Preparation and characterization of p–n heterojunction CuBi₂O₄/CeO₂ and its photocatalytic activities under UVA light irradiation

2.1 Materials and methods

α-Bi₂O₃ (99.99%), CuO (99.99%) and CeO₂ (99.99%) materials were obtained from Aldrich Chemical Company Ltd. Congo red (CR) azoic dye (C.I. 22,020, MW = 696.67 g mol⁻¹, C₃₂H₂₄N₆O₆S₂.2Na, λ_max = 497 nm and pKa = 4) and other chemicals used in the experiments (NH₄OH and H₂SO₄) were purchased from C.I.S.A. Espagne.

2.2 Preparation of p-CuBi₂O₄

The p-CuBi₂O₄ powder was prepared according to the previously reported procedure (Liu et al., 2006; Chen et al., 1999; Takeo et al., 2007). The stoichiometric proportion mixture of α-Bi₂O₃ and CuO oxides was previously ground for a period of time in an agate mortar, and then heated at the rate of 5 °C/min in a muffle oven (Linn High Therm) and thermally treated at 750 °C for 72 h in air. After the muffle oven was naturally cooled to room temperature, the black CuBi₂O₄ powder was ground in the agate mortar and then was collected as the precursor to prepare the CuBi₂O₄/CeO₂ photocatalysts.

2.3 Preparation of CuBi₂O₄/CeO₂ photocatalyst

CuBi₂O₄/CeO₂ nanocomposite photocatalysts were prepared by the solid state technique with the CuBi₂O₄:CeO₂ mass ratio of 5:95, 10:90, 20:80, 30:70, 40:60 and 50:50. The corresponding precursors of CuBi₂O₄/CeO₂ were milled in an agate mortar for 30 min to form the nanosized photocatalysts.

2.4 Characterization

X-ray diffraction patterns of the powders were recorded at room temperature using an automatic D8 Bruker AXS diffractometer with CuKα radiation (λ = 1.5406 Å) over the 2θ collection interval of 10–70° with a scan speed of 10°/min. The mean grain size (d_DRX) was assessed using the Debye–Scherrer equation (Cullity, 1956; Pullar et al., 1988; Azàroff, 1968) as follows Eq. (1):

where β is the corrected full-width at half maximum (FWHM) (radian), λ is the X-ray wavelength (1.5406 Å) and θ is the Bragg angle (radian). The lattice constants of the samples calculated from their corresponding XRD pattern data are obtained by Fullprof program. UV–Vis DRS measurements were carried out at room temperature using a Perkin Elmer Lambda 650 spectrophotometer equipped with an integrating sphere attachment. The analysis range was from 200 to 900 nm, and polytetrafluoroethylene (PTFE, Teflon) was used as a reflectance standard. The band gap values were estimated by extrapolation of the linear part of the plot of absorbance versus the wavelength and E_g = 1240/λ_{Absorp. Edge} equation assuming that all the prepared photocatalysts are direct crystalline semiconductors. Scanning electron microscopy observations (SEM) were performed by using Hitachi S-4800N.

2.5 Photocatalytic study measurements

The photodegradation of CR catalyzed by the CuBi₂O₄/CeO₂ samples was investigated under UV-light irradiation. 100 mg of catalyst was suspended in a CR solution (200 mL, 20 mg/L) in quartz cell tube. The suspension pH value was previously adjusted at 7 using NaOH/H₂SO₄ solutions using (Hanna HI 210) pH meter. Prior to UVA light irradiation, the suspension was stirred with a magnetic stirrer (Speedsafe™ Hanna) for 30 min under dark conditions at 298 K to ensure the establishment of adsorption/desorption equilibrium between the catalyst and CR. The sample was then irradiated at 298 K using 6 W ultraviolet (λ = 365 nm, BLX-E365) photoreactor under continuous stirring. As the reaction proceeded, a 5 mL suspension was taken at 20 min intervals during the catalytic reaction and was centrifuged using centrifuge (EBA-Hettich) at 3500 rpm for 15 min to completely remove photocatalyst particles. The residual RC concentrations during the course of degradation were monitored with a UV mini-1240 Spectrophotometer (Shimadzu UV mini-1240) in the range of 200–800 nm, using 1 cm optical pathway cells.

The effect of initial pH on the photocatalytic degradation of CR was conducted in the pH range of 6–12. The experiments were also performed by varying the amount of CuBi₂O₄ from 0 to 100 wt%.

The data obtained from the photocatalytic degradation experiments were then used to calculate the degradation efficiency η′ (%) of the substrate Eq. (2):

According to Planck’s Law and some further calculation, we can find that the absorption wavelength of the photoreactor can be done by determining its band gap value Eq. (3):

The photocatalytic degradation efficiency of catalyst for the degradation of CR was quantified by measurement of dye apparent first order rate constants under operating parameters.

