Effects of Anionic Polymer Modification of Dye‐Sensitized Niobate Photocatalysts on Solar‐Driven Z‐Scheme Overall Water Splitting

Pt‐intercalated calcium niobate nanosheets (Pt/HCa2Nb3O10) sensitized by a Ru(II) complex dye are good photocatalysts for producing H2 from aqueous solutions containing I− as a reversible electron donor. These materials are applicable to Z‐scheme overall water splitting in combination with a WO3‐based O2‐evolving photocatalyst under simulated sunlight. In this work, the effects of anionic polymer modification of the dye‐sensitized nanosheets are examined by adsorbing sodium poly(styrenesulfonate) (PSS), sodium polyacrylate, sodium polymethacrylate (PMA), or sodium poly(4‐styrenesulfonic‐co‐maleic acid) onto the dye‐sensitized nanosheet surface. For half‐cell H2‐evolution reaction in the presence of NaI, all of the polymers have a positive impact on the activity under visible light at lower light intensity, whereas only PMA is effective under high light‐intensity condition. For Z‐scheme overall water splitting with PtOx/H‐Cs‐WO3, PSS and PMA give almost the same solar‐to‐hydrogen energy conversion efficiencies (0.12% ± 0.01%) under optimized conditions. However, PMA operates better than PSS at relatively low and high NaI concentrations, which are in general disadvantageous for the H2‐ and O2‐evolving components of the Z‐scheme, respectively.


