Effects of La-doping on charge separation behavior of ZnO:GaN for its enhanced photocatalytic performance

Abstract

In this work, lanthanum (La) has been proven as an effective space charge layer modifier to promote efficient photogenerated charge carrier separation for ZnO:GaN solid solution photocatalysts with enhanced photocatalytic water-splitting performance. Photogenerated electron–hole pairs are effectively separated due to the increase of the thickness of the space charge layer after La-doping. The stable photocatalytic activity of the ZnO:GaN solid solution photocatalysts for overall water-splitting was remarkably enhanced by about 3.4 times when the La-doping concentration was 3% in atom ratio. The maximum apparent quantum efficiency of the La-doped (3% in atom percentage) ZnO:GaN solid solution photocatalyst for photocatalytic water oxidation was up to 14.5% at 350 nm without loading of any co-catalyst.

---

1 Introduction

Direct conversion of solar energy to chemical energy via artificial photosynthesis has gained intense attention due to wide application in photodegradation of pollutants,^1,2 organic synthesis,^3,4 H₂O splitting^5,6 and CO₂ reduction.⁷ Overall water-splitting with high efficiency under sunlight using semiconductor photocatalysts is one of the most promising methods to fuel this world.^8–10 Over the past decades, a large number of photocatalysts have been developed successfully for stoichiometric overall water-splitting into H₂ and O₂ in the absence of sacrificial agents.^11–14 Among them, ZnO:GaN solid solution (GZNO) photocatalysts have been demonstrated as one of the most potential photocatalysts for stoichiometric overall water-splitting due to the highest efficiency and steady performance under visible light illumination.^15–17 However, the efficiency is relatively low to meet the industrial requirement mainly due to the ultrafast recombination of photogenerated charge carriers.^17–19 Therefore, preventing the recombination of photogenerated charge carriers is one of the most efficient ways to enhance the photocatalytic efficiency of GZNO photocatalysts.

Various kinds of methods have been developed to improve the separation efficiency of photogenerated charge carriers in semiconductor photocatalysts for enhanced photocatalytic performance. Among them, co-catalyst loading is a common method due to the efficient separation of charge carriers between the photocatalysts and co-catalysts as well as the abundant active sites on the co-catalysts.^20–22 For GZNO photocatalysts, various types of co-catalysts, including single co-catalysts such as Rh–Cr mixed oxides^23,24 and dual co-catalysts,²⁵ have been developed to promote the separation of photogenerated electronic–hole pairs and suppress the back-reaction. Semiconductor combination is another effective method to enhance the photocatalytic activity by overcoming the fast charge recombination through semiconductor–semiconductor heterojunctions.²⁶ Wang and coworkers theoretically presented a GaN/(Ga_1−xZn_x)(N_1−xO_x)/ZnO core–shell solar cell model to improve the visible light absorption ability and carrier collection efficiency.²⁷ Experimentally, photo-generated charge carriers were efficiently separated on a ZnO–(Ga_1−xZn_x)(N_1−xO_x) nanowire-array-on-a-film structure.²⁸ Exotic metal ions can also act as electron (or hole) traps by altering the e⁻/h⁺ pair recombination rate.^29,30 The photoluminescence properties and photocatalytic activity were investigated, respectively, when Mn ref. 31 and In ref. 32 were introduced into ZnO:GaN solid solution photocatalysts. Band bending induced by cation doping also helps to prevent the photogenerated charge carriers from recombination because the gradient potential in the space charge layer shows positive influence on the separation of electrons and holes.³³

The correlations between photo penetration depth (D_p) of incident light and space charge layer thickness (D_s) show great effects on the separation behavior of the photogenerated charge carriers in semiconductors (Scheme 1). When D_p > D_s, only the photogenerated electron–hole pairs in the space charge layer can be separated and those beyond the space charge layer will recombine, which may result in the high recombination rate of the charge carriers and low photocatalytic efficiency (Scheme 1a). However, almost all the photogenerated electron–hole pairs can be separated efficiently when D_p < D_s, leading to the low recombination rate of the charge carriers and high photocatalytic efficiency (Scheme 1b). It can be inferred that the larger D_s could help to promote the higher separation efficiency of the photogenerated carriers under the same incident light illumination, namely, under a certain value of D_p. Therefore, appropriate modification of D_s is a desirable method to reduce the recombination of photogenerated electron–hole pairs in semiconductor photocatalysts for the aim of enhanced photocatalytic efficiency.

Scheme 1 Illustration of the effects of space charge layer thickness (D_s) and photo penetration depth (D_p) of the incident light on the separation behavior of photogenerated carriers in a semiconductor. (a) D_p > D_s, (b) D_p < D_s.

