Semiconductor-based photocatalysts have attracted considerable attention owing to their great potential in solving worldwide environmental and energy-related issues with abundant solar light ^{[1, 2, 3, 4]}. Recently, Ye and coworkers ^[5] presented a pioneering work exploring the photocatalytic properties of Ag₃PO₄, which exhibits extremely high photooxidative ability for O₂ evolution from water and organic dye decomposition under visible light irradiation. Moreover, this novel photocatalyst has been reported to achieve a quantum efficiency of up to 90% at wavelengths greater than 420 nm, which is significantly higher than previously reported values. Since then, our group has devoted much effort to the further improvement and optimization of the photoelectric and photocatalytic properties of Ag₃PO₄, including morphology and crystal facet control, fabrication of hetero-composites, and plasmon effects ^{[6, 7, 8, 9, 10, 11, 12, 13]}. However, there are still some limitations to the enhancement of photocatalytic performance by the above strategies. Particularly, the challenges of long-term stability, electrons-hole recombination, and photocorrosion greatly restrict the future application of this material. Accordingly, it is highly crucial to develop new strategies to improve the photocatalytic activity and stability of Ag₃PO₄ photocatalysts. The doping of metal ions into semiconductor materials has been widely studied in recent years, and can not only retard charge pair recombination, but also enable visible light absorption by providing defect states in the bandgap ^[14]. However, until now there have been no investigations of metal ion dopants in the Ag₃PO₄ lattice.

Furthermore, the fabrication of photocatalysts with hollow or porous structures has been paid increasing attention owing to their unique multifunctional physical and chemical properties such as high specific surface area, low effective density, and good surface permeability ^{[15, 16, 17, 18, 19]}. Recently, the ion exchange strategy has paved a new path for chemical transformation of inorganic nanostructured materials which largely preserves the morphology and structure of the precursors ^{[20, 21, 22]}. Hollow structures can be obtained as a result of a diffusion rate discrepancy between two of the components. For the ion exchange fabrication of Ag-based photocatalysts, extensive efforts have been focused on anion exchange using Ag₂MoO₄, Ag₂CO₃, Ag₃VO₄, and Ag₈W₄O₁₆ as the starting materials ^{[23, 24, 25, 26]}. In these transformation processes, Ag⁺ ions are released into an aqueous solution and then reacted with free anions to obtain Ag_nX (X = Cl, Br, I, S). Unfortunately, no reports have been made on the design and fabrication of hollow Ag₃PO₄ structures using such an ion exchange process because of its intrinsic solubility limitations. If controlled synthesis of a Ag₃PO₄ photocatalyst with hollow structure could be achieved, it is expected that a metal ion doped, hollow structured Ag₃PO₄ photocatalyst would exhibit both high photocatalytic activity and light- harvesting efficiency.

Selectivity is a very important factor in the evaluation of the performance of photocatalyst. Most attention has been paid to the photocatalytic selectivity of TiO₂ systems towards the degradation of alcohols and azo dyes ^{[27, 28, 29, 30]}. Generally, photocatalytic reactions on TiO₂ are often accompanied by formation of highly reactive radical species such as ^•OH which are typically nonselective. Modification of the adsorption selectivity is a powerful technique for the tuning catalytic selectivity of TiO₂. Ag₃PO₄ shows very poor photocatalytic selectivity, and there have been no reported attempts to design a Ag₃PO₄ system with adjustable selective photocatalytic properties.

In this paper, we have developed a facile and novel cation exchange route for the one-step synthesis of Ba-doped nanoporous Ag₃PO₄ hollow sheets (AHS) at room temperature using Ba₃(PO₄)₂ as the starting material. Constituent-dependent changes in phase and morphology were comprehensively investigated to elucidate the catalyst formation mechanism. Photocatalytic performance experiments clearly revealed that the Ba-doped AHS exhibited superior photocatalytic properties and stability for the degradation of organic contaminants. Intriguingly, the products showed preferential decomposition of methyl orange (MO) in comparison to that of rhodamine B (RhB) under visible light irradiation as a result of the Ba-doping.

All chemicals were purchased from Sinopharm Group Chemical Reagent Co. Ltd., China, were of analytical grade and were used as received without further purification. Deionized water with a resistivity of 18.25 MΩ cm was used in all reactions.

