Effect of Porphyrin Molecular Structure on Water Splitting Activity of a Ktao 3 Photocatalyst

Photocatalytic water splitting is one of the ideal methods for solving the global energy crisis and its associated environmental problems. In this study, the effect of altering the molecular structure of porphyrins was investigated to improve the water splitting activity of Zr-doped KTaO 3 (KTa(Zr)O 3) modified with porphyrin dyes. UV-vis spectra indicated that porphyrins with long alkoxy chains tended to form well-developed H-aggregates on the KTa(Zr)O 3 surface. The photocatalytic activity of Pt-loaded KTa(Zr)O 3 was improved by using porphyrins with longer alkoxy chains because of the improvement in the charge migration between porphyrin dye molecules. While the charge transfer between the inorganic semiconductor and porphyrin dye interface is important, it was found that the formation of H-aggregation was more effective in improving the water splitting activity of the porphyrin-modified photocatalysts.


Introduction
Photocatalytic water splitting is one of the most promising strategies for renewable and clean production of hydrogen.This method of hydrogen production has attracted much attention not only for the conversion of light energy but also from a scientific viewpoint, because photocatalytic water splitting is regarded as an artificial photosynthetic system.In photosynthesis, two photons are required to facilitate NADPH synthesis from NADP + and water [1].This two-step excitation process, the so-called "Z-scheme", has been widely studied in both fundamental and applied fields of research [2][3][4][5].In the past decade, various systems capable of Z-scheme water splitting (Figure 1b), such as Pt or Ru/SrTiO 3 :Rh-Fe 3+ /Fe 2+ -WO 3 , BiVO 4 , or Bi 2 MoO 6 [6,7], Pt/TaON-IO 3 ´/I ´-PtO X /WO 3 or RuO 2 /TaON [8,9], Pt/ZrO 2 /TaON-IO 3 ´/I ´-Ir/TiO 2 /Ta 3 N 5 or RuO 2 /TaON [10,11], and other systems [12][13][14][15], have been reported.Our research group has also reported Z-scheme water splitting using porphyrin-dye-modified inorganic semiconductors, such as KTa(Zr)O 3 [16][17][18], GaN:ZnO [19] and TaON [20].In these dye-modified photocatalysts, the dyes used for modification improve the utilization efficiency of visible light and increase the charge separation lifetime owing to the electron transfer between the excited inorganic semiconductors and the dyes, as shown in Figure 1c.In our previous results, it was found that 5,10,15,20-tetrakis(tetraphenyl)porphyrin-Cr(III) chloride (Cr-TPPCl, Dye 1 in this study) had a positive effect on the photocatalytic activity of KTa(Zr)O 3 for overall water splitting [17].Cr-TPPCl formed aggregates on the surface of KTa(Zr)O 3 , which is similar to the structure of the chlorophyll special pair in the naturally occurring photosynthetic reaction center [21].
Thus, we have an interest in the relation between the aggregation states of porphyrins and their effects on the photocatalytic activity of the Cr-TPPCl/KTa(Zr)O 3 system.Therefore, in this study, we have synthesized various porphyrins that contain chromium-coordinated porphyrin structures with different alkoxy chain lengths and used these compounds as modification dyes for the KTa(Zr)O 3 photocatalyst.The aggregation states of the porphyrins were investigated by using UV-vis absorption spectroscopy; photocatalytic water splitting on KTa(Zr)O 3 modified with various porphyrins was also demonstrated.
Catalysts 2016, 6, 42 2 of 10 Thus, we have an interest in the relation between the aggregation states of porphyrins and their effects on the photocatalytic activity of the Cr-TPPCl/KTa(Zr)O3 system.Therefore, in this study, we have synthesized various porphyrins that contain chromium-coordinated porphyrin structures with different alkoxy chain lengths and used these compounds as modification dyes for the KTa(Zr)O3 photocatalyst.The aggregation states of the porphyrins were investigated by using UV-vis absorption spectroscopy; photocatalytic water splitting on KTa(Zr)O3 modified with various porphyrins was also demonstrated.