Rare earth dopants have been proven to be effective elements in tuning the depletion layer thickness of TiO₂.³⁴ Among them, La(III) has also been introduced to the depletion layer of Cd_0.6Zn_0.4S photocatalysts for efficient H₂ evolution with a high efficiency of 93.3% at 350 nm.³⁵ La(III) is also used to construct active sites for NiO/NaTaO₃ photocatalysts, which hold the highest apparent quantum yield for overall water-splitting.^36,37 In addition, the La₂O₃ layer has also been demonstrated to be the most efficient modifier among the lanthanoid oxide layers as the co-catalyst for ZnO:GaN solid solution photocatalysts.³⁸

In this work, La(III) was introduced into ZnO:GaN solid solution photocatalysts to modify their space charge layer thickness. The effects of La-doping on the space charge layer thickness are investigated experimentally. The correlations between space charge layer thickness and charge carrier separation efficiency were also discussed. The function of the La dopant was also assessed on the basis of the effects of La-doping on the crystal structure, optical properties and photocatalytic activities. This work provides a new insight to enhance the photogenerated charge carrier efficiency in semiconductor materials for artificial photosynthesis and other uses of solar energy conversion applications.

2 Experimental

2.1 Preparation of photocatalysts

The pristine and La-doped ZnO:GaN solid solution photocatalyst samples were prepared by a high temperature solid state method. Briefly, a mixture of 0.47 g of Ga₂O₃ (99.99%, Aladdin Chemistry Co. Ltd), 0.54 g of ZnO (99.99%, Aladdin Chemistry Co. Ltd) powders and a certain amount of La₂O₃ (99.99%, Aladdin Chemistry Co. Ltd) were nitrided at 1123 K for 15 h under NH₃ flow (~100 mL min⁻¹) after being ground for 30 min in an agate mortar. The amount of La₂O₃ was determined by setting the percentage of La atoms to the total amount of Ga atoms in Ga₂O₃ and Zn atoms in ZnO as different amounts. The samples were collected after being cooled to room temperature under NH₃ flow.

2.2 Characterization

The crystal structure of the samples was characterized using a powder X-ray diffractometer (XRD, Rigaku D/max-2000) with Cu-Kα radiation (λ = 0.15406 nm, 45 kV, 50 mA). The Raman spectra were recorded on a Renishaw inVia micro Raman spectroscopy system with a TE air-cooled 576 × 400 CCD array in a confocal Raman system (wavelength: 633 nm). A field-emission scanning electron microscope (FESEM, FEI, Quanta 200 F) was employed to observe the morphology of the samples. A transmission electron microscope equipped with an EDX system (TEM, FEI Tecnai G²) working at 300 kV was employed to observe the morphology and characterize the chemical composition of the samples. The high angle annular dark field (HAADF) detector forms a Z-contrast image and a scanning transmission electron microscopy (STEM) image operated in the mode. UV-vis absorption spectra ranging from 300 to 600 nm were recorded on a spectrophotometer (TU-1901) using Ba₂SO₄ as reference. The photoluminescence (PL) measurements were carried out at room temperature with a luminescence spectrometer (Perkin-Elmer, LS-55) using 325 nm as the excitation wavelength. Time-resolved PL spectra were recorded on a FluoroMax®-4 fluorescence spectrophotometer from HORIBA Scientific under 301 nm excitation at room temperature. X-ray photoelectron spectroscopy (XPS) measurements were carried out on a K-Alpha XPS spectrometer (PHI 5700 ESCA System), using Al Kα X-ray radiation (1486.6 eV) for excitation. The carbon C 1s line with position at 284.6 eV was used as a reference to correct the charging effect. The X-ray absorption data (XAS) at the Ga K-edge, Zn K-edge and La L₃-edge of the sample were recorded at room temperature in transmission mode using ion chambers or in fluorescence mode with a silicon drift fluorescence detector at beam line BL14W1 of the Shanghai Synchrotron Radiation Facility (SSRF), China. The XAS data were processed using the Athena and Artemis software packages.³⁹ BET specific surface areas were characterized using an AUTOSORB-1-MP surface analyzer. The surface area of the materials was analyzed by the Brunauer–Emmett–Teller (BET) method with a Micromeritics Accelerated Surface Area and Porosimetry System (ASAP) 2020.

2.3 Photocatalytic reactions

The photocatalytic reactions were performed in a vacuum-closed gas-circulation system with a top window. The photocatalyst powder (50 mg) was ultrasonically dispersed for 10 min in ultrapure water (18 MΩ) or in an aqueous solution (100 mL) containing methanol (10 vol%) or AgNO₃ (0.1 mol L⁻¹) for H₂ or O₂ evolution, respectively. For the overall water-splitting reactions, the pH value of the aqueous solution was adjusted to pH 4.5 (adjusted using H₂SO₄). The reaction was carried out at 278 K by irradiating the suspension with light from a 300 W Xe lamp (PLS-SXE300/300UV). Prior to irradiation, the reaction mixture was evacuated using a vacuum pump (Oerlikon Leybold Vacuum, Germany) that was connected to the vacuum-closed gas-circulation system for 30 min in order to remove dissolved gases. The amount of produced gases were measured using a gas chromatograph (Agilent 7890) with a thermal conductivity detector (TCD), taking Ar as the carrier gas.

Apparent quantum yields (AQY) for water oxidation were measured under the same experimental conditions, except for the equipment of band pass filters with different wavelengths. The apparent quantum yields are defined by the following equation:

The number of incident photons was measured by a radiometer with different models for the UV and visible light regions (Photoelectric Instrument Factory, Beijing Normal University).

