified EY. Fig. S1 shows photographs of dye modified TiO₂ and Fig. S1 shows photographs of dye modified TiO₂. Almost the same color is observed; however, in agreement with UV-vis, pyridine modified TiO₂ shows red color. Fig. S2 shows 1 M aqueous solution of modified EY and UV-vis and almost the same order of absorbance with that on TiO₂ shown in Fig. 7a was observed. So, the molecular excitation coefficient changes based on the modified group and PyE has generally a small molecular excitation coefficient. This could be explained by the expansion of the conjugation length caused by the pyridine group.

Fig. 7 (a) UV-vis and (b) photoluminescence spectra of dye/TiO₂ nanoparticles.

Photoluminescence spectra are effective for the analysis of the annealing efficiency of photoexcited charge. Photoexcited charge generally recombines with the hole and electron formed, and recombination of charge forms photoluminescence or heat. As shown in Fig. 7b, a broad photoluminescence peak was observed around 600 nm for pristine EY modified TiO₂. On the other hand, the peak intensity decreased in the following order OPyE > MPyE > PPyE > MEY > PEY > EY. This peak intensity represents the charge transfer efficiency and the weaker intensity suggests that photoexcited charge is transferred to the conduction band of TiO₂ more efficiently. Since the molecular excitation efficiency changed, normalized photoluminescence (PL) intensity by UV-vis absorbance is shown in Table S8. The normalized PL intensity is as follows: MPyE > OPyE = EY > PEY > PPyE > MEY, demonstrating that the charge transfer occurs easily on PPyE and MEY which is suggested by theoretical calculations.

The life of photoexcited charge was further studied using transient photoluminescence spectra (Fig. S3 and Table S9 in the SI). Time dependence of the photoluminescence decay period varies with the modified group and its position. In agreement with the order of PL peak intensity shown in Fig. 7b, the PL decay period is also longer. The estimated decay rate for each EY modified TiO₂ is also summarized in Table 2. Since the fast backward reaction shows fast decay of PL, a longer PL for pyridine modified EY (τ = 3.03–3.59 ns) suggests that photoexcited charge has a long life which means photoexcited electrons can transfer to the TiO₂ conduction band and cause charge separation. This long life of photoexcited electrons may be effectively transferred to the active site of the photocatalytic reaction. As predicted by theoretical calculation, para-modified pyridine EY shows fast charge transfer to TiO₂.

Table 2 Spectrophotometric measurements of eosin Y and its derivatives

Dye/TiO2	EY	MEY	PEY	OPyE	MPyE	PPyE
λmax (nm)	527	539	537	544	535	537
Abs (a.u.)	0.83	0.85	0.87	0.29	0.18	0.26
ε (L mol^−1 cm^−1)	88 000	79 500	86 350	29 110	17 500	26 490

---

3.3. Photocatalytic reduction of methyl viologen

The MV˙⁺ reduction experiment was performed to demonstrate the transport properties of photoexcited electrons generated by photocatalysts under light illumination. In the photocatalytic MV˙⁺ reduction by dye/TiO₂ nanoparticles, L-cysteine was employed as a sacrificial electron donor, dye/TiO₂ composites served as the photocatalytic material, and oxidized methyl viologen (colorless MV²⁺) acted as an electron acceptor, with its reduced form (dark blue MV˙⁺) detectable by spectrophotometry. It was found that no MV˙⁺ reduction activity occurred with pure TiO₂, indicating that pure TiO₂ cannot absorb visible light energy to excite electrons. In contrast, the formation of reduced MV˙⁺ occurs in the presence of dye on the TiO₂ surface, implying that eosin and its derivatives can absorb light energy at 520 nm to excite electrons to the LUMO level while simultaneously receiving electrons from L-cysteine at the HOMO level. The reaction mechanism of photocatalytic MV²⁺ reduction by dye/TiO₂ in this system can be proposed as follows:

e⁻ + MV²⁺ (colorless) → MV˙⁺ (blue) (2)

