Crystal Facet-Dependent Intrinsic Charge Separation on Well-Defined Bi 4 TaO 8 Cl Nanoplate for Efficient Photocatalytic Water Oxidation

The development of photocatalysts with wide spectral absorption and high charge separation e ﬃ ciency has always been a pursued objective for photocatalytic solar energy conversion. Herein, we reported a wide-range visible-light-active Bi 4 TaO 8 Cl (BTOC) single crystal nanoplate with dominating {110} and {001} facets for enhancing the intrinsic charge separation e ﬃ ciency. Insitu selective photodeposition of metals and metal oxides provides evidences of photogenerated electrons and holes spatially separated on {110} and {001} coexposed facets of BTOC, respectively. The intrinsic charge separation e ﬃ ciency was demonstrated to be closely dependent on the crystal facets, which can be modulated by tuning the coexposed crystal facet ratio. Further surface modi ﬁ cation of BTOC with suitable dual cocatalyst Ag and RuO x enables remarkable improvement of charge separation e ﬃ ciency and photocatalytic water oxidation performance. Investigation by comparison between well-de ﬁ ned BTOC nanoplate and BTOC nanoparticles con ﬁ rmed the signi ﬁ cance of coexposed crystal facets for e ﬃ cient spatial charge separation and the blocking of reverse reaction from Fe 2+ to Fe 3+ ions during water oxidation reaction, indicating that rational modulation of exposed crystal facets is signi ﬁ cant for controlling the intrinsic charge separation e ﬃ ciency on Bi 4 TaO 8 Cl photocatalyst for e ﬃ cient photocatalytic water splitting.