Introduction
Renewable energy resources are increasingly needed today because of global warming and the depletion of energy resources.
Hydrogen has attracted renewed attention as an energy-storage vehicle because of its high energy density.[3][4][5][6][7] Metal oxides are generally good photocatalysts for overall water splitting, and some of them give very high apparent quantum yields (AQYs). [8,9]owever, most of them absorb only high-energy UV light and not visible or near-infrared light, the main parts of solar spectrum, because of their wide bandgap.Therefore, it is important to develop photocatalysts that are capable of absorbing visible light so as to realize solar-driven overall water splitting with high efficiency.
[12] The mechanism of dye-sensitized H 2 evolution is shown in Scheme 1.[15][16][17][18] First, the dye sensitizer is excited by visible light to create an excited state.An electron is injected from the excited-state dye to the conduction band of the semiconductor.
Pt-intercalated calcium niobate nanosheets (Pt/HCa 2 Nb 3 O 10 ) sensitized by a Ru(II) complex dye are good photocatalysts for producing H 2 from aqueous solutions containing I À as a reversible electron donor.These materials are applicable to Z-scheme overall water splitting in combination with a WO 3 -based O 2 -evolving photocatalyst under simulated sunlight.In this work, the effects of anionic polymer modification of the dye-sensitized nanosheets are examined by adsorbing sodium poly(styrenesulfonate) (PSS), sodium polyacrylate, sodium polymethacrylate (PMA), or sodium poly(4-styrenesulfonic-co-maleic acid) onto the dye-sensitized nanosheet surface.For half-cell H 2 -evolution reaction in the presence of NaI, all of the polymers have a positive impact on the activity under visible light at lower light intensity, whereas only PMA is effective under high light-intensity condition.For Z-scheme overall water splitting with PtO x /H-Cs-WO 3 , PSS and PMA give almost the same solar-to-hydrogen energy conversion efficiencies (0.12% AE 0.01%) under optimized conditions.However, PMA operates better than PSS at relatively low and high NaI concentrations, which are in general disadvantageous for the H 2 -and O 2 -evolving components of the Z-scheme, respectively.
The injected electron moves to the cocatalyst, at which H 2 O (or H þ ) is reduced to H 2 .A solution-phase electron donor transfers an electron to the oxidized dye sensitizer to regenerate its reduced form.Although overall water splitting by using a dye-sensitized photocatalyst only is difficult to achieve because of low oxidizing power of the dye component, the dye-sensitized semiconductor can work as a H 2 -evolving photocatalyst in Z-scheme overall water splitting when it is combined with an O 2 -evolving photocatalyst and a shuttle redox mediator (i.e., reversible donor/acceptor pair). [19,20]In this complex photoinduced electron-transfer scheme, it is essential to suppress the back electron-transfer reactions that compete with the forward reactions. [4]For example, the oxidized form of electron donor is thermodynamically more susceptible to reduction than H 2 O (or H þ ).Therefore, the oxidized electron donor must be rapidly reduced by the paired O 2 -evolving photocatalyst to achieve efficient overall water splitting.
We have developed dye-sensitized H 2 -evolving photocatalysts for Z-scheme overall water splitting by using Pt-intercalated metal oxide nanosheets (Pt/HCa 2 Nb 3 O 10 ) and [Ru(dmb) 2 (4,4'-(PO 3 H 2 ) 2 bpy)] 2þ (dmb = 4,4'-dimethyl-2,2'-bipyridine, bpy = 2,2'-bipyridine), abbreviated as RuP. [20]The dye-sensitized photocatalyst produces H 2 by using I À as a reversible electron donor and achieves Z-scheme overall water splitting in combination with PtO x /H-Cs-WO 3 [21] as the O 2 -evolving photocatalyst.Surface modification has proven to be essential for realizing efficient photochemical H 2 generation from RuP/Pt/HCa 2 Nb 3 O 10 .Surface modification with poly(sodium 4-styrenesulfonate) (PSS), an anionic polymer, suppresses the reduction of I 3 À , the oxidized form of the electron donor, which is represented as the A/D pathway in Scheme 1.Thus, anionic polymer modification increases the quantum yield of photochemical H 2 production. [22,23]By adding an Al 2 O 3 modifier, which suppresses back electron transfer from the semiconductor to the dye sensitizer, [24,25] with PSS, the activity of the dye-sensitized photocatalyst can be dramatically improved, giving a solar-to-hydrogen (STH) energy conversion efficiency of 0.12%, which is the highest value yet reported for dye-sensitized photocatalysts. [23]ur previous experiments with this system have shown that the inhibition of the back electron-transfer reaction by PSS becomes less effective when the concentration of I 3 À in solution is high. [23]While a good O 2 -evolving photocatalyst that efficiently reduces I 3 À is required to keep the I 3 À concentration low under Z-scheme water-splitting conditions, such photocatalysts are unfortunately as yet unknown.However, if one can find a modifier that inhibits the back electron-transfer reaction even at high concentrations of I 3 À , it may be possible to pair other O 2 -evolving photocatalysts with the dye-sensitized H 2 -evolving photocatalyst.It is also important to make the Z-scheme system workable at lower I À concentration, which can minimize the rate of I À oxidation at the O 2 -evolving photocatalyst. [4]ased on these considerations, we examine anionic polymers other than PSS as modifiers for RuP/Pt/HCa 2 Nb 3 O 10 to study the effects of I À concentration and to improve the efficiency of the Z-scheme overall water splitting.First, KCa 2 Nb 3 O 10 was synthesized by a flux method. [26]K 2 SO 4 (≥99.0%;Kanto Chemical Co.), CaCO 3 (≥99.99%;Kanto Chemical Co.), and Nb 2 O 5 (≥99.95%;Kanto Chemical Co.) were ground by using an agate motor and pestle for 15 min.The molar ratio of K/Ca/Nb was 5/2/3.The mixture was put into Pt crucible and heated in an electric furnace.The mixture was heated to 1173 K at a ramp rate of 300 K h À1 , and then to 1573 K at 100 K h À1 and kept at that temperature for 24 h.It was cooled down to 1073 K at 25 K h À1 and then cooled ambiently.The sample was washed with water and centrifuged several times, and finally dried at 343 K.