Ba-doped Ag₃PO₄ hollow nanosheets were fabricated by cation exchange reaction using Ba₃(PO₄)₂ as the precursor. First, Ba₃(PO₄)₂ nanosheets were prepared by a simple precipitation process. In a typical synthesis, an aqueous solution of Na₃PO₄ (0.05 mol/L) was added to an aqueous solution of BaCl₂ (0.15 g), and a white precipitate was formed. The obtained Ba₃(PO₄)₂ was collected, washed with distilled water to remove the Cl^-, and dried under atmosphere before morphological analyses and subsequent use. In a typical cation exchange reaction, the obtained Ba₃(PO₄)₂ nanosheets were distributed evenly in water under vigorous magnetic stirring, and AgNO₃ aqueous solution (0.1 mol/L) was added dropwise to form a yellow suspension. The obtained Ag₃PO₄ product was collected by centrifugation, washed with deionized water three times, and then dried at 80 °C in air for 5 h.

The morphology of the samples was observed using a field-emission scanning electron microscope (FE-SEM; JSM-6701F, JEOL) operated at an accelerating voltage of 5.00 kV and equipped with an energy dispersive spectrometer (EDS). Transmission electron microscopy (TEM) analyses were conducted with a JEM-1200EX electron microscope using a 200 kV accelerating voltage. X-ray diffraction (XRD) measurements were performed on an X' pert PRO diffractometer using Cu K_α radiation at 40 keV and 40 mA. The XRD patterns were recorded from 10° to 90° with a scanning rate of 0.067°/s. X-ray photoelectron spectroscopy (XPS) was performed using an ESCALAB250Xi photoelectron spectrometer with an Al K_α X-ray excitation source to analyze the elemental composition of the samples. Ultravidet-visible (UV-vis) absorption spectra were collected using a Shimadzu UV-2550 spectrophotometer at room temperature.

RhB and MO were employed as representative pollutants to investigate the photocatalytic properties of the Ba-doped Ag₃PO₄ hollow nanosheets. A 100 mL RhB or MO solution with an initial concentration of 8 mg/L was mixed in the presence of the catalyst (0.2 g). Before light irradiation, the suspension was magnetically stirred in the dark for 20 min to ensure adsorption-desorption equilibrium was achieved. Then, the reaction system was irradiated with a 300W Xe arc lamp equipped with an ultraviolet cutoff filter to provide visible light of λ ≥ 420 nm. The concentration of RhB or MO remaining in the solution was monitored using a UV-vis spectrometer at 554 or 464 nm, respectively. Before UV-vis measurements, the photocatalyst was removed from the photocatalytic reaction system using a dialyzer (Millipore, Millex-LH 0.45 µm).

The photoelectric conversion properties of the photocatalyst were investigated using a conventional three-electrode cell with a computer-controlled electrochemical workstation (CHI 650A). First, 50 mg catalyst was suspended in 2 mL of aqueous Nafion solution (1 wt%), and the mixture was ultrasonically treated for 10 min to form a homogeneous suspension. Then, 0.1 mL of the photocatalyst suspension was dropped on fluorine doped tin oxide (FTO) glass (0.5 cm × 4.0 cm). After evaporation of the water in air, the catalyst was attached to the FTO glass surface. The FTO glass, a Pt wire, a saturated calomel electrode (SCE), and Na₂SO₄ (0.1 mol/L) were used as the working electrode, counter-electrode, reference electrode, and the electrolyte, respectively. Current-time (i-t) curves were collected at 1 V vs SCE. The light source was a 300W Xe lamp, and a cutoff filter of 420 nm was employed for visible-light irradiation.

Figure 1(a) shows typical SEM images of the Ba₃(PO₄)₂ nanosheets synthesized by directly mixing aqueous solutions of BaCl₂ and Na₃PO₄. As can be seen, a large quantity of Ba₃(PO₄)₂ nanosheets with an average thickness of about 250 nm were produced. The Ag₃PO₄ product was obtained through a facile cation exchange reaction between the as-prepared Ba₃(PO₄)₂ nanosheets and an aqueous solution of AgNO₃ at room temperature. As shown in Fig. 1(b), the as-converted Ag₃PO₄ product retained the original shape of the precursor well and were about 500 nm in thickness and 5−8 µm long, larger than the Ba₃(PO₄)₂ nanosheet precursor. A side view (inset of Fig. 1(b)) of the broken nanosheets shows the presence of a cavity between every upper and lower layer, indicating a double-layer hollow structure and porous morphology. Magnified SEM images of the Ag₃PO₄ nanosheets surface clearly showed that they were composed of primary nanoparticles, which made their surface rough and porous.