Introduction
Photocatalytic water splitting for converting solar energy into clean and sustainable hydrogen energy is a promising approach to solve energy and environmental problems, while efficient photogenerated charge separation on a semiconductor photocatalyst is the key of the heterogeneous photocatalysis process [1][2][3][4][5]. As it is well known, surface atomic arrangement and coordination of semiconductors significantly affect the physicochemical properties and photogenerated charge separation [6,7]. Therefore, rational regulation of the exposed crystal facets and surface atomic configurations of semiconductor-based photocatalysts by precisely optimizing the preparation process is essential for modulating the intrinsic charge separation properties [8][9][10][11].
Recently, the strategy for enhancing charge separation is focused on spatial charge separation between coexposed anisotropic facets of a low symmetry single crystal semiconductor [12][13][14][15][16][17][18]. Realizing spatial charge separation on photocatalysts is dependent on the adopted preparation processes for the construction of crystal facets with dissimilar surface work function [19,20]. The crystal facet-dependent charge separations have been reported on several photocatalysts, such as BiVO 4 , TiO 2 , BiOCl, SrTiO 3 , BaTaON, and NaTaO 3 [21][22][23][24][25][26]. Nevertheless, exploring wide-range visiblelight-active photocatalysts with excellent charge separation properties and deepening understandings of the relationship between spatial charge separation and coexposed crystal facets are needed for further enhancing intrinsic charge separation efficiency and promoting the development of photocatalysis.
Silleń-Aurivillius perovskite (Bi 4 MO 8 X) compounds with narrow band gap are promising for solar water splitting [27][28][29][30][31]. More importantly, their layered structure characteristics are beneficial for constructing anisotropic facets with low symmetry by precisely regulating the preparation procedures [32]. Bismuth tantalum oxychloride (Bi 4 TaO 8 Cl) as a typical example was recently reported for the preparation of {001} facet-dominated orthorhombic nanoplates [33], and the {001} facets are considered to be the photoactive surface [34]. In addition, Bi 4 TaO 8 X was also recently reported with improved photocatalytic CO 2 reduction activity due to the surface modification by Ag nanoparticles [35]. Additionally, photoinduced surface modification by generating reactive species on the Bi 4 TaO 8 Cl surface was also reported [36]. Despite most the of literatures having demonstrated improved photocatalytic activities on Bi 4 TaO 8 Cl through surface modification [37], or loading of -suitable cocatalyst [38], the relationship between the intrinsic photoreactivity and the exposed facets is still unclear, and there is a lack of in-depth exploration about the crystal facet-dependent charge separation properties of Bi 4 TaO 8 Cl.
Herein, aiming to explore the crystal facet-dependent photoreactivity and modulate the intrinsic charge separation through regulation of highly exposed facets, Bi 4 TaO 8 Cl (BTOC) was chosen as a visible-light-active photocatalyst for water oxidation. Well-defined BTOC nanoplates with exposed {110} side facets and {001} top facets were synthesized using a facile flux method. The BTOC nanoplate displays the capacity of spatial charge separation proven by in situ photochemical deposition, where photogenerated electrons and holes are selectively accumulated on the {110} facets and {001} facets, respectively. The BTOC-0.8 displays a higher charge separation efficiency and AQE for photocatalytic O 2 evolution than BTOC-0.5, BTOC-2.1, and BTOC-2.5, which may be due to the modulation of the coexposed crystal facet ratio. Furthermore, the charge separation efficiency and photocatalytic activity of BTOC-0.8 were further improved by loading dual cocatalysts Ag and RuO x , and the reverse oxidation of Fe 2+ to Fe 3+ ions was also blocked owing to spatial charge separation between coexposed facets.
2.2. Synthesis of Samples. BTOC well-defined photocatalyst crystal samples, BTOC-0.5, BTOC-0.8, BTOC-2.1, and BTOC-2.5, were prepared using flux method. In a typical synthesis, sample BTOC-0.5 was prepared by weighing stoichiometric ratios of Bi 2 O 3 , BiOCl, and Ta 2 O 5 in the ratio of 5 : 4 : 2 in mmol with the eutectic mixture of NaCl : KCl (10 mmol : 10 mmol) as flux agents; the mixture was thoroughly mixed using mortar and pestle. Thereafter, the obtained mixture was transferred into a high-temperature stable evacuated silica crucible and calcined at 973 K for 3 h in a temperature-programed oven. Afterward, the obtained cooled yellow product was washed several times with deionized water and dried in the oven at 353 K overnight. Likewise, sample BTOC-0.8 was prepared by weighing stoichiometric ratios of Bi 2 O 3 , BiOCl, and Ta 2 O 5 (5 : 4 : 2 in mmol) with eutectic mixture of NaCl and KCl (36 mmol : 36 mmol) of the precursor as flux agents and following suit. Similarly, samples BTOC-2.1 and BTOC-2.5 were prepared using a similar preparation process, except that the precursor concentration was adjusted. Samples BTOC-2.1 and BTOC-2.5 were prepared by weighing stoichiometric ratios of Bi 2 O 3 , BiOCl, and Ta 2 O 5 (2 : 1.5 : 0.8 in mmol) and Bi 2 O 3 , BiOCl, and Ta 2 O 5 (1.5 : 1 : 0.5 in mmol) with eutectic mixture of NaCl and KCl (1 : 1) as flux agents, respectively. The precursors were thoroughly mixed together using mortar and pestle for each sample; thereafter, the obtained mixture was transferred into a high-temperature stable evacuated silica crucible and calcined at 1023 K for 3 h in a temperatureprogramed oven to obtain samples BTOC-2.1 and BTOC-2.5. The obtained cooled yellow powder was washed several times with deionized water and dried in the oven at 353 K.

Characterizations.
The as-prepared BTOC sample crystal structure was characterized using scanning electron microscopy (SEM, Quanta 200 FEG, FEI) and high-resolution transmission electron microscopy (HRTEM, JEM-F200). X-ray diffraction (XRD) was recorded on Rigaku D/Max-2500/ PCXRD, and UV-visible (UV-vis) diffuse reflectance spectra were measured on a UV-vis spectrophotometer (JASCO V-550). Transient photocurrent response and electrochemical impedance spectroscopy were measured using an electrochemical workstation and VersaStudio CV (CHI-760D, Chenhua Instruments Co., Ltd.). Mott-Schottky measurements were carried out on a Princeton Applied Research PARSTAT 2273, and the frequency and amplitude of AC potential used were 1 kHz and 100 mV, respectively.