2.4 Photoelectrochemistry characterization

The photoelectrochemical measurements were carried out using a standard three-electrode cell with a Ag/AgCl (3.0 M KCl) reference electrode and platinum foil as a counter electrode on an Autolab PGSTAT302N (ECO-Chemie) electrochemical workstation. Na₂SO₄ (0.5 M, pH = 4.5, adjusted using H₂SO₄) was used as the electrolyte solution. The working electrode was prepared by the spin coating method on a 20 mm × 30 mm fluorine doped tin oxide (FTO). The Mott–Schottky plots of the samples were determined by applying potentials in the range of −2.0 to 1.0 V versus the Ag/AgCl reference. Photocurrent curves were collected using 0.0 V bias under visible light (λ > 400 nm) irradiation using Na₂SO₄ (0.5 M, pH = 4.5) containing methanol (10 vol%) as the electrolyte solution. AC impedance measurement was carried out using a 5.0 mV ac voltage signal in the 100 KHz to 1 Hz frequency range The obtained potential (vs. Ag/AgCl) was converted to RHE (NHE at pH = 0) using the following equation:

3. Results and discussion

The effects of La-doping on the crystal structure of GZNO are examined by XRD. Both pristine GZNO and 3% La GZNO showed the same hexagonal phase crystal structure (Fig. 1). The peaks are indexed by taking the hexagonal phase GaN as the reference (JCPDS 50-0792). The intensity of all the peaks in the XRD patterns of GZNO decreased remarkably after La-doping, indicating the relatively lower crystallinity of the 3% La GZNO sample (Table S1†). A magnified view of the (001) peak shows a shift to lower angle and further slight shift is observed when increasing the amount of La (Fig. S1†), suggesting that La(III) was inserted into the lattice of GZNO. However, no peaks assigned to La₂O₃ appear in the XRD pattern of La-doped GZNO even with the doping concentration of 5 atom%, which means that La was homogeneously doped into the lattice.

Fig. 1 XRD patterns (a) and Raman spectra (b) of GZNO (red) and 3% La GZNO (black).

The local atomic structure of GZNO before and after La-doping was also analyzed by Raman spectroscopy. The Raman-active modes of A₁(TO) (464 cm⁻¹), E₂ (565 cm⁻¹), and A₁(LO) (730 cm⁻¹) corresponding to the hexagonal wurtzite-type structure are shown in both spectra (Fig. 1b).⁴⁰ No shift of these peak positions and no additional modes are observed in the Raman spectra of 3% La GZNO. Bands assigned to crystalline La₂O₃ (104, 191, and 411 cm⁻¹) are not shown in the Raman spectrum, indicating that La₂O₃ is present probably as a highly dispersed phase in the 3% La GZNO photocatalyst.⁴¹ However, all of these peaks become broader, which suggests that the lattice structure turned to be more disordered after La-doping, in accordance with the results from XRD. The stronger Raman band at 265 cm⁻¹ attributed to the wurtzite disorder also indicates the lower crystallinity of the La-doped sample compared to that of pristine GZNO.^42,43

Fig. 2a presents the TEM image of 3% La GZNO in which no regular shape of the particles is observed. Furthermore, the composition of the particles analyzed by EDX point analysis at the point marked in Fig. 2a reveals that the as-synthesized sample is composed of Ga, Zn, N, O and La (Fig. S2†). Contrast in an HADDF image is greatly sensitive to the atomic number of the elements. As shown Fig. 2b, no La segregation happens in 3% La GZNO as confirmed by the homogeneous contrast observed, suggesting the homonogeous distribution of La. Crystal lattice fringes with a distance of 0.263 nm matched with the (001) plane of GZNO are observed in Fig. 2c. Otherwise, the fuzzy crystal lattice fringes and a large amount of defects are also observed in the HRTEM image, indicating the low crystallinity of 3% La GZNO.

Fig. 2 TEM and HADDF images of 3% La GZNO. (a) TEM image, (b) HAADF (Z contrast) image and (c) HRTEM image.

The chemical states of Ga, Zn, N and O before and after La-doping in GZNO were investigated by XPS. As shown in Fig. 3, the binding energies of Zn 2p_3/2, N 1s and O 1s shift to lower binding energies in 3% La GZNO compared to those of GZNO, while the binding energy of Ga 3d remains the same in both samples. La was also determined in the +3 state in 3% La GZNO by XPS (Fig. S3†). These shifts in binding energy further confirm that La was doped into the lattice of GZNO. The lower binding energies of N 1s and O 1s in 3% La GZNO were mainly due to the lower electronegativity of La compared to that of lattice Zn and Ga in GZNO (Pauling electronegativity: 1.81 for Ga, 1.65 for Zn and 1.1 for La). Furthermore, no shift in the binding energy of Ga 3d is observed indicating that La-doping shows little effects on the chemical environment of the Ga atoms.

Fig. 3 XPS spectra for the as-prepared GZNO (red) and 3% La GZNO (black): (a) Ga 3d, Zn 2p_3/2 and (b) N 1s, O 1s.

