The semiconductors TiO₂ and ZnO are commonly used as photocatalysts and require activation by short wavelength such as UV-rays ^[1], which make up only a small portion of available sunlight compared with visible light. Hence, the development of visible-light-driven photocatalysts has received much attention from many research groups and Ag₃PO₄ has been found to possess excellent photocatalytic properties. It has a direct band gap of 2.36 eV and can therefore absorb solar energy with wavelengths shorter than 530 nm ^[2]. The photocatalytic activity of Ag₃PO₄ is strongly influenced by its morphology, which can be controlled by the precipitation parameters used for its preparation, including: the type of complexing agent ^[3]; the type of precipitating agent ^[4]; and the pH of the reaction solution ^[5]. Although control of the Ag₃PO₄ particle shape and size can enhance its photocatalytic activity ^{[3, 4, 5, 6, 7, 8]}, it still decomposes to Ag⁰ under illumination with visible light and this leads to a reduction of its photocatalytic performance ^[6]. Surface modification by coupling with a promoter to give a heterostructure has been shown to reduce the solubility of Ag₃PO₄ and also inhibit charge recombination that leads to increase photocatalytic activity ^{[9, 10]}. It has been reported that combining photosensitive AgBr with Ag₃PO₄ facilitates its photocatalytic properties, increases particle stability, and requires relatively simple preparation techniques ^{[9, 10, 11]}. AgBr/Ag₃PO₄ is usually synthesized using a two-step ion-exchange method, by first preparing Ag₃PO₄ and then transferring into a Br⁻ solution. In AgBr/Ag₃PO₄ system, photogenerated holes (h⁺) in the Ag₃PO₄ can be transferred to the AgBr, whilst the photogenerated electrons in the AgBr can be transferred to Ag₃PO₄, inhibiting the electron-hole pair recombination ^[9]. Therefore, the amount of AgBr formed at the Ag₃PO₄ surfaces and the particle morphology of Ag₃PO₄ before exchange with Br⁻ ion strongly influence the photocatalytic activity of AgBr/Ag₃PO₄.

Altering the pH of precipitation solution is one of the effective parameters that is simple and easy to handle but there have been few reports describing the effect of precipitation pH on the photocatalytic activity of Ag₃PO₄. Wang et al. ^[5] studied the pH dependence of Ag₃PO₄ growth by using hexamethylenetetramine (HMT) to increase the pH of the reaction system. In this work, the Ag₃PO₄ particles were grown in solutions containing Na₃PO₄ and H₃PO₄. Furthermore, the number of preparation steps for AgBr/Ag₃PO₄ was reduced by using a one-pot synthesis. The photocatalytic activities of AgBr/Ag₃PO₄ powders were investigated through the degradation of methylene blue (MB), rhodamine B (RhB), reactive orange (RO) and methyl orange (MO) dyes and the presence of the reactive species OH^•, O₂^•−, and h⁺ was determined using appropriate scavengers. Finally, the ecotoxicity of the treated and untreated dye solutions were studied by testing their inhibition of the growth of the bioindicator Chlorella vulgaris.

Ag₃PO₄ was prepared by a precipitation method between Ag⁺ and PO₄³⁻ ions. The pH of precipitation solution was adjusted by addition of H₃PO₄ into Na₃PO₄ solution. Precipitation solutions were prepared with pH values of 5-10 because the formation of Ag₂O was observed when the pH of the solution was more than 10 ^[4]. Upon addition of aqueous AgNO₃ solution to the precipitation solutions, a yellow precipitate formed immediately and the mixture was stirred at room temperature for 1 h. The solids were filtered off, rinsed with distilled water, and dried at 100 ℃ in a hot air oven for 1 h.

AgBr/Ag₃PO₄ powders were prepared using a one-pot synthesis. Aqueous AgNO₃ solution was added to the precipitation solution at pH = 6 and this mixture was stirred at room temperature for 30 min. Aqueous KBr solution was added to the mixture and the resulting solution was stirred at room temperature for a further 30 min. The solids were filtered off, rinsed with distilled water, and dried at 100 ℃ in a hot air oven for 1 h.