In general, several factors can affect the performance of photocatalysts which should be studied. As shown in Fig. 8a, the effect of pH reveals that carboxyl-eosin/TiO₂ composites exhibit higher activity under acidic conditions, while pyridine-eosin/TiO₂ composites perform better under alkaline conditions. This work reveals the range of optimal pH for each dye/TiO₂ nanoparticle in practical applications. In addition to the pH effect, stability is another key factor to consider. Generally, pure dye molecules dissolved in the reaction solution cannot be reused after experiments. Therefore, dyes anchored on TiO₂ surfaces known as dye/TiO₂ composites have been developed to improve their reusability or more strong bonding. In this study, the reusability of dye/TiO₂ nanoparticles was evaluated based on MV˙⁺ reduction activities at pH 8. Each dye/TiO₂ nanoparticle was reused three times for the reaction as shown in Fig. 8b. Among the different photocatalysts, AQY analysis clearly shows that only MPyE/TiO₂ and PPyE/TiO₂ remain stable after three reuse cycles, while the efficiency of other dye/TiO₂ photocatalysts decreased by more than 50%. It is known that dyes bound to the hydroxy site of titanium dioxide by dye-carboxy groups are easily desorbed by hydrolysis in the presence of water.^60,61 Although computer modeling suggests that both carboxyl and pyridine eosin could bind on TiO₂, the experimental results indicate that only MPyE and PPyE are stable on the TiO₂ surface in our reaction system. Although reasonably high photoexcitation properties were achieved for EY without modification, the decrease in activity with time is the most serious drawback; however, MPyE and PPyE show no degradation during recycling. So, in addition to the high efficiency, high stability was achieved by strong bonding of the pyridine group of MPyE and PPyE to TiO₂. Therefore, pyridine-meta and pyridine-para eosin are the most suitable candidates for further development in future studies.

Fig. 8 Methyl viologen (MV˙⁺) reduction of different dye/TiO₂ nanoparticles. (a) Effect of pH on MV˙⁺ reduction and (b) stability of the nanoparticles in the reaction solution determined by testing their reusability up to three times. The raw data are shown in Fig. S4 and S5.

3.4. Dye-modified TiO₂ in photobiocatalytic H₂ production

To demonstrate the usability of EY modified anatase photocatalytic for H₂ production, we typically combined with a biocatalytic enzyme, hydrogenase, in the photobiocatalytic process. In our system, each dye/TiO₂ nanoparticle serves as a photocatalyst, while a hydrogenase-expressing bacterial cell acts as a biocatalyst in the presence of L-cysteine solution as a sacrificial electron donor and MV²⁺ as an electron mediator that transfers photoexcited electrons from the photocatalyst to the biocatalyst for H₂ production.⁵⁰ The mechanism of the photobiocatalytic system in this work is illustrated below:

Photocatalytic reaction (light reaction):

Biocatalytic reaction (dark reaction):

According to the results shown in Fig. 9, all dye/TiO₂ composites can be utilized under visible light in the photobiocatalytic system when coupled with hydrogenase-expressing recombinant E. coli for H₂ production (Fig. 9a). Among the five different photocatalysts, PPyE/TiO₂ demonstrated the highest activity, achieving an AQY of 1.97%, followed by OPyE/TiO₂, PEY/TiO₂, MPyE/TiO₂, EY/TiO₂, and MEY/TiO₂, respectively, (Fig. 9b) upon the addition of the sacrificial agent. In general, pyridine-eosin/TiO₂ composites exhibit higher efficiency than carboxyl-eosin/TiO₂ composites corresponding to their photocatalytic reduction of MV²⁺, while pure TiO₂ cannot absorb light energy at 520 nm for H₂ production. This indicates that dye-sensitized molecules on the surface of TiO₂ can absorb visible light energy at 520 nm, generating and transferring photoexcited electrons to the biocatalytic reaction in recombinant E. coli for H₂ production. In accordance with prediction, this high AQY was assigned to the efficient charge injection of electrons from the dye to the conduction band of TiO₂. This could be related to the expanded π conjugated orbital.