Results and Discussion
Various porphyrins (Figure 2 The 1 H NMR, 13 C NMR and CSI-MS spectra of the porphyrins are summarized in the Supplementary Materials (Figures S1-S9).From these results, we could confirm that the porphyrins were successfully synthesized.Figure 3 shows the UV-vis spectra of the synthesized porphyrins dissolved in CH2Cl2.In all spectra, the Soret and Q-band absorptions, which are typical porphyrin absorption bands, were observed at 450 and 560-610 nm, respectively.Compared with the spectrum of Dye 1 (Figure 3a), the -OC 12 H 25 5 -OC 16 H 33 6 Figure 1.Various heterogeneous photocatalyst systems for water splitting reaction.(a) A one-photon system; (b) the Z-scheme system; and (c) the Cr-TPPCl-modified KTa(Zr)O 3 system.

Results and Discussion
Various porphyrins (Figure 2  13 C NMR and CSI-MS spectra of the porphyrins are summarized in the Supplementary Materials (Figures S1-S9).From these results, we could confirm that the porphyrins were successfully synthesized.
Thus, we have an interest in the relation between the aggregation states of porphyrins and their effects on the photocatalytic activity of the Cr-TPPCl/KTa(Zr)O3 system.Therefore, in this study, we have synthesized various porphyrins that contain chromium-coordinated porphyrin structures with different alkoxy chain lengths and used these compounds as modification dyes for the KTa(Zr)O3 photocatalyst.The aggregation states of the porphyrins were investigated by using UV-vis absorption spectroscopy; photocatalytic water splitting on KTa(Zr)O3 modified with various porphyrins was also demonstrated.

Results and Discussion
Various porphyrins (Figure 2 The 1 H NMR, 13 C NMR and CSI-MS spectra of the porphyrins are summarized in the Supplementary Materials (Figures S1-S9).From these results, we could confirm that the porphyrins were successfully synthesized.Figure 3 shows the UV-vis spectra of the synthesized porphyrins dissolved in CH2Cl2.In all spectra, the Soret and Q-band absorptions, which are typical porphyrin absorption bands, were observed at 450 and 560-610 nm, respectively.Compared with the spectrum of Dye 1 (Figure 3a), the  Figure 3 shows the UV-vis spectra of the synthesized porphyrins dissolved in CH 2 Cl 2 .In all spectra, the Soret and Q-band absorptions, which are typical porphyrin absorption bands, were observed at 450 and 560-610 nm, respectively.Compared with the spectrum of Dye 1 (Figure 3a), the absorption peaks of Dyes 2-6 were slightly shifted to longer wavelengths because of electron donation from alkoxy groups.The peak intensities in the absorption spectra of Dyes 2-6 appeared at the same positions; thus, it appears that the porphyrin dyes 2-6 exhibit similar energy states.To investigate the energy states of the porphyrin dyes, cyclic voltammetry was performed (Figure S10).The first oxidation-reduction potential was observed at ´1.48 V (vs.Ag/Ag + ) for Dye 1, while the potentials for Dyes 2-6 were observed at ´1.51 V (vs.Ag/Ag + ).The small difference between Dye 1 and the other dyes was also attributed to electron donation from the alkoxy groups.The ability of the alkoxy chain to donate electron density to the porphyrin ring appears to be similar, regardless of the length of the chain.
Catalysts 2016, 6, 42 3 of 10 absorption peaks of Dyes 2-6 were slightly shifted to longer wavelengths because of electron donation from alkoxy groups.The peak intensities in the absorption spectra of Dyes 2-6 appeared at the same positions; thus, it appears that the porphyrin dyes 2-6 exhibit similar energy states.To investigate the energy states of the porphyrin dyes, cyclic voltammetry was performed (Figure S10).