The identification of crystalline phases and morphological studies on the synthesized powders were carried out using an X-ray diffractometer (XRD, X′Pert MPD, Philips) with Cu K_aradiation (l = 0.15406 nm) and a scanning electron microscope (SEM, QUANTA 400, FEI), respectively. Elemental analysis of the AgBr/Ag₃PO₄ powders was performed using an energy dispersive X-ray spectrometer (EDS, Oxford) operating in a point scan mode coupled with the SEM. The surface areas of samples were evaluated by the Brunauer-Emmett-Teller (BET) method using a surface area analyzer (Autosorb 1 MP, Quantachrome). X-ray photoelectron spectroscopy (XPS) was performed on an AXIS Ultra DLD (Kratos Analytical Ltd.) electron spectrometer and all binding energies were calibrated to the C 1s line at 284.8 eV.

The photocatalytic activities were investigated through the degradation of the cationic dyes MB and RhB and the anionic dyes RO and MO. In a typical procedure, the photocatalyst (0.1 g) was dispersed in an aqueous dye solution (10⁻⁵ mol/L, 150 mL) for 30 min to study the adsorption of dye and ensure that the system had reached equilibrium. After this period, the suspension was irradiated with visible light by three parallel fluorescent tubes (15 W) with a long pass filter (l > 420 nm, Edmund) and 3 mL of the dye solution were removed at 10 min intervals during the irradiation. The remaining dye concentration after irradiation was determined by the absorption technique using a UV-vis spectrophotometer (UV-2600, Shimadzu) and the decolorization was calculated using the following equation:

The reactive species involved in the photocatalytic degradation of dyes over AgBr/Ag₃PO₄ powder were determined using appropriate scavengers. The disodium salt of ethylenediaminetetraacetic acid (Na₂EDTA), benzoquinone (BQ), and t-butyl alcohol (TBA) were used as effective trapping agents for h⁺, O₂^•− and OH^•, respectively. In a typical procedure, the scavenger (1 mmol) was added to aqueous RhB and RO solutions (dye concentration = 10⁻⁵ mol/L, 150 mL) containing the AgBr/Ag₃PO₄ photocatalyst .

The total carbon content of the dye solutions before and after irradiation was analyzed by a CHNO analyzer (Flash 1112 EA Series EA, Thermoquest). The degradation products were detected by liquid chromatograph-mass spectrometry (LC-MS, 2690-LCT, Waters, Micromass) coupled with an electrospray ionizer operating in the positive mode (ESI⁺) for the cationic dyes and in the negative mode (ESI⁻) for the anionic dyes. Ecotoxicity was determined by growth inhibition of the unicellular green algae Chlorella vulgaris (C. vulgaris) using a previously reported method ^[12].

Na₃PO₄ produces OH⁻ when dissolved in water and this can result in solution with pH value more than 12. In this work, the pH of the precipitation solution was reduced by the addition of H₃PO₄ into the Na₃PO₄ solution. Fig. 1 shows the XRD patterns of the products using the precipitation solutions with different pH values (pH = 5-10). It was observed that all diffraction patterns matched well with the JCPDS NO. 06-0505 with no additional impurity peaks; thus, it was concluded that all the prepared powders were pure Ag₃PO₄. The lattice parameters of all the prepared Ag₃PO₄ powders are shown in Table 1.

Fig. 1. XRD patterns of the Ag3PO4 powders prepared from the precipitation solutions with various pH values.

Table 1 Yields, lattice parameters (a), band gap energies (Eg) and surface areas (SBET) evaluated using the BET method for Ag3PO4 powders prepared using precipitation solutions with different pH values. pH of precipitation solution Yield (%) a (Å) Eg (eV) SBET (m2/g) 5 8.36 6.01386(5) 2.417 2.04 6 18.01 6.01357(3) 2.417 2.56 7 26.45 6.01282(4) 2.425 3.95 8 38.14 6.01275(4) 2.425 4.01 9 41.36 6.01273(5) 2.425 4.03 10 43.80 6.01273(5) 2.425 4.03

Table 1 Yields, lattice parameters (a), band gap energies (Eg) and surface areas (SBET) evaluated using the BET method for Ag3PO4 powders prepared using precipitation solutions with different pH values.

