Hydrogen Production Improvement on Water Decomposition Through Internal Interfacial Charge Transfer in M3(PO4)2-M2P2O7 Mixed-Phase Catalyst (M = Co, Ni, and Cu)

In this study, three types of Nasicon-type materials, Co3(PO4)2-CO2P2O7, Ni3(PO4)2-Ni2P2O7, and Cu3(PO4)2-Cu2P2O7, were synthesized as mixed-phase catalysts (MPCs) for evaluating their potential as new photocatalytic candidates (called Co3(PO4)2-CO2P2O7mpc, Ni3(PO4)2-Ni2P2O7mpc, and Cu3(PO4)2-Cu2P2O7mpc herein). Based on various physical properties, it was confirmed that there are two phases, M3(PO4)2 and M2P2O7, in which a similar phase equilibrium energy coexists. These colored powders showed UV and visible light responses suitable to our aim of developing 365-nm light-response photocatalysts for overall water-splitting. The photocatalytic performance of Ni2(PO4)3-Ni2P2O7 MPC showed negligible or no activity toward H2 evolution. However, Co2(PO4)3-Co2P2O7 MPC and Cu3(PO4)2-Cu2P2O7 MPC were determined as interesting materials because of their ability to absorb visible light within a suitable band. Moreover, an internal interface charge transfer was suggested to occur that would lower the recombination rate of electrons and holes. For Cu3(PO4)2-Cu2P2O7 MPC, the charge separation between the electron and hole was advantageously achieved, a water-splitting reaction was promoted, and hydrogen generation was considerably increased. The performance of a catalyst depended on the nature of the active metal added. In addition, the performance of the catalyst was improved when electrons migrated between the inter-phases despite the lack of a heterojunction with other crystals.


Introduction
Since the hydrogen production performance of a TiO 2 photocatalyst was first reported by Honda-Fujishima [1][2][3][4], numerous studies have been conducted to improve the performance of photocatalysts using efficient light harvesting. As a result, researchers have concluded that if electrons and holes generated by light can be advantageously separated, their recombination could be delayed to increase the frequency of participation in the oxidation-reduction reaction occurring at the interface between the catalyst and reactant [5,6]. To slow down the recombination of electrons and holes, several methods are available, including a one-step excitation system [7], a photosensitized semiconductor system [8], a two-step excitation semiconductor heterojunction system [9], and a Z-scheme system [10]. Catalysts based on such systems have achieved unexpectedly good results, and many related studies have been published [11][12][13]. Currently, various photocatalysts ranging from widely used TiO 2 particles 2 of 14 to metal-sulfide [14], metal-nitride [15], and metal-tungstate [16] with a slightly smaller bandgap have been studied.
Nasicon-type materials such as metal phosphate particles, which are relatively small with a bandgap of approximately 2-3 eV compared to TiO 2 , have recently attracted the interest of a few researchers [17,18]. In general, Nasicon is a crystalline solid of A 1 B 2 (PO 4 ) 3 , where A is a monovalent cation and B is a single ion or a combination of tri-, tetra-, and penta-atomic ions. Here, the monovalent A ions can move within the lattice with low activation energy [19]. During the early 1980s, Susman et al. succeeded in synthesizing a compound with a formula of Na 1+x Zr 2 Si x P 3-x O 12 (0 < x < 3) [20], and because Na, Zr, and Si ions exhibit unique ion conductivity by substituting other elements, this compound has been widely applied as a secondary battery material [21,22]. To date, many ion-exchanged derivatives and substituted Nasicon frameworks are known, e.g., AM 2 (PO 4 ) 3 (A = Na, Ca, Sr, Ba; M = Ti, Mo). These structural-type materials consist of a three-dimensional network made up of PO 4 tetrahedra, sharing corners with MO 6 octahedra and forming interconnected tunnels, where M + (Na + , K + , Ag + ) or oxygen ions have freedom of movement [23]. During a photocatalytic reaction, the transfer of electrons through the formation of interconnected tunnels or the transfer of oxygen atoms can increase the adsorption of the reactants, effectively isolating the photo-induced charge, and consequently contributing to an improved photoactivity [24]. As examples, Fu, et al. reported the photocatalytic activity of MgTi 4 (PO 4 ) 6 and CaTi 4 (PO 4 ) 6 glass-ceramics containing Nasicon-type crystals [25], and Palla et al. suggested the photocatalytic degradation of organic dyes with Sn 2+and Ag + -substituted K 3 Nb 3 WO 9 (PO 4 ) 2 under visible light irradiation [26]. However, this remains in the early stage of research, and has thus, not shown a remarkable catalytic activity. Thus, in this study, M 3 (PO 4 ) 2 particles as Nasicon-type materials were prepared by binding phosphate ions to a Co, Ni, or Cu metal in a divalent oxidation state with 7, 8, and 9 electrons in a 3d-orbit of the transition metal. These particles were used as a water-splitting catalyst to compare their hydrogen production performance.