Experimental Section
The KCa 2 Nb 3 O 10 powder was acid exchanged by stirring in an aqueous 1 M HNO 3 solution (100 mL) for 3 days.The solution was refreshed on the second day.After that, the sample was washed with water until the pH of the supernatant was 6-7.
The HCa 2 Nb 3 O 10 product of the acid exchange reaction was stirred in an aqueous tetra(n-butyl)ammonium hydroxide (TBA þ OH À ) solution (40 wt% in H 2 O; Sigma-Aldrich) for 1 week to obtain TBA þ /Ca 2 Nb 3 O 10 À nanosheets. [27]The molar ratio of TBA þ /Ca 2 Nb 3 O 10 À was 1/1 in the solution.The suspension was separated by decantation and unreacted solid was removed.
After adjusting the concentration of the suspension to 5 g L À1 , a 1 mM aqueous solution of dissolved [Pt(NH 3 ) 4 ]Cl 2 (Wako Pure Chemicals) was dropped at a rate of 1 to 2 mL min À1 into the Ca 2 Nb 3 O 10 À nanosheet suspension (1 wt% Pt). [28]The solution was stirred for 1 day, and the colloidal nanosheets were restacked by adding an aqueous 2 M HCl solution.The precipitate was washed with water and centrifuged several times, and then dried at 343 K.The resulting powder was then heated at 473 K for 1 h under a H 2 flow (20 mL min À1 ).Finally, the obtained sample was stirred in aqua regia at its boiling temperature for 15 min to remove Pt from the external surface of the restacked HCa 2 Nb 3 O 10 , washed with water, and dried at 343 K. [20]

Al 2 O 3 Modification
Pt/HCa 2 Nb 3 O 10 was modified with Al 2 O 3 by a sol-gel method. [20]100 mg of Pt/HCa 2 Nb 3 O 10 was suspended in EtOH (20 mL) containing H 2 SO 4 solution (0.1 M, 100 μL) and aluminum isopropoxide (2 wt%, ≥98.0%,TCI).The suspension was subjected to sonication for 30 min.After that, the suspension was stirred for 1 day, and the solid was filtered and washed with water, and dried at room temperature under vacuum.It has been confirmed by X-ray photoelectron spectroscopy and Fourier transform infrared (FT IR) spectroscopy that the deposited Al species was Al 2 O 3 , although the location of Al 2 O 3 could not be visualized primarily due to its low concentration. [20]
The suspension was stirred for 1 h at room temperature in the dark.After that, the solid was filtered and washed with water, and finally dried at room temperature under vacuum.

Characterization
FT IR spectra were measured using an FT/IR-4600 (JASCO) by the ATR method with a diamond prism.X-ray diffraction (XRD) patterns were measured by using a Rigaku MiniFlex600 powder diffractometer operating at 15 mA and 40 kV (Cu Kα radiation).Scanning electron microscopy (SEM) images were obtained by using a JSM-IT100 InTouchScope SEM (JEOL).IR mapping analysis was conducted using an infrared Raman microscope (AIRsight, Shimadzu).Zeta potentials were measured using an SZ-100 (HORIBA).UV-vis diffuse reflectance spectra (DRS) were obtained using a spectrophotometer (V-770, JASCO).Emission quantum yields of RuP on solid materials were measured using a Quantaurus-QY Plus (Hamamatsu, C13534).

Photocatalytic Reactions
The experimental detail of photocatalytic reactions was essentially the same as reported previously. [23]The dye-sensitized photocatalyst of 20 mg was suspended in a NaI (≥99.5%;Kanto Chemical Co.) aqueous solution (10 mM, 100 mL).The pH of the reaction solution was adjusted to 4 with aqueous HCl solution.When Z-scheme overall water splitting was performed, 50 mg of PtO x /H-Cs-WO 3 photocatalyst [21] was used together as an O 2 -evolving photocatalyst.A 300 W xenon lamp (Cermax, PE300BF) fitted with a CM-1 cold mirror and an L42 cutoff filter to allow for visible light irradiation (λ > 400 nm) was used as the light source.The irradiation area was 44 cm 2 .The light intensity was measured using a calibrated silicon photodiode in the wavelength range of 400-600 nm.These photoreaction experiments were conducted at room temperature using a top irradiation-type cell connected to a closed gas circulation system made of glass.The evolved gases were analyzed by gas chromatography (Shimadzu, GC-8 A with a TCD detector and an MS-5 A column, argon carrier gas).
STH conversion efficiency was measured using an HAL-320 solar simulator as the light source and was estimated according to the following equation where R H , R O , ΔG°, P, and S are the rates of hydrogen/oxygen evolution (mol s À1 ) in Z-scheme water splitting, the standard Gibbs free energy of liquid water (237 Â 10 3 J mol À1 ), the intensity of simulated sunlight (100 mW cm À2 ), and the irradiation area (9 cm 2 ).
The AQY for H 2 evolution was measured using a band-pass filter (λ = 420 nm) and was estimated according to the following equation where A, R, and I represent the reaction coefficient (H 2 evolution, 4: O 2 evolution, 8), the H 2 or O 2 -evolution rate, and the rate of incident photons, respectively.The rate of incident photons (1.8 Â 10 15 photons cm À2 s À1 ) was measured by using a calibrated silicon photodiode.