The first oxidation-reduction potential was observed at −1.48 V (vs.Ag/Ag + ) for Dye 1, while the potentials for Dyes 2-6 were observed at −1.51 V (vs.Ag/Ag + ).The small difference between Dye 1 and the other dyes was also attributed to electron donation from the alkoxy groups.The ability of the alkoxy chain to donate electron density to the porphyrin ring appears to be similar, regardless of the length of the chain.The morphology of the prepared samples was determined using FIB-SEM.Figure 4 shows SEM images of the various KTa(Zr)O3 photocatalysts modified with Dyes 1-6.All the samples exhibited small particles of various size distributions (<10 nm) deposited on the surface of larger KTa(Zr)O3 particles.These smaller particles appear to be Pt co-catalyst and/or aggregates of porphyrin dyes loaded on the KTa(Zr)O3 surface.The particle size of KTa(Zr)O3 was approximately 1 μm, which corresponded well with the results of a laser diffraction analysis of the particle diameters (Figure S11).TEM images of the dye-modified KTa(Zr)O3 samples are shown in Figure 5.In all images, the Pt-co-catalysts were observed as black small particles loaded on the larger KTa(Zr)O3 particles.The average diameter of Pt-co-catalysts was approximately 2-3 nm.Compared to a TEM image of KTa(Zr)O3 without dye modification (Figure S12), the dye-modified KTa(Zr)O3 surface seemed to be dappled, which suggested that the porphyrin dyes partially coated the KTa(Zr)O3 surface.The morphology of the prepared samples was determined using FIB-SEM.Figure 4 shows SEM images of the various KTa(Zr)O 3 photocatalysts modified with Dyes 1-6.All the samples exhibited small particles of various size distributions (<10 nm) deposited on the surface of larger KTa(Zr)O 3 particles.These smaller particles appear to be Pt co-catalyst and/or aggregates of porphyrin dyes loaded on the KTa(Zr)O 3 surface.The particle size of KTa(Zr)O 3 was approximately 1 µm, which corresponded well with the results of a laser diffraction analysis of the particle diameters (Figure S11).TEM images of the dye-modified KTa(Zr)O 3 samples are shown in Figure 5.In all images, the Pt-co-catalysts were observed as black small particles loaded on the larger KTa(Zr)O 3 particles.The average diameter of Pt-co-catalysts was approximately 2-3 nm.Compared to a TEM image of KTa(Zr)O 3 without dye modification (Figure S12), the dye-modified KTa(Zr)O 3 surface seemed to be dappled, which suggested that the porphyrin dyes partially coated the KTa(Zr)O 3 surface.
Catalysts 2016, 6, 42 3 of 10 absorption peaks of Dyes 2-6 were slightly shifted to longer wavelengths because of electron donation from alkoxy groups.The peak intensities in the absorption spectra of Dyes 2-6 appeared at the same positions; thus, it appears that the porphyrin dyes 2-6 exhibit similar energy states.To investigate the energy states of the porphyrin dyes, cyclic voltammetry was performed (Figure S10).