Surface catalyzed reactions can often be adequately described by a monomolecular Langmuir–Hinshelwood mechanism, in which an adsorbed substrate with fractional surface coverage θ is consumed at an initial rate given as follows Eq. (4) (Vasanth Kumar et al., 2008):

Thus, a plot of reciprocal of the apparent first order rate constant 1/K_app against initial concentration of the dye C₀ should be a straight line with a slope of 1/K₁ and an intercept of 1/K₁K₂. Such analysis allows one to quantify the photocatalytic activity of catalyst through the specific rate constant K₁ (with larger K₁ values corresponding to higher photocatalytic activity) and adsorption equilibrium constant K₂ (K₂ expresses the equilibrium constant for fast adsorption–desorption processes between surface of catalyst and substrates). The integrated form of the above equation (Eq. (5)) yields to the following Eq. (6):

The half-life of dye degradation at various process parameters was raised from Eq. (8).

where half-life time, t_1/2, is defined as the amount of time required for the photocatalytic degradation of 50% of CR dye in an aqueous solution by catalyst.

3.1 XRD analysis of (x wt%)CuBi₂O₄/CeO₂ composites

Fig. 1 shows the XRD patterns of the as-synthesized (30 wt%) CuBi₂O₄/CeO₂ composite in comparison with those of precursor CuBi₂O₄ and pure CeO₂. Diffraction peaks of pure CeO₂ (Fig. 1a) at 2θ of 28.02°, 33.11°,47.45°, and 56.3°can be indexed as the (1 1 1), (2 0 0), (2 2 0), and (3 1 1) planes of pure fluorite phase CeO₂, which is in good agreement with standard value (Fm.3m, JCPDS file No. 34–0394) with lattice constant a = 5.4110 (2) Å. This is in agreement with the reported previous work (Keren, 2011; Truffault, 2010). The diffraction peaks of the Cubi₂O₄ precursor (Fig. 1b) at 2θ of 28.03°, 29.73°, 30.73°, 32.54°, 33.36° and 46.71° were respectively indexed as (2 1 1), (2 2 0), (0 0 2), (1 0 2), (3 1 0), and (4 1 1) planes of pure tetragonal phase of crystalline Cubi₂O₄, according to the Joint Committee Powder Diffraction Standards (P4₂/mnm, JCPDS file No. 42–0334). The lattice constants (a = 8.5004 Å, c = 5.819 Å) were calculated from their corresponding XRD pattern data obtained by Fullprof program. Both precursor CuBi₂O₄ and pure CeO₂ show preferred (0 0 2) crystallographic orientation owing to the preparation route of the sample during the XRD analysis.

Close

Figure 1

XRD patterns of pure CeO₂ (a) precursor CuBi₂O₄ (b) and the synthesized (30 wt%) CuBi₂O₄/CeO₂ (c).

Export to PPT

The crystallite sizes of pure CeO₂ deduced from the XRD patterns by calculation of the Scherrer equation showed that crystalline size of the composite, d_XRD was calculated to 100 nm.

On the other hand, the XRD patterns of (30 wt%) CuBi₂O₄/CeO₂ composite exhibited characteristic diffraction peaks of both Cubi₂O₄ and CeO₂ crystalline phases. It can be seen from Fig. 1c that at 30 wt% mass concentration of Cubi₂O₄, the diffraction pattern of the materials was quite similar to that of pure CeO₂. This is probably due to the high crystallinity of the CeO₂ phase, thus appearing as the dominant peaks in the XRD spectra of the composite sample.

Here, we observe that the XRD patterns (Fig. 2) in the 2θ range from 25° to 40° show that (30 wt%) CuBi₂O₄/CeO₂ sample exhibits broadened peaks with a little shift toward higher intensities. Based on the Scherrer equation, the crystallite size of a sample is inversely proportional to the full-width-half-maximum (FWHM), indicating that a broader peak represents smaller crystallite size (Hu et al., 2006). Thus the presence of Cubi₂O₄ promotes the crystallinity and a consequent broadening of the diffraction peaks of the (30 wt%) CuBi₂O₄/CeO₂ composite sample.

Close

Figure 2

XRD patterns of pure CeO₂ (a) precursor CuBi₂O₄ (b) and the synthesized (30 wt%) CuBi₂O₄/CeO₂ (c) in the 2θ range from 2 ° to 40°.

