Constructing Co3O4/g-C3N4 Ultra-Thin Nanosheets with Z-Scheme Charge Transfer Pathway for Efficient Photocatalytic Water Splitting

Photocatalytic water splitting for hydrogen generation is a significant pathway for sustainable energy conversion and production. The photocatalysts with a Z-scheme water splitting charge transfer pathway is superior due to the good separation and migration ability of photoexcited charge carriers. Herein, Co3O4/g-C3N4 photocatalysts with Z-scheme charge transfer pathway were successfully constructed by an electrostatic interaction-annealing method. The as-prepared Co3O4/g-C3N4 ultra-thin nanosheets were tested and analyzed by XRD, EA, ICP, SEM, TEM, AFM, XPS, UV-Vis DRS, PL and photoelectrochemical measurements. Moreover, the influences of fabrication parameters on performance of Co3O4/g-C3N4 catalysts were investigated, and 0.5% Co3O4/g-C3N4 exhibited the optimal activity. Based on the characterization and catalytic performance, the Z-scheme charge transfer pathway of Co3O4/g-C3N4 was established and put forward. To further improve the catalytic performance of Co3O4/g-C3N4, 0.5% Pt was added as a co-catalyst. The obtained Pt/0.5% Co3O4/g-C3N4 was recyclable and remained the original catalytic water splitting performance within 20 h. The modification of Co3O4 and Pt improved the separation and migration of e− and h+, and induced the increased hydrogen evolution rate of g-C3N4.


Introduction
Photocatalytic hydrogen generation is a high efficiency, environmentally friendly and economically practical technology for utilizing solar energy [1][2][3]. It promises a sustainable alternative via semiconductors to address environmental issues and energy shortages all around the world [4][5][6]. The strong optical absorption ability, high separation and migration efficiency and stability are major factors for the photocatalysts' activity for hydrogen evolution [7,8]. Now efforts have been devoted to fabricating more and more effective semiconductors for solar-energy utilization and conversion, majorly categorized as metal oxides, amorphous (oxy)-hydroxides, (oxy)nitrides, (oxy)sulphides and polymeric catalysts [9,10].
Graphitic carbon nitride (g-C 3 N 4 ) is a robust and nontoxicity polymeric catalyst with good properties [11]. It was constructed by the polymerization method as a photocatalytic water splitting catalyst in 2009 [12]. Up to now, g-C 3 N 4 has been widely developed as an emerging and prospective catalyst in different areas, for example photocatalytic degradation of pollution [13], hydrogen evolution through photocatalytic water splitting [14] and CO 2 reduction [15]. Nowadays g-C 3 N 4 with various morphologies have emerged,
For comparison, the pure Co 3 O 4 was fabricated by calcination of CoCl 2 ·6H 2 O at 600 • C for 4 h.

Characterization
The X-ray powder diffraction (XRD) was characterized by a diffractometer from Rigaku Corporation. Scanning electron microscopy image was observed by a SU3500 (Hitachi, Tokyo, Japan) microscope. Elemental analysis was explored by Elemental Analyzer (Vario EL cube, Langenselbold, Germany). The ICP analysis was performed by Inductive Coupled Plasma Emission Spectrometer (SpectroBlue, Kleve, Germany). Transmission electron microscopy (TEM) was observed by a FEI TalosF200X instrument. EDS analyses were characterized by an energy dispersive X-ray spectroscope (Bruker, Billerica, MA, USA). Atomic force microscopy (AFM) images were performed by the Bruker icon AFM instrument. The X-ray photoelectron spectroscopy (XPS, Thermo escalab 250Xi) analysis was characterized on a photoemission spectroscopy, using a mono AlKα radiation source. The UV-Vis diffuse reflection spectra (UV-Vis DRS) was explored by a UV-Vis spectrophotometer (UV3600-Plus, Shimadzu, Kyoto, Japan). Photoluminescence spectra (PL) was conducted by a spectrophotometer (PTI, New York, NY, USA). Electrochemical experiments were explored with a workstation (Zahner Zenium, Kronach, Germany).

Photocatalytic Activity
The H 2 production tests were implemented with the catalyst (50 mg), which was dispersed in triethanolamine solution (10% vol TEOA) in the Labsolar-6A online system. 0.5 wt.% Pt, which was used as co-catalyst, was in-situ reduced by light on the catalysts by adding H 2 PtCl 6 aqueous solution in the above suspension of the reaction. After vacuuming, the Argon gas was added. A 300 W Xenon lamp was served to achieve the system irradiation during tests. The solution temperature was controlled at 5 • C to avoid the thermal effect produced by light. Additionally, the tests were carried out using a gas chromatograph and the carrier gas is Argon (GC D7900 II, Tianmei, Shanghai, China).

