Photoelectrochemical Degradation of Organic Pollutants on a La3+ Doped BiFeO3 Perovskite

Towards nonconventional wastewater treatment methods for the degradation of organic pollutants in wastewater, a perovskite-based photoelectrochemical system was developed. Bismuth ferrite doped with lanthanum (La-BiFeO3, La-BFO) perovskite was synthesised through a hydrothermal method with low calcination temperature for the photoelectrochemical degradation of orange II dye and other cocktails of dyes. Photoanodes were prepared by the deposition of the perovskites on a fluorine-doped tin oxide (FTO) substrate. The photoanodes were characterised using XRD, FESEM, FTIR and UV-vis diffuse reflectance. The photoelectrochemical properties of the synthesised photoanodes were investigated with chronoamperometry and electrochemical impedance spectroscopy (including Mott–Schottky analysis). The results show that all La3+-doped BFO photoanodes exhibited a higher absorption edge in the visible light region than the undoped BFO. The photocurrent response of 10% La-BFO (the best performing electrode) exhibited a three times higher current response than the pure BFO. In addition, the electrode exhibited a good degradation efficiency of 84.2% within 120 min with applied bias potential of 2 V at a pH of 7. EIS studies showed a significant enhancement of the interfacial electron transfer of the charge carriers. The enhancements in electrode performances were attributed to the synergistic effect of the applied bias potential and the introduction of La3+ into the BFO matrix. This study therefore shows that the photoelectrocatalytic performance of BFO for water treatment can be improved by the introduction of perovskites-doping ions such as La3+.