Fig. 9 Photobiocatalytic H₂ production of the dye/TiO₂ composite coupled with recombinant E. coli expressing [FeFe]-hydrogenase. (a) A time-course and (b) AQY analysis of H₂ production obtained from the reaction mixture including 100 mM L-cysteine pH 8, 3 mg per mL of the dye/TiO₂ composite, 5 mM MV²⁺ and 5 mg per mL cell mass of recombinant E. coli under visible light illumination at 520 nm.

3.5. Dye-modified TiO₂ in photobiocatalytic NH₃ production

In nature, the ability to synthesize NH₃ is observed in various species of bacteria, archaea, and cyanobacteria under ambient conditions. Nitrogenase is the only biocatalytic enzyme capable of fixing nitrogen by breaking the strong N≡N triple bond of dinitrogen (N₂) gas to NH₃, as shown in eqn (5):⁶²

Although N₂-fixing microorganisms can produce extracellular NH₃ without generating greenhouse gases or requiring high energy input, nitrogen metabolism in these organisms is tightly regulated to maintain intracellular C–N balance and amino acid pool homeostasis.⁶³ Due to the low productivity of natural NH₃ synthesis through nitrogenase activity and ATP-dependent photosystem I, the replacement of the photosystem with photocatalysts is the focus of this study.

In the filamentous cyanobacterium A. variabilis, nitrogenase is naturally expressed from the nif gene cluster within differentiated heterocysts, which function as specialized sites for N₂ reduction.⁶⁴ Notably, it has been reported that methyl viologen (MV²⁺), as an electron mediator, can transfer electrons to nitrogenase, mimicking the natural electron flow through the P-cluster to FeMoco at the enzyme's active site.^65,66 This finding suggests the feasibility of developing an artificial photobiocatalytic system to enhance NH₃ production efficiency under mild conditions.

In this study, dye-modified TiO₂ was applied in a photobiocatalytic system coupled with cyanobacteria A. variabilis for NH₃ production in the presence of MV²⁺ and a sacrificial electron donor under visible light (520 nm). During the reaction, the kinetics of H₂ production and N₂ consumption were simultaneously monitored, as shown in Fig. S6 (SI). The rates of NH₃ production, H₂ production, and N₂ consumption are presented in Fig. 10. Among the different dye/TiO₂ composites tested, pyridine-EY/TiO₂ (particularly PPyE/TiO₂ and OPyE/TiO₂) demonstrated higher efficiency in this system, with superior rates and AQY compared to carboxyl-EY/TiO₂, as shown in Table S10 (SI). However, no significant differences were observed in the rates of N₂ consumption.

Fig. 10 Rate of (a) H₂ production, (b) N₂ consumption and (c) NH₃ production produced from the photobiocatalytic system of dye/TiO₂ coupled with cyanobacteria in the reaction system including 100 mM L-cysteine pH 7, 2 mg per mL dye/TiO₂, 10 mM MV²⁺ and 50 mg per mL cyanobacterial cell mass under visible light illumination at 520 nm. The raw data are also shown in Fig. S6 and AQY analysis in Table S10.

The results demonstrate that L-cysteine acts as an effective sacrificial reagent, donating electrons to deplete the HOMO level of dye/TiO₂ particles. The excited electrons are transferred to MV²⁺, reducing it to MV˙⁺ during the photocatalytic reaction, which proceeds similarly to the light reaction of H₂ production described in eqn (3). The simultaneous production of NH₃ and H₂, along with N₂ consumption, indicates that the reduced MV˙⁺ molecules serve as electron mediators, transferring electrons to cyanobacterial nitrogenase for N₂ reduction and NH₃/H₂ generation under visible light (520 nm). The proposed reaction mechanism for artificial photobiocatalytic NH₃ production using dye/TiO₂ and cyanobacteria is outlined below:

Photocatalytic reaction (light reaction):

Biocatalytic reaction (dark reaction):