XRD Analysis
The XRD measurement was characterized to confirm the phase of the catalysts, as shown in Figure 1. The peaks at 2θ = 13. 0 and 27.7 • were indexed to the characteristic peaks of g-C 3 N 4 [5,35] (1 1 1), (2 2 0), (3 1 1), (2 2 2), (4 0 0), (4 2 2), (5 1 1) and (4 4 0) plane of cubic Co 3 O 4 (PDF#42-1467), respectively. It was evident that the peaks of Co 3 O 4 and g-C 3 N 4 were both existed in the 6% and 12% Co 3 O 4 /g-C 3 N 4 photocatalysts' XRD patterns. No obvious Co 3 O 4 peaks were observed in Co 3 O 4 /g-C 3 N 4 samples with 0.3%, 0.5%, 1% and 3% Co 3 O 4 . That might be owing to the relatively low content and high distribution of Co 3 O 4 [36]. What is more, as clearly demonstrated, the peaks of the photocatalysts were sharp and strong, illustrating that the samples possessed an excellent crystallization. There were no peaks of other phases in Figure 1, which indicated the good purity of the synthesized photocatalysts. irradiation during tests. The solution temperature was controlled at 5 °C to avoid the thermal effect produced by light. Additionally, the tests were carried out using a gas chromatograph and the carrier gas is Argon (GC D7900 II, Tianmei, Shanghai, China).

EA and ICP
For the g-C3N4 sample, the atomic ratio of C element and N element was investigated by EA. The value exhibited a result that the atomic ratio of C/N = 3.00: 4.27, which was very close to the value of g-C3N4. Based on XRD analysis results, the as-synthesized products were pure g-C3N4.
The amounts of Co element in the as-prepared catalysts were monitored by the ICP. The actual contents of 0.3%, 0.5%, 1%, 3%, 6% and 12% Co3O4/g-C3N4 is 0.3540, 0.5354, 1.146, 3.038, 6.253, 12.44 wt.%, respectively. The corresponding contents of Co3O4 in the series Co3O4/g-C3N4 photocatalysts are close to theoretical values. Due to the negative zeta potentials of g-C3N4, there was strong electrostatic interaction between Co 2+ and g-C3N4 nanosheets [37]. As a result, the added Co 2+ should have no considerable loss. Actually, the actual weight ratio of Co3O4 to g-C3N4 in the prepared Co3O4/g-C3N4 catalysts was a little higher than original CoCl2·6H2O to g-C3N4. It was assignable to the slight weight loss after the second calcination process of g-C3N4.

Morphology Analysis
The microstructure of as-fabricated catalysts was presented in Figure 2. From Figure  2a, Co3O4 had polyhedral structure, and the 0.5% Co3O4/g-C3N4 showed an irregular and crinkly structure. Pt/0.5% Co3O4/g-C3N4 exhibited 2D morphology with a crinkly nanosheet structure, which seem to be a loose and soft product with a diameter of several micrometers. As observed in the images, the microstructures of 0.5% Co3O4/g-C3N4 and

EA and ICP
For the g-C 3 N 4 sample, the atomic ratio of C element and N element was investigated by EA. The value exhibited a result that the atomic ratio of C/N = 3.00: 4.27, which was very close to the value of g-C 3 N 4 . Based on XRD analysis results, the as-synthesized products were pure g-C 3 N 4 .
The amounts of Co element in the as-prepared catalysts were monitored by the ICP. The actual contents of 0.3%, 0.5%, 1%, 3%, 6% and 12% Co 3 O 4 /g-C 3 N 4 is 0.3540, 0.5354, 1.146, 3.038, 6.253, 12.44 wt.%, respectively. The corresponding contents of Co 3 O 4 in the series Co 3 O 4 /g-C 3 N 4 photocatalysts are close to theoretical values. Due to the negative zeta potentials of g-C 3 N 4 , there was strong electrostatic interaction between Co 2+ and g-C 3 N 4 nanosheets [37]. As a result, the added Co 2+ should have no considerable loss. Actually, the actual weight ratio of Co 3 O 4 to g-C 3 N 4 in the prepared Co 3 O 4 /g-C 3 N 4 catalysts was a little higher than original CoCl 2 ·6H 2 O to g-C 3 N 4 . It was assignable to the slight weight loss after the second calcination process of g-C 3 N 4 .