Results and Discussion
Characteristics of M 3 (PO 4 ) 2 -M 2 P 2 O 7 Mixed-Phase Catalysts Figure 1 shows X-ray diffraction (XRD) patterns of the as-synthesized catalysts. A sample of cobalt phosphate was shown to have a consistent pattern with the crystal structure of Co 3 (PO 4 ) 2 in a monoclinic crystal system [27], and a small XRD peak, belonging to the cobalt phosphate hydrate (monoclinic, P21/n), was also found at 33 • . Nickel phosphate particles mostly show an XRD pattern of Ni 3 (PO 4 ) 2 in a monoclinic crystal system with a space group of P21/a, although the Ni 2 P 2 O 7 phase of the monoclinic crystal system with the P21/c space group was also mixed [28]. The Cu 2 P 2 O 7 phase (C2/c space group) was also mixed with the Cu 3 (PO 4 ) 2 monoclinic crystal (C2/c space group) [29]. From this result, we confirmed that the M 3 (PO 4 ) 2 and M 2 P 2 O 7 crystal phases coexist because the energy phase equilibrium of the two crystals lies at approximately the same positions [30]. Jain et al. [31] observed that the mixed-phased M x (P 2 O 7 )(PO 4 ) 2 is energetically stable and can be decomposed into M x P 2 O 7 and M x (PO 4 ) 3 , although the overall energy of the decomposition is unfavorable by 5 meV/atom. The authors concluded that M x (P 2 O 7 )(PO 4 ) 2 has a stable phase. Furthermore, in this study, we did not intentionally attempt to produce two separate hetero-type phases, but the two phases are instead naturally produced during synthesis. Transmission electron microscopy (TEM) images are shown in Figure 2. In most studies [32,33], TEM images of M2P2O7 are expressed in sheet form. We also interpreted them in the same context. In the photograph showing cobalt phosphate, we can see uniform particles of an elliptical shape with a width of 300 nm and a length of 150 nm. However, it was confirmed in the images that nickel phosphate and copper phosphate were mixed with unclear particles: Ni3(PO4)2 with small particles of 50-100 nm was clustered, and a wide sheet of Ni2P2O7 particles could also be seen. In the copper phosphate image, it was confirmed that rectangular Cu3(PO4)2 particles with a width of 500 nm and a height of 300 nm were mixed with sheet-type Cu2P2O7 particles. From the XRD and TEM results, we recognized that the synthesized catalysts coexisted as two crystal phases. Herein, we refer to the M3(PO4)2-M2P2O7 mixed-phase catalysts (MPCs) as Co3(PO4)2-CO2P2O7mpc, Ni3(PO4)2-Ni2P2O7mpc, and Cu3(PO4)2-Cu2P2O7mpc. Transmission electron microscopy (TEM) images are shown in Figure 2. In most studies [32,33], TEM images of M 2 P 2 O 7 are expressed in sheet form. We also interpreted them in the same context. In the photograph showing cobalt phosphate, we can see uniform particles of an elliptical shape with a width of 300 nm and a length of 150 nm. However, it was confirmed in the images that nickel phosphate and copper phosphate were mixed with unclear particles: Ni 3 (PO 4 ) 2 with small particles of 50-100 nm was clustered, and a wide sheet of Ni 2 P 2 O 7 particles could also be seen. In the copper phosphate image, it was confirmed that rectangular Cu 3 (PO 4 ) 2 particles with a width of 500 nm and a height of 300 nm were mixed with sheet-type Cu 2 P 2 O 7 particles. From the XRD and TEM results, we recognized that the synthesized catalysts coexisted as two crystal phases. Herein, we refer to the M 3 (PO 4 ) 2 -M 2 P 2 O 7 mixed-phase catalysts (MPCs) as Co 3 (PO 4 ) 2 -CO 2 P 2 O 7 mpc, Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc, and Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc. An X-ray photoelectron spectroscopy analysis was conducted on Cu3(PO4)2-Cu2P2O7mpc as a representative catalyst, the results of which are shown in Figure 3. The peaks at binding energies of 935.7, 133.1 and 530.6 eV refer to the Cu2p3/2, P2p3/2 and O1s spectra, respectively. All core peaks in Cu3(PO4)2-Cu2P2O7mpc were recorded, and as shown in Figure 3, the core peak of Cu2p showed two main spin-orbit components at 935.7 and 955.6 eV, corresponding to Cu2p3/2 and Cu2p3/2, which confirmed the presence of Cu 2+ [34]. Two peaks located at 942.5 and 962.5 eV are the "shake up" satellite peaks of Cu2p. The high-resolution spectra of P2p show two main spin-orbit components at 133.1 and 134.8 eV, corresponding to Cu2p3/2 and Cu2p3/2, respectively, thereby confirming the presence of P 5+ . The O1s regions for Cu3(PO4)2 located at 530.6 and 532.4 eV are assigned to the M-O-P and P-O-P bonds, respectively [35]. In particular, a peak at 532.5 eV is shown, which was assigned to the P = O of Cu2P2O7 [36]. This is similar to the XRD result, and it was concluded that Cu3(PO4)2 coexists with the Cu2P2O7 phase in Cu3(PO4)2-Cu2P2O7mpc.  Figure 4 shows the DIR-UV-visible absorbance of three M3(PO4)2-M2P2O7mpc particles. In general, according to a Sugano-Tanabe diagram [37], three absorption curves are allowed in Co 2+ and Ni 2+ complexes with d 7 -and d 8 -electron configurations. The first curve is observed in the UV region below 300 nm (T1g↔A2g), the second curve is shown in the visible region of 400-750 nm (T1g↔T1g), and the last curve can be seen in the IR region at above 800 nm (T1g↔T2g). In general, within the visible region, the blue Cu3(PO4)2-Cu2P2O7mpc particles absorb a red wavelength of approximately 650 nm and the yellow Ni3(PO4)2-Ni2P2O7mpc particles absorb a blue wavelength of 450 nm and a red wavelength within the vicinity of 700 nm. The purple Co3(PO4)2-Co2P2O7mpc particles absorb a yellow-green wavelength of approximately 550-600 nm. In particular, Cu3(PO4)2-Cu2P2O7mpc particles show a strong absorption band at above 650 nm. It is known that the diluted Cu 2+ ions inside a glass matrix exhibit a broad optical absorption of approximately 700 nm, which is assigned to the 2 B2g 2 B1g transition owing to the Jahn-Teller splitting of the 3d levels of Cu 2+ ions in a ligand field [38]. The broad absorption band was observed in the NIR region, which corresponds to three possible d-d electronic absorption transitions for distorted octahedral coordination rather than perfect octahedral coordination. Thus, the broadening