Fig. 1. SEM images of Ba3(PO4)2 nanosheets (a) and AHS (b).

The crystallographic structures of the original Ba₃(PO₄)₂ and the as-synthesized AHS were analyzed by XRD. Figure 2(a) shows that all the diffraction peaks of AHS could be indexed to the body-centered cubic structure of Ag₃PO₄ (JCPDS 06-0505), and no differences in crystal structure were observed between the Ba-doped Ag₃PO₄ product and cubic structured Ag₃PO₄, indicating a good dispersion of Ba²⁺ ions in the Ag₃PO₄ host crystal lattice. The UV-vis diffuse reflectance spectra of Ba₃(PO₄)₂ and AHS are shown in Fig. 2(b). Ba₃(PO₄)₂ exhibited no absorption throughout the whole ultraviolet-visible light region, while an absorbance peak edge around 510 nm was observed for the AHS. The bandgap of the AHS was estimated using the equation (αhν)^1/2 = A(hν−E_g), in which α, ν, and A are the absorption coefficient, light frequency, and proportionality constant, respectively. To estimate the bandgap, (αhν)^1/2 was plotted against hν and then the straight line was extrapolated to the hν axis intercept. As shown in the inset of Fig. 2(b), the bandgap of the AHS thereby obtained was 2.45 eV.

Fig. 2. XRD patterns (a) and UV-visible diffuse absorption spectra (b) of the Ba3(PO4)2 nanosheets (1) and AHS (2). The inset in Fig. 2(b) shows a plot of (αhν)1/2 vs hν for AHS.

XPS measurements were carried out to identify the chemical environment of each element present in the photocatalyst surface. High-resolution XPS spectra of the Ag₃PO₄ products are presented in Fig. 3. As shown in Fig. 3(a), the nature of Ba element in the doped AHS was determined from the Ba 3d spectrum. The peaks at 778.38 and 793.67 eV could be indexed to Ba 3d_5/2 and Ba 3d_3/2, which indicates the existence of the Ba²⁺ state ^[31]. In addition, a slight shift to lower binding energy was observed in Ag 3d, P 2p, and O 1s XPS spectra of the Ba-doped Ag₃PO₄ compared with the undoped Ag₃PO₄ sample (Fig. 3(b)-(d)), which is expected to have been caused by the relatively lower electronegativity of Ba²⁺ compared with that of Ag⁺.

Fig. 3. High-resolution XPS spectra of as-prepared Ba-doped AHS and undoped Ag3PO4.

To investigate the formation mechanism of the hollow structured Ag₃PO₄ product, four intermediates at different reaction stages were obtained and examined by SEM to determine the morphological and structural evolution of the product. The results are shown in Fig. 4, which demonstrates the process of growth from the outside to the interior in detail. As can be seen from Fig. 4(a), the Ba₃(PO₄)₂ nanosheet was first coated by a thin layer of Ag₃PO₄ nanoparticles with a rough surface. Gradually, Ba₃(PO₄)₂-Ag₃PO₄ sandwich-like structures were formed, and the thickness of the Ba₃(PO₄)₂ cores became thinner as the exchange reaction progressed (Fig. 4(b) and (c)). When most of the Ba₃(PO₄)₂ cores were rendered nearly invisible, the hollow interiors and porous surfaces started to form (Fig. 4(d)).

Fig. 4.Typical SEM images and corresponding compositions of intermediates obtained at different stages of the growth process. The compositions were determined by EDS.