Morphology Analysis
The microstructure of as-fabricated catalysts was presented in Figure 2. From Figure 2a, Co 3 O 4 had polyhedral structure, and the 0.5% Co 3 O 4 /g-C 3 N 4 showed an irregular and crinkly structure. Pt/0.5% Co 3 O 4 /g-C 3 N 4 exhibited 2D morphology with a crinkly nanosheet structure, which seem to be a loose and soft product with a diameter of several micrometers. As observed in the images, the microstructures of 0.5% Co 3 O 4 /g-C 3 N 4 and Pt/0.5% Co 3 O 4 /g-C 3 N 4 were nanosheets. The composition and elements distribution of 0.5% Co 3 O 4 /g-C 3 N 4 and Pt/0.5% Co 3 O 4 /g-C 3 N 4 were further investigated by EDS maps, as displayed in Figures 3 and 4. All of the elements were uniformly dispersed in the photcatalysts, and no signals of other elements were observed. Furthermore, the existence of C, N, Co, O and Pt were highly coincided with the area of g-C 3 N 4 . It seems that Co 3 O 4 and Pt could be uniformly deposited on the g-C 3 N 4 .
Pt/0.5% Co3O4/g-C3N4 were nanosheets. The composition and elements distribution of 0.5% Co3O4/g-C3N4 and Pt/0.5% Co3O4/g-C3N4 were further investigated by EDS maps, as displayed in Figures 3 and 4. All of the elements were uniformly dispersed in the photcatalysts, and no signals of other elements were observed. Furthermore, the existence of C, N, Co, O and Pt were highly coincided with the area of g-C3N4. It seems that Co3O4 and Pt could be uniformly deposited on the g-C3N4.

AFM Analysis
To further investigate the structure and height profiles of the photocatalysts, AFM was performed. The images and corresponding thickness of samples were presented in Figure 5. The architectures of the g-C3N4 were 2D ultra-thin nanosheets, with 1-2 nm step height. As illustrated in Figure 5c,d, there is no obvious step height change for 0.5% Co3O4/g-C3N4 ultra-thin nanosheets. Additionally, g-C3N4 and 0.5% Co3O4/g-C3N4 nanosheets both exhibited few-layer ultra-thin nanosheets with uniform thickness. Ultrathin nanosheets had a short bulk diffusion length [22,38]. Compared with bulk g-C3N4, ultrathin g-C3N4 nanosheets and 0.5% Co3O4/g-C3N4 photocatalysts had a shorter electron transfer path, which can increase the lifetime of photoexcited electrons. The unique structure of few-layer g-C3N4 photocatalyst generated numerous charge transfer nanochannels

AFM Analysis
To further investigate the structure and height profiles of the photocatalysts, AFM was performed. The images and corresponding thickness of samples were presented in Figure 5. The architectures of the g-C3N4 were 2D ultra-thin nanosheets, with 1-2 nm step height. As illustrated in Figure 5c,d, there is no obvious step height change for 0.5% Co3O4/g-C3N4 ultra-thin nanosheets. Additionally, g-C3N4 and 0.5% Co3O4/g-C3N4 nanosheets both exhibited few-layer ultra-thin nanosheets with uniform thickness. Ultrathin nanosheets had a short bulk diffusion length [22,38]. Compared with bulk g-C3N4, ultrathin g-C3N4 nanosheets and 0.5% Co3O4/g-C3N4 photocatalysts had a shorter electron transfer path, which can increase the lifetime of photoexcited electrons. The unique structure of few-layer g-C3N4 photocatalyst generated numerous charge transfer nanochannels

AFM Analysis
To further investigate the structure and height profiles of the photocatalysts, AFM was performed. The images and corresponding thickness of samples were presented in Figure 5. The architectures of the g-C 3 N 4 were 2D ultra-thin nanosheets, with 1-2 nm step height. As illustrated in Figure 5c,d, there is no obvious step height change for 0.5% Co 3 O 4 /g-C 3 N 4 ultra-thin nanosheets. Additionally, g-C 3 N 4 and 0.5% Co 3 O 4 /g-C 3 N 4 nanosheets both exhibited few-layer ultra-thin nanosheets with uniform thickness. Ultrathin nanosheets had a short bulk diffusion length [22,38]. Compared with bulk g-C 3 N 4 , ultrathin g-C 3 N 4 nanosheets and 0.5% Co 3 O 4 /g-C 3 N 4 photocatalysts had a shorter electron transfer path, which can increase the lifetime of photoexcited electrons. The unique structure of few-layer g-C 3 N 4 photocatalyst generated numerous charge transfer nanochannels and provided a short transfer path for electron transfer, which would promote the efficient separation and migration of e − and h + [22]. and provided a short transfer path for electron transfer, which would promote the efficient separation and migration of e − and h + [22].