4. Conclusion

This study explored the possibility of improving the photocatalytic performance of EY/anatase TiO₂ by anchoring modification. Pyridine ring was used to replace the conventional carboxyl group linker. The structures, adsorption strengths, and electron transfer properties for the systems of anatase and EY anchored by the carboxyl and pyridine ring position were investigated through DFT and TD-DFT calculations and confirmed with experimental support. The adsorption of EY with the carboxyl acid was found to be much stronger than that with the pyridine ring, irrespective of the anchor positions (ortho, para, and meta). While pyridine-linked EY and anatase showed prominent optical properties, the calculated oscillator strength of anatase/EY with pyridine-ortho was the highest among the six investigated systems, which was over ten times higher than that with carboxyl-para (the highest form among carboxyl anchors). NEGF-DFT calculations were used to elucidate the electron transport in the pyridine-based device and the carboxyl-based device. It was found that the electric current in the pyridine-based device was higher compared to that in the carboxyl-based device. In the case of the pyridine-based device, the transmission spectra at the applied bias showed that two peaks entered the bias window due to the strong electron transfer in the conducting channel, contributing to the high oscillator strength and enhanced catalytic activity. The results imply the promising design of utilizing a pyridine anchor for the EY/anatase cluster compared to the carboxyl anchor. Experimental validations support this design strategy in good agreement. In particular, by modifying TiO₂ with para pyridine MY, AQY for H₂ and NH₃ formation under irradiation of 520 nm visible light was achieved to be ca. 2 and 0.67% upon the addition of the sacrificial agent, respectively. No decrease in activity over three times repetition suggests that the pyridine linker is effective not only for charge transfer but also for stable adsorption on anatase. This study demonstrated that molecular design for dye modification by DFT and NEGT-DFT calculation is highly efficient and effective for finding new organic dyes with more efficient charge transfer.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

All data are available upon request from the corresponding author.

Supplementary information: atomic coordinates of the optimized ortho, meta, and para isomers of EY anchored by the carboxyl group as well as pyridine group on anatase, and the total energies of the isolated EY/anatase. See DOI: https://doi.org/10.1039/d5ta05712d.

Acknowledgements

This study was financially supported by the Mitsui Chemical Inc Carbon Neutral Research Center (MCI-CNRC). All authors acknowledge the support of the World Premier International Research Center Initiative (WPI), Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT).