Transient Absorption Measurements
Transient absorption spectroscopy measurements were performed by using an enVISion transient absorption system (Magnitude Instruments, State College, PA).The measurements were made as described in previous papers. [20,23,24]The modified dye-sensitized sample of 10 mg was suspended in an aqueous NaI (10 or 100 mM, 4 mL) in a quartz cuvette, and transient spectra were recorded in diffuse reflectance mode.The pH of the solution was adjusted to about 4 with aqueous H 2 SO 4 solution.The suspension was purged with Ar at least for 20 min before the measurements.

Characterization
To avoid redundant discussion, we do not describe here the characterization details of HCa 2 Nb 3 O 10 , its modified analogues (with Pt/Al 2 O 3 /RuP loadings), or PtO x /H-Cs-WO 3 , because they have already been reported elsewhere. [23,26]Figure 2a shows the FT IR spectra of the unmodified and polymer-modified RuP/Pt/ HCa 2 Nb 3 O 10 samples measured by the ATR method.[33] Therefore, it was confirmed that these samples were modified with the anionic polymers as intended.
As shown in Figure S1, Supporting Information, XRD patterns of the polymer-modified RuP/Pt/HCa 2 Nb 3 O 10 nanosheets were nearly identical to those with no polymer modification.In addition, SEM images in Figure S2, Supporting Information, show that the morphology is almost identical.Previously, it is reported that there is no change in morphology before and after PSS modification. [23]Thus, it was difficult to identify the loaded polymer by electron microscopy.Therefore, the distribution of polymer species on the surface of RuP/Pt/HCa 2 Nb 3 O 10 was investigated by infrared Raman microscopy.Figure 2b displays a typical IR mapping image of PMA/RuP/Pt/HCa 2 Nb 3 O 10 , which was made on the basis of a characteristic peak at 1183 cm À1 derived from the adsorbed PMA species.These images indicate that the PMA species were unevenly distributed on the surface at the resolution of the microscope.
Figure 2c shows the ζ-potential of the unmodified and modified RuP/Pt/HCa 2 Nb 3 O 10 samples.All the values of modified samples were shifted negative relative to the unmodified one.The pH-independent ζ-potential of the PSS-modified sample is consistent with the low pK a of the sulfonate group, whereas the negative shift of the ζ-potential of PMA and PAA with increasing pH is consistent with the weak acid behavior of their carboxylate groups.The intermediate behavior of PSSMA is consistent with its structure as a copolymer of strongly and weakly acidic monomers.The suppression of the back electron-transfer reaction involving I 3 À by PSS results from electrostatic repulsion between the anionic polymers and I 3 À . [22,23]hus, PAA, PMA, and PSSMA modification are all expected to suppress back electron transfer at pH 4, where all the polymers have negative ζ-potentials.
Figure 2d shows the UV-visible DRS of the unmodified and modified samples.In the DRS after modification with the polymer, a peak around 460 nm, which is attributed to the singlet metalto-ligand charge transfer ( 1 MLCT) absorption of RuP, appears.It can also be seen that the peak of the 1 MLCT absorption of RuP is redshifted by the polymer modification.This suggests that the excited state is stabilized or the ground state is destabilized due to the interaction of the anionic polymers with RuP. [34,35]igure 2e shows the steady-state emission spectra of the unmodified and modified samples.RuP on Al 2 O 3 exhibited an emission peak at around 670 nm, and the emission quantum yield was 4.1%.In contrast, the emission of RuP was almost completely quenched on Pt/HCa 2 Nb 3 O 10 due to excited-state electron transfer from RuP to the HCa 2 Nb 3 O 10 . [20]When the sample was modified by anionic polymers, the emission was also efficiently quenched.The emission quantum yield was 0.2% or less.Therefore, even though a slight peak shift of the 1 MLCT absorption of RuP was seen after polymer modification, the electron injection efficiency from the excited-state sensitizer was largely unaffected.