XPS Analysis
To confirm the element compositions and chemical states in the as-synthesized 0.5% Co3O4/g-C3N4, the XPS analysis was conducted. From XPS survey spectrum (Figure 6a), C 1s, N 1s and O 1s peaks were evident. What is more, no signals assigned to other elements were displayed in this survey spectrum, suggesting that the purity of the samples is very high. As displayed in Figure 6b, the peak at 287.7 eV is due to sp 2 C atoms, which can bond to N atoms in catalysts [39]. Additionally, the peak at 293.1 eV is owing to π-excitation [40]. Evidently, N 1s has three different peaks at 398.1 eV, 400.0 eV and 404.0 eV in Figure 5c assigning to sp 2 N atoms of triazine rings, bridging nitrogen atoms in (N-(C)3) and the charging effects of the hetero-cycles or positive charge localization [41], respectively. The Co 2p at the peaks of 780.4 and 796.1 eV are ascribed to Co 2p3/2 and Co 2p1/2, suggesting that Co3O4 exists in the photocatalysts [42,43]. Compared with g-C3N4, there was negative energy shift of C 1s and N 1s in 0.5% Co3O4/g-C3N4 photocatalysts, indicating that the eare transferred from Co3O4 to g-C3N4 [31]. As discussed in the Z-scheme charge transfer system, eon CB of Co3O4 can migrate to VB of the other semiconductor. Therefore, Z-scheme charge transfer pathway of Co3O4/g-C3N4 system agreed with the results of negative binding energy shift in 0.5% Co3O4/g-C3N4 photocatalysts. These results take firm evidence that the Co3O4 and g-C3N4 were co-existed in the samples and Z-scheme charge transfer pathway was constructed in photocatalysts.

XPS Analysis
To confirm the element compositions and chemical states in the as-synthesized 0.5% Co 3 O 4 /g-C 3 N 4 , the XPS analysis was conducted. From XPS survey spectrum (Figure 6a), C 1s, N 1s and O 1s peaks were evident. What is more, no signals assigned to other elements were displayed in this survey spectrum, suggesting that the purity of the samples is very high. As displayed in Figure 6b, the peak at 287.7 eV is due to sp 2 C atoms, which can bond to N atoms in catalysts [39]. Additionally, the peak at 293.1 eV is owing to π-excitation [40]. Evidently, N 1s has three different peaks at 398.1 eV, 400.0 eV and 404.0 eV in Figure 5c assigning to sp 2 N atoms of triazine rings, bridging nitrogen atoms in (N-(C) 3 ) and the charging effects of the hetero-cycles or positive charge localization [41], respectively. The Co 2p at the peaks of 780.4 and 796.1 eV are ascribed to Co 2p 3/2 and Co 2p 1/2 , suggesting that Co 3 O 4 exists in the photocatalysts [42,43]. Compared with g-C 3 N 4 , there was negative energy shift of C 1s and N 1s in 0.5% Co 3 O 4 /g-C 3 N 4 photocatalysts, indicating that the e − are transferred from Co 3 O 4 to g-C 3 N 4 [31]. As discussed in the Z-scheme charge transfer system, e − on CB of Co 3 O 4 can migrate to VB of the other semiconductor. Therefore, Z-scheme charge transfer pathway of Co 3 O 4 /g-C 3 N 4 system agreed with the results of negative binding energy shift in 0.5% Co 3 O 4 /g-C 3 N 4 photocatalysts. These results take firm evidence that the Co 3 O 4 and g-C 3 N 4 were co-existed in the samples and Z-scheme charge transfer pathway was constructed in photocatalysts.

UV-Vis DRS Analysis
The UV-Vis DRS spectrophotometer analysis was performed to measure optical absorption properties of the catalysts. 0.5% Co3O4/g-C3N4 possessed a strong absorption and its edge located at 443 nm in Figure 6a. Additionally, although the absorption edges and bandgap energies of g-C3N4 and 0.5% Co3O4/g-C3N4 had no apparent distinction, optical absorption intensity of 0.5% Co3O4/g-C3N4 was apparently increased, which made it easier to harvest light. The spectra (Figure 7a) of the catalysts suggested that the samples displayed a strong absorption all between 200-440 nm, which covered ultraviolet and visible regions. These photocatalysts can respond to the visible light and their responsiveness was enhanced, which has probably a more important advantage to photocatalytic water splitting. Bandgaps of Co3O4, g-C3N4 and 0.5% Co3O4/g-C3N4 photocatalysts were achieved by the Tauc plots, and results are presented in Figure 7b. The bandgap (Eg) of catalysts can be received by the Tauc Equation [44]. Because the Co3O4 [45] and g-C3N4 [46] are typical direct semiconductors, n in the Tauc equation should be all 1/2. The Eg of as-prepared g-C3N4 and Co3O4 are 2.8 eV and 2.0 eV, respectively. The ECB values of Co3O4 and g-C3N4 are obtained via this equation. Figure 6. XPS spectra (a-d) survey, C 1s, N 1s and Co 2p of 0.5 % Co 3 O 4 /g-C 3 N 4 , (e,f) C 1s and N 1s of g-C 3 N 4 , respectively.