References

1. M. Yue, H. Lambert, E. Pahon, R. Roche, S. Jemei and D. Hissel, Hydrogen Energy Systems: A Critical Review of Technologies, Applications, Trends and Challenges, Renew. Sustain. Energy Rev., 2021, 146, 111180.
2. P.-A. Le, V. D. Trung, P. L. Nguyen, T. V. B. Phung, J. Natsuki and T. Natsuki, The Current Status of Hydrogen Energy: An Overview, RSC Adv., 2023, 13, 28262–28287.
3. F. Dawood, M. Anda and G. M. Shafiullah, Hydrogen Production for Energy: An Overview, Int. J. Hydrogen Energy, 2020, 45, 3847–3869.
4. R. S. El-Emam and H. Özcan, Comprehensive Review on the Techno-Economics of Sustainable Large-Scale Clean Hydrogen Production, J. Clean. Prod., 2019, 220, 593–609.
5. A. Fujishima and K. Honda, Electrochemical Photolysis of Water at a Semiconductor Electrode, Nature, 1972, 238, 37–38.
6. H. Tong, N. Enomoto, M. Inada, Y. Tanaka and J. Hojo, Mater. Lett., 2015, 141, 259–262.
7. S. Nishioka, K. Yanagisawa, D. T. Lu, J. J. M. Vequizo, A. Yamakata, K. Kimoto, M. Inada and K. Maeda, Sustain. Energy Fuels, 2019, 3, 2337–2346.
8. R. Abe, K. Hara, K. Sayama, K. Domen and H. Arakawa, Steady Hydrogen Evolution from Water on Eosin Y-Fixed TiO₂ Photocatalyst Using a Silane-Coupling Reagent under Visible Light Irradiation, J. Photochem. Photobiol., A, 2000, 137, 63–69.
9. H. Eidsvåg, S. Bentouba, P. Vajeeston, S. Yohi and D. Velauthapillai, TiO₂ as a Photocatalyst for Water Splitting—An Experimental and Theoretical Review, Molecules, 2021, 26, 1687.
10. C.-H. Liao, C.-W. Huang and J. C. S. Wu, Hydrogen Production from Semiconductor-Based Photocatalysis via Water Splitting, Catalysts, 2012, 2, 490–516.
11. M. Bledowski, L. Wang, A. Ramakrishnan, O. V. Khavryuchenko, V. D. Khavryuchenko, P. C. Ricci, J. Strunk, T. Cremer, C. Kolbeck and R. Beranek, Visible-Light Photocurrent Response of TiO₂–Polyheptazine Hybrids: Evidence for Interfacial Charge-Transfer Absorption, Phys. Chem. Chem. Phys., 2011, 13, 21511–21519.
12. M. Murdoch, G. I. N. Waterhouse, M. A. Nadeem, J. B. Metson, M. A. Keane, R. F. Howe, J. Llorca and H. Idriss, The Effect of Gold Loading and Particle Size on Photocatalytic Hydrogen Production from Ethanol over Au/TiO₂ Nanoparticles, Nat. Chem., 2011, 3, 489–492.
13. M. Anpo and M. Takeuchi, The Design and Development of Highly Reactive Titanium Oxide Photocatalysts Operating under Visible Light Irradiation, J. Catal., 2003, 216, 505–516.
14. Z. Li, L. Shi, D. Franklin, S. Koul, A. Kushima and Y. Yang, Drastic Enhancement of Photoelectrochemical Water Splitting Performance over Plasmonic Al@TiO₂ Heterostructured Nanocavity Arrays, Nano Energy, 2018, 51, 400–407.
15. J. Venturini, F. Bonatto, W. C. Guaglianoni, T. Lemes, S. Arcaro, A. K. Alves and C. P. Bergmann, Cobalt-Doped Titanium Oxide Nanotubes Grown via One-Step Anodization for Water Splitting Applications, Appl. Surf. Sci., 2019, 464, 351–359.
16. Z. Dong, D. Ding, T. Li and C. Ning, Ni-Doped TiO₂ Nanotubes Photoanode for Enhanced Photoelectrochemical Water Splitting, Appl. Surf. Sci., 2018, 443, 321–328.
17. S. Y. Sawant, M. S. Sayed, T. H. Han, M. R. Karim, J.-J. Shim and M. H. Cho, Bio-Synthesis of Finely Distributed Ag Nanoparticle-Decorated TiO₂ Nanorods for Sunlight-Induced Photoelectrochemical Water Splitting, J. Ind. Eng. Chem., 2019, 69, 48–56.