Introduction
Water pollution is a global issue, partly because of the discharge of effluents into the environment from industries such as pharmaceutical, textiles and metallurgical. Over the years, several methods have been employed in the treatment of water. However, most of these conventional methods are associated with drawbacks, such as the generation of secondary pollution, incomplete mineralisation and, in some cases, cost intensiveness. Photoelectrochemical oxidation is one of the more advanced oxidation processes, which uses both light and applied potential to facilitate the production of strong oxidants that nonselectively oxidise organics in wastewater [1]. The photoelectrochemical (PEC) degradation of organic pollutants using different semiconducting materials has attracted considerable interest over the past few years owing to its advantages over photocatalysis. Photocatalysis (PC), which has been widely utilised in non-conventional water treatment, is found to be associated with drawbacks such as difficulty in catalyst recovery and electron-hole recombination [2]. However, in PEC, as photons illuminate the catalyst, electrons are migrated to the cathode via the applied bias potential for efficient separation of photogenerated charges to combat electron-hole recombination.
Semiconductors, such as ZnO, TiO 2 , and WO 3 , have been utilised as photocatalyts in PEC degradation. However, their photocatalytic efficiencies are limited by their large band gaps, which inhibit their absorption of visible light [3,4]. Therefore, research has been geared towards other catalysts that are visible light-active with minimal electron-hole recombinations and narrow band gaps. In addition, improved techniques, such as metal doping [5,6], metal decoration [7], and heterojunction structure formation [8,9], have been explored in the preparation of the semiconductors to harness their efficiency in water treatment. Perovskites, which are binary metal oxides with a formula ABO 3 , have been sought-after materials owing to their interesting properties, such as thermal stability, low band gap, visible light activity, ionic conductivity and excellent electron mobility [10,11]. Bismuth ferrite (BiFeO 3 , BFO) is an established multiferroic material that exhibits both ferroelectric and antiferromagnetic properties at room temperature. BFO has found wide applicability in memory devices, data storage, sensors, water splitting and photocatalysis, owing to its narrow band gap (2.2-2.8 eV), chemical stability and low cost [12][13][14]. With regards to its application in water treatment, Gao et al. [15] synthesised BFO for the degradation of methyl red via the sol-gel method involving different calcination temperatures. They recorded well-dispersed BFO nanoparticles with tunable sizes and bandgaps. Additionally, Wang et al. [16] studied the effects of BFO for the degradation of dye and phenols. Recently, studies have shown that the photocatalytic performance of pure BFO is not impressive due to the recombination of photogenerated electron-hole pairs, high current leakage and impurity phases formed during preparation [12]. The poor photocatalytic performance of pure BFO has led to the preparation of heterojunction structures of BFO in photocatalysis, such as BFO/BVO [8], N-rGO/BiFeO 3 [17], BiFeO 3 /CuO [18], BiFeO 3 /Pt [19], BiFeO 3 /Sm/Pd [12], BiFeO 3 /La 3+ [20] BiFeO 3 /Zr [5] and BiFeO 3 -gC 3 N 4 -WO 3 [9] for enhancements in the degradation of organics in water. Many studies have also shown that the partial substitution of A site Bi 3+ in BFO with rare earth metals, such as La 3+ , Nd 3+ , Eu 3+ and Gd 3+ , can lead to enhanced ferroelectric and magnetisation properties, which leads to a magnetoelectric (ME) effect via the sinking of impurity phases and the disparity in ionic radii, since lanthanides have smaller ionic radii compared to Bi 3+ [21][22][23].
Meng et al. [24] synthesised La-doped BFO for the photocatalysis degradation of organic pollutants in water. They reported that La-doped BFO photocatalysts manifested greater photocatalytic enhancement, which they attributed to reductions in electron-hole recombination, narrow band gap, low charge transfer resistance, and the smaller radius of La 3+ (RLa 3+ = 0.116 nm) than Bi 3+ (RBi 3+ = 0.117 nm). Singh and co-workers [20] also reported the substitution of La 3+ in BFO for the photocatalysis degradation of methylene blue. They recorded enhancements in photocatalytic degradation efficiency, increased surface area, and improved ferromagnetic behaviour. However, the authors did not report on the performance of the bare BFO photocatalyst. Dhanalakshmi [25] demonstrated the enhanced photocatalytic degradation of phenol red using La 3+ -doped BiFeO 3 .
Reports on the photoelectrocatalytic degradation of recalcitrant organics using perovskites are sparse, and to the best of our knowledge, the photoelectrocatalytic application of La-doped BFO has not been reported. We thus report the effects of Bi 1-x La x FeO 3 (mole fraction x = 0.00, 0.05, 0.10, 0.15) prepared through the hydrothermal method with a low calcination temperature. The perovskite was prepared through the hydrothermal method with a low calcination temperature. The powdered perovskite was deposited onto an FTO substrate using the drop-cast technique. The photoanodes were characterised with XRD, FE-SEM/ EDS and UV diffusive reflectance spectroscopy (UV-DRS). Chronoamperometry for photocurrent response and electrochemical impedance spectroscopy (EIS) were employed to study the electrochemical properties of BFO and La-BFO photoanodes. Additionally, the influence of lanthanum on bismuth ferrite was also studied using Mott-Schottky plots to understand the charge separation efficiency and suppression of recombination of charged carriers, and a mechanism for improved charge separation was proposed. For the photoelectrocatalytic degradation experiments, Orange II, Congo red and methylene blue were selected as model organic dyes, and acetaminophen and sulfamethoxazole were selected as model pharmaceutical pollutants (See chemical structures of all the selected organic pollutants in Supplementary Figure S1).