Half-Cell Reaction
Figure 3 shows the results of the half-cell H 2 -evolution reactions of each modified sample in 10 mM NaI solution.The reaction is described as follows.
As mentioned in Introduction, the accumulation of I 3 À in the reaction solution lowers the H 2 -evolution rate, as the reduction of reduction, as clearly seen in any of time course data in Figure 3.This was also confirmed by conducting a reaction in the presence of I 2 (see Supporting Information, Figure S3).The initial rates of H 2 evolution from the PAA-and PSSMAmodified samples were not significantly different from that of the unmodified sample at a light intensity of 28 mW cm À2 .In contrast, when the light intensity was 15 mW cm À2 (i.e., at lower light intensity), the initial H 2 -evolution rate was markedly increased compared to the unmodified sample.This is similar to the results of PSS modification reported previously. [23]At lower light intensity, the concentration of I 3 À , which is the oxidation product of the I À electron donor, [21] is lower than under higher light-intensity conditions.Thus, the effect of PAA-and PSSMA-modification on the H 2 -evolution activity appears only when the concentration of I 3 À is sufficiently low, as it does with PSS modification.In contrast, for the PMA-modified sample, the initial evolution rate increased even under at high light intensity and the largest increase was seen at the low light intensity.This suggests that PMA is effective in suppressing the back electrontransfer reaction even at relatively high concentration of I 3 À .