Export to PPT

3.2 SEM analysis

Fig. 3a, illustrates typical SEM images of CuBi₂O₄ powder synthesized by solid-state reaction of CuO and α-Bi₂O₃ at 750 °C for 24 h, pure CeO₂ and (30 wt%) CuBi₂O₄/CeO₂ composite. It can be seen that, for the CuBi₂O₄, the appearance is a shapeless sheet, and the particle size of the CuBi₂O₄ is about 10–20 μm. Fig. 3b shows typical high-resolution SEM image of CuBi₂O₄ powder to further show the details of the particles. As shown in Fig. 3b, it clearly shows two different crystal shapes on the CuBi₂O₄ surface, corresponding to two different particle sizes of CuBi₂O₄. The appearance of CuBi₂O₄ is a shape sheet and a well-defined tetragonal phase with the crystallite diameter of the CuBi₂O₄ being 5 μm, whereas groups of smaller particles do not have any specific shape with size up to 500 nm tend to cover the bigger particles. However, pure CeO₂ from SEM analysis (Fig. 3c) clearly shows two different spherical-shaped nanoparticle structures on the CeO₂ surface, which can be assigned to CeO₂ with a particle size in the range of 100 nm and Ce₂O₃ with approximately 200 nm dimensions, which agrees with the UV–Vis diffuse reflectance of Ceria (see Fig. 4 in the UV–Vis DRS Spectra and Band Gap Energy section). Both nanoparticles are close to each other in the form of chains. The as synthesized (30 wt%) CuBi₂O₄/CeO₂ composite (Fig. 3d) clearly shows the presence of CeO₂ nanoparticles deposited onto the CuBi₂O₄ surface, displaying a particle size of 100–200 nm and strong assembly of the nanoparticles measuring from 200 nm to 1 μm. Such aggregation can be explained by the solid-state synthesis route, which generally requires repeated mechanical mixing process and a high temperature process, often give rise to particle agglomeration with severe loss of effective surface area.

Close

Figure 3

SEM images of (a) low-resolution of CuBi₂O₄ (b) high-resolution of precursor CuBi₂O₄ (c) pure CeO₂ (d) (30 wt%)CuBi₂O₄/CeO₂ composite.

Export to PPT

Close

Figure 4

UV–Visible absorbance spectra of pure CeO₂ and CuBi₂O₄ synthesized by solid-state reaction.

Export to PPT

3.3 UV–Vis diffuse reflectance spectra and band gap energy

Fig. 4 shows the UV–Vis absorbance spectra of CuBi₂O₄ synthesized by solid-state reaction at 750 °C for 24 h and pure CeO₂. It is clear from the recorded UV–Visible spectrum of CeO₂ that two absorption bands are observed in the UV region at 345 and 245 nm. Generally, the absorption of ceria in the UV region originates from the charge-transfer transition between the O 2p and Ce 4f states in O²⁻ and Ce⁴⁺. This spectral profile indicates that charge-transfer transition of Ce⁴⁺ overlaps with the 4f¹ → 5d¹ transition of Ce³⁺ (Lin et al., 2010a,b). The UV–Visible spectrum of CuBi₂O₄ sample is presented in Fig. 4. It can be seen that it has strong and broad absorption in the range of 200–900 nm. This suggests that the prepared sample absorbs both UV and visible light. Obviously, for CuBi₂O₄ sample, the broad absorption band observed in the UV–Visible region was attributed to the charge-transfer transition between the O 2p and Cu 3dx² − y² states in O²⁻ and Cu²⁺ respectively (Hahn et al., 2012). Fig. 5 shows UV–Vis diffuse reflectance spectra of a series of photocatalysts (x wt%) CuBi₂O₄/CeO₂. From Fig. 5, it can be seen that the absorption wavelength range of the (x wt%) CuBi₂O₄/CeO₂ photocatalysts is extended greatly toward visible light and its absorption intensity is also increased in comparison with pure CeO₂. The red-shift observed in the nanocrystalline CeO₂ would be explained by the formation of localized states within the band gap owing to oxygen vacancies and increase in Ce³⁺ ion concentration (Lu et al., 2009).

Close

Figure 5

UV–Visible absorbance spectra of a series of (x wt%) CuBi₂O₄/CeO₂ composites (x = 0–40 wt%).

Export to PPT

The onset absorption edges and band gap energies of CuBi₂O₄ particle, CeO₂ nanoparticle and (x wt%) CuBi₂O₄/CeO₂ composite are shown in Figs. 4 and 5 respectively. The as-synthesized CuBi₂O₄ exhibits an absorption onset at 900 nm, which corresponds to band gap energy of 1.38 eV. This value is lower than that reported in the literature (1.5 eV) (Arai et al., 2007). It is clear from the recorded spectrum (Fig. 4) that the pure CeO₂ nanocrystalline has two absorption onsets at 390 and 520 nm, which match to band gap energies of 3.18 and 2.38 eV, attributing to CeO₂ dioxide and Ce₂O₃ sesquioxide respectively. These results are in well agreement with values reported in the literature (Xu and Schoonen, 2000; Magesh et al., 2009). The optical properties of the as-synthesized CuBi₂O₄ and pure CeO₂ nanoparticles are reported in Table 1.

It is widely accepted that electronic transport properties depend on the physical and structural characteristics of photocatalyst, such as crystallite size, morphology, phase structure and amount of CuBi₂O₄ loaded (Li et al., 2009; Liu et al., 2009; Yu et al., 2008). As reported from the UV–Vis DRS in Fig. 5 and Table 2, for the series of (x wt%) CuBi₂O₄/CeO₂ composites, the band gap energy decreased from 3.18 to 3.12 eV as the amount of CuBi₂O₄ was increased up to 40% on the CeO₂ matrix, suggesting that the physical preparation of composite powders will result in good particle-to-particle connections, especially in cases where there is a high electrical conductivity (Marunsek, 2009). So, the decrease of the band gap energy with an enhanced absorption intensity of the (30 wt%) CuBi₂O₄/CeO₂ composite upon loading the amount of CuBi₂O could be ascribed to the homogeneous dispersion of CuBi₂O₄ within the CeO₂ matrix in the bulk of the catalyst and the formation of conducting network at very low temperature.