The first oxidation-reduction potential was observed at −1.48 V (vs.Ag/Ag + ) for Dye 1, while the potentials for Dyes 2-6 were observed at −1.51 V (vs.Ag/Ag + ).The small difference between Dye 1 and the other dyes was also attributed to electron donation from the alkoxy groups.The ability of the alkoxy chain to donate electron density to the porphyrin ring appears to be similar, regardless of the length of the chain.The morphology of the prepared samples was determined using FIB-SEM.Figure 4 shows SEM images of the various KTa(Zr)O3 photocatalysts modified with Dyes 1-6.All the samples exhibited small particles of various size distributions (<10 nm) deposited on the surface of larger KTa(Zr)O3 particles.These smaller particles appear to be Pt co-catalyst and/or aggregates of porphyrin dyes loaded on the KTa(Zr)O3 surface.The particle size of KTa(Zr)O3 was approximately 1 μm, which corresponded well with the results of a laser diffraction analysis of the particle diameters (Figure S11).TEM images of the dye-modified KTa(Zr)O3 samples are shown in Figure 5.In all images, the Pt-co-catalysts were observed as black small particles loaded on the larger KTa(Zr)O3 particles.The average diameter of Pt-co-catalysts was approximately 2-3 nm.Compared to a TEM image of KTa(Zr)O3 without dye modification (Figure S12), the dye-modified KTa(Zr)O3 surface seemed to be dappled, which suggested that the porphyrin dyes partially coated the KTa(Zr)O3 surface.To investigate the aggregation states of Dyes 1-6, UV-vis absorption spectra were measured.Figure 6 shows the UV-vis spectra of the KTa(Zr)O3 photocatalyst modified with Dyes 1-6.As shown in Figure S13, KTa(Zr)O3 absorbed UV light at wavelengths shorter than 350 nm.Broad absorption peaks corresponding to Dyes 1-6 were observed in the visible region from 420 to 700 nm.In comparison with the absorption spectra of the dye solutions (Figure 3), the Soret band absorption peaks of Dyes 1-6 became broad and blue-shifted (6-26 nm), while the Q-band absorption peaks were red-shifted (11-38 nm).This tendency was most obviously observed in the porphyrins bearing longer alkoxy chains, such as Dye 5 and Dye 6.These peak shifts in the absorption spectra occurred as a result of the differences in the aggregation states of the dye compounds.In general, a red-shift is assigned to J-aggregations and a blue-shift to H-aggregations [22].The peak-shift of J-and  To investigate the aggregation states of Dyes 1-6, UV-vis absorption spectra were measured.Figure 6 shows the UV-vis spectra of the KTa(Zr)O3 photocatalyst modified with Dyes 1-6.As shown in Figure S13, KTa(Zr)O3 absorbed UV light at wavelengths shorter than 350 nm.Broad absorption peaks corresponding to Dyes 1-6 were observed in the visible region from 420 to 700 nm.In comparison with the absorption spectra of the dye solutions (Figure 3), the Soret band absorption peaks of Dyes 1-6 became broad and blue-shifted (6-26 nm), while the Q-band absorption peaks were red-shifted (11-38 nm).This tendency was most obviously observed in the porphyrins bearing longer alkoxy chains, such as Dye 5 and Dye 6.These peak shifts in the absorption spectra occurred as a result of the differences in the aggregation states of the dye compounds.In general, a red-shift is assigned to J-aggregations and a blue-shift to H-aggregations [22].The peak-shift of J-and To investigate the aggregation states of Dyes 1-6, UV-vis absorption spectra were measured.Figure 6 shows the UV-vis spectra of the KTa(Zr)O 3 photocatalyst modified with Dyes 1-6.As shown in Figure S13, KTa(Zr)O 3 absorbed UV light at wavelengths shorter than 350 nm.Broad absorption peaks corresponding to Dyes 1-6 were observed in the visible region from 420 to 700 nm.In comparison with the absorption spectra of the dye solutions (Figure 3), the Soret band absorption peaks of Dyes 1-6 became broad and blue-shifted (6-26 nm), while the Q-band absorption peaks were red-shifted (11-38 nm).This tendency was most obviously observed in the porphyrins bearing longer alkoxy chains, such as Dye 5 and Dye 6.These peak shifts in the absorption spectra occurred as a result of the differences in the aggregation states of the dye compounds.In general, a red-shift is assigned to J-aggregations and a blue-shift to H-aggregations [22].The peak-shift of Jand H-aggregates are dependent on the differences in transition dipoles: transition dipoles of J-aggregates are aligned parallel, while those of H-aggregates are aligned perpendicular.In the case of porphyrin derivatives, it has been reported that the aggregation state of the porphyrins can be determined by the shift direction of the Soret band [23,24].Thus, all porphyrins used in this study appear to form H-aggregates on the KTa(Zr)O 3 surface.Furthermore, the porphyrins that have longer alkoxy chains formed well-developed H-aggregates.In comparison with amorphous aggregates, dye aggregates have an advantage when transporting photogenerated charge carriers.