The morphologies of the Ag₃PO₄ particles prepared from the precipitation solutions with pH values between 5 and 10 are shown in Fig. 2. This indicated that the particles of Ag₃PO₄ prepared from the precipitation solution of pH 5 had the biggest size, while the Ag₃PO₄ particles prepared from the precipitation solutions of pH 7-10 were all similar in size. The product yield was very low in acidic solution (Table 1) because PO₄³⁻ is a strong base that reacts with H⁺ to form the HPO₄²⁻, thereby increasing the solubility of Ag₃PO₄ in acidic solution. In general, when the Ag⁺ ions reacted with the PO₄³⁻ ions, it produced the Ag₃PO₄ nuclei that become suspended in the reaction solutions. In the case of acidic precipitation solutions, some of these nuclei can dissolve and this suppresses the overall rate of nucleation and growth. So, the prepared Ag₃PO₄ in an acidic condition has a larger sizes and a low-yield. The dissolution rate is negligible in basic precipitation solution so more Ag₃PO₄ nuclei form rapidly, leading to smaller particles and high yield.

Fig. 2. SEM images of the Ag3PO4 powders prepared from the precipitation solutions with various pH values.

The band gap energies for all the prepared Ag₃PO₄ powders were determined as the interceptions at the hv axis from the linear portion of (ahv)² vs photon energy (hv) plots as shown in Fig. 3 and their values are shown in Table 1. The particle sizes of all samples were larger than the Bohr radius so the changes in band gap energies with the sizes of the Ag₃PO₄ particles cannot be explained by quantum size effects. However, the changes could result from the lattice contraction as the lattice parameters of the Ag₃PO₄ powders trended to decrease with increasing pH values of the precipitation solutions as shown in Table 1.

Fig. 3. Evolution of the band gap energies of the Ag3PO4 powders prepared from the precipitation solutions with various pH values.

The photocatalytic decolorization of MB and RhB dyes by the Ag₃PO₄ powders prepared from precipitation solutions with different pH values are shown in Fig. 4. It was clearly observed that the Ag₃PO₄ prepared at pH 6 showed the best activity for decolorization of MB and RhB. The efficiency of semiconductor-based photocatalysts has traditionally been correlated to their surface area; however, this was not observed in this case because the photocatalytic activities of the prepared Ag₃PO₄ powders were not related to their surface areas, as shown in Table 1. The results of Indra et al. ^[13] have shown that the photocatalytic activity of Ag₃PO₄ particles with well-defined multifaceted morphologies, larger particle sizes, and lower surface areas were better than particles with an irregular shape and a correspondingly high surface area. Furthermore, Dong et al. ^[14] reported that irregular sphere-like Ag₃PO₄ particles had the worst photocatalytic activity when compared with Ag₃PO₄ with other shapes. As the Ag₃PO₄ powders prepared from the precipitation solutions of pH 7-10 had irregular sphere-like shapes and the particles prepared from the precipitation solution of pH 5 did not possess a well-defined multifaceted structure, these Ag₃PO₄ particles had lower photocatalytic activities than those prepared from the precipitation solution of pH 6.

Fig. 4. Photocatalytic decolorization percentages of MB (a) and RhB (b) dyes by Ag3PO4 powders prepared from the precipitation solutions with various pH values as a function of irradiation time.

Fig. 5 shows the XRD patterns of AgBr/Ag₃PO₄ powders prepared at pH = 6 after stirring for 30 min in solutions containing different amounts of dissolved KBr. When the concentration of Br⁻ ions in the solution was increased, the products changed from being yellow to chartreuse in color and additional diffraction peaks corresponding to AgBr were present. The intensities of these diffraction peaks increased with the amount of KBr added to the reaction solutions, while these assigned to Ag₃PO₄ decreased. This resulted from the exchange of Br⁻ ions for the PO₄^3- ions at the surface of the Ag₃PO₄ to produce AgBr/Ag₃PO₄. If the product was formed through the co-precipitation of Ag₃PO₄ and AgBr, the diffraction peaks of Ag₃PO₄ would not be expected to disappear at a high Br⁻ concentrations and the XRD should demonstrate the co-existence of Ag₃PO₄ and AgBr. Fig. 6 shows the EDS spectra of Ag₃PO₄ and AgBr present in a synthesized powder and this confirmed the heterostructure of AgBr/Ag₃PO₄.