The crystallographic structure and optical properties of the intermediates were also studied. The XRD patterns of the intermediates (Fig. 5(a)) clearly revealed that the peaks of Ba₃(PO₄)₂ gradually reduced in intensity and that of Ag₃PO₄ crystals were significantly enhanced as the growth progressed, indicating that the Ag₃PO₄ crystals formed gradually. The corresponding optical absorption spectra (Fig. 5(b)) clearly show that the intensity of the absorption in the UV range gradually increased during the growth process, and no significant changes in the absorption edges in the visible light region were detected. Additionally, no diffraction peaks and absorptions corresponding to impurities were detected during the entire growth process. It is worth mentioning that these intermediates composed of Ag₃PO₄ and Ba₃(PO₄)₂ showed very high photocatalytic activities for dye decomposition even though Ba₃(PO₄)₂ has no photocatalytic activity owing to its total spectral reflection. We speculate that this was caused by light reflection and reutilization by the photocatalyst. Therefore, the present catalysts will significantly reduce the consumption of noble metal, which would normally be used to enhance photocatalytic activity, owing to the affordability of the Ba₃(PO₄)₂, which is beneficial to their industrial production.

Fig. 5. XRD patterns (a) and UV-vis diffuse absorption spectra (b) of the intermediates obtained at different stages of the growth process.

The Ba₃(PO₄)₂-based replacement reaction in the one-step synthesis of the Ba-doped hollow structured Ag₃PO₄ should be studied first. We identified a potential reaction process based on the experimental observations and crystal growth habits of the AHS, a schematic illustration of which is shown in Scheme 1. Owing to the lower solubility of Ag₃PO₄ relative to that of Ba₃(PO₄)₂, the transformation of Ba₃(PO₄)₂ into Ag₃PO₄ by reaction with Ag⁺ ions is thermodynamically favored. In this process, the initial anion exchange between Ba₃(PO₄)₂ and Ag⁺ ions generates thin porous Ag₃PO₄ shells around the Ba₃(PO₄)₂ surface, gradually forming intermediate Ba₃(PO₄)₂-Ag₃PO₄ sandwich-like structures. Subsequent reaction leads to the formation of hollow Ag₃PO₄ nanosheets with a double-layer structure, which is likely caused by diffusion of the PO₄³⁻ ions out of the core region and the precipitation of more Ag₃PO₄ particles on the shells, leading to continuous growth of the Ag₃PO₄ shells until complete reaction of the Ba₃(PO₄)₂ cores. It should be noted that the dissociated Ba²⁺ ions released into the aqueous solution simultaneously doped the deposited Ag₃PO₄. This cation exchange reaction method represents a simple and effective route for large-scale synthesis of photocatalysts with hollow structures as well as metal ion doping.

Scheme. 1.Schematic illustration of the possible cation exchange process for the nanoporous Ba-doped Ag3PO4 hollow nanosheets.

The photocatalytic activity of the as-prepared Ba-doped AHS was evaluated by monitoring the degradation of RhB and MO under visible light irradiation. The activities of Ag₃PO₄ cubes and spherical particles were also investigated for comparison. From Fig. 6(a), it can be clearly seen that the AHS exhibited the highest photocatalytic activity and completely degraded the RhB within 5 min. In contrast, the cubes and the spherical particles required 8 and 14 min for complete RhB degradation, respectively. A similar order of photoreactivity was observed for the degradation of MO (Fig. 6(b)). Specifically, the AHS completely degraded the MO dye in 2.5 min, compared with 14 and 28 min for the cubes and the spherical particles. To clarify the origin of this trend, the surface-to-volume ratio of the three samples was determined from their surface areas: AHS 3.8 m²/g, Ag₃PO₄ cubes 2.6 m²/g, and spherical particles 3.5 m²/g. This confirms that the surface-to-volume ratio of the Ag₃PO₄ photocatalyst was not crucial in determining its photocatalytic properties. Instead, we attribute the improved performance of the Ba-doped AHS to its metal ion doping and its unique hollow and porous structure. However, the photocatalytic mechanism needs to be further researched. Overall, our results show that Ba-doped AHS is a very efficient photocatalyst and that its photocatalytic activity is significantly improved by the Ba²⁺ cation doping and tailoring of its shape and surface structure.

Fig. 6.Photocatalytic activity of Ag3PO4 sub-microcubes, spherical particles, and as-prepared and hollow structured Ag3PO4 nanosheets for RhB (a) and MO (b) degradation under visible-light irradiation (λ > 420 nm).