of the absorption band observed at approximately 700 An X-ray photoelectron spectroscopy analysis was conducted on Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc as a representative catalyst, the results of which are shown in Figure 3. The peaks at binding energies of 935.7, 133.1 and 530.6 eV refer to the Cu2p 3/2 , P2p 3/2 and O1s spectra, respectively. All core peaks in Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc were recorded, and as shown in Figure 3, the core peak of Cu2p showed two main spin-orbit components at 935.7 and 955.6 eV, corresponding to Cu2p 3/2 and Cu2p 3/2 , which confirmed the presence of Cu 2+ [34]. Two peaks located at 942.5 and 962.5 eV are the "shake up" satellite peaks of Cu2p. The high-resolution spectra of P2p show two main spin-orbit components at 133.1 and 134.8 eV, corresponding to Cu2p 3/2 and Cu2p 3/2 , respectively, thereby confirming the presence of P 5+ . The O1s regions for Cu 3 (PO 4 ) 2 located at 530.6 and 532.4 eV are assigned to the M-O-P and P-O-P bonds, respectively [35]. In particular, a peak at 532.5 eV is shown, which was assigned to the P = O of Cu 2 P 2 O 7 [36]. This is similar to the XRD result, and it was concluded that Cu 3 (PO 4 ) 2 coexists with the Cu 2 P 2 O 7 phase in Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc. An X-ray photoelectron spectroscopy analysis was conducted on Cu3(PO4)2-Cu2P2O7mpc as a representative catalyst, the results of which are shown in Figure 3. The peaks at binding energies of 935.7, 133.1 and 530.6 eV refer to the Cu2p3/2, P2p3/2 and O1s spectra, respectively. All core peaks in Cu3(PO4)2-Cu2P2O7mpc were recorded, and as shown in Figure 3, the core peak of Cu2p showed two main spin-orbit components at 935.7 and 955.6 eV, corresponding to Cu2p3/2 and Cu2p3/2, which confirmed the presence of Cu 2+ [34]. Two peaks located at 942.5 and 962.5 eV are the "shake up" satellite peaks of Cu2p. The high-resolution spectra of P2p show two main spin-orbit components at 133.1 and 134.8 eV, corresponding to Cu2p3/2 and Cu2p3/2, respectively, thereby confirming the presence of P 5+ . The O1s regions for Cu3(PO4)2 located at 530.6 and 532.4 eV are assigned to the M-O-P and P-O-P bonds, respectively [35]. In particular, a peak at 532.5 eV is shown, which was assigned to the P = O of Cu2P2O7 [36]. This is similar to the XRD result, and it was concluded that Cu3(PO4)2 coexists with the Cu2P2O7 phase in Cu3(PO4)2-Cu2P2O7mpc.  Figure 4 shows the DIR-UV-visible absorbance of three M3(PO4)2-M2P2O7mpc particles. In general, according to a Sugano-Tanabe diagram [37], three absorption curves are allowed in Co 2+ and Ni 2+ complexes with d 7 -and d 8 -electron configurations. The first curve is observed in the UV region below 300 nm (T1g↔A2g), the second curve is shown in the visible region of 400-750 nm (T1g↔T1g), and the last curve can be seen in the IR region at above 800 nm (T1g↔T2g). In general, within the visible region, the blue Cu3(PO4)2-Cu2P2O7mpc particles absorb a red wavelength of approximately 650 nm and the yellow Ni3(PO4)2-Ni2P2O7mpc particles absorb a blue wavelength of 450 nm and a red wavelength within the vicinity of 700 nm. The purple Co3(PO4)2-Co2P2O7mpc particles absorb a yellow-green wavelength of approximately 550-600 nm. In particular, Cu3(PO4)2-Cu2P2O7mpc particles show a strong absorption band at above 650 nm. It is known that the diluted Cu 2+ ions inside a glass matrix exhibit a broad optical absorption of approximately 700 nm, which is assigned to the 2 B2g 2 B1g transition owing to the Jahn-Teller splitting of the 3d levels of Cu 2+ ions in a ligand field [38]. The broad absorption band was observed in the NIR region, which corresponds to three possible d-d electronic absorption transitions for distorted octahedral coordination rather than perfect octahedral coordination. Thus, the broadening of the absorption band observed at approximately 700  Figure 4 shows the DIR-UV-visible absorbance of three M 3 (PO 4 ) 2 -M 2 P 2 O 7 mpc particles. In general, according to a Sugano-Tanabe diagram [37], three absorption curves are allowed in Co 2+ and Ni 2+ complexes with d 7 -and d 8 -electron configurations. The first curve is observed in the UV region below 300 nm (T 1g ↔A 2g ), the second curve is shown in the visible region of 400-750 nm (T 1g ↔T 1g ), and the last curve can be seen in the IR region at above 800 nm (T 1g ↔T 2g ). In general, within the visible region, the blue Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc particles absorb a red wavelength of approximately 650 nm and the yellow Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc particles absorb a blue wavelength of 450 nm and a red wavelength within the vicinity of 700 nm. The purple Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc particles absorb a yellow-green wavelength of approximately 550-600 nm. In particular, Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc particles show a strong absorption band at above 650 nm. It is known that the diluted Cu 2+ ions inside a glass matrix exhibit a broad optical absorption of approximately 700 nm, which is assigned to the 2 B 2g → 2 B 1g transition owing to the Jahn-Teller splitting of the 3d levels of Cu 2+ ions in a ligand field [38]. The broad absorption band was observed in the NIR region, which corresponds to three possible d-d electronic absorption transitions for distorted octahedral coordination rather than perfect octahedral coordination. Thus, the broadening of the absorption band observed at approximately 700 nm is attributed to the two electronic transitions in the d orbital corresponding to 2 A 1g → 2 B 1g and 2 B 2g → 2 B 1g [39]. In general, this means that the bandgap narrows as the longer wavelength is absorbed. In this case, it is known that the excitation of electrons easily occurs even under weak light, thereby increasing the activity of the photocatalyst. Based on the absorption peak, which is the largest among the absorption wavelengths, the bandgaps decrease in the order of Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc > Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc > Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc. The wavelengths of 430, 580, and 670 nm were calculated for the Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc, Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc, and Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc samples using the following equation: where E g is the bandgap energy, h is Planck's constant, c is the speed of light, and λ is the given wavelength [40]. Bandgaps of 2.88, 2.14, and 1.77 eV are then obtained. However, in the Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc sample, the yellow-green color at approximately 410 nm and the dark blue color at 700 nm must be combined to obtain a light-blue colored Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc compound, which we synthesized. Thus, we recognize that both absorption peaks must be considered, and the bandgap at 410 nm is approximately 3.0 eV. In the literature, it was confirmed that Cu 3 (PO 4 ) 2 absorbs light at wavelengths of 380 nm and 650 nm [41], and Cu 2 P 2 O 7 does so at a wavelength of 700 nm or longer [42].
Catalysts 2019, 9, x FOR PEER REVIEW 5 of 13 nm is attributed to the two electronic transitions in the d orbital corresponding to 2 A1g 2 B1g and 2 B2g 2 B1g [39]. In general, this means that the bandgap narrows as the longer wavelength is absorbed. In this case, it is known that the excitation of electrons easily occurs even under weak light, thereby increasing the activity of the photocatalyst. Based on the absorption peak, which is the largest among the absorption wavelengths, the bandgaps decrease in the order of Ni3(PO4)2-Ni2P2O7mpc > Co3(PO4)2-Co2P2O7mpc > Cu3(PO4)2-Cu2P2O7mpc. The wavelengths of 430, 580, and 670 nm were calculated for the Ni3(PO4)2-Ni2P2O7mpc, Co3(PO4)2-Co2P2O7mpc, and Cu3(PO4)2-Cu2P2O7mpc samples using the following equation: where Eg is the bandgap energy, h is Planck's constant, c is the speed of light, and λ is the given wavelength [40]. Bandgaps of 2.88, 2.14, and 1.77 eV are then obtained. However, in the Cu3(PO4)2-Cu2P2O7mpc sample, the yellow-green color at approximately 410 nm and the dark blue color at 700 nm must be combined to obtain a light-blue colored Cu3(PO4)2-Cu2P2O7mpc compound, which we synthesized. Thus, we recognize that both absorption peaks must be considered, and the bandgap at 410 nm is approximately 3.0 eV. In the literature, it was confirmed that Cu3(PO4)2 absorbs light at wavelengths of 380 nm and 650 nm [41], and Cu2P2O7 does so at a wavelength of 700 nm or longer [42]. By contrast, electrons exciting from the valence band (VB) to the conduction band (CB) are recombined with holes, and photoluminescence (PL) is a useful method for predicting the degree of recombination. The PL results for the three samples when the electrons are excited by light with a wavelength of 365 nm are shown in Figure 5. The photoluminescence spectra observed at a 365 nm excitation were found to be dependent on both the structural type and the transition metal ions. The PL spectra of Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc, excited at 365 nm, show only one strong broad emission band centered at approximately 440 nm. It has been fairly accepted that cobalt exists mostly in a divalent state with two coordinations, namely, octahedral and tetrahedral [43]. A Co 2+ ion has a d 7 electronic configuration, and in a tetrahedral crystal field, presents the splitting of energy levels of a d 3 electronic configuration in an octahedral field. In octahedral coordination (Co 2+ ), the free ion ground state 4F splits into 4 T1, 4 T2, and 4 A2 states with the 4 T1 state being the lowest. In a tetrahedral symmetry, the energy levels of Co 2+ ions are 4 T2(4F), 4 T1(4F), 2 E(2G), and 4 T1(4P), with a ground state of 4 A2(4F). The emission band of Co 2+ ions within the region of 630-670 nm is assigned to 2 E(2G) → 4 A2(4F) of Co 2+ ions in tetrahedral coordination [44]. The PL spectra Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc show three strong broad By contrast, electrons exciting from the valence band (VB) to the conduction band (CB) are recombined with holes, and photoluminescence (PL) is a useful method for predicting the degree of recombination. The PL results for the three samples when the electrons are excited by light with a wavelength of 365 nm are shown in Figure 5. The photoluminescence spectra observed at a 365 nm excitation were found to be dependent on both the structural type and the transition metal ions. The PL spectra of Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc, excited at 365 nm, show only one strong broad emission band centered at approximately 440 nm. It has been fairly accepted that cobalt exists mostly in a divalent state with two coordinations, namely, octahedral and tetrahedral [43]. A Co 2+ ion has a d 7 electronic configuration, and in a tetrahedral crystal field, presents the splitting of energy levels of a d 3 electronic configuration in an octahedral field. In octahedral coordination (Co 2+ ), the free ion ground state 4F splits into 4 T 1 , 4 T 2 , and 4 A 2 states with the 4 T 1 state being the lowest. In a tetrahedral symmetry, the energy levels of Co 2+ ions are 4 T 2 (4F), 4 T 1 (4F), 2 E(2G), and 4 T 1 (4P), with a ground state of 4 A 2 (4F). The emission band of Co 2+ ions within the region of 630-670 nm is assigned to 2 E(2G) → 4 A 2 (4F) of Co 2+ ions in tetrahedral coordination [44]. The PL spectra Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc show three strong broad emission bands centered at approximately 450, 465, 510, and 610 nm. Ni 2+ ions (3d 8 ) are expected to exist as octahedral and tetrahedral coordination sites. The luminescence of Ni 2+ ion-doped phosphate can be associated with the d-d optical transitions. It was reported [45] that the energy levels of Ni 2+ ions in an octahedral symmetry are 3 A 2g (F) → 4 T 2g (F), 3 A 2g (F) → 3 T 1g (F), and 3 A 2g (F) → 3 T 1g (P). In addition to these three spin-allowed transitions, a spin-forbidden transition 3 A 2g (F) → 1 E g (D) could be observed at 610 nm. Many authors have