Structural and Morphological Characterization.
Bismuth tantalum oxychloride Bi 4 TaO 8 Cl (BTOC) was prepared using flux preparation method. Firstly, the precursor ratio and the synthesis temperature were extensively studied to determine the optimal synthesis conditions. For BTOC samples prepared using different ratios of Bi 2 O 3 , BiOCl, and Ta 2 O 5 , the X-ray diffraction (XRD) results show no significant difference and the diffraction peaks of all samples were indexed to Bi 4 TaO 8 Cl with orthorhombic phase symmetry, and the UV-visible absorption patterns also exhibit similar optical properties ( Figure S1). In addition, the SEM images shows that all samples are nanoplate-like in shape except for thickness differences that affect the proportion of exposed facets ( Figure S2). To investigate the validity of precursor ratio regulation, photocatalytic water oxidation was conducted on the BTOC samples using Fe(NO 3 ) 3 as electron scavengers ( Figure S3). It was observed that photocatalytic water oxidation activity varied significantly with the precursor  Figure S4a). In addition, UV-visible absorption spectra display similar absorption band edges with exception of the sample at 600°C ( Figure S4b). SEM images exhibit that the apparent shape of the as-prepared BTOC changes from regular nanoplates to nanoparticles over temperatures ranging from 900°C to 600°C ( Figure S5). Therefore, the flux temperature has significant effect on the growth of BTOC nanoplates crystal. Similarly, the photocatalytic water oxidation activity was conducted, and the BTOC synthesized at 700°C show the highest photocatalytic O 2 evolution activity ( Figure S6).
In order to further explore the reasons for the differences in photoreactivity of different BTOC nanoplates, several typical BTOC samples were selected for a more detailed characterization and analysis. For convenient identification, these samples are named BTOC-0.5, BTOC-0.8, BTOC-2.1, and BTOC-2.5, according to the diffraction peak intensity ratios of (004) and (220) crystal facets that reflect the modulation of the coexposed facet ratio to some extent. Figure 1(a) shows that there is no obvious difference in the XRD patterns, and they were indexed to the BTOC with orthorhombic phase symmetry. SEM images of BTOC-0.5, BTOC-0.8, BTOC-2.1, and BTOC-2.5 show a square nanoplate-like shape with difference in the thickness (Figure 1(b), Figure S7), which further shows the diversity in the exposed ratio of the top and lateral surfaces. To identify the corresponding exposed facets, high-resolution transmission electron microscopy (HRTEM) analysis was performed on the BTOC-0.8 (Figure 1(c)). Selected-area electron diffraction (SAED) pattern gives a diffraction spot with identified (220) and (200) planes of orthorhombic BTOC (Figure 1(d)). HRTEM image conducted on the top facet shows the lattice fringe spacing of 0.387 nm and 0.274 nm indexes to (110) and (020) planes of BTOC, respectively (Figure 1(e), Figure S8). Therefore, given the orthorhombic BTOC phase symmetry, the highly exposed facets of the as-prepared BTOC samples can be recognized as {001} top facets and {110} side facets; the difference of these samples is the variation in the exposed facet ratio.

Photogenerated Charge Distribution. The photogenerated charge distribution on the coexposed {001} and {110}
facets of the as-prepared BTOC samples was further explored by in situ photochemical deposition of metal oxides and metals [22]. As depicted in Figure 3, after photooxidation reactions taking Mn 2+ and Pb 2+ ions as precursors and IO 3 ion as electron scavenger, MnO x and PbO 2 particles were found mainly deposited on the top {001} surface, showing that photogenerated holes are specifically selectively accumulated on the {001} facets (Figures 3(b) and 3(c), Figure S9). Also, after photoreduction reactions of metals in the presence of methanol (CH 3 OH) as the hole scavenger, it is revealed that Au, Ag, and Pd particles were These results confirm that precisely controlled exposure of anisotropic crystal facets could achieve effective spatial charge separation by properly optimizing the synthesis parameters during BTOC sample preparation.