Fig. 5. XRD patterns of AgBr/Ag3PO4 powders prepared using KBr concentrations of 0 (1), 0.001 (2), 0.005 (3), 0.01 (4), 0.03 (5), and 0.05 mol (6). The indexed peaks correspond to the AgBr phase.

Fig. 6. SEM image of the AgBr/Ag3PO4 prepared using a 0.001 mol KBr solution with the EDS patterns from the corresponding locations that comprised Ag3PO4 (1) and AgBr (2).

When the concentration of KBr in the preparation solution was increased, the morphologies of the photocatalysts changed from a rhombic dodecahedral-like shape (pure Ag₃PO₄ at pH 6), to more irregular shapes initially and then to large sphere-like shapes (pure AgBr) as shown in Fig. 7.

Fig. 7. SEM images of AgBr/Ag3PO4 particles prepared using KBr concentrations of 0.001 (a), 0.005 (b), 0.01 (c), 0.03 (d), and 0.05 mol (e).

The diffuse reflectance absorption spectra of the AgBr/Ag₃PO₄ powders are shown in Fig. 8(a). The intensity of the characteristic peak corresponding to the direct band gap of Ag₃PO₄ at 455 nm decreased as the amount of KBr in the preparation solution was increased. This was in agreement with the XRD results, which showed that the number of Ag₃PO₄ phases decreased as the amount of KBr added increased. Moreover, the band gap energy also increased with the amount of KBr added as shown in Fig. 8(b).

Fig. 8. Absorption spectra (a) and the evolution of band gap energies (b) of the AgBr/Ag3PO4 powders prepared using KBr concentrations of 0 (1), 0.001 (2), 0.005 (3), 0.01 (4), 0.03 (5), and 0.05 mol (6).

The effect of AgBr on the photocatalytic decolorization of dyes over AgBr/Ag₃PO₄ powders is shown in Fig. 9. Pure Ag₃PO₄ did not degrade the anionic dyes RO and MO because of the repulsion between the anionic dye and the negatively charged Ag₃PO₄ surface ^[15]. The presence of AgBr on the surface of Ag₃PO₄ enabled the photocatalyst to decolorize the anionic dyes as shown in Fig. 9. The photocatalytic degradations of the anionic dyes increased as the KBr concentrations in the preparation solutions were increased from 0 to 0.005 mol, but their activities for degradation of the cationic dyes became lower than that of pure Ag₃PO₄ powder (KBr < 0.005 mol). The presence of small amounts of AgBr on the surface of Ag₃PO₄ (KBr < 0.005 mol) clearly inhibited the photocatalytic degradation of the cationic dyes. When the KBr concentration in the preparation solution was 0.005 mol, the prepared photocatalyst showed the best photodegradation efficiency for all four dyes. At higher KBr concentrations, the decolorization efficiency for the four dyes also continuously decreased with the amounts of KBr added because the excess of AgBr would also reduce the photocatalytic efficiency of the AgBr/Ag₃PO₄ powders. The oxidation states of AgBr/Ag₃PO₄ (pH = 6, KBr = 0.005 mol) before and after photocatalysis were analyzed by XPS as shown in Fig. 10 and this indicated partial decomposition of Ag⁺ to Ag⁰ during the photocatalytic process.

Fig. 9. Photocatalytic decolorization of MB, RhB, RO, and MO dyes by AgBr/Ag3PO4 powders prepared using various KBr concentrations of 0 (1), 0.0005 (2), 0.001 (3), 0.005 (4), 0.01 (5), 0.03 (6) and 0.05 mol (7).

Fig. 10. Ag 3d XPS spectra of AgBr/Ag3PO4 powder prepared using a KBr concentration of 0.005 mol before (a) and after (b) photocatalytic dye degradation.

During photocatalysis the dye molecules are degraded by photogenerated reactive species. Hydroxyl radicals (OH^•), superoxide radical anions (O₂^•−), and photogenerated holes (h⁺) have been identified as playing an important role in the degradation of dye molecules by AgX/Ag₃PO₄ (X = Cl, Br and I) powders ^{[16, 17]}. Fig. 11 shows the decolorization activity of AgBr/Ag₃PO₄ (KBr = 0.005 mol) towards RhB and RO dyes in the presence of Na₂EDTA, BQ and TBA. It was observed that the addition of TBA did not affect the decolorization activity towards the cationic dye RhB and the anionic dye RO and this indicated that OH^• was not the main reactive species in this process. In contrast, the photocatalytic activity was strongly suppressed by the addition of Na₂EDTA and BQ so both h⁺ and O₂^•− were main reactive species involved in the photocatalytic decolorization of RhB and RO dyes by AgBr/Ag₃PO₄ (KBr = 0.005 mol) powder under illumination with visible light.