proposed that the luminescence properties of Ni-doped samples exist in two regions, namely, the green (510 nm) and red (610 nm) regions. Hence, according to the energy levels of Ni 2+ ion transitions in octahedral sites, the emissions in the green and red regions are assigned to the 1 T 2g (D) → 3 A 2g (F) and 1 T 2g (D) → 3 T 2g (F) transitions. Finally, the PL spectra of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc show two strong emission bands extending from 450 to 510 nm within the visible light range. According to the octahedral crystal field, Cu 2+ (3d 9 ) loses its degeneracy and splits into 2 E g and 2 T 2g , with 2 E g being the lower level. The luminescence spectra of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc exhibit emission peaks at 450, 465, and 510 nm, which are assigned to 3d 9 4s → 3d 10 triplet transitions in Cu 2+ ions [46]. In general, the lower the PL intensity, the smaller the number of electrons recombined. Therefore, it is expected that the photoactivity of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc and Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc particles are good; in particular, Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc particles with a high interfacial transition and defects have high photocatalytic activity. A recent study described the hetero-phase junction phenomenon in association with crystal defects [47]. That is because positive or negative ion defects are generated in crystals, they act as electron or hole capture sites, slowing the recombination rate and increasing the activity of the photocatalyst [48]. The M 3 (PO 4 ) 2 -M 2 P 2 O 7 mpc catalyst synthesized in this study is expected to facilitate the separation of electrons and holes owing to an internal interfacial charge transfer between the inter-phases.  [46]. In general, the lower the PL intensity, the smaller the number of electrons recombined. Therefore, it is expected that the photoactivity of Cu3(PO4)2-Cu2P2O7mpc and Co3(PO4)2-Co2P2O7mpc particles are good; in particular, Cu3(PO4)2-Cu2P2O7mpc particles with a high interfacial transition and defects have high photocatalytic activity. A recent study described the hetero-phase junction phenomenon in association with crystal defects [47]. That is because positive or negative ion defects are generated in crystals, they act as electron or hole capture sites, slowing the recombination rate and increasing the activity of the photocatalyst [48]. The M3(PO4)2-M2P2O7mpc catalyst synthesized in this study is expected to facilitate the separation of electrons and holes owing to an internal interfacial charge transfer between the inter-phases. In general, effectively separated charges have shown a significant effect on the photoactivity of a catalyst [49]. If the excited electrons flow well on the surface of the catalyst without loosening (that is if the electron and hole are separated well), the reduction reaction in CB and the oxidation reaction in VB will be maintained, and the photocatalytic activity will avoid deterioration. Here, the photocurrent density of the three catalysts was measured, the results of which are shown in Figure  6. For the photocurrent density cycle, the sample to be measured was prepared as a paste and coated onto a cell-type FTO glass. This cell was used as a working electrode, and a Pt-coated FTO glass was used as a counter electrode. The two electrodes were connected to form a single system. An iodolyte electrolyte (AN-50, Solarnonix) was used and sufficiently immersed in the electrode, and a potential of 0 V was applied. The current was measured while illuminating one sunlight of a 2000 solar simulator (IVIUM STAT, ABET Technologies). After the initial 1 min stabilization period, the current was measured repeatedly after the light was applied and removed at intervals of 30 s. The cell area (0.4 cm 2 ) was divided by the measured current, and the current density was finally calculated. The higher the current density is, the more electrons flowing through the surface of the catalyst without being coupled with the holes. The current density of the Cu3(PO4)2-Cu2P2O7mpc particles was the In general, effectively separated charges have shown a significant effect on the photoactivity of a catalyst [49]. If the excited electrons flow well on the surface of the catalyst without loosening (that is if the electron and hole are separated well), the reduction reaction in CB and the oxidation reaction in VB will be maintained, and the photocatalytic activity will avoid deterioration. Here, the photocurrent density of the three catalysts was measured, the results of which are shown in Figure 6. For the photocurrent density cycle, the sample to be measured was prepared as a paste and coated onto a cell-type FTO glass. This cell was used as a working electrode, and a Pt-coated FTO glass was used as a counter electrode. The two electrodes were connected to form a single system. An iodolyte electrolyte (AN-50, Solarnonix) was used and sufficiently immersed in the electrode, and a potential of 0 V was applied. The current was measured while illuminating one sunlight of a 2000 solar simulator (IVIUM STAT, ABET Technologies). After the initial 1 min stabilization period, the current was measured repeatedly after the light was applied and removed at intervals of 30 s. The cell area (0.4 cm 2 ) was divided by the measured current, and the current density was finally calculated. The higher the current density is, the more electrons flowing through the surface of the catalyst without being coupled with the holes. The current density of the Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc particles was the largest at 2.75 mA/cm 2 after five cycles, followed by Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc and Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc particles. This result led us to expect that the photocatalytic activity would be best for Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc, which has the highest current density. The results are also consistent with the PL results shown in Figure 5.