3.4 Photocatalytic activity tests

3.4.1

3.4.1 Effect of pH solution on the photocatalytic activity of (30 wt%) CuBi₂O₄/CeO₂ composite

In order to study the effect of initial pH on the degradation efficiency of (30 wt%) CuBi₂O₄/CeO₂ composite on photodecomposition of CR, experiments were carried out at various pH, ranging from 6 to 12 for avoiding dye aggregation. The results showed that the pH significantly affected the photocatalytic degradation efficiency of both CR. As shown in Fig. 6 and Table 3, for CR, the degradation rate increased from 64.96% to 83.05% as the pH value was increased from 6 to 7, and then decreased to 26.53 at pH = 12. The maximum degradation rate of CR (83.05%) was achieved at pH = 7. For this reason, the pH = 7 was selected for subsequent experiments.

Close

Figure 6

Effect of the pH solution on the photocatalytic redox of CR under UVA-light irradiation ([(30 wt%) CuBi₂O₄/CeO₂] = 0.5 g/L, [CR] = 20 mg/L, T = 298 K, λ_max = 365 nm, I = 90 J/cm² and irradiation time = 100 min).

Export to PPT

Table 3 Results of the effect of the pH solution on the photocatalytic redox of CR under UVA-light irradiation ([(30 wt%) CuBi₂O₄/CeO₂] = 0.5 g/L, [CR] = 20 mg/L,T = 298 K, λ_max = 365 nm, I = 90 J/cm² and irradiation time = 100 min).

pH initial	Adsorption activity η (%)	Photocatalytic activity η′ (%)
2	Dye aggregation
4
6	11.85	64.96
7	17.30	83.05
8	1.73	55.98
10	17.14	29.50
12	9.41	26.54

It is commonly accepted that in photocatalyst/aqueous systems, the potential of the surface charge is determined by the activity of ions (e.g. H⁺ or pH). A convenient index of the tendency of a surface to become either positively or negatively charged as a function of pH is the value of the pH required to give zero net charge (pH_PZC) (Zhang et al., 1998; Yates et al., 1974). pH_PZC is a critical value for determining the sign and magnitude of the net charge carried on the photocatalyst surface during adsorption and the photocatalytic degradation process. Most of the semiconductor oxides are amphoteric in nature, can associate Eq. (15) or dissociate Eq. (17) proton. To explain the relationship between the layer charge density and the adsorption, so-called Model of Surface Complexation (SCM) was developed (Fernandez et al., 2002), which consequently affects the sorption–desorption processes as well as the separation and transfer of the photogenerated electron–hole pairs at the surface of the semiconductor particles. In the 2-pK approach we assume two reactions for surface protonation. The zero point charge pH _PZC for CeO₂ (about 7.5) is approximately identical to that of (30 wt%) CuBi₂O₄/CeO₂ sample, since there is no adsorption of CR ions than the potential determining H⁺/OH⁻ at the surface of CuBi₂O₄ particles. This is often the case for pure (“pristine surface”) oxides in water.

When the pH is lower than the pH _PZC value, the system is said to be “below the PZC.” Below the PZC, the acidic water donates more protons than hydroxide groups, and so the adsorbent surface is positively charged (attracting anions/repelling cations), according to the following reaction Eqs. (9) and (10):

(9)

$$\text{pH}<{\text{pH}}_{\text{PZC}}(30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}_2+{\text{H}}^{+}\to (30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}_2{\text{H}}^{+}$$

(10)

$$(30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}_2{\text{H}}^{+}+{\text{CR}}^{-}\to [{\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}_2{\text{H}}^{+}\text{,}{\text{CR}}^{-}](\text{electrostatic interaction})$$

Conversely, above pH _PZC the surface is negatively charged (attracting cations/repelling anions), given by the following reaction Eqs. (11) and (12):

(11)

$$\text{pH}>{\text{pH}}_{\text{PZC}}(30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}_2+{\text{OH}}^{-}\to (30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}^{-}+{\text{H}}_2\text{O}$$

(12)

$$(30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}^{-}+{\text{CR}}^{-}\to [(30\text{wt}\%){\text{CuBi}}_2{\text{O}}_4/{\text{CeO}}^{-}\text{,}{\text{CR}}^{-}](\text{electrostatic repulsion})$$

CR contains an azo (–N⚌N–) chromophore and an acidic auxochrome (–SO₃H) associated with the benzene structure. CR is also called acidic diazo dye. The pKa value of CR is 4.1, thus CR would be negatively charged at pH range 5.0–10.0 (Ahmad and Kumar, 2010; Zhang et al., 2011). At pH below the pKa value, a dye exists predominantly in the molecular form.