18. W. Choi, A. Termin and M. R. Hoffmann, The Role of Metal Ion Dopants in Quantum-Sized TiO₂: Correlation between Photoreactivity and Charge Carrier Recombination Dynamics, J. Phys. Chem., 2002, 98, 13669–13679.
19. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Visible-Light Photocatalysis in Nitrogen-Doped Titanium Oxides, Science, 2001, 293, 269–271.
20. R. Abe, T. Takata, H. Sugihara and K. Domen, Photocatalytic Overall Water Splitting under Visible Light by TaON and WO₃ with an IO³⁻/I⁻ Shuttle Redox Mediator, Chem. Commun., 2005, 30, 3829–3831.
21. Q. Gao, F. Si, S. Zhang, Y. Fang, X. Chen and S. Yang, Hydrogenated F-Doped TiO₂ for Photocatalytic Hydrogen Evolution and Pollutant Degradation, Int. J. Hydrogen Energy, 2019, 44, 8011–8019.
22. M. H. Elbakkay, W. M. El Rouby, S. I. El-Dek and A. A. Farghali, S-TiO₂/S-Reduced Graphene Oxide for Enhanced Photoelectrochemical Water Splitting, Appl. Surf. Sci., 2018, 439, 1088–1102.
23. Y. Honda, M. Watanabe, H. Hagiwara, S. Ida and T. Ishihara, Inorganic/Whole-Cell Biohybrid Photocatalyst for Highly Efficient Hydrogen Production from Water, Appl. Catal., B, 2017, 210, 400–406.
24. F. De Angelis, Direct vs. Indirect Injection Mechanisms in Perylene Dye-Sensitized Solar Cells: A DFT/TDDFT Investigation, Chem. Phys. Lett., 2010, 493, 323–327.
25. W. Chu, K. H. Chan, C. T. Jafvert and Y. S. Chan, Removal of Phenylurea Herbicide Monuron via Riboflavin-Mediated Photosensitization, Chemosphere, 2007, 69, 177–183.
26. C. Chen, X. Qi and B. Zhou, Photosensitization of Colloidal TiO₂ with a Cyanine Dye, J. Photochem. Photobiol., A, 1997, 109, 155–158.
27. M. M. Ibrahim, U. I. Gaya, M. M. Ibrahim and U. I. Gaya, Synthesis of Eosin Y-sensitized Ag-TiO₂ Nano-hybrid for Optimized Photocatalytic Deradation of Aqueous Caffeine, J. Chil. Chem. Soc., 2019, 64, 4275–4284.
28. P. Wang, Z. Guan, Q. Li and J. Yang, Efficient Visible-Light-Driven Photocatalytic Hydrogen Production from Water by Using Eosin Y-Sensitized Novel g-C₃N₄/Pt/GO Composites, J. Mater. Sci., 2018, 53, 774–786.
29. (a) F. Zhang, F. Shi, W. Ma, F. Gao, Y. Jiao, H. Li, J. Wang, X. Shan, X. Lu and S. Meng, Controlling Adsorption Structure of Eosin Y Dye on Nanocrystalline TiO₂ Films for Improved Photovoltaic Performances, J. Phys. Chem. C, 2013, 117, 14659–14666; (b) I. Gillaizeau-Gauthier, F. Odobel, M. Alebbi, R. Argazzi, E. Costa, C. A. Bignozzi, P. Qu and G. J. Meyer, Phosphonate-Based Bipyridine Dyes for Stable Photovoltaic Devices, Inorg. Chem., 2001, 40, 6073–6079.
30. (a) K. Murakoshi, G. Kano, Y. Wada, S. Yanagida, H. Miyazaki, M. Matsumoto and S. Murasawa, Importance of Binding States between Photosensitizing Molecules and the TiO₂ Surface for Efficiency in a Dye-Sensitized Solar Cell, J. Electroanal., 1995, 396, 27–34; (b) S. M. Zakeeruddin, M. K. Nazeeruddin, P. Pechy, F. P. Rotzinger, R. Humphry-Baker, K. Kalyanasundaram, M. Grätzel, V. Shklover and T. Haibach, Molecular Engineering of Photosensitizers for Nanocrystalline Solar Cells: Synthesis and Characterization of Ru Dyes Based on Phosphonated Terpyridines, Inorg. Chem., 1997, 36, 5937–5946.
31. (a) M. Nilsing, S. Lunell, P. Persson and L. Ojamäe, Phosphonic Acid Adsorption at the TiO₂ Anatase (101) Surface Investigated by Periodic Hybrid HF-DFT Computations, Surf. Sci., 2005, 582, 49–60; (b) Y. Ooyama, T. Nagano, S. Inoue, I. Imae, K. Komaguchi, J. Ohshita and Y. Harima, Dye-sensitized Solar Cells Based on Donor-π-acceptor Fluorescent Dyes with a Pyridine Ring as an Electron-withdrawing-injecting Anchoring Group, Chem.–Eur. J., 2011, 17, 14837–14843.