Z-Scheme Overall Water-Splitting Reactions
Figure 4 shows the results of the Z-scheme overall water-splitting reactions of each modified sample in combination with PtO x /H-Cs-WO 3 as an O 2 -evolving photocatalyst in 5 mM NaI aqueous solution.All samples produced H 2 and O 2 at a ratio of 2:1, and the polymer-modification improved both H 2 and O 2 evolution.The difference between PMA and other polymers, which appeared in the half-cell reaction, was not observed in the Z-scheme water splitting.We hypothesize that this occurred because the I 3 À electron acceptor was consumed efficiently by PtO x /H-Cs-WO 3 in the Z-scheme reaction and thus the I 3 À remained low. [21]Under these conditions, we would expect similar activity for all anionic polymers used to modify RuP/Pt/HCa 2 Nb 3 O 10 .
Modification with a monomer (SS salt) also resulted in an increase in the H 2 -and O 2 -evolution rates, but it was not as effective as that observed with PSS modification.As reported previously, [22] the monomer is unlikely to cover the surface as effectively as the PSS polymer.
Z-scheme water-splitting reactions were also conducted using Al 2 O 3 -modified nanosheet samples (Figure 4b).Al 2 O 3 modification can suppress back electron-transfer reaction from the conduction band of the semiconductor to the oxidized form of the Ru dye in Ru-dye-sensitized metal oxide systems. [24]In all cases, it is clear that Al 2 O 3 modification improved the water-splitting activity, compared to non-Al 2 O 3 cases.
Considering the results of both half-cell H 2 evolution and Z-scheme water splitting, we concluded that PMA-modified RuP/Al 2 O 3 /Pt/HCa 2 Nb 3 O 10 nanosheets were the most suitable for further investigation as photocatalysts.We therefore measured the STH energy conversion efficiencies under 1 sun (100 mW cm À2 ) conditions, in combination with different amounts of PtO x /H-Cs-WO 3 and different NaI concentrations.The Z-scheme water-splitting activity is known to vary with the photocatalyst amount and the redox mediator concentration. [23,36]The STH values under various conditions are displayed in Table 1 and Figure 5, along with data for the PSS-modified sample for comparison. [23]The corresponding time course data are shown in Figure S4, Supporting Information.In 1 mM NaI solution (entry 1), the PMA-modified material exhibited 4-7 times higher STH value than that of the PSS-modified one.This result indicates that the PMA system could work better than the PSS system at low concentrations of NaI.This is consistent with the results of the half-cell H 2 -evolution experiment under low light-intensity conditions (Figure 3b).Increasing the concentration of NaI from 1 to 5-10 mM resulted in an increase of the STH.As the result, the highest STH value of 0.12% AE 0.01%, calculated based on the H 2 -evolution rate, was obtained (entries 2 and 3), which was identical to the previously reported value using PSS.Note that the O 2 -evolution-based STH values were slightly lower than those based on H 2 evolution because of slower O 2 evolution.
A further increase in the NaI concentration to 20 mM led to a drop in STH values (entry 4).This is basically because at high I À concentration, oxidation of I À occurs in competition to oxidation of water. [37]We note that the STH of the PMA-modified sample was higher than that of the PSS-modified sample, with the H 2 /O 2 ratio of 2.1, close to the stoichiometric value.This result again confirms that the PMA-modified sample is workable even under high I 3 À concentration conditions, where it is more efficient than the PSS-modified sample.This suggests that the paired O 2 -evolving photocatalyst could work more efficiently than in the PSS system.
When the amount of O 2 -evolving photocatalyst was varied, the STH was higher for the PSS-modified photocatalyst at 25 mg (entry 6), but the stoichiometry of H 2 /O 2 evolution was not significantly different, and at 50 mg (entry 2), the activity was almost the same.In contrast, at 10 mg (entry 5), the H 2 /O 2 was 7.1 for PSS modification and 3.8 for PMA modification, closer to the stoichiometric ratio, even though the STH on the H 2 -evolving side was not significantly different.This may be because PMA inhibits the back reaction with I 3 À even at higher I 3 À concentration, which is more favorable for the O 2 -evolving photocatalyst.
The AQY was also measured to be 5.1% based on the O 2 -evolution rate at 420 nm (see Figure S5, Supporting Information).This value is a little higher than that of PSS-modified sample (4.1%), [23] and the highest among Z-scheme overall water-splitting systems that are using a dye-sensitized H 2 -evolution photocatalyst.

Transient Absorption Spectroscopy
We carried out laser flash photolysis/transient absorption measurements to investigate the difference between the PMA-and PSS-modified systems.Figure 6a shows the time-dependent absorbance change in diffuse reflectance of RuP-sensitized HCa 2 Nb 3 O 10 nanosheets in 10 mM aqueous NaI solutions monitored at 475 nm.By monitoring at 475 nm, bleaching and recovery of the 1 MLCT absorption of ground-state RuP could be observed.For both the PMA-and PSS-modified samples, the injection yield was lower and the 1 MLCT bleaching recovery of RuP was slower than with the unmodified sample.From these results, we can conclude that the PMA modifier can suppress the access of not only I 3 À but also I À to the photocatalyst surface, as previously reported for PSS. [23]The bleaching recovery measured in aqueous NaI was much faster than that in pure water, [24] consistent with fast electron donation from I À ions to the oneelectron-oxidized RuP complex.Comparing the results of PMA-and PSS-modified samples, the bleaching recovery of RuP was faster with PMA, suggesting that it blocks access of I À to the sensitizer less effectively than PSS.This is reasonable considering the results of the ζ-potential measurements, which showed that PMA provided less negative surface charge than PSS at pH 4 (Figure 2c); therefore, electrostatic repulsion with I À should be weaker in the PMA-modified sample than in the PSS-modified one.
To investigate the suppression of the I 3 À reduction rate, timedependent absorbance changes were monitored at 380 nm, where I 3 À absorbs strongly.Figure 6b shows the time-dependent absorbance changes at 380 nm, and the measured lifetimes are tabulated in Table 2.All the profiles could be fitted by double-or triple-exponential functions (Equation ( 5) and ( 6)) In all samples, there was an increase and subsequent decay of ΔAbs, corresponding to an increase and decrease in I 3 À concentration, respectively.Compared to the unmodified sample, the PMA-and PSS-modified samples showed a smaller initial increase.In addition, τ 1 , corresponding to generation of I 3 À , of the PMA-and PSS-modified samples was an order of magnitude longer than that of the unmodified sample.This result indicates again that the production of I 3 À is suppressed; that is, the reaction between the photocatalyst and I À is suppressed by anionic polymer modification to a certain extent.τ 2 corresponds to the reaction of I 3 À with the photocatalyst. [23] was confirmed that τ 2 became longer after PMA or PSS modification, attributable to the repulsive interaction between the negatively charged nanosheet surface and I 3 À ions. [22]We note that τ 1 and τ 2 of the PMA-modified sample were both slightly shorter than those of the PSS-modified sample.Therefore, the PMA modifier suppresses the access of both I À and I 3 À less effectively than PSS, but a faster rate of reaction of I À with the one-electron-oxidized RuP was guaranteed.Consequently, higher STH efficiencies were obtained with the PMA-modified sample in aqueous NaI solutions at relatively low and high concentrations, compared to the PSS system.It is also noted that, since the ability to donate an electron from I À is different in each case (PMA vs. PSS) as revealed by transient absorption spectroscopy (Figure 6), any of the forward reaction paths are considered to be the rate limiting.This would be a plausible reason why the STHs of PMA-and PSS-modified samples were almost the same in NaI concentrations of 5-10 mM.