Catalysts 2016, 6, 42 5 of 10 H-aggregates are dependent on the differences in transition dipoles: transition dipoles of J-aggregates are aligned parallel, while those of H-aggregates are aligned perpendicular.In the case of porphyrin derivatives, it has been reported that the aggregation state of the porphyrins can be determined by the shift direction of the Soret band [23,24].Thus, all porphyrins used in this study appear to form H-aggregates on the KTa(Zr)O3 surface.Furthermore, the porphyrins that have longer alkoxy chains formed well-developed H-aggregates.In comparison with amorphous aggregates, dye aggregates have an advantage when transporting photogenerated charge carriers.Figure 7 shows the time course for water splitting on Pt-loaded KTa(Zr)O3 modified with Dye 2. A 1-2 h induction period was observed after starting photoirradiation, and then the amounts of H2 and O2 formed linearly increased with time.This tendency was observed in all samples (Figure S14).This behavior might be attributed to state changes of the photocatalysts, such as changes in the Pt co-catalysts' oxidation states or oxidative dye decomposition.The results of a recycle test of the photocatalyst are also shown in Figure 7.The circulating Ar gas was evacuated and exchanged every 5 h.Liner gas formation was observed during each run, and formation rates were slightly decreased over the 15 h period of the experiment.After the photocatalytic reaction, the amount of Dye 2 remaining on the KTa(Zr)O3 surface had decreased to 40% of the original loading (Figure S15).Although the porphyrin dyes remained in the reaction mixture for at least 15 h under the photocatalytic water splitting conditions, improving porphyrin stability to allow for their use in prolonged reactions is necessary.Table 1 shows H2 and O2 formation rates for photocatalytic water splitting on Pt-loaded KTa(Zr)O3 modified with Dyes 1-6.In comparison with the result observed for the unmodified Pt-loaded KTa(Zr)O3, the photocatalytic activity of the KTa(Zr)O3 catalysts was improved by using Figure 7 shows the time course for water splitting on Pt-loaded KTa(Zr)O 3 modified with Dye 2. A 1-2 h induction period was observed after starting photoirradiation, and then the amounts of H 2 and O 2 formed linearly increased with time.This tendency was observed in all samples (Figure S14).This behavior might be attributed to state changes of the photocatalysts, such as changes in the Pt co-catalysts' oxidation states or oxidative dye decomposition.The results of a recycle test of the photocatalyst are also shown in Figure 7.The circulating Ar gas was evacuated and exchanged every 5 h.Liner gas formation was observed during each run, and formation rates were slightly decreased over the 15 h period of the experiment.After the photocatalytic reaction, the amount of Dye 2 remaining on the KTa(Zr)O 3 surface had decreased to 40% of the original loading (Figure S15).Although the porphyrin dyes remained in the reaction mixture for at least 15 h under the photocatalytic water splitting conditions, improving porphyrin stability to allow for their use in prolonged reactions is necessary.H-aggregates are dependent on the differences in transition dipoles: transition dipoles of J-aggregates are aligned parallel, while those of H-aggregates are aligned perpendicular.In the case of porphyrin derivatives, it has been reported that the aggregation state of the porphyrins can be determined by the shift direction of the Soret band [23,24].Thus, all porphyrins used in this study appear to form H-aggregates on the KTa(Zr)O3 surface.Furthermore, the porphyrins that have longer alkoxy chains formed well-developed H-aggregates.In comparison with amorphous aggregates, dye aggregates have an advantage when transporting photogenerated charge carriers.Figure 7 shows the time course for water splitting on Pt-loaded KTa(Zr)O3 modified with Dye 2. A 1-2 h induction period was observed after starting photoirradiation, and then the amounts of H2 and O2 formed linearly increased with time.This tendency was observed in all samples (Figure S14).This behavior might be attributed to state