EXAFS has also been carried out to investigate the local chemical environment of Ga, Zn and La in 3% La GZNO. For the Fourier transform of the Ga K-edge and Zn K-edge, the EXAFS spectra are shown in Fig. 4a and b; the peaks near 1.9 Å, 3.0 Å and 3.7 Å correspond to the first-, second- and third-shell scattering of Ga and Zn in hexagonal GZNO.^44,45 However, the La L₃-edge EXAFS spectrum displayed in Fig. 4c is remarkably different from those of the Ga K-edge and Zn K-edge EXAFS spectra, indicating the much more different chemical environment of La in the La-doped GZNO. In addition, the La L₃-edge EXAFS spectrum is also different from that of lanthanide oxide which further means that La was not present as lanthanide oxide but was doped into the GZNO lattice. For instance, in La₂O₃, a high intense peak is present around 4 Å due to a shell of twelve La cations at 3.881 Å.⁴⁶ To confirm the coordination geometry and chemical environment of La in the GZNO lattice, a hexagonal structure in which a (Ga,Zn) cation is coordinated with four anions (N,O) to form a (Ga,Zn)(N,O)₄ regular tetrahedron⁴⁷ was used to generate the theoretical EXAFS spectra for the first-shell fitting. In combination with the results from XPS, La was set in the lattice sites to replace the (Ga,Zn) cation in the model. As shown in Fig. 4d, the experimental data fitted well with the hexagonal model, suggesting that La tends to occupy the tetrahedron sites in the doped samples. Otherwise, the much low intensity for the second- and third-shells in the La L₃-edge EXAFS spectrum suggests that the dopant is moving to disordered regions, which is in accordance with the results from XRD and Raman spectroscopy.

Fig. 4 Fourier transforms of the (a) Ga K-edge, (b) Zn K-edge, and (c) La L₃-edge EXAFS spectra and corresponding fit of the first shell of the La L₃-edge (d).

Effects of La-doping on the band structure of GZNO were also investigated. No obvious difference in the absorption edge (about 2.55 eV) as well as in the optical band gap is shown between the pristine GZNO and 3% La GZNO samples (Fig. 5a). Additionally, both pristine GZNO and 3% La GZNO display almost the same edge in the XPS valence band of about 1.81 eV and no additional diffusive electronic states are observed above the valence band edge after La-doping (Fig. 5b). Therefore, it can be inferred that the La dopant shows little influence on the optical bandgap and electronic structure of GZNO.

Fig. 5 (a) UV-vis spectra (the inset image gives the Tauc plots) and (b) VB XPS spectra of pristine GZNO and 3% La GZNO.

The slight change in the conduction band after La-doping was also revealed by the Mott–Schottky plots (Table 1). Meanwhile, the positive slope of the Mott–Schottky plots for the pristine and La-doped samples implies their n-type character (Fig. S4†). However, the value of the slope for the Mott–Schottky plots increases gradually after La-doping. The difference in the value of the slope also indicates the difference in the thickness of the space charge region. According to the MS equation:⁴⁸

the slope of the linear part of the Mott–Schottky plots (1/C²vs. V plot) is expressed aswhere ε represents the dielectric constant of the film, ε₀ = 8.854 × 10⁻¹⁴ F cm⁻¹ is the vacuum permittivity, e = 1.6 × 10⁻¹⁹ C denotes the elementary charge, N_D is the donor density, V_fb is the flatband potential, k_B = 1.3806488 × 10⁻²³ m² kg s⁻² K⁻¹ is the Boltzmann constant, A is the sample area (in this work, A = 4 cm²) and T is the temperature. According to the Schottky approach, the space charge layer thickness can be expressed as^49,50where E is the applied potential. Therefore, by combining eqn (2) and eqn (3), the space charge layer thickness can be calculated through the following equation:

D_s = Aεε₀{k(E − E_fb)}^1/2. (4)

Table 1 Flatband potentials (E_fb) and the values of the slope calculated from the Mott–Schottky plots and corresponding depths of the depletion layer of all samples

Samples	E fb (V vs. NHE)	k (10^11 F^−2 cm^4 V^−1)	D s (nm)
GZNO	−0.25	3.0	9.2
1% La GZNO	−0.23	3.6	9.7
2% La GZNO	−0.22	8.9	14.9
3% La GZNO	−0.24	12.0	18.1
4% La GZNO	−0.21	15.2	19.0
5% La GZNO	−0.23	18.9	22.2

---

In this work, E was zero. ε is set to be 9.5 for both pristine GZNO and 3% La GZNO due to the relatively low concentration of La. The space charge layer thickness was calculated according to eqn (4), and the results are listed in Table 1. As shown in Fig. 6a, the space charge layer thickness is linearly related to the amount of La in La-doped GZNO. This means that the La can act as an effective dopant to modify the D_s of GZNO photocatalysts, which plays a significant role in the separation of the photogenerated electron–hole pairs.

Fig. 6 (a) The correlation of D_s with the inventory of La; (b) PL spectra of all samples; (c) correlation of PL emission intensity (I_p) with D_s; (d) time-resolved PL spectra under 301 nm excitation at room temperature.