Furthermore, it should be noted that the AHS exhibited preferential decomposition of MO in comparison to RhB under visible light irradiation. As can be seen in Fig. 7(a), the photodecomposition rate constant k of RhB over AHS, as calculated from the slope of a lnC₀/C vs time plot, was 0.9920 min⁻¹, much lower than that of the MO degradation (1.3725 min⁻¹). However, as shown in Fig. 7(b) and (c), the photodegradation of RhB was clearly much faster than that of MO over the Ag₃PO₄ cubes and spherical particles. This is not surprising because the positively charged RhB molecules would be preferentially absorbed on the negatively charged surface of the Ag₃PO₄ cubes and particles, which favors the degradation of positively charged contaminants. However, this consistency between adsorption selectivity and photocatalytic selectivity is not applicable to the AHS, because it demonstrated the same adsorption selectivity as that of the Ag₃PO₄ cubes and spherical particles. We conclude that the doping of Ba²⁺ ions improved the photocatalytic properties of the Ag₃PO₄ in two aspects. First, because the Ba²⁺ cation is larger than Ag⁺, Ba doping may have greatly increased the number of oxygen defects as the Ba²⁺ ions were doped into the Ag₃PO₄ lattices, which are necessary to restrain the recombination of the electrons and holes. Second, the Ba doping would have created a charge imbalance in the Ag₃PO₄, causing more OH^- to be adsorbed on the surface of the photocatalyst. The greater amount of OH^- would have trapped more holes, retarding charge pair recombination, and generated a larger number of hydroxyl radicals with high oxidizability, thus resulting in enhanced photocatalytic activity. The above observations suggest that the photocatalytic properties of Ag₃PO₄ photocatalysts can be tuned by modifying their surface by Ba doping. We expect that this strategy may be generally applicable to other semiconductor materials.

Fig. 7. Comparison of photocatalytic decomposition of RhB and MO by as-prepared nanoporous structured AHS (a), Ag3PO4 spherical particles (b), and Ag3PO4 sub-microcubes (c).

Durability is an important factor for photocatalysts in practical applications. We carried out durability experiments by recycling and reusing the hollow structured Ba-doped Ag₃PO₄ samples for five cycles. The corresponding photocatalytic performance is shown in Fig. 8. The as-prepared AHS maintained a high photocatalytic activity for the degradation of both RhB and MO dye throughout all five cycles. The time required to fully degrade the RhB dye was in the range of 5-7 min, while that for MO was 2.5-4 min. These results clearly reveal that the present AHS can serve as a stable and efficient visible-light photocatalyst. At present, we attribute the enhanced photocatalytic activity and stability of the nanoporous structured hollow Ag₃PO₄ sheets to their unique hollow structure and the porous morphology as well as the Ba doping.

Fig. 8. Photocatalytic activity of a Ba-doped Ag3PO4 sample for the degradation of RhB (a) and MO (b) under visible light irradiation (λ > 420 nm) over five cycles.

Next, the photoconversion efficiency of Ag₃PO₄ samples deposited on FTO glass was explored in detail. As can be seen from the time dependence of the photocurrent generated under intermittent light irradiation shown in Fig. 9, the hollow structured Ba-doped Ag₃PO₄ sheets exhibited a higher photoelectric current and conversion efficiency than the Ag₃PO₄ cubes and spherical particles. The trends in performance had the same order as that observed for the photodegradation of organic dyes shown in Fig. 6. Therefore, these results clearly show that Ba-doping and hollow structuring are also very feasible for enhancing the photoelectric performance of Ag₃PO₄.

Fig. 9. Photoelectric conversion performance of Ba-doped AHS (1), Ag3PO4 cubes (2), and spherical particles (3) in aqueous Na2SO4 solution (0.1 mol/L).

We have demonstrated a facile and efficient cation exchange process for the large-scale synthesis of Ba-doped nanoporous Ag₃PO₄ hollow nanosheets. The well-defined structures obtained exhibited excellent photocatalytic properties and photoelectric conversion performance compared with those of Ag₃PO₄ cubes and spherical particles under visible light irradiation. Furthermore, the products showed preferential decomposition of MO in comparison to RhB as a result of the Ba doping. We have a great expectation that this chemical transformation strategy from a nanotemplate to a metal ion doped hollow and porous product may represent a new universal route for the preparation of similar highly active and stable photocatalysts based on other semiconductor materials.