UV-Vis DRS Analysis
The UV-Vis DRS spectrophotometer analysis was performed to measure optical absorption properties of the catalysts. 0.5% Co 3 O 4 /g-C 3 N 4 possessed a strong absorption and its edge located at 443 nm in Figure 6a. Additionally, although the absorption edges and bandgap energies of g-C 3 N 4 and 0.5% Co 3 O 4 /g-C 3 N 4 had no apparent distinction, optical absorption intensity of 0.5% Co 3 O 4 /g-C 3 N 4 was apparently increased, which made it easier to harvest light. The spectra (Figure 7a) of the catalysts suggested that the samples displayed a strong absorption all between 200-440 nm, which covered ultraviolet and visible regions. These photocatalysts can respond to the visible light and their responsiveness was enhanced, which has probably a more important advantage to photocatalytic water splitting. Bandgaps of Co 3 O 4, g-C 3 N 4 and 0.5% Co 3 O 4 /g-C 3 N 4 photocatalysts were achieved by the Tauc plots, and results are presented in Figure 7b. The bandgap (E g ) of catalysts can be received by the Tauc Equation [44]. Because the Co 3 O 4 [45] and g-C 3 N 4 [46] are typical direct semiconductors, n in the Tauc equation should be all 1/2. The E g of as-prepared g-C 3 N 4 and Co 3 O 4 are 2.8 eV and 2.0 eV, respectively. The E CB values of Co 3 O 4 and g-C 3 N 4 are obtained via this equation.

PL
As displayed in Figure 8, the recombination ability of eand h + was speculated from the photoluminescence intensity of the samples. All the photocatalysts were evaluated at the excitation wavelength, which was 320 nm. Generally, the higher intensity of the PL emission spectra means a worse separation ability of carriers [49]. The spectrum of 0.5% Co3O4/g-C3N4 catalyst showed lower emission intensity than the intensity of g-C3N4, leading to a lower recombining frequency of eand h + in 0.5% Co3O4/g-C3N4. The lower the peak of photoluminescence spectra presented, the more excellent catalytic performance can be anticipated. It seems that 0.5% Co3O4/g-C3N4 catalyst may have a better photocatalytic performance than pure g-C3N4.

Photoelectrochemical Analysis
The photocurrent measurement was produced under the irradiation of 429 nm wavelength light, as illustrated in Figure 9. The photocurrent immediately increased to a value and remained when the light irradiated on the catalysts. The photocurrent decreased as soon as the light turned off. It manifested that 0.5% Co3O4/g-C3N4 has a photocurrent response and the photocurrent value is around 10 nA at 0.1 V. However, no photocurrent existed in g-C3N4. It indicated that 0.5% Co3O4/g-C3N4 is visible to light-responsive photocatalysts. After Co3O4 was introduced, the g-C3N4 catalyst can respond to visible light, and

PL
As displayed in Figure 8, the recombination ability of e − and h + was speculated from the photoluminescence intensity of the samples. All the photocatalysts were evaluated at the excitation wavelength, which was 320 nm. Generally, the higher intensity of the PL emission spectra means a worse separation ability of carriers [49]. The spectrum of 0.5% Co 3 O 4 /g-C 3 N 4 catalyst showed lower emission intensity than the intensity of g-C 3 N 4 , leading to a lower recombining frequency of e − and h + in 0.5% Co 3 O 4 /g-C 3 N 4 . The lower the peak of photoluminescence spectra presented, the more excellent catalytic performance can be anticipated. It seems that 0.5% Co 3 O 4 /g-C 3 N 4 catalyst may have a better photocatalytic performance than pure g-C 3 N 4 .

PL
As displayed in Figure 8, the recombination ability of eand h + was speculated from the photoluminescence intensity of the samples. All the photocatalysts were evaluated at the excitation wavelength, which was 320 nm. Generally, the higher intensity of the PL emission spectra means a worse separation ability of carriers [49]. The spectrum of 0.5% Co3O4/g-C3N4 catalyst showed lower emission intensity than the intensity of g-C3N4, leading to a lower recombining frequency of eand h + in 0.5% Co3O4/g-C3N4. The lower the peak of photoluminescence spectra presented, the more excellent catalytic performance can be anticipated. It seems that 0.5% Co3O4/g-C3N4 catalyst may have a better photocatalytic performance than pure g-C3N4.

Photoelectrochemical Analysis
The photocurrent measurement was produced under the irradiation of 429 nm wavelength light, as illustrated in Figure 9. The photocurrent immediately increased to a value and remained when the light irradiated on the catalysts. The photocurrent decreased as soon as the light turned off. It manifested that 0.5% Co3O4/g-C3N4 has a photocurrent response and the photocurrent value is around 10 nA at 0.1 V. However, no photocurrent existed in g-C3N4. It indicated that 0.5% Co3O4/g-C3N4 is visible to light-responsive photocatalysts. After Co3O4 was introduced, the g-C3N4 catalyst can respond to visible light, and