Catalysts 2019, 9, x FOR PEER REVIEW 7 of 13 largest at 2.75 mA/cm 2 after five cycles, followed by Co3(PO4)2-Co2P2O7mpc and Ni3(PO4)2-Ni2P2O7mpc particles. This result led us to expect that the photocatalytic activity would be best for Cu3(PO4)2-Cu2P2O7mpc, which has the highest current density. The results are also consistent with the PL results shown in Figure 5.  Figure 7 shows the hydrogen production achieved by water-splitting under 365 nm light radiation. The amount of hydrogen and oxygen generated gradually increased over time in all catalysts. As expected from the optical properties, a cumulative hydrogen production of 80 μmol/cat.g was obtained after 10 h on the Cu3(PO4)2-Cu2P2O7mpc, and hydrogen production of 30 μmol/cat.g on the Co3(PO4)2-Co2P2O7mpc catalyst was obtained. However, Ni3(PO4)2-Ni2P2O7mpc showed an extremely low hydrogen content of 5 μmol/g, which is within 10% of Cu3(PO4)2-Cu2P2O7mpc. Copper, in particular, has high reduction potential and strong electron attracting power, which accelerates the movement of electrons. Furthermore, it can easily hydrogenate protons through a redox oxidation-reduction (Cu 1+ /Cu 2+ ) during the reaction. By contrast, the amount of oxygen generated in each sample was exactly half the amount of hydrogen. This result shows that the water decomposition reaction took place quite stoichiometrically on the M3(PO4)2-M2P2O7mpc catalysts. During the water-splitting reaction, the potentials of the CB and VB of the semiconductors are essential to the electron transfer mechanism, and thus, the photocatalytic performance. To establish an energy potential diagram of Cu3(PO4)2-Cu2P2O7mpc particles, the VB spectra were obtained from  Figure 7 shows the hydrogen production achieved by water-splitting under 365 nm light radiation. The amount of hydrogen and oxygen generated gradually increased over time in all catalysts. As expected from the optical properties, a cumulative hydrogen production of 80 µmol/cat.g was obtained after 10 h on the Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc, and hydrogen production of 30 µmol/cat.g on the Co 3 (PO 4 ) 2 -Co 2 P 2 O 7 mpc catalyst was obtained. However, Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc showed an extremely low hydrogen content of 5 µmol/g, which is within 10% of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc. Copper, in particular, has high reduction potential and strong electron attracting power, which accelerates the movement of electrons. Furthermore, it can easily hydrogenate protons through a redox oxidation-reduction (Cu 1+ /Cu 2+ ) during the reaction. By contrast, the amount of oxygen generated in each sample was exactly half the amount of hydrogen. This result shows that the water decomposition reaction took place quite stoichiometrically on the M 3 (PO 4 ) 2 -M 2 P 2 O 7 mpc catalysts.
Catalysts 2019, 9, x FOR PEER REVIEW 7 of 13 largest at 2.75 mA/cm 2 after five cycles, followed by Co3(PO4)2-Co2P2O7mpc and Ni3(PO4)2-Ni2P2O7mpc particles. This result led us to expect that the photocatalytic activity would be best for Cu3(PO4)2-Cu2P2O7mpc, which has the highest current density. The results are also consistent with the PL results shown in Figure 5.  Figure 7 shows the hydrogen production achieved by water-splitting under 365 nm light radiation. The amount of hydrogen and oxygen generated gradually increased over time in all catalysts. As expected from the optical properties, a cumulative hydrogen production of 80 μmol/cat.g was obtained after 10 h on the Cu3(PO4)2-Cu2P2O7mpc, and hydrogen production of 30 μmol/cat.g on the Co3(PO4)2-Co2P2O7mpc catalyst was obtained. However, Ni3(PO4)2-Ni2P2O7mpc showed an extremely low hydrogen content of 5 μmol/g, which is within 10% of Cu3(PO4)2-Cu2P2O7mpc. Copper, in particular, has high reduction potential and strong electron attracting power, which accelerates the movement of electrons. Furthermore, it can easily hydrogenate protons through a redox oxidation-reduction (Cu 1+ /Cu 2+ ) during the reaction. By contrast, the amount of oxygen generated in each sample was exactly half the amount of hydrogen. This result shows that the water decomposition reaction took place quite stoichiometrically on the M3(PO4)2-M2P2O7mpc catalysts. During the water-splitting reaction, the potentials of the CB and VB of the semiconductors are essential to the electron transfer mechanism, and thus, the photocatalytic performance. To establish an energy potential diagram of Cu3(PO4)2-Cu2P2O7mpc particles, the VB spectra were obtained from During the water-splitting reaction, the potentials of the CB and VB of the semiconductors are essential to the electron transfer mechanism, and thus, the photocatalytic performance. To establish an energy potential diagram of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc particles, the VB spectra were obtained from an XPS analysis, the results of which are shown in Figure 8A. The VB of Cu 3 (PO 4 ) 2 and Cu 2 P 2 O 7 was found to be 2.48 and 1.20 eV, respectively. Based on the bandgap and VB, the CB was obtained using the following equation [50]: Energy potential diagrams for the Cu 3 (PO 4 ) 2 and Cu 2 P 2 O 7 particles are shown in Figure 8B. According to the photocatalytic water-splitting reaction, holes produced in a VB generate oxygen by oxidizing water. In addition, photoelectrons excited in the CB generate hydrogen through water reduction. It is well known that the energy required for a water-splitting reaction is approximately 1.23 eV or more, including both the reduction potential (H + /H 