32. J. P. Perdew, K. Burke and M. Ernzerhof, Generalized, Gradient Approximation Made Simple, Phys. Rev. Lett., 1996, 77, 3865–3868.
33. A. D. Becke, Density-functional Thermochemistry. I. The Effect of the Exchange-only Gradient Correction, J. Chem. Phys., 1992, 96, 2155–2160.
34. C. Lee, W. Yang and R. G. Parr, Development of the Colle-Salvetti Correlation-Energy Formula into a Functional of the Electron Density, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785.
35. F. Weigend and R. Ahlrichs, Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
36. P. J. Hay and W. R. Wadt, Ab Initio Effective Core Potentials for Molecular Calculations. Potentials for the Transition Metal Atoms Sc to Hg, J. Chem. Phys., 1985, 82, 270–283.
37. D. T. Cromer and K. Herrington, The Structures of Anatase and Rutile, J. Am. Chem. Soc., 1955, 77, 4708–4709.
38. M. Addamo, M. Bellardita, A. Di Paola and L. Palmisano, Preparation and Photoactivity of Nanostructured Anatase, Rutile and Brookite TiO₂ Thin Films, Chem. Commun., 2006, 47, 4943–4945.
39. Q. Tay, X. Liu, Y. Tang, Z. Jiang, T. C. Sum and Z. Chen, Enhanced Photocatalytic Hydrogen Production with Synergistic Two-Phase Anatase/Brookite TiO2 Nanostructures, J. Phys. Chem. C, 2013, 117, 14973–14982.
40. F. H. Tian, X. Wang, W. Zhao, L. Zhao, T. Chu and S. Yu, Adsorption of 2-Propanol on Anatase TiO₂ (101) and (001) Surfaces: A Density Functional Theory Study, Surf. Sci., 2013, 616, 76–84.
41. TURBOMOLE V6.4, TURBOMOLE GmbH, Karlsruhe, Germany, 2007, http://www.turbomole.com.
42. S. Datta, Quantum Transport: Atom to Transistor, Cambridge University Press, Cambridge, U.K., 2005.
43. QuantumATK S-2021.06-SP2; Quantumwise, Copenhagen, Denmark, 2021, http://www.quantumwise.com.
44. W. Kohn and L. J. Sham, Self-Consistent Equations Including Exchange and Correlation Effects, Phys. Rev., 1965, 140, A1133–A1138.
45. M. Fuchs and M. Scheffler, Ab Initio Pseudopotentials for Electronic Structure Calculations of Poly-Atomic Systems Using Density-Functional Theory, Comput. Phys. Commun., 1999, 119, 67–98.
46. E. R. Davidson and D. Feller, Basis Set Selection for Molecular Calculations, Chem. Rev., 1986, 86, 681–696.
47. X. Liu, W. Huang, Y. Lei, Y. Li, Y. Xue, F. Wang and S. Min, Effective Hydrothermal Grafting of Eosin Y onto TiO₂ Nanoparticles Towards Stable Photocatalysts for Efficient Visible-Light-Driven Photocatalytic H₂ Evolution, New J. Chem., 2018, 42, 6631–6635.
48. N. Kosem, M. Watanabe, J. T. Song, A. Takagaki and T. Ishihara, A Comprehensive Study on Rational Biocatalysts and Individual Components of Photobiocatalytic H₂ Production Systems, Appl. Catal., A, 2023, 651, 119019.
49. N. Kosem, X.-F. Shen, Y. Ohsaki, M. Watanabe, J. T. Song and T. Ishihara, Photobiocatalytic Conversion of Solar Energy to NH₃ from N₂ and H₂O under Ambient Condition, Appl. Catal., B, 2024, 342, 123431.
50. N. Kosem, Y. Honda, M. Watanabe, A. Takagaki, Z. P. Tehrani, F. Haydous, T. Lippert and T. Ishihara, Photobiocatalytic H₂ Evolution of GaN:ZnO and [FeFe]-Hydrogenase Recombinant Escherichia coli, Catal. Sci. Technol., 2020, 10, 4042–4052.
51. H. Hirakawa, M. Hashimoto, Y. Shiraishi and T. Hirai, Photocatalytic Conversion of Nitrogen to Ammonia with Water on Surface Oxygen Vacancies of Titanium Dioxide, J. Am. Chem. Soc., 2017, 139, 10929–10936.