Photoelectrochemical Analysis
The photocurrent measurement was produced under the irradiation of 429 nm wavelength light, as illustrated in Figure 9. The photocurrent immediately increased to a value and remained when the light irradiated on the catalysts. The photocurrent decreased as soon as the light turned off. It manifested that 0.5% Co 3 O 4 /g-C 3 N 4 has a photocurrent response and the photocurrent value is around 10 nA at 0.1 V. However, no photocurrent existed in g-C 3 N 4 . It indicated that 0.5% Co 3 O 4 /g-C 3 N 4 is visible to light-responsive photocatalysts. After Co 3 O 4 was introduced, the g-C 3 N 4 catalyst can respond to visible light, and a photocurrent was exhibited. That should be attributed to the formation of Z-scheme charge transfer pathway, which resulted in improved separation and migration of the photo-induced charges. As a result, a photocurrent with higher intensity can be obtained.
Nanomaterials 2021, 11, x FOR PEER REVIEW 10 of 14 a photocurrent was exhibited. That should be attributed to the formation of Z-scheme charge transfer pathway, which resulted in improved separation and migration of the photo-induced charges. As a result, a photocurrent with higher intensity can be obtained. Figure 9. The photocurrent measurement spectra of the 0.5% Co3O4/g-C3N4 catalysts.

Photocatalytic Activity and Cycling Tests
As illustrated in Figure 10a, g-C3N4 exhibits no activity. However, the series Co3O4/g-C3N4 catalysts all present good photocatalytic performance, and 0.5% Co3O4/g-C3N4 has the fine H2 evolution rate of 8.774 μmol•g −1 •h −1 . Obviously, the modification of Co3O4 can greatly promote the water splitting reaction rate of g-C3N4. The results of photocatalytic activity tests can be proved by photoelectrochemical analysis. As indicated in the photoelectrochemical analysis, the g-C3N4 can respond to 429 nm only after Co3O4 was modified. Therefore, the quantum yield of g-C3N4 is low. Catalytic activity of g-C3N4 is also low. The photoluminescence spectra analysis showed that eand h + in the g-C3N4 have higher recombination rates. It suggested that g-C3N4 has a lower photocatalytic hydrogen production. According to the band structure of Co3O4 and g-C3N4 as we noted above, the reasonable photocatalytic charge transfer pathway was preliminarily put forward to explain the reason for catalytic improved-performance of 0.5% Co3O4/g-C3N4 in Figure 11. The electrons of Co3O4 and g-C3N4 are photo-generated from VB to CB in response to irradiation. After calcination, interaction between Co3O4 and g-C3N4 was enhanced. The eon CB of Co3O4 would quickly migrate to VB of g-C3N4 and combine with h + , which replaced inner self-combination. Some holes were reacted with TEOA, which worked as an oxidation sacrificial agent. Therefore, the electric field between Co3O4 and g-C3N4 was formed, with electrons and holes accelerated at different parts. The electric field can promote the directional transfer of charges and introduction of Co3O4 can boost the electric field of Co3O4/g- Figure 9. The photocurrent measurement spectra of the 0.5% Co 3 O 4 /g-C 3 N 4 catalysts.

Photocatalytic Activity and Cycling Tests
As illustrated in Figure 10a, g-C 3 N 4 exhibits no activity. However, the series Co 3 O 4 /g-C 3 N 4 catalysts all present good photocatalytic performance, and 0.5% Co 3 O 4 /g-C 3 N 4 has the fine H 2 evolution rate of 8.774 µmol·g −1 ·h −1 . Obviously, the modification of Co 3 O 4 can greatly promote the water splitting reaction rate of g-C 3 N 4 . The results of photocatalytic activity tests can be proved by photoelectrochemical analysis. As indicated in the photoelectrochemical analysis, the g-C 3 N 4 can respond to 429 nm only after Co 3 O 4 was modified. Therefore, the quantum yield of g-C 3 N 4 is low. Catalytic activity of g-C 3 N 4 is also low. The photoluminescence spectra analysis showed that e − and h + in the g-C 3 N 4 have higher recombination rates. It suggested that g-C 3 N 4 has a lower photocatalytic hydrogen production. a photocurrent was exhibited. That should be attributed to the formation of Z-scheme charge transfer pathway, which resulted in improved separation and migration of the photo-induced charges. As a result, a photocurrent with higher intensity can be obtained. Figure 9. The photocurrent measurement spectra of the 0.5% Co3O4/g-C3N4 catalysts.