2 ) and the oxidation potential (O 2 /H 2 O). Based on the results, the Cu 3 (PO 4 ) 2 particle satisfies the required energy potential for the water-splitting reaction. It is expected that Cu 2 P 2 O 7 will not cause a water decomposition reaction because the CB value contains the reduction potential of water, whereas the VB value does not include the oxidation potential. However, Cu 2 P 2 O 7 is thought to contribute to a reduction of protons in hydrogen by receiving electrons relaxed from Cu 3 (PO 4 ) 2 . Furthermore, the internal electron transfer system facilitates the separation of electrons and holes, and thus, their recombination is suppressed, thereby improving the catalytic activity. an XPS analysis, the results of which are shown in Figure 8A. The VB of Cu3(PO4)2 and Cu2P2O7 was found to be 2.48 and 1.20 eV, respectively. Based on the bandgap and VB, the CB was obtained using the following equation [50]: Energy potential diagrams for the Cu3(PO4)2 and Cu2P2O7 particles are shown in Figure 8B. According to the photocatalytic water-splitting reaction, holes produced in a VB generate oxygen by oxidizing water. In addition, photoelectrons excited in the CB generate hydrogen through water reduction. It is well known that the energy required for a water-splitting reaction is approximately 1.23 eV or more, including both the reduction potential (H + /H2) and the oxidation potential (O2/H2O). Based on the results, the Cu3(PO4)2 particle satisfies the required energy potential for the watersplitting reaction. It is expected that Cu2P2O7 will not cause a water decomposition reaction because the CB value contains the reduction potential of water, whereas the VB value does not include the oxidation potential. However, Cu2P2O7 is thought to contribute to a reduction of protons in hydrogen by receiving electrons relaxed from Cu3(PO4)2. Furthermore, the internal electron transfer system facilitates the separation of electrons and holes, and thus, their recombination is suppressed, thereby improving the catalytic activity. Scheme 1a,b show the expected water-splitting mechanisms based on optical properties and hydrogen production performance when the Cu2P2O7 mixed-phase in the Cu3(PO4)2 catalyst (Cu3(PO4)2-Cu2P2O7mpc) was present. According to Figure 8, the conductor band and valence band of Cu3(PO4)2 (and Cu2P2O7) had been measured to −0.52 eV (−0.57 eV) and 2.48 eV (1.20 eV), respectively. In Scheme 1a, the bandgap position of Cu2P2O7 particle does not contain the redox potential for water decomposition, and it has the reduction potential of hydrogen but has not the oxidation potential of oxygen [51]. Eventually, the Scheme 1a mechanism is undesirable in this study, even though the Cu3(PO4)2 includes the redox potential for water decomposition [52]. Thus, like Scheme 1b, excited electrons can only move. The Z-scheme mechanism is well-known and refers to when electrons are moved to the Z-type [53]. In particular, when the light of a wavelength of 365 nm is irradiated, the charge transfer can be performed like the Z-scheme system as shown in the red dotted line in Scheme 1b. In particular, when two phases are internally connected rather than heterogeneous, and electrons can move through the internal interfacial transition between these two phases. Furthermore, Wu et al. reported that an interfacial internal electric field is formed in a direct Z-scheme photocatalyst synthesized by an in situ growth method and the photocatalytic activity is enhanced by an internal charge transfer mechanism [54]. Although their catalysts are somewhat different from the catalysts in this work, it is assumed that almost similar forms of internal interfacial Scheme 1a,b show the expected water-splitting mechanisms based on optical properties and hydrogen production performance when the Cu 2 P 2 O 7 mixed-phase in the Cu 3 (PO 4 ) 2 catalyst (Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc) was present. According to Figure 8, the conductor band and valence band of Cu 3 (PO 4 ) 2 (and Cu 2 P 2 O 7 ) had been measured to −0.52 eV (−0.57 eV) and 2.48 eV (1.20 eV), respectively. In Scheme 1a, the bandgap position of Cu 2 P 2 O 7 particle does not contain the redox potential for water decomposition, and it has the reduction potential of hydrogen but has not the oxidation potential of oxygen [51]. Eventually, the Scheme 1a mechanism is undesirable in this study, even though the Cu 3 (PO 4 ) 2 includes the redox potential for water decomposition [52]. Thus, like Scheme 1b, excited electrons can only move. The Z-scheme mechanism is well-known and refers to when electrons are moved to the Z-type [53]. In particular, when the light of a wavelength of 365 nm is irradiated, the charge transfer can be performed like the Z-scheme system as shown in the red dotted line in Scheme 1b. In particular, when two phases are internally connected rather than heterogeneous, and electrons can move through the internal interfacial transition between these two phases. Furthermore, Wu et al. reported that an interfacial internal electric field is formed in a direct Catalysts 2019, 9, 602 9 of 14 Z-scheme photocatalyst synthesized by an in situ growth method and the photocatalytic activity is enhanced by an internal charge transfer mechanism [54]. Although their catalysts are somewhat different from the catalysts in this work, it is assumed that almost similar forms of internal interfacial barriers have been formed, and thus, water can be expected to be decomposed by an interfacial charge transfer mechanism in the same context.