Charge Separation Efficiency.
To investigate the intrinsic charge separation efficiency of BTOC-0.5, BTOC-0.8, BTOC-2.1, and BTOC-2.5 samples, photocatalytic reaction was conducted in the presence of CH 3 OH as hole scavengers and Fe 3+ ions as electron acceptors under visible light (λ ≥ 420 nm). Noticeably, BTOC-0.8 displays much higher charge separation efficiency than BTOC-0.5, BTOC-2.1, and BTOC-2.5 (Figure 4(a), Figure S10), although electrons and holes are spatially separated onto different coexposed facets for all samples, indicating that charge separation and migration from the bulk to the surface in BTOC are largely affected by the difference in the coexposed facet ratio. Furthermore, photocatalytic water oxidation and the corresponding apparent quantum efficiency (AQE) tests were conduted to explore the effect of charge separation property on photocatalytic water oxidation performance. The results show that BTOC-0.8 still displays the higher photocatalytic O 2 evolution activity than other samples regardless of whether AgNO 3 and Fe(NO 3 ) 3 are electron acceptors (Figure 4(b), Figure S11) and achieves the highest AQE for photocatalytic water oxidation ( Figure S11). This is consistent with the results of charge separation efficiency, indicating that the variation of photocatalytic activity between BTOC samples mainly originates from the difference of charge separation properties. Moreover, photoelectrochemical (PEC) measurement was also performed to corroborate the above results, and the BTOC-0.8 shows a markedly higher photocurrent density than others, a result comparable with the charge separation efficiency trend (Figure 4(c)). Also, electrochemical impedance spectroscopy (EIS) analysis was adopted to explore the charge transfer process (Figure 4(d), Table S1). It is shown that the BTOC-0.8 possess the smaller charge transfer resistance indicating faster charge separation and transfer. The above results specifically confirm that charge separation is readily available on these samples, but the difference in charge carrier dynamics owing to the different coexposed facet ratio led to the observed significant difference in the photocatalytic performance between the as-prepared samples.
3.5. Photocatalytic Activity. As the reduction and oxidation cocatalysts has been widely demonstrated to promote surface reaction, we further optimized the photocatalytic activity by loading appropriate cocatalysts. BTOC is modified by means of in situ photodeposition of Ag and RuO x as a single and or dual cocatalyst, and the photocatalytic water oxidation performance was tested in the present of Fe(NO 3 ) 3 as electron scavengers. As shown in Figure 5(a), the cocatalyst-modified BTOC-0.8 show higher photocatalytic activity regardless of single or dual cocatalyst modification than the pristine BTOC-0.8, especially the sample loaded with a dual cocatalyst (Ag/RuO x /BTOC-0.8); the photocatalytic activity almost doubled. In addition, (Fe,Ru)O x as dual cocatalyst was also demonstrated to enhance the photocatalytic activity as shown in Figure S12 [39]. Moreover, the improved surface reaction in turn further promotes the enhancement of charge separation efficiency ( Figure 5(b)), which indicates that the loading of the cocatalyst significantly accelerates the extraction and utilization of photogenerated charges. Finally,  5 Energy Material Advances the AQE of photocatalytic oxygen evolution improved to~10% at 420 nm ( Figure 5(b)). As it has been demonstrated previously that achieving spatial charge separation between coexposed facets on well-define particulate photocatalysts could inhibit the reverse oxidation reaction from Fe 2+ to Fe 3+ due to the existence of coulomb repulsion force between the Fe 2+ ions and positively charged facets [40,41], time-dependent photocatalytic water oxidation was conducted in the presence of the Fe 3+ /Fe 2+ shuttle to corroborate the effect of spatial charge separation on blocking of reverse reaction. Notably, Fe 2+ ions formed during the photocatalytic water oxidation could reach the theoretical value of Fe 3+ ions in the original solution, regardless of the initial concentration of Fe 3+ (Figure 5(c), Figure S13), and the cocatalysts promoted the water oxidation reaction rate and accelerated the consumption of Fe 3+ ions ( Figure 5(d)). This illustrates that the oxidation reverse reaction, from Fe 2+ to Fe 3+ , was completely blocked by the achievement of spatial charge separation, which was also observed in BTOC-2.1 ( Figure S14). In order to further confirm the significant role of spatial charge separation between coexposed facets in improving photocatalytic water oxidation activity, BTOC nanoparticle was prepared under the same reaction conditions and explored for photocatalytic water oxidations ( Figure S15-S16). The results clearly show that BTOC nanoparticles display a very poor photocatalytic activity 6 Energy Material Advances compared to the well-defined BTOC nanoplates, indicating that the spatial charge separation within BTOC nanocrystals plays a substantial role on the intrinsic charge separation efficiency and photocatalytic performance ( Figure S17).

Discussion
In summary, we have synthesized well-defined Bi 4 TaO  crystal facet ratio. The BTOC-0.8 displays a significant high charge separation efficiency and AQE for photocatalytic water oxidation, owing to the modulation of the coexposed crystal facet ratio. The photocatalytic water oxidation activity was further improved by loading dual cocatalysts Ag and RuO x , and the oxidation reverse reaction of Fe 2+ to Fe 3+ ions was totally blocked owing to spatial charge separation between coexposed facets. Our work unravels the crystal facet-dependent intrinsic photoreactivity, which provides a feasible strategy to fabricate semiconductor-based photocatalyst for solar energy conversion.

Data Availability
All data presented in the paper and the supporting information are available from the corresponding author upon reasonable request.