The experimental results revealed that higher degradation rate of CR was observed at pH = 7. Since CR is an anionic dye, its adsorption is mainly performed via electrostatic interactions between the positive (30 wt%) CuBi₂O₄/CeO₂H⁺ surface (pH > pH_PZC) and CR⁻ anionic form (pH > pKa), leading to a maximum value in lower pH_PZC (i.e. pH = 7). Thus, the activity of an adsorbent is due to the presence of sulfonated groups (–SO₃⁻). The presence of the tightly physically bonded or close contact interfaces between the two semiconductors, by which the photoinduced charge transfer from one particle to the other through interfaces spatially is available, can lead to a strong photocatalytic redox of CR over the combined catalysts.

At acidic medium (i.e. pH = 6), higher adsorption extent of CR onto (30 wt%) CuBi₂O₄/CeO₂H⁺ was observed. Such an occurrence could be explained via van der Waals forces, H-bonding and hydrophobic–hydrophobic interactions (Ahmad and Kumar, 2010). Although the electrostatic interaction between the positively charged (30 wt%) CuBi₂O₄/CeO₂H⁺ surface and CR⁻ anionic dye was detected, the photocatalytic activity of the (30 wt%) CuBi₂O₄/CeO₂ catalyst was significantly reduced. This can be explained by the following causes: Assuming most of the reactions take place at the surface of the catalyst, with decreasing pH medium (i.e. pH = 6), Congo red (CR) has a propensity to aggregate in acidic or highly acidic pH ranges. The proposed mechanisms suggest hydrophobic interactions between the aromatic rings of the dye molecules, leading to a π–π stacking phenomenon. CR decolorization has been limited by the available surface area. Moreover, due to this, only fewer photons reach the surface of the photocatalyst. This results in a decrease in concentration of hydroxyl radicals (^•OH) and superoxide (O₂^•−) radicals, thereby decreasing the photocatalytic activity.

At pH higher than pH PZC value (i.e. pH = 10–12), the total surface of the (30 wt%) CuBi₂O₄/CeO₂ catalyst is negatively charged. Hence due to the electrostatic repulsion forces between the negatively charged (30 wt%) CuBi₂O₄/CeO₂ surface and CR anionic dye, mainly sulfonated groups (–SO₃⁻), affecting strongly the accessibility of the surface reducing species to the CR photocatalytic oxidation/reduction kinetics. But appreciable adsorption extent in this pH range suggested strong involvement of physical forces such as hydrogen bonding, van der Waals force, etc. in the adsorption process (Chatterjee et al., 2007). Thus, the observed degradation is primarily taking place in the solution. Further, under alkaline conditions (high concentration of hydroxide ions), more hydroxyl radical (^•OH) formation is possible from the abundant hydroxide ions, which also decline the degradation. There were similar results in the previous reports (Laouedj et al., 2011; Elaziouti et al., 2011; Elaziouti et al., 2012).

3.4.2

3.4.2 Effect of the amount of CuBi₂O₄ on the photocatalytic activity of (x wt%) CuBi₂O₄/CeO₂

The effect of the amount of CuBi₂O₄ on photocatalytic degradation of CR was conducted over a range of catalyst amounts from x = 0 to x = 100 wt%. As observed in Fig. 7 and Table 4, it is evident that the photocatalytic redox of CR greatly depends on the amount of CuBi₂O₄ loaded. The photocatalytic activity increased drastically from 14. 928% to 83.054% as the catalyst amount was raised from x = 0 to x = 30 wt%. On further increase in the CuBi₂O₄ amount beyond x = 30 wt%, the photocatalytic activity decreased gradually, almost reaching 3.13% at x = 100wt%. The highest photocatalytic activity of (x wt%) CuBi₂O₄/CeO₂ (83.054%) under UVA-light irradiation was achieved within 100 min when the amount of CuBi₂O₄ loaded x was 30 wt%, which is obviously about 6 times higher than that of pure CeO₂ and 28 times superior than that of the synthesized CuBi₂O₄.

Close

Figure 7

Effect of the amount of CuBi₂O₄ on the photocatalytic redox of CR under UVA-light irradiation ([Catalyst] = 0.5 g/L, [CR] = 20 mg/L, pH = 7, T = 298 K, λ_max = 365 nm, I = 90 J/cm² and irradiation time = 100 min).

Export to PPT

Table 4 Results of the effect of the amount of CuBi₂O₄ on the photocatalytic redox of CR under UVA- light irradiation ([Catalyst] = 0.5 g/L, [CR] = 20 mg/L, pH = 7, T = 298 K, λ_max = 365 nm, I = 90 J/cm² and irradiation time = 100 min).

Amount of CuBi2O4 x (%)	Adsorption activity η (%)	Photocatalytic activity η′ (%)
0	8.17	14.92
5	20.84	29.33
10	21.515	39.03
20	13.71	60.98
30	17.30	83.05
40	4.024	75.00
50	17.25	53.53
100	0.00	3.13

On the other hand, both CuBi₂O₄ on CeO₂ precursors showed poor adsorption affinity toward organic pollutant among the CuBi₂O₄ loadings. Within the range of CuBi₂O₄ amounts from 0 to 30 wt%, the observed increase in CR decolorization may be due to an increased number of available adsorption and catalytic sites on the surface of (x wt%) CuBi₂O₄/CeO₂ catalyst. So there is an optimum CuBi₂O₄ content for high dispersion morphology of particles CuBi₂O₄ on the CeO₂ surface with high activity.