Catalysts 2019, 9, x FOR PEER REVIEW 9 of 13 barriers have been formed, and thus, water can be expected to be decomposed by an interfacial charge transfer mechanism in the same context. Scheme 1. The expected water degradation mechanism when the initial electron excitation occurs in the valence band of (a) Cu2P2O7 and (b) Cu3(PO4)2.

Characterizations
Scheme 1. The expected water degradation mechanism when the initial electron excitation occurs in the valence band of (a) Cu 2 P 2 O 7 and (b) Cu 3 (PO 4 ) 2 .

Synthesis of Catalysts
The synthesis procedures of the M 3 (PO 4 ) 2 -M 2 P 2 O 7 mpc particles are shown in Figure 9, the specific method of which is as follows: Water in the solvent was placed in an Erlenmeyer flask, and metal nitrates (99.95%, Co(NO 3 ) 2 ·xH 2 O, Ni(NO 3 ) 2 ·xH 2 O, Cu(NO 3 ) 2 ·xH 2 O, Junsei Co., Tokyo, Japan) were quantitatively added and uniformly stirred for 1 h until completely dissolved. Sodium hydrogen phosphate (99%, Na 2 HPO 4 , Junsei Co., Japan) as a phosphate source was added such that the ratio of metal to PO 4 was 3:2, followed by stirring for 2 h. The chemical reaction at this time was MNO 3 + Na 2 HPO 4 → MHPO 4 + Na 2 NO 3 , and the amorphous MHPO 4 was precipitated. The precipitate was washed and then dried at 70 • C for 24 h. The dried powder was sintered in an electric furnace at 700 • C for 4 h under air conditions. At this time, a chemical reaction occurred as 6MHPO 4 + 3/2O 2 → 2M 3 (PO 4 ) 2 + 3H 2 O, and finally, we obtained the three types of crystallized Co 3 (PO 4 ) 2 -CO 2 P 2 O 7 mpc, Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc, and Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc particles.

Hydrogen Production through Water Photo Splitting
The photocatalytic decomposition of DI water was carried out using a liquid photoreactor prepared in our laboratory, as shown in Figure 10. First, the photocatalytic decomposition of an aqueous solution without scavengers using a UV light source was conducted using a Pyrex reactor. Next, 1.0 L of distilled water was placed into the reactor, and 0.5 g of synthesized Co3(PO4)2-Co2P2O7mpc, Ni3(PO4)2-Ni2P2O7mpc, and Cu3(PO4)2-Cu2P2O7mpc catalyst powder was added. The light was irradiated using a UV lamp (3 × 6 cm 2 = 18 W cm 2 , length of 30 cm, a diameter of 2.0 cm, Shinan, Pochon, Korea) at a wavelength of 365 nm, and the reaction was conducted for a total of 10 h. To investigate the oxidation states of the Cu2p, P2p, and O1s components and the valence band values of the prepared sample, X-ray photoelectron spectroscopy (XPS) (AXIS Nov, Kratos, Inc., San Diego, CA, USA) was used. The resulting gas was analyzed through gas chromatography (GC, DS7200, Donam Co., Gwangju, Korea). For the GC conditions, a thermal conductivity detector and a Carboxen-1000 column (Bruker, Billerica, MA, USA) were applied, along with injection, oven, and detector temperatures of 423, 393 and 473 K, respectively.

Conclusions
This study focused on improving the performance of the catalyst by promoting the charge transfer through the internal interface without heterojunctions between other crystals. Metal (Co, Ni, and Cu) phosphates having 3d 7 , 3d 8 , and 3d 9 valence electrons were prepared and their photocatalytic activities were compared during a water-splitting reaction. An XRD analysis confirmed that the

Conclusions
This study focused on improving the performance of the catalyst by promoting the charge transfer through the internal interface without heterojunctions between other crystals. Metal (Co, Ni, and Cu) phosphates having 3d 7 , 3d 8 , and 3d 9 valence electrons were prepared and their photocatalytic activities were compared during a water-splitting reaction. An XRD analysis confirmed that the monoclinic octahedral structured Co 3 (PO 4 ) 2 , Ni 3 (PO 4 ) 2 , and Cu 3 (PO 4 ) 2 partially co-exist with Co 2 P 2 O 7 , Ni 2 P 2 O 7 , and Cu 2 P 2 O 7 phases (called M 3 (PO 4 ) 2 -M 2 P 2 O 7 mpc) because their phase equilibrium energies are in similar locations. Absorption peaks for Co 3 (PO 4 ) 2 -CO 2 P 2 O 7 mpc, Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc, and Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc were observed at various wavelengths depending on the d-d electron transition of the metallic components. The intensity of the PL peak related to the recombination between electrons and holes increased in the order of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc < Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc < Co 3 (PO 4 ) 2 -CO 2 P 2 O 7 mpc. The photocurrent density correlated with the charge separation between the electrons and the holes decreased in the order of Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc > Ni 3 (PO 4 ) 2 -Ni 2 P 2 O 7 mpc > Co 3 (PO 4 ) 2 -CO 2 P 2 O 7 mpc. The water-splitting performance was found to be the best in Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc, and cumulative hydrogen production of 80.0 µmol/g was observed during a 10 h period. This is probably attributed to Cu ions in Cu 3 (PO 4 ) 2 -Cu 2 P 2 O 7 mpc with high reduction potential easily capturing electrons and decomposing H 2 O into H + and OH -. In addition, it was confirmed that the performance of a catalyst can be improved through the effective separation of electrons and holes induced by electrons migrating between the internal interfaces of M 3 (PO 4 ) 2 -M 2 P 2 O 7 mpc.