Photoluminescence (PL) spectroscopy was carried out to clarify the effects of the increase of D_s on the separation behavior of the photogenerated charge carriers in the photocatalysts. As shown in Fig. 6b, the I_p gradually decreased after increasing the La ratio, which means the efficient photogenerated charge carrier separation in the photogenerated electrons and holes due to the increase of D_s. Taking the value of I_p as a measurement of the photogenerated charge carrier separation efficiency, the correlation between D_s and I_p is intuitively represented. As shown in Fig. 6c, the photogenerated charge carrier separation efficiency increased exponentially along with the D_s. It can also be concluded that adjustment of the D_s is an effective way to promote the separation of charge carriers in GZNO photocatalysts.

Furthermore, time-resolved photoluminescence spectroscopy was applied to better understand the roles of La-doping on the dynamics of the photogenerated carriers (Fig. 6d). A double first-order exponential decay “biexponential” model was used to fit both decay curves as follows:⁵¹

The time-resolved PL decay curves of all the samples can be fitted well into the biexponential function with the shorter decay lifetime attributed to the nonradiative relaxation process relevant to the defects and the longer decay lifetime coming from the radiative process which is relevant to the recombination of photogenerated holes and electrons.^52,53

Average lifetime was calculated using the equation below:

in which B₁ and B₂ represent the relative percentages of each component.

All the average photoluminescence lifetimes of the La-doped samples are prolonged compared to that of the pristine sample (Table 2). It has been intensively demonstrated that efficient separation of charge carriers induced by incident light in semiconductor photocatalysts has significant effects on their photocatalytic performance. Simultaneously, photogenerated charge carriers with longer lifetime on the surface of the catalysts are beneficial for the photocatalytic reactions since the longer lifetime means more chance for the charge carriers to react with the water molecules or other reactants rather than the fast recombination of electrons and holes.

Table 2 Time-resolved PL decay fit parameters of pristine GZNO and La-doped GZNO samples under 301 nm excitation

Samples	Component	Lifetime (ns)	Relative percentage (%)	Average lifetime (ns)
GZNO	τ 1	0.54	56.83	2.52
	τ 2	5.13	43.17
1% La GZNO	τ 1	0.66	58.26	2.70
	τ 2	5.54	41.74
2% La GZNO	τ 1	0.64	55.19	2.78
	τ 2	5.40	44.81
3% La GZNO	τ 1	0.69	57.27	3.08
	τ 2	6.28	42.73
4% La GZNO	τ 1	0.81	55.43	3.47
	τ 2	6.78	44.57
5% La GZNO	τ 1	0.76	54.32	3.49
	τ 2	6.73	45.68

---

The effects of charge carrier separation behavior were also investigated by photoelectrochemical measurements (Fig. 7). The transient photocurrent responses of the catalysts were also investigated under visible light irradiation. Obvious improvement in the photocurrent is observed when the La-doping amount is increased, which indicates that the separation efficiency of the photogenerated electrons and holes is significantly improved in the La-doped GZNO samples (Fig. 7a). Meanwhile, the effect of La dopant on the kinetics of interfacial charge immigration was further investigated by electrochemical impedance spectroscopy (EIS) analysis. According to recent research, a smaller arc radius in EIS Nyquist plots corresponds to a more effective separation of photogenerated electron–hole pairs in semiconductor-based photoelectrodes.^54,55 As shown in Fig. 7b, compared to the pristine GZNO, a significantly decreased diameter in the EIS Nyquist plots is observed for the La-doped GZNO samples, indicating more efficient separation and immigration of the photogenerated electron–hole pairs in La-doped GZNO, which is in good accordance with the results of the PL tests and photocurrent measurements.

Fig. 7 (a) Transient photocurrent responses under visible light irradiation (λ > 400 nm) and (b) EIS Nyquist plots of samples doped with different amounts of La.

The effects of La-doping on the performance of photocatalytic water-splitting of GZNO were further investigated. A maximum performance for photocatalytic water oxidation was obtained when the doping concentration of La was 3% in atom ratio (Fig. 8a), a possible reason of the combined effects of the efficient charge carrier separation and the redox-inactive character of lanthanum compounds in the doped samples.³⁸ As discussed previously, the separation behavior of the charge carriers in semiconductors is sensitive to the correlation between D_p and D_s. The D_p can be adjusted by changing the wavelength of the incident monochromatic light since the penetration depth of the incident light increases when increasing the incident light wavelength to some extent for a certain semiconductor.^56,57 Therefore, it is reasonable to use monochromatic light with a different wavelength to investigate the separation of the charge carriers in the shallow region of the photocatalysts. In this work, the charge carrier separation efficiency was also assessed using the AQY (%) for photocatalytic water oxidation. The dependence of AQY (%) on the incident light wavelength for GZNO and 3% La GZNO was investigated (Fig. 8b). The variation in AQY (%) with respect to the wavelength for both samples exhibits a similar trend to that of the UV-vis spectroscopy results. The 3% La GZNO shows a high AQY (%) of 14.5% for the photocatalytic water oxidation at 325 nm without loading of any co-catalyst. Interestingly, the enhancing effect of AQY (%) after La-doping is greater in the short wavelength region and decreased with the increase of the incident light wavelength (~1.3 times at 325 nm, ~1.1 times at 420 nm and ~1.0 times at 450 nm). These results further confirmed the effects of the increase of the width of the space charge region on the efficient charge carrier separation in the photocatalysts. Namely, when the wavelength of the incident monochromatic light is short, that is D_p is smaller than D_s or D_p is slightly bigger than D_s, the photogenerated electron–hole pairs can be more effectively separated due to the higher gradient potential in the space charge region. When the incident monochromatic light goes to a longer wavelength region whose D_p is far beyond the D_s of the photocatalyst,^58,59 the effects of band bending on the charge separation behavior are negligible and most of the excited electrons/holes tend to recombine, resulting in almost the same apparent quantum yields in the absorption edge for the pristine and La-doped GZNO photocatalysts.