The effective electron–hole separation both at the physically bonded interfaces and in the two semiconductors as well as charge defect during the physical mixing method was believed to be mainly responsible for the remarkably enhanced photocatalytic activity of (30 wt%) CuBi₂O₄/CeO₂ in the course of the photocatalytic redox conversion of CR.

But until now, there are no reports about synergistic effect between CeO₂ and CuBi₂O₄ in the (30 wt%) CuBi₂O₄/CeO₂ catalyst under UVA-light excitation. From Fig. 7, it is clear that the photocatalytic activity of CeO₂ is drastically increased in the presence of an amount of CuBi₂O₄ (30 wt%) compared to pure CeO₂ and the CuBi₂O₄ samples. These results strongly suggest the existence of a synergistic effect between CeO₂ and the CuBi₂O₄ in the (30 wt%) CuBi₂O₄/CeO₂ catalyst under UVA light excitation.

A further increase in catalyst amount (i.e. >30 wt%), however, may cause an increase in the overlapping of adsorption sites of CeO₂ particles as a result of overcrowding of the CuBi₂O₄ granule owing to the decrease in screening effect and interfering of light. Furthermore, at higher catalyst amount, it is difficult to maintain a homogeneous suspension due to agglomeration of the particles, which decreases the number of active sites. An exception was observed for (50 wt%) CuBi₂O₄/CeO₂ catalyst sample owing to the overestimating value in the experimental data. Thus, results indicate that an optimized catalyst amount (30 wt%) is necessary for enhancing the decolorization efficiency. An analogous trend was reported in the reduction of Cr₂O₇^2- and photocatalytic oxidation of methylene blue orange (MB) using p–n heterojunction photocatalyst CuBi₂O₄/Bi₂WO₆ (Liu et al., 2011).

3.4.3

3.4.3 Effect of UVA-light and catalyst

The photocatalytic activities of all three CuBi₂O₄, CeO₂, (30 wt%) CuBi₂O₄/CeO₂ catalysts were assessed by the photocatalytic redox reaction of Congo red (CR) aqueous solution under UVA-light irradiation. Variations of CR reduced concentration (C/C₀) versus UVA-light irradiation time (t) over different catalysts under different experimental conditions through UV-A alone, UVA/CuBi₂O₄, UVA/CeO₂, (30 wt%) CuBi₂O₄/CeO₂ and UVA/(30 wt%)CuBi₂O₄/CeO₂ are presented in Fig. 8. Results showed that (30 wt%)CuBi₂O₄/CeO₂ sample exhibited higher photocatalytic performance, as compared to the single phases CuBi₂O₄ and CeO₂. The highest efficiency was obtained, under UVA-light irradiation over (30 wt%)CuBi₂O₄/CeO₂, as a result of 83.05% degradation of CR for 100 min of irradiation time. However, the photocatalytic degradation of CR over single phases CuBi₂O₄ and CeO₂ was only 3.13% and 14.92% respectively. When 20 mg/L of CR along with (30 wt%) CuBi₂O₄/CeO₂ was magnetically stirred for the same optimum irradiation time in the absence of light, lower (21.48%) degradation was observed, whereas, disappearance of dye was negligible (0.49%) in the direct photolysis. On the basis of these results, the high decomposition of CR dye in the presence of (30 wt%) CuBi₂O₄/CeO₂ catalyst is exclusively attributed to the photocatalytic reaction of the combined semiconductor particles under UVA-light irradiation. Thus, such an above occurrence in the present experiment is primarily assigned to the charge defect during the physical mixing method, which is advantageous for the effective electron–hole separation and the suppression of the recombination rate of the photogenerated charge carriers, hence result in an improvement of the probability of light-generated carrier transfer via interfaces spatially available. Thus, enhancing the effectiveness of the photocatalytic redox conversion of CR over (30 wt%) CuBi₂O₄/CeO₂ composite under UV light irradiation. A similar result was reported in the heterojunction semiconductor SnO₂/SrNb₂O₆ with an enhanced photocatalytic activity (Liu and Yu, 2008).

Close

Figure 8

Photocatalytic degradation kinetics of CR at different experimental conditions ([Catalyst] = 0.5 g/L, [CR] = 20 mg/L, pH = 7, T = 298 K, λ_max = 365 nm, I = 90 J/cm² and irradiation time = 100 min).

Export to PPT

3.5 Discussion of mechanism

The above analysis shows that the migration direction of the photogenerated charge carrier depends on the band edge positions of the two semiconductors. There are three methods to determine the band edge positions: experiments based on photoelectrochemical techniques, calculation according to the first principle, and predicting theoretically from the absolute (or Mulliken) electronegativity (Kim et al., 1993; Butler and Ginley, 1978; Xu and Schoonen, 2000). The first one is not always easy to handle, and the second one cannot obtain the absolute energy of band edges with respect to vacuum and always has large discrepancies between calculated and measured values. The third one is a simple approach with reasonable results for many oxide photocatalysts (Xu and Schoonen, 2000).