Photocatalytic Activity and Cycling Tests
As illustrated in Figure 10a, g-C3N4 exhibits no activity. However, the series Co3O4/g-C3N4 catalysts all present good photocatalytic performance, and 0.5% Co3O4/g-C3N4 has the fine H2 evolution rate of 8.774 μmol•g −1 •h −1 . Obviously, the modification of Co3O4 can greatly promote the water splitting reaction rate of g-C3N4. The results of photocatalytic activity tests can be proved by photoelectrochemical analysis. As indicated in the photoelectrochemical analysis, the g-C3N4 can respond to 429 nm only after Co3O4 was modified. Therefore, the quantum yield of g-C3N4 is low. Catalytic activity of g-C3N4 is also low. The photoluminescence spectra analysis showed that eand h + in the g-C3N4 have higher recombination rates. It suggested that g-C3N4 has a lower photocatalytic hydrogen production. According to the band structure of Co3O4 and g-C3N4 as we noted above, the reasonable photocatalytic charge transfer pathway was preliminarily put forward to explain the reason for catalytic improved-performance of 0.5% Co3O4/g-C3N4 in Figure 11. The electrons of Co3O4 and g-C3N4 are photo-generated from VB to CB in response to irradiation. After calcination, interaction between Co3O4 and g-C3N4 was enhanced. The eon CB of Co3O4 would quickly migrate to VB of g-C3N4 and combine with h + , which replaced inner self-combination. Some holes were reacted with TEOA, which worked as an oxidation sacrificial agent. Therefore, the electric field between Co3O4 and g-C3N4 was formed, with electrons and holes accelerated at different parts. The electric field can promote the directional transfer of charges and introduction of Co3O4 can boost the electric field of Co3O4/g- According to the band structure of Co 3 O 4 and g-C 3 N 4 as we noted above, the reasonable photocatalytic charge transfer pathway was preliminarily put forward to explain the reason for catalytic improved-performance of 0.5% Co 3 O 4 /g-C 3 N 4 in Figure 11. The electrons of Co 3 O 4 and g-C 3 N 4 are photo-generated from VB to CB in response to irradiation. After calcination, interaction between Co 3 O 4 and g-C 3 N 4 was enhanced. The e − on CB of Co 3 O 4 would quickly migrate to VB of g-C 3 N 4 and combine with h + , which replaced inner self-combination. Some holes were reacted with TEOA, which worked as an oxidation sacrificial agent. Therefore, the electric field between Co 3 O 4 and g-C 3 N 4 was formed, with electrons and holes accelerated at different parts. The electric field can promote the directional transfer of charges and introduction of Co 3 O 4 can boost the electric field of Co 3 O 4 /g-C 3 N 4 . Hence, the existence of Co 3 O 4 effectively promoted the interfacial charges to separate and transfer and the hydrogen generation, leading to superior photocatalytic performance. What is more, the XPS analysis indicated that e − are transferred from Co 3 O 4 to g-C 3 N 4 nanosheets. It suggested that the Z-scheme charge transfer pathway was in agreement with experiment results.
C3N4. Hence, the existence of Co3O4 effectively promoted the interfacial charges to separate and transfer and the hydrogen generation, leading to superior photocatalytic performance. What is more, the XPS analysis indicated that e − are transferred from Co3O4 to g-C3N4 nanosheets. It suggested that the Z-scheme charge transfer pathway was in agreement with experiment results. Figure 11. Proposed mechanism of photocatalytic system.
The probable reactions process of photocatalytic hydrogen generation is as follows: Thence, the proposed Z-scheme charge transfer pathway of photocatalysts showed good separation and migration of eand h + . The obvious enhancement of the hydrogen generation ability of the 0.5% Co3O4/g-C3N4 ultra-thin nanosheets may attribute to the significant and ideal Z-scheme photocatalytic charge transfer pathway.
After Pt was added, further efforts were made to improve the performance of Co3O4/g-C3N4. As presented in Figure 10a, Pt/0.5% Co3O4/g-C3N4 exhibited a predominant rate and the average rate was up to 1620 μmol•g −1 •h −1 . This is 2.1 times the photocatalytic performance of Pt/g-C3N4.
To explore the influence of Co3O4 contents on the performance of the Pt/Co3O4/g-C3N4, results of Pt/Co3O4/g-C3N4 with different Co3O4 contents were shown in Figure 10b. From this figure, it can be observed that all photocatalysts manifested improved hydrogen generation activity compared with Pt/g-C3N4, except Pt/12% Co3O4/g-C3N4. Owning to the relatively low content of Co3O4 (0.5 % Co3O4), Pt was mainly deposited on g-C3N4 nanosheets. When the catalyst was illuminated, the electrons migrated from Co3O4 to g-C3N4, and were finally transferred to Pt. The H + in the solution obtained the electrons from the Pt particles to form H2. When the Co3O4 content was relatively low (0.3%), Z-scheme charge transfer pathway is still the main pathway in this photocatalytic system. However, the