Conclusions
We have investigated the modification of Ru-dye-sensitized photocatalysts with various anionic polymers to improve photocatalytic activity for solar-driven Z-scheme overall water splitting from an aqueous NaI solution.Transient absorption spectroscopy showed that both PMA and PSS salt had similar effects on the suppression of back electron transfer from the conduction band of HCa 2 Nb 3 O 10 to I 3 À , whereas the electron donation process from I À was more efficient in the PMA system than in the PSS system.The use of PMA as the modifier provided higher water-splitting activity, compared to the previously reported PSS, when the NaI concentration was relatively low or high.This may be advantageous in Z-schemes with other water oxidation photocatalysts that do not have good tolerance toward the oxidation of I À (i.e., the reverse reaction that competes with water oxidation).We also note that the highest STH energy conversion efficiencies recorded by the two systems were similar to each other (0.12%), strongly suggesting that there is a bottleneck other than the back electron-transfer reaction to I 3 À .

Scheme 1 .
Scheme 1. Photoredox scheme for Z-scheme water splitting by using a dye-sensitized wide-gap metal oxide semiconductor as the H 2 -evolution photosystem.HOMO: highest occupied molecular orbital, LUMO: lowest unoccupied molecular orbital, D: electron donor, A: electron acceptor.

Figure 1 .
Figure 1.Anionic polymers used in this study as modifiers.

Figure 2 .
Figure 2. Characterization of polymer-modified RuP/Pt/HCa 2 Nb 3 O 10 nanosheets."None" means unmodified RuP/Pt/HCa 2 Nb 3 O 10 .a) FT IR spectra.b) Optical microscope and IR mapping images of sodium polymethacrylate (PMA)/RuP/Pt/HCa 2 Nb 3 O 10 made using the characteristic IR peak at 1183 cm À1 .The color codes of the mapping image represent the intensity of the signal; density: high (red) and low (blue).c) ζ-potentials as a function of pH.d) UV-visible diffuse reflectance spectra.The inset shows an enlarged view.e) Steady-state emission spectra.The samples were suspended in H 2 O saturated with Ar.Excitation wavelength was 460 nm.

Figure 6 .
Figure 6.Time-dependent absorbance change in transient diffuse reflectance of RuP-sensitized HCa 2 Nb 3 O 10 nanosheets with and without polymer modification recorded in aqueous NaI solutions suspension (pH 4) monitored at a) 475 nm (10 mM NaI) and b) 380 nm (100 mM NaI).

Table 2 .
Absorption decay lifetimes of I 3