changes of the photocatalysts, such as changes in the Pt co-catalysts' oxidation states or oxidative dye decomposition.The results of a recycle test of the photocatalyst are also shown in Figure 7.The circulating Ar gas was evacuated and exchanged every 5 h.Liner gas formation was observed during each run, and formation rates were slightly decreased over the 15 h period of the experiment.After the photocatalytic reaction, the amount of Dye 2 remaining on the KTa(Zr)O3 surface had decreased to 40% of the original loading (Figure S15).Although the porphyrin dyes remained in the reaction mixture for at least 15 h under the photocatalytic water splitting conditions, improving porphyrin stability to allow for their use in prolonged reactions is necessary.Table 1 shows H2 and O2 formation rates for photocatalytic water splitting on Pt-loaded KTa(Zr)O3 modified with Dyes 1-6.In comparison with the result observed for the unmodified Pt-loaded KTa(Zr)O3, the photocatalytic activity of the KTa(Zr)O3 catalysts was improved by using Table 1 shows H 2 and O 2 formation rates for photocatalytic water splitting on Pt-loaded KTa(Zr)O 3 modified with Dyes 1-6.In comparison with the result observed for the unmodified Pt-loaded KTa(Zr)O 3 , the photocatalytic activity of the KTa(Zr)O 3 catalysts was improved by using the porphyrins.In all cases, the gas formation ratio is slightly different from the expected stoichiometry.The errors in the values of formation rates were evaluated using the Pt/Dye 2/KTa(Zr)O 3 photocatalyst.As shown in Figure S16, the error observed in the oxygen formation rate was larger than that of the hydrogen formation rate.This difference may come from the low sensitivity for oxygen detection in gas chromatography with an Ar carrier and the hole consumption by oxidative dye decomposition.From the viewpoint of dye structure, the porphyrins which have small functional groups (Dye 1 and 2) tended to have a larger positive influence on photocatalytic activity.The porphyrins with small alkoxy groups closely adsorbed onto the KTa(Zr)O 3 surface; thus, the charge transfer between the photo-excited electrons in KTa(Zr)O 3 and the porphyrin dyes occurred easily compared with the porphyrins that contained large and bulky alkoxy groups.In our previous study, it was found that efficient electron transfer between the inorganic semiconductor surface and a porphine ring (or its central metal) is important for prolonging the charge separation lifetime and improving the water splitting activity of dye-modified photocatalysts [17,18].Therefore, Dye 2-modified KTa(Zr)O 3 , containing a methoxy group, showed the highest photocatalytic activity for water splitting in this study.On the contrary, the gas formation rates gradually increased with increasing alkoxy chain length (i.e., from Dye 3 to Dye 6).As mentioned above, the porphyrins with longer alkoxy chains formed well-developed H-aggregates on the KTa(Zr)O 3 surface.The photogenerated electrons (holes) are transported via redox of the porphyrin units; thus, the porphyrin aggregations have great advantages for charge migration because of their overlapping π electron clouds.Improvements in the electronic conductivity of the dye aggregates lead to an increase in photogenerated charges that reach the reaction sites on the photocatalyst surface.Therefore, the photocatalytic activity of KTa(Zr)O 3 was improved by using the porphyrins with longer alkoxy chains.As a result of these two different factors, the modification effects of Dyes 1-6 were observed, as shown in Table 1.Although the porphyrin H-aggregates effectively improved the water splitting activity of the KTa(Zr)O 3 photocatalyst, the charge transfer between the inorganic semiconductor and porphyrin dye interface appears to be the most important factor for achieving efficient water splitting.
The water was evaporated to dryness with stirring.The resultant mixture was ground in an alumina mortar for 10 min and calcined in an alumina crucible at 900 ˝C for 10 h under an air atmosphere.The crystal structure of the resultant sample was determined by using powder X-ray diffraction (Figure S17).