amount of electrons and holes in Co3O4 was relatively low than that in the 0.5% one. Therefore relatively large amounts of holes in g-C3N4 were reacted with TEOA instead of combining with the electrons in Co3O4. As a result, e − and h + in the Pt/0.3% Co3O4/g-C3N4 have higher recombination rates than in Pt/0.5% Co3O4/g-C3N4, and the Pt/0.3% Co3O4/g-C3N4 exhibited lower photoactivity. When Co3O4 contents increased, Co3O4 will accumulate. More and more Pt/Co3O4 interfaces formed, and the electrons migrated from Co3O4 to Pt, instead of combining with the holes in g-C3N4. Consequently, the Z-scheme charge transfer pathway was blocked, and the activity of Co3O4/g-C3N4 decreased with the increasing Co3O4 contents. The probable reactions process of photocatalytic hydrogen generation is as follows: Thence, the proposed Z-scheme charge transfer pathway of photocatalysts showed good separation and migration of e − and h + . The obvious enhancement of the hydrogen generation ability of the 0.5% Co 3 O 4 /g-C 3 N 4 ultra-thin nanosheets may attribute to the significant and ideal Z-scheme photocatalytic charge transfer pathway.
After Pt was added, further efforts were made to improve the performance of Co 3 O 4 /g-C 3 N 4 . As presented in Figure 10a, Pt/0.5% Co 3 O 4 /g-C 3 N 4 exhibited a predominant rate and the average rate was up to 1620 µmol·g −1 ·h −1 . This is 2.1 times the photocatalytic performance of Pt/g-C 3 N 4 .
To explore the influence of Co 3 O 4 contents on the performance of the Pt/Co 3 O 4 /g-C 3 N 4 , results of Pt/Co 3 O 4 /g-C 3 N 4 with different Co 3 O 4 contents were shown in Figure 10b. From this figure, it can be observed that all photocatalysts manifested improved hydrogen generation activity compared with Pt/g-C 3 N 4 , except Pt/12% Co 3 O 4 /g-C 3 N 4 . Owning to the relatively low content of Co 3 O 4 (0.5 % Co 3 O 4 ), Pt was mainly deposited on g-C 3 N 4 nanosheets. When the catalyst was illuminated, the electrons migrated from Co 3 O 4 to g-C 3 N 4 , and were finally transferred to Pt. The H + in the solution obtained the electrons from the Pt particles to form H 2 . When the Co 3 O 4 content was relatively low (0.3%), Z-scheme charge transfer pathway is still the main pathway in this photocatalytic system. However, the amount of electrons and holes in Co 3 O 4 was relatively low than that in the 0.5% one. Therefore relatively large amounts of holes in g-C 3 N 4 were reacted with TEOA instead of combining with the electrons in Co 3 O 4 . As a result, e − and h + in the Pt/0.3% Co 3 O 4 /g-C 3 N 4 have higher recombination rates than in Pt/0.5% Co 3 O 4 /g-C 3 N 4 , and the Pt/0.3% Co 3 O 4 /g-C 3 N 4 exhibited lower photoactivity. When Co 3 O 4 contents increased, Co 3 O 4 will accumulate. More and more Pt/Co 3 O 4 interfaces formed, and the electrons migrated from Co 3 O 4 to Pt, instead of combining with the holes in g-C 3 N 4 . Consequently, the Z-scheme charge transfer pathway was blocked, and the activity of Co 3 O 4 /g-C 3 N 4 decreased with the increasing Co 3 O 4 contents.
Cycling stability tests of the optimal Pt/0.5% Co 3 O 4 /g-C 3 N 4 were evaluated by a recycle experiment, which was presented in Figure 12. It could suggest that after four successive cycles, no noticeable diminution was observed within 20 h with TEOA as sacrificing agents, suggesting that the Pt/0.5% Co 3 O 4 /g-C 3 N 4 photocatalyst had an excellent stability and it is reusable for practical applications.
Cycling stability tests of the optimal Pt/0.5% Co3O4/g-C3N4 were evaluated by a recycle experiment, which was presented in Figure 12. It could suggest that after four successive cycles, no noticeable diminution was observed within 20 h with TEOA as sacrificing agents, suggesting that the Pt/0.5% Co3O4/g-C3N4 photocatalyst had an excellent stability and it is reusable for practical applications.

Conclusions
In this paper, g-C3N4 ultra-thin nanosheets were fabricated via the polymerization method. The Co3O4/g-C3N4 series were fabricated by an electrostatic interaction-calcination treatment. The various characterization results indicated that the Co3O4/g-C3N4 catalysts exhibit Z-scheme charge transfer pathway, and possessed excellent photocatalytic hydrogen evolution performance with high stability and fine charge transfer ability. After Pt is added, the Co3O4/g-C3N4 exhibited better activity, and Pt/0.5% Co3O4/g-C3N4 exhibited the best activity, with a hydrogen generation rate of 1620 μmol•g −1 •h −1 . Based on the characterization and environment data, the catalytic reaction charge transfer pathway of Pt/Co3O4/g-C3N4 was proposed. Therefore, the work provided a promising method to design efficient and reusable photocatalytic systems with Z-scheme charge transfer pathways for solar-energy conversion. Further practical applications of this Z-scheme charge transfer pathway were anticipated.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.