52. L. Zhang, X. Liu, W. Rao and J. Li, Multilayer Dye Aggregation at Dye/TiO₂ Interface via π-π Stacking and Hydrogen Bond and Its Impact on Solar Cell Performance: A DFT Analysis, Sci. Rep., 2016, 6, 35893.
53. L. Zhang and J. M. Cole, Adsorption Properties of P-Methyl Red Monomeric-to-Pentameric Dye Aggregates on Anatase (101) Titania Surfaces: First-Principles Calculations of Dye/TiO₂ Photoanode Interfaces for Dye-Sensitized Solar Cells, ACS Appl. Mater. Interfaces, 2014, 6, 15760–15766.
54. F. Labat, I. Ciofini, H. P. Hratchian, M. Frisch, K. Raghavachari and C. Adamo, First Principles Modeling of Eosin-Loaded ZnO Films: A Step toward the Understanding of Dye-Sensitized Solar Cell Performances, J. Am. Chem. Soc., 2009, 131, 14290–14298.
55. F. De Angelis, C. Di Valentin, S. Fantacci, A. Vittadini and A. Selloni, Theoretical Studies on Anatase and Less Common TiO2 Phases: Bulk, Surfaces, and Nanomaterials, Chem. Rev., 2014, 114, 9708–9753.
56. Y. Ie, T. Hirose, H. Nakamura, M. Kiguchi, N. Takagi, M. Kawai and Y. Aso, Nature of Electron Transport by Pyridine-Based Tripodal Anchors: Potential for Robust and Conductive Single-Molecule Junctions with Gold Electrodes, J. Am. Chem. Soc., 2011, 133, 3014–3022.
57. X. Li, L. Cao, H.-L. Li, H. Wan and G. Zhou, Spin-Resolved Transport Properties of a Pyridine-Linked Single Molecule Embedded between Zigzag-Edged Graphene Nanoribbon Electrodes, J. Phys. Chem. C, 2016, 120, 3010–3018.
58. M. Kamenetska, S. Y. Quek, A. C. Whalley, M. L. Steigerwald, H. J. Choi, S. G. Louie, C. Nuckolls, M. S. Hybertsen, J. B. Neaton and L. Venkataraman, Conductance and Geometry of Pyridine-Linked Single-Molecule Junctions, J. Am. Chem. Soc., 2010, 132, 6817–6821.
59. O. Adak, R. Korytár, A. Y. Joe, F. Evers and L. Venkataraman, Impact of Electrode Density of States on Transport through Pyridine-Linked Single Molecule Junctions, Nano Lett., 2015, 15, 3716–3722.
60. H.-O. Son, C. Prasittichai, J. E. Mondloch, L. Luo, J. Wu, D. W. Kim, O. K. Farha and J. T. Hupp, Dye Stabilization and Enhanced Photoelectrode Wettability in Water-Based Dye-Sensitized Solar Cells through Post-assembly Atomic Layer Deposition of TiO₂, J. Am. Chem. Soc., 2013, 135, 11529–11532.
61. J. Willkomm, K. L. Orchard, A. Reynal, E. Pastor, J. R. Durrant and E. Reisner, Dye-Sensitised Semiconductors Modified with Molecular Catalysts for Light-Driven H₂ Production, Chem. Soc. Rev., 2016, 45, 9–23.
62. C. J. M. van der Ham, M. T. M. Koper and D. G. H. Hetterscheid, Challenges in Reduction of Dinitrogen by Proton and Electron Transfer, Chem. Soc. Rev., 2014, 43, 5183–5191.
63. M. I. Muro-Pastor and F. J. Florencio, Regulation of Ammonium Assimilation in Cyanobacteria, Plant Physiol. Biochem., 2003, 41, 595–603.
64. L. Wang, M. Xia, H. Wang, K. Huang, C. Qian, C. T. Maravelias and G. A. Ozin, Greening Ammonia Toward the Solar Ammonia Refinery, Joule, 2018, 2, 1055–1074.
65. D. G. Boucher, E. Carroll, Z. A. Nguyen, R. G. Jadhav, O. Simoska, K. Beaver and S. D. Minteer, Bioelectrocatalytic Synthesis: Concepts and Applications, Angew. Chem., Int. Ed., 2023, 62, e202307780.
66. K. Sengupta, J. P. Joyce, L. Decamps, L. Kang, R. Bjornsson, O. Rüdiger and S. DeBeer, Investigating the Molybdenum Nitrogenase Mechanistic Cycle Using Spectroelectrochemistry, J. Am. Chem. Soc., 2025, 147, 2099–2114.

---