Fig. 8 (a) Photocatalytic water oxidation performance of photocatalysts with different La-doping concentrations (at%) under simulated sunlight using AgNO₃ as sacrificial agent. (b) Wavelength dependence of AQY (%) for GZNO and 3% La GZNO.

Table 3 shows the BET surface areas and photocatalytic activities for the overall water-splitting into H₂ and O₂ using the pristine and doped GZNO. The photoactivity is normalized per unit mass (TOR_m) and per unit surface area (TOR_s), respectively, in this work. The BET surface areas show a slight increase as a possible reason for the formation of the hollow structure as observed in the electron microscopy images after La-doping (Fig. S5†). Phivilay and coworkers found that La₂O₃ worked only as a textual promoter to construct an active structure for photocatalytic water-splitting but not as an electron promoter in a La-doped NiO/NaTaO₃ photocatalyst.³⁷ Interestingly, both the TOR_m and TOR_s for the 3% La GZNO were enhanced about 3.4 times compared to those of pristine GZNO, which means that La₂O₃ not only acts as a textual promoter for the formation of the hollow structure, but also as the electron promoter in the GZNO photocatalysts.

Table 3 Photoactivities of GZNO and 3% La GZNO for overall water-splitting under 300 W Xe lamp illumination

Photocatalyst	BET (m^2 g^−1)	TORm (mol h^−1 g^−1)		TORs (mol h^−1 m^−2)
		H2	O2	H2	O2
GZNO	7.9	18	8.4	2.3	1.1
3% La GZNO	8.2	80	42	9.8	5.1

---

The redox potentials in pristine and La-doped GZNO are efficient to drive both water reduction and oxidation (Fig. S6†). Fig. 9a shows the overall water-splitting over pristine GZNO and 3% La GZNO in pure water. Both samples exhibit capability for stoichiometric overall water-splitting into H₂ and O₂. It can be clearly seen that the amount of gas generated from 3% La GZNO after 8 h was ~3.4 times that for GZNO. When excited by light, the excited charge carriers will be separated efficiently in samples with larger D_s due to the gradient potential in the space charge layer. Then the photogenerated charge carriers transfer to the particles' surfaces where reduction and oxidation reactions of surface species can occur (Fig. S6†). To further investigate the function of La-doping on the half reaction of water-splitting, the performance of H₂ and O₂ evolution in the presence of methanol and silver nitrate as sacrificial reagents, respectively, was further studied. As shown in Fig. 9b, La-doping shows positive effects on both photocatalytic water-splitting for H₂ and O₂ evolution for GZNO. Furthermore, La-doping shows a higher promoting effect for O₂ evolution (~2.3 times) than that for H₂ evolution (~1.6 times). A possible reason for the low H₂ evolution rate is the relatively low reduction capability of the electrons in the conduction band as well as the high overpotential for H₂ evolution on the surface of the catalysts. Otherwise, 3% La GZNO shows stable capability for stoichiometric overall water-splitting into H₂ and O₂ even after 24 h (Fig. 9c).

Fig. 9 Photocatalytic performance of GZNO and 3% La GZNO. (a) Overall water-splitting performance after 8 h of GZNO and 3% La GZNO, (b) water-splitting using methanol and AgNO₃ as electron and hole donors for H₂ and O₂ evolution, respectively, of GZNO and 3% La GZNO, and (c) the time course of overall water-splitting over 3% La GZNO under a 300 W Xe lamp illumination.

Based on the results and discussion above, it can be concluded that the low recombination rate of the photogenerated electron–hole pairs with long lifetime at room temperature is achieved in GZNO photocatalysts when taking La(III) as a modifier of the width of the space charge layer. As a result, the La-doped GZNO photocatalysts showed higher photocatalytic water-splitting performance compared to that of pristine GZNO.

4 Conclusions

In summary, an La dopant has been demonstrated to be an effective modifier for the adjusting of the space charge layer thickness for GZNO photocatalysts in this work. Enhanced photocatalytic water-splitting activity of the GZNO photocatalysts has been achieved after La-doping. The higher charge carrier separation efficiency was demonstrated by photoluminescence (PL) spectroscopy and wavelength dependence of AQY (%). The stable photocatalytic activity for overall water-splitting of GZNO was remarkably enhanced after La-doping. The maximum apparent quantum efficiency for photocatalytic water oxidation was up to 14.5% for La-doped GZNO at 350 nm without loading of any co-catalyst. The enhanced photocatalytic performance was mainly attributed to the enlarged thickness of the space charge layers which plays an important role in the separation of the photogenerated charge carriers. These findings are expected to provide new guidance for the design and preparation of highly efficient photocatalysts by tuning the width of the space charge layers for water-splitting as well as other applications.