The conduction band edge of a semiconductor at the point of zero charge (pH zpc) can be predicted by Eq. (14):

The as-prepared CuBi₂O₄ is a p-type semiconductor, which always exhibits good stability under UV–Visible illumination, and CeO₂ is determined as an n-type material. Fig. 9 depicts reaction schemes of CuBi₂O₄ (a) and CeO₂ (b) as the p and n type respectively for charge separation for the reductivity/oxidizability improvement model. According to Fig. 9, when the CuBi₂O₄ and CeO₂ photocatalysts are irradiated under UVA (365 nm) light, both catalysts CuBi₂O₄ and CeO₂ can be activated since the band gap energies of CuBi₂O₄ observed in this study were 3.18 and 1.38 eV respectively.

Close

Figure 9

Reaction schemes of CuBi₂O₄ (a) and CeO₂ (b) as the p and n type respectively for charge separation for the reductivity/oxidizability improvement model (electron and hole ).

Export to PPT

For the p-CuBi₂O₄ (Fig. 9a), the electronic potential of the CB of p-CuBi₂O₄ is around −0.44 eV/NHE which is more negative than that of superoxide radical (E₀ (O₂/O₂^•−) = −0.28 V/NHE. This indicated that the electron photoproduced at the CB directly reduced O₂ into O₂^•−. These reduced O₂^•− can subsequently transfer the charge to the species present in the reaction medium that are preferentially adsorbed onto the p-CuBi₂O₄ particles. Hence, the superoxide radical (O₂^•−) reduces the recombination of the charge carriers enhancing the activity in the UVA light. However, the p-CuBi₂O₄ VB of +0.94 eV/NHE, which is too negative than the potential of hydroxyl radical (E₀ (H₂O/^•OH)) = +1.9 V/NHE. The photogenerated holes in the VB of p-CuBi₂O₄ are not able to oxidize H₂O to ^•OH radicals.

p-CuBi₂O₄ powder formed in our laboratory exhibits a black color. The presence of non stoichiometric regions of the nominally p-CuBi₂O₄ particles or small domains of binary oxide phases of Cu_xO or Bi_xO, undetected by XRD data, as unstable impurity phases which could be originated from a number of processes such as reduction of the p-CuBi₂O₄, could be responsible for higher recombination rates. Thus, the result is consistent with the previous study in electrochemical synthesis and characterization of p-CuBi₂O₄ thin film photocathodes (Hahn et al., 2012). Therefore, CuBi₂O₄ alone shows negligible photocatalytic activity under UVA light. As a result, less efficient charge-carrier separation, and thus the increment of photocatalytic activity was restricted.

On the other hand, pure CeO₂ (Fig. 9b) shows little photocatalytic activity under UVA light. Since the VB of CeO₂ is around +2.65 eV/NHE and the CB of CeO₂ is around −0.53 eV/NHE, we expect that photogenerated electrons at the CB of CeO₂ can directly reduce O₂ into superoxide (O₂^•−) because electronic potential of the CB of CeO₂ (−0.53 eV/NHE) is more negative than that of superoxide radical (E₀ (O₂/O₂^•−) = −0.28 V/NHE). In contrast, the CeO_{2 VB} of +2.65 eV/NHE is more positive than that of hydroxyl radical (E₀ (H₂O/^•OH)) = +1.9 V/NHE, indicating that the photogenerated holes in the CeO₂ can oxidize H₂O to ^•OH radicals and CR dye molecule directly forming the organic cation-radicals (R^•+_ads). These (O₂^•−) superoxide and ^(•OH) and organic cation (R^•+_ads) radical species can subsequently transfer the charge to the present in the reaction medium. Thus, the superoxide radical (O₂^•−) suppresses the recombination of the charge carriers enhancing the photocatalytic activity in the UVA light as well. Moreover, the redox potential for one-electron oxidoreduction of cerium Ce⁺⁴/Ce³⁺ (1.3–1.8 V) is more negative than that of CeO₂ VB (+2.65 eV/NHE) and more positive than that of CeO₂ CB (−0.53 eV/NHE). Hence, the photogenerated electrons at the CB and VB of CeO₂ are able to reduce Ce⁺⁴ to Ce³⁺ and can oxidize Ce⁺³ to Ce⁴⁺, respectively, reducing the recombination of the charge carriers.

In a contrast experiment, p-CuBi₂O₄/n-CeO₂ composite exhibits higher activity than phases p-CuBi₂O₄ and n-CeO₂. So we should continue with a further discussion on the mechanism in the photocatalysis. The possible reason for the remarkably enhanced photocatalytic performance of p-CuBi₂O₄/n-CeO₂ in the course of the photocatalytic redox of Congo red can be explained by p–n type heterojunction formation model of the electron–hole separation process under UV light irradiation. The schematic diagram p–n heterojunction formation model is depicted in Fig. 10.