Preparation of Porphyrin-Modified KTa(Zr)O 3 Photocatalysts
An "evaporation to dryness" method was employed to modify the KTa(Zr)O 3 surface using pyridine as the solvent at 130 ˝C.In this study, the dye amount was unified at 0.8 wt.%.For loading the Pt co-catalyst onto the porphyrin/KTa(Zr)O 3 , the "evaporation to dryness" method was also used using water as the solvent.Pt(NH 3 ) 4 (NO 3 ) 2 (Sigma-Aldrich Co. LLC., Tokyo, Japan) was used as the Pt co-catalyst precursor.The loading amount of the Pt co-catalyst was 0.2 wt.%.

Photocatalytic Water Splitting Reaction
The photocatalytic activity for water splitting was evaluated using a closed-circulation-type glass reactor with a dead volume of 450 mL (Figure S18).The catalyst (50 mg) was dispersed in 30 mL of distilled water.A quartz reaction cell was illuminated by a 300 W Xe lamp (Cermax, PE300BUV, Excelitas Technologies Corp., Waltham, MA USA).The spectrum of the excitation light source is shown in Supplementary Materials (Figure S19).Argon gas (10.67 kPa) was used as a carrier gas in the reactor.Gas leakage was checked for 1 h in the dark prior to reaction initiation.Gas evolution amounts were measured using a gas chromatograph (GC-8A, Shimadzu Corp., Kyoto, Japan).

Photocatalyst Characterization
TEM (JEM-2100TI, JEOL, Tokyo, Japan) and FIB-SEM (Versa 3D HiVac, FEI, Hillsboro, OH, USA) were used to observe the surface morphology of the dye-modified KTa(Zr)O 3 samples.UV-vis absorption spectra of the porphyrin-pyridine solutions and the dye-modified KTa(Zr)O 3 were measured using a spectrophotometer (U-3310, Hitachi, Tokyo, Japan).The 1 H and 13 C NMR spectra were recorded on a Bruker AVANCE-400 NMR spectrometer (400 MHz).The chemical shifts are reported as δ values (ppm) relative to internal tetramethylsilane.The coupling constants (J) are given in hertz.Absorption spectra were recorded on a Shimadzu UV-3600 spectrometer.Mass spectra were recorded on either a Bruker Autoflex mass spectrometer or a JEOL JMS-T100LP mass spectrometer.Silica gel 60 F254 (Merck, Frankfurt, Germany) was used for analytical thin-layer chromatography.Cyclic voltammetry was performed using a three-electrode cell, with a glassy carbon electrode as the working electrode, a platinum (Pt) wire as the counter electrode and 0.01 M Ag/AgNO 3 in 0.1 M tetrabutylammonium perchlorate as the reference electrode.An X-ray diffractometer (RINT-2500, Rigaku, Tokyo, Japan) was used to determine the crystal structures of the samples.

Conclusions
The effect of porphyrin molecular structure on the water splitting activity of porphyrin-modified KTa(Zr)O 3 was investigated in this study.Although the surface morphologies of the Pt/porphyrin dye/KTa(Zr)O 3 photocatalysts were similar in SEM and TEM observations, UV-vis spectra indicated that the porphyrins with long alkoxy chains formed well-developed H-aggregates on the KTa(Zr)O 3 surface.The photocatalytic activity of Pt-loaded KTa(Zr)O 3 was improved by the porphyrins with longer alkoxy chains because of the improvement in the charge migration between the porphyrin dye molecules.Although the charge transfer between the inorganic semiconductor and porphyrin dye interface is important for water splitting, this study revealed that the formation of H-aggregates was more effective in improving the activity of porphyrin-modified photocatalysts.

Figure 2 .
Figure 2. Molecular structures of the porphyrin dyes used for the modification of KTa(Zr)O3.

Figure 2 .
Figure 2. Molecular structures of the porphyrin dyes used for the modification of KTa(Zr)O3.

Figure 2 .
Figure 2. Molecular structures of the porphyrin dyes used for the modification of KTa(Zr)O 3 .