Fig. 11. Photocatalytic degradation of dyes by AgBr/Ag3PO4 powder (KBr = 0.005 mole) in the presence of different scavengers.

Table 2 shows the total carbon content (TCC) values of the dye solutions before and after photocatalysis using AgBr/Ag₃PO₄ powder (KBr = 0.005 mol) as the photocatalyst. The observed TCCs of the treated dye solutions were significantly lower than those of the corresponding untreated dye solutions. This implied that the carbon compounds were probably converted to CO₂ that was lost from the system. However, the dyes did not completely degrade to CO₂ after photocatalysis and some residual organic compounds were present in the treated solutions. The degraded products after photocatalysis were investigated by electrospray ionization mass spectrometry (ESI-MS) and the mass spectra of the treated dye solutions are presented in Fig. 12.

Table 2 Comparison of total carbon content (TCC) values (%) of dye solutions before and after photocatalysis. Sample type MB RhB MO RO Untreated solution 0.96 1.51 0.89 0.91 Treated solution 0.30 0.32 0.29 0.30

Table 2 Comparison of total carbon content (TCC) values (%) of dye solutions before and after photocatalysis.

Fig. 12. Mass spectra of the treated dye solutions. Positive ionization mode ESI+ was used for the cationic dyes (MB and RhB) and negative ionization mode ESI− was used for the anionic dyes (RO and MO).

Because most of the detected products were aromatic compounds that may be toxic, the ecotoxicities of the treated and untreated solutions were further evaluated. The ecotoxicities were determined through growth inhibition of C. vulgaris used as an ecological indicator and the results were also presented in Fig. 13. This showed that the MB dye strongly inhibited the growth of C. vulgaris when compared with the three other dyes. Fig. 14 shows microphotographs of C. vulgaris after 5 days of incubation. It indicated that only MB diffused into the C. vulgaris cells as the pigment of the algae changed color from green to blue. From Fig. 13, the number of individual C. vulgaris cells in all of the treated dye solutions was greater than for their corresponding untreated solutions but these solutions were more toxic than the deionized water (D.I.) used as a control. It cannot be concluded that the degraded products were less toxic than their parent dye molecules because the concentrations of these species were not the same. However, by comparing C. vulgaris growth in the same solution volume before and after photocatalysis, it was possible to conclude that the overall toxicity of the solution containing the degraded products after photocatalysis was lower.

Fig. 13. Individual C. vulgaris cell counts in dye solutions before and after (denoted as T-[dye]) photocatalysis and in deionized water (D.I.).

Fig. 14. Light microscopic observations of C. vulgaris after 5 days for incubation in untreated dye solutions and D.I.

Ag₃PO₄ powders prepared from the precipitation solution at pH 6 had a multi-faceted structure and showed the best activity for decolorizing MB and RhB solutions under irradiation by visible light. AgBr/Ag₃PO₄ powders were successfully synthesized using a one-pot ion-exchange method. The photocatalyst prepared from the precipitation solution with a KBr concentration of 0.005 mole showed the best photocatalytic degradation activity for all of the dyes tested under visible light. Using appropriate scavengers, it was concluded that the photodegradation was most likely due to the generation of h⁺ and O₂^•− by the AgBr/Ag₃PO₄ photocatalyst. These reactive species degraded the dyes into smaller organic molecules during the photocatalytic process and some aromatic compounds remained in the treated dye solution after photocatalysis. Although the dyes were not completely converted to CO₂, the overall ecotoxicities of the degradation products in the treated dye solutions were shown to be much lower than the corresponding untreated dye solutions using C. vulgaris as a bioindicator.

This work was supported from Prince of Songkla University under contract number SCI570276S. The author acknowledges the Center of Excellence for Innovation in Chemistry (PERCH-CIC), Office of the Higher Education Commission, Ministry of Education. We would also like to thank Dr. Brian Hodgson for assistance with the English.