Acknowledgements

This work was financially supported by the projects of the National Natural Science Foundation of China (21271055, 21471040), the Fundamental Research Funds for the Central Universities (HIT. IBRSEM. A. 201410) and the Open Project of the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (QAK201304). The authors thank the staff at beamline BL14W1 (Shanghai Synchrotron Radiation Facility) for providing the beam time.

Notes and references

1. W.-J. Ong, S.-Y. Voon, L.-L. Tan, B. T. Goh, S.-T. Yong and S.-P. Chai, Ind. Eng. Chem. Res., 2014, 53, 17333–17344.
2. W.-J. Ong, J.-J. Yeong, L.-L. Tan, B. T. Goh, S.-T. Yong and S.-P. Chai, RSC Adv., 2014, 4, 59676–59685.
3. B. Han, S. Liu, Y.-J. Xu and Z.-R. Tang, RSC Adv., 2015, 5, 16476–16483.
4. L. Yuan, Q. Yu, Y. Zhang and Y.-J. Xu, RSC Adv., 2014, 4, 15264.
5. X. Yu, J. Zhang, Z. Zhao, W. Guo, J. Qiu, X. Mou, A. Li, J. P. Claverie and H. Liu, Nano Energy, 2015, 16, 207–217.
6. Y.-J. Yuan, Z.-T. Yu, Y.-H. Li, H.-W. Lu, X. Chen, W.-G. Tu, Z.-G. Ji and Z.-G. Zou, Appl. Catal., B, 2016, 181, 16–23.
7. W.-J. Ong, L. K. Putri, L.-L. Tan, S.-P. Chai and S.-T. Yong, Appl. Catal., B, 2016, 180, 530–543.
8. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570.
9. J. Ran, J. Zhang, J. Yu, M. Jaroniec and S. Z. Qiao, Chem. Soc. Rev., 2014, 43, 7787–7812.
10. D. Kim, K. K. Sakimoto, D. Hong and P. Yang, Angew. Chem., Int. Ed., 2015, 54, 3259–3266.
11. K. Maeda, T. Takata, M. Hara, N. Saito, Y. Inoue, H. Kobayashi and K. Domen, J. Am. Chem. Soc., 2005, 127, 8286–8287.
12. J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz, S. T. Lee, J. Zhong and Z. Kang, Science, 2015, 347, 970–974.
13. T. Oshima, D. Lu, O. Ishitani and K. Maeda, Angew. Chem., Int. Ed., 2015, 54, 2698–2702.
14. L. Liao, Q. Zhang, Z. Su, Z. Zhao, Y. Wang, Y. Li, X. Lu, D. Wei, G. Feng, Q. Yu, X. Cai, J. Zhao, Z. Ren, H. Fang, F. Robles-Hernandez, S. Baldelli and J. Bao, Nat. Nanotechnol., 2014, 9, 69–73.
15. K. Maeda, K. Teramura and K. Domen, J. Catal., 2008, 254, 198–204.
16. T. Ohno, L. Bai, T. Hisatomi, K. Maeda and K. Domen, J. Am. Chem. Soc., 2012, 134, 8254–8259.
17. Y. Li, L. Zhu, Y. Yang, H. Song, Z. Lou, Y. Guo and Z. Ye, Small, 2015, 11, 871–876.
18. F. E. Osterloh and B. A. Parkinson, MRS Bull., 2011, 36, 17–22.
19. B. Adeli and F. Taghipour, ECS J. Solid State Sci. Technol., 2013, 2, Q118–Q126.
20. J. Yang, D. Wang, H. Han and C. Li, Acc. Chem. Res., 2013, 46, 1900–1909.
21. Y. Zhou, G. Chen, Y. Yu, Z. Han, J. Pei, J. Sun and F. He, RSC Adv., 2014, 4, 29555.
22. A. Xiong, T. Yoshinaga, T. Ikeda, M. Takashima, T. Hisatomi, K. Maeda, T. Setoyama, T. Teranishi and K. Domen, Eur. J. Inorg. Chem., 2014, 2014, 767–772.
23. K. Maeda, K. Teramura, H. Masuda, T. Takata, N. Saito, Y. Inoue and K. Domen, J. Phys. Chem. B, 2006, 110, 13107–13112.
24. T. Hisatomi, K. Maeda, K. Takanabe, J. Kubota and K. Domen, J. Phys. Chem. C, 2009, 113, 21458–21466.
25. K. Maeda, A. Xiong, T. Yoshinaga, T. Ikeda, N. Sakamoto, T. Hisatomi, M. Takashima, D. Lu, M. Kanehara, T. Setoyama, T. Teranishi and K. Domen, Angew. Chem., Int. Ed., 2010, 49, 4096–4099.
26. H. Wang, L. Zhang, Z. Chen, J. Hu, S. Li, Z. Wang, J. Liu and X. Wang, Chem. Soc. Rev., 2014, 43, 5234–5244.