Close

Figure 10

Reaction scheme of CuBi₂O₄/CeO₂ as the p–n type charge separation for the reductivity/oxidizability improvement model (electron and hole ).

Export to PPT

CuBi₂O₄ is a p-type semiconductor, which always exhibits good stability under UV visible illumination, and CeO₂ is determined as an n-type semiconductor. The band gap of p-CuBi₂O₄ was 1.38 eV, which could be excited by photons with wavelengths below 900 nm, whereas n-CeO₂ with band gap is about 3.18 eV which can be excited by photons with wavelengths of 390 nm. So at the interfaces of p-CuBi₂O₄ loaded n-CeO₂ composite, a p–n heterostructure would be formed.

According to the band edge position (Table 1), the electronic potential of the CB of n-CeO₂ is slightly more anodic than that of p-CuBi₂O₄, whereas, the hole potential of the VB of n-CeO₂, is more positive than that of p-CuBi₂O₄. Under UVA (λ_UVA = 355–375 nm → E_hν = 3.30–3.49 eV) light irradiation, the energy of the excitation light was large enough to yield an excited state of both p-CuBi₂O₄ (λ_CuBi2O4 = 900 nm → E_g = 1.38 eV) and n-CeO₂ (λ_CeO2 = 390 nm → E_g = 3.18 eV) semiconductors. A part of the photogenerated charge carriers, free electron (e⁻) and electronic vacancy-a hole (h⁺), recombines in the bulk of the semiconductors, while the rest transfers in the photocatalyst surfaces being partially localized on structural defective centers of its crystalline lattice.

So, when p-type semiconductor CuBi₂O₄ and n-type semiconductor CeO₂ were connected to each other, p–n heterostructure will be formed between p-CuBi₂O₄ and n-CeO₂, and at the equilibrium the inner electric field will be also produced at the same time in the interface. So a number of micro p–n heterostructure CuBi₂O₄/CeO₂ photocatalysts will be formed after doping p-CuBi₂O₄ powder into n-CeO₂ granule. The electron–hole pairs will be created under UVA light illumination. With the effect of the inner electric field, the holes can transfer from n-CeO₂ to p-CuBi₂O₄ easily, because it is thermodynamically more favorable than to go to H₂O (larger driving force). Photogenerated electrons in p-CuBi₂O₄ will recombine with excess holes that are injected from CeO₂. But the electrons cannot move from n-CeO₂ to p-CuBi₂O₄. If electrons are transferred to p-CuBi₂O₄, the photocatalytic activity would decrease because of recombination (Li et al., 2004). Although the transfer of electrons is feasible for the potential between the two CB, it is blocked because of the inner electric field. So the minor carrier in n-CeO_2, which is the control factor of recombination in this n-CeO₂ semi-conductor, can transfer out. In this way, the photoinduced electron (e⁻)–hole (h⁺) pairs in the two semiconductors were effectively separated by the p–n heterostructure formed in the CuBi₂O₄/CeO₂ catalyst and transferred to the semiconductor/substrate interfaces efficiently, thus the probability of electron–hole recombination was reduced. As a result, the net effect of holes in p-CuBi₂O₄ surface acting as powerful oxidants Eqs. (15) and (16). The stepwise photocatalytic mechanism is illustrated below:

The photogenerated electrons in the CB of n-CeO₂ as well as holes in the VB of p-CuBi₂O₄ act as powerful oxidants, respectively. The electrons at the CB of n-CeO₂ react with the adsorbed molecular O_2ads on the p-CuBi₂O₄/n-CeO₂ catalyst sites, reducing it to superoxide anion (^•−_2ads), hydroperoxide (HO_2ads) radicals, hydrogen peroxide (H₂O_2ads) and hydroxide (^•OH_ads) radicals Eqs. (17)–(19), while the holes at the VB of p-CuBi₂O₄ are not able to oxidize the CR dye molecule. These processes could be represented in the following equations:

The superoxide anion (O^•−_2ads) and the hydroxide radicals (^•OH_ads) formed from n-CeO₂ on the illuminated p-CuBi₂O₄/n-CeO₂ catalyst surface are highly effective oxidizing agents in the p-CuBi₂O₄/n-CeO₂ mediated photocatalytic oxidation of Congo red Eq. (20).

The primary reason for the observed photocatalytic activity of the p-CuBi₂O₄/n-CeO₂ composites can be attributed to p-CuBi₂O₄ being less active than n-CeO₂. At 30 wt% p-CuBi₂O₄ loading, the amount of Ce⁺⁴/Ce⁺³ present on the p-CuBi₂O₄/n-CeO₂ composites surface is favorable for faster charge transfer and at the same time allows light to reach the p-CuBi₂O₄/n-CeO₂ surface. A similar trend was reported in the efficient photocatalytic degradation of phenol over Co₃O₄/BiVO₄ composite under Visible Light Irradiation (Mingce et al., 2006).