Structural and Morphology Characterisation of the Electrodes
We utilised the powdered XRD technique to analyse the crystal structure of FTO, BFO, 5% La-BFO, 10% La-BFO and 15% La-BFO. All the major reflection peaks in the XRD pattern are indexed to perovskite BFO with space group R3c confirmed by JCPDS card no. 01-084-7216 with a rhombohedral structure [26]. The peaks around 28 • and 30.3 • in Figure 1a correspond to the presence of secondary phase Bi 25 FeO 40 [24]. This could be due to the kinetics of formation and the metastable nature of BiFeO 3 . These impurities disappeared as the lanthanum doping increased, confirming the stability of lanthanum and its ability to suppress impurities [20,24]. The enlarged view of diffraction peaks (104) and (110) shows a split as the lanthanum doping increases, and these splits broadened slightly to a higher angle, implying that substituting La 3+ into Bi 3+ brings about distortions in the lattice structure (Figure 1a insert). These results suggest the effect of doping lanthanum into the bismuth ferrite lattice, and they agree with the findings of the Singh and Weng groups [20,24]. The field emission scanning electron microscope (FESEM) image of BFO shows an irregular mixture of rods and globules, with some falling into the nanoscale. The effect of La 3+ doping could not be observed from the FESEM; however, the EDS image confirms the presence of lanthanum in the 10% La-BFO composite (Figure 1c). The FTIR spectra of Bi 1-x La x FeO 3 (x = 0.00, 0.05, 0.10, 0.15) are shown in Figure 1d. The pronounced absorptive peaks around 400-600 cm −1 are the characteristics of metal oxide formation in all as-prepared samples. The peaks at 434 and 547 cm −1 could be the vibration along the Fe-O axis and O-Fe-O bending, which are both attributes of the FeO 6 group in perovskites [21,27]. The peak at 632 cm −1 disappeared as lanthanum was incorporated into the matrix. The peaks around 695 cm −1 and 857 cm −1 could be a result of the vibration of Bi-O bonds [28]. These peaks confirm the formation of the perovskite structure in all the prepared samples.

UV-VIS Diffuse Reflectance Spectroscopic Characterisation of the Photoanodes
The optical properties of the as-prepared BFO, 5% La-BFO, 10% La-BFO and 15% La-BFO photoanodes were studied using UV-visible diffuse reflectance spectroscopy, and the results are shown in Figure 2a. All the materials show absorbance in the visible range. The effect of La doping can be seen by the increase in the absorbance in the visible range of 550 to 800 nm, with the 10% La doping giving the highest absorbance. The Tauc plot equation was then used to extract information with regards to the band gap energy of the semiconductors. The band gap energy was calculated using the Kubelka-Munk (K-M) where α, h, υ, Eg, and A are the absorption coefficient, Planck constant, light frequency, band gap energy, and absorption constant, respectively. The parameter "n" in the equation depends on the characteristics of the transition in any semiconductor, i.e., direct transition (n = 1) or indirect transition (n = 2). Bismuth ferrite has a direct band gap, hence n = 1 [29]. The plot of (αhυ) 2 against hυ in Figure 2b shows that the band gaps of pure BFO, 5%, 10% and 15% La-BFO were 2.4, 2.37, 2.28 and 2.33 eV, respectively. This band gap engineering may be attributed to the formation of an impurity level or the formation of localised electronic states due to the incorporation of La 3+ , which brought about a shift in the Fermi level. The 10% La-BFO photoanode absorbed more visible light than other photoanodes owing to the reduced band gap of 2.28 eV in comparison to the other materials. This will likely favour the photoelectrocatalytic degradation of the water pollutants [20,30,31]. Additionally, these band gaps are within the range of the previously reported studies of BFO [18,32].

UV-VIS Diffuse Reflectance Spectroscopic Characterisation of the Photoanodes
The optical properties of the as-prepared BFO, 5% La-BFO, 10% La-BFO and 15% La-BFO photoanodes were studied using UV-visible diffuse reflectance spectroscopy, and the results are shown in Figure 2a. All the materials show absorbance in the visible range. The effect of La doping can be seen by the increase in the absorbance in the visible range of 550 to 800 nm, with the 10% La doping giving the highest absorbance. The Tauc plot equation was then used to extract information with regards to the band gap energy of the semiconductors. The band gap energy was calculated using the Kubelka-Munk (K-M) Equation (1) where α, h, υ, Eg, and A are the absorption coefficient, Planck constant, light frequency, band gap energy, and absorption constant, respectively. The parameter "n" in the equation depends on the characteristics of the transition in any semiconductor, i.e., direct transition (n = 1) or indirect transition (n = 2). Bismuth ferrite has a direct band gap, hence n = 1 [29]. The plot of (αhυ) 2 against hυ in Figure 2b shows that the band gaps of pure BFO, 5%, 10% and 15% La-BFO were 2.4, 2.37, 2.28 and 2.33 eV, respectively. This band gap engineering may be attributed to the formation of an impurity level or the formation of localised electronic states due to the incorporation of La 3+ , which brought about a shift in the Fermi level. The 10% La-BFO photoanode absorbed more visible light than other photoanodes owing to the reduced band gap of 2.28 eV in comparison to the other materials. This will likely favour the photoelectrocatalytic degradation of the water pollutants [20,30,31]. Additionally, these band gaps are within the range of the previously reported studies of BFO [18,32].

Electrochemical and Photoelectrochemical Characterisation
The photocurrent response of BFO and 10% La-BFO deposited on the FTO substrate was investigated using chronoamperometry in 0.1 M Na2SO4 solution with a bias potential of 0.6 V. Figure 3a shows the photocurrent response with respect to time under the

Electrochemical and Photoelectrochemical Characterisation
The photocurrent response of BFO and 10% La-BFO deposited on the FTO substrate was investigated using chronoamperometry in 0.1 M Na 2 SO 4 solution with a bias potential of 0.6 V. Figure 3a shows the photocurrent response with respect to time under the ON/OFF chopped light illumination mode for BFO and 10% La-BFO. The photocurrents generated show that the electrode is responsive to photoelectrochemical stimulus, hence its suitability as a visible light-responsive photoanode [33]. Moreover, the 10% La-BFO displayed a greater photoelectrochemical response, which is three times higher (0.1180 mAcm −2 ) than the pure BFO (0.0372 mAcm −2 ). This is an indication that 10% La-BFO will promote greater current mobility and reduction when in recombination with photogenerated electron-hole pairs. This equally shows the positive effect of the La 3+ ion in the BFO perovskite matrix for photocatalytic applications. To further understand the charge transfer properties of BFO and 10% La-BFO samples, EIS experiments were performed. The semi-circular arc in the Nyquist plot in Figure 3b denotes the charge transfer resistance (Rct) at the interface of the electrodes. The BFO has an Rct of 4500 Ω, while the incorporation of La 3+ ion into BFO markedly reduced the Rct to 406 Ω, suggesting a greater interfacial charge transfer efficiency, the low recombination of photogenerated electron-hole pairs, and the greater mobility and diffusion of electrons [2]. The bode plot (Figure 3b inset) was used to calculate the lifetime (τ) of the electrons generated in the photoanodes of the BFO and 10% La-BFO composites from Equation (2). The lifetime is inversely proportional to the peak frequency, as seen in Equation (2) [34].
where f max is the maximum frequency. The calculated electron lifetimes (τ) for BFO and 10% La-BFO are 21.67 ms and 117.66 ms, respectively. These results show that an appropriate doping amount of lanthanum in the bismuth ferrite lattice increased the electron lifetime in the photoanode. Mott-Schottky analysis was carried out to estimate the donor concentrations in the BFO and 10% La-BFO films [35]. The flat band potential V FB of BFO and 10% La-BFO was estimated by the Mott-Schottky (MS) measurement using Equation (3).
where C is the semiconductor capacitance, e is elementary charge, ε is the dielectric constant of the semiconductor (assumed to be 52) [36], ε o is the permittivity of the vacuum 8.85 × 10 −12 Fm −1 , N D is the donor density, E app is the applied potential, K is the Boltzmann constant 1.38 × 10 −23 Fm −1 , E FB is the flat band potential, and T is the absolute temperature [2]. From the slope of the plot of 1/C 2 against E (potential), the N D is estimated, while the E FB is estimated from the intercept on the x-axis [34]. Figure 3c shows the MS plot for