27. Z. Wang, M. Zhao, X. Wang, Y. Xi, X. He, X. Liu and S. Yan, Phys. Chem. Chem. Phys., 2012, 14, 15693.
28. M. Zhong, Y. Ma, P. Oleynikov, K. Domen and J.-J. Delaunay, Energy Environ. Sci., 2014, 7, 1693.
29. T. Takata and K. Domen, J. Phys. Chem. C, 2009, 113, 19386–19388.
30. X. Liu and M. T. Swihart, Chem. Soc. Rev., 2014, 43, 3908–3920.
31. Y. Watanabe and S.-I. Uekusa, Vacuum, 2009, 84, 514–517.
32. K. Kamata, K. Maeda, D. Lu, Y. Kako and K. Domen, Chem. Phys. Lett., 2009, 470, 90–94.
33. Z. Zhang and J. T. Yates, Jr., Chem. Rev., 2012, 112, 5520–5551.
34. A. Xu, J. Catal., 2002, 207, 151–157.
35. Y. G. Yu, G. Chen, L. X. Hao, Y. S. Zhou, Y. Wang, J. Pei, J. X. Sun and Z. H. Han, Chem. Commun., 2013, 49, 10142–10144.
36. H. Kato, K. Asakura and A. Kudo, J. Am. Chem. Soc., 2003, 125, 3082–3089.
37. S. P. Phivilay, A. A. Puretzky, K. Domen and I. E. Wachs, ACS Catal., 2013, 3, 2920–2929.
38. M. Yoshida, K. Maeda, D. Lu, J. Kubota and K. Domen, J. Phys. Chem. C, 2013, 117, 14000–14006.
39. B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537–541.
40. P. J. Pauzauskie, D. Talaga, K. Seo, P. Yang and F. Lagugne-Labarthet, J. Am. Chem. Soc., 2005, 127, 17146–17147.
41. G. Burns and F. Dacol, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 4747–4748.
42. V. B. Ram Boppana, D. J. Doren and R. F. Lobo, J. Mater. Chem., 2010, 20, 9787.
43. V. B. Boppana, H. Schmidt, F. Jiao, D. J. Doren and R. F. Lobo, Chem. – Eur. J., 2011, 17, 12417–12428.
44. H. Chen, W. Wen, Q. Wang, J. C. Hanson, J. T. Muckerman, E. Fujita, A. I. Frenkel and J. A. Rodriguez, J. Phys. Chem. C, 2009, 113, 3650–3659.
45. S. RajaAmbal, A. K. Yadav, S. N. Jha, D. Bhattacharyya and C. S. Gopinath, Phys. Chem. Chem. Phys., 2014, 16, 23654–23662.
46. F. Ali, A. V. Chadwick and M. E. Smith, J. Mater. Chem., 1997, 7, 285–291.
47. M. Yashima, K. Maeda, K. Teramura, T. Takata and K. Domen, Chem. Phys. Lett., 2005, 416, 225–228.
48. R. van de Krol, J. Electrochem. Soc., 1997, 144, 1723.
49. W. Schottky, Z. Phys., 1939, 113, 367–414.
50. W. Schottky, Z. Phys., 1942, 118, 539–592.
51. T. I. Quickenden, T. A. Green and D. Lennon, J. Phys. Chem., 1996, 100, 16801–16807.
52. J. Yuan, Q. Wu, P. Zhang, J. Yao, T. He and Y. Cao, Environ. Sci. Technol., 2012, 46, 2330–2336.
53. Y. Yu, Y. Tang, J. Yuan, Q. Wu, W. Zheng and Y. Cao, J. Phys. Chem. C, 2014, 118, 13545–13551.
54. S. J. Hong, S. Lee, J. S. Jang and J. S. Lee, Energy Environ. Sci., 2011, 4, 1781.
55. Y. Zhou, G. Chen, Y. Yu, Y. Feng, Y. Zheng, F. He and Z. Han, Phys. Chem. Chem. Phys., 2015, 17, 1870–1876.
56. K. Okamoto, I. Niki, A. Shvartser, Y. Narukawa, T. Mukai and A. Scherer, Nat. Mater., 2004, 3, 601–605.
57. M. Holtz, W. M. Duncan, S. Zollner and R. Liu, J. Appl. Phys., 2000, 88, 2523.
58. Z. Li, W. Luo, M. Zhang, J. Feng and Z. Zou, Energy Environ. Sci., 2013, 6, 347–370.
59. I. S. Jeong, J. H. Kim and S. Im, Appl. Phys. Lett., 2003, 83, 2946–2948.

---

Footnote

† Electronic supplementary information (ESI) available: XRD patterns of 1–5% La GZNO, XPS spectra of La 3d5/2 in 3% La GZNO, Mott–Schottky plots of all samples and SEM images of pristine GZNO and 3% La GZNO. See DOI: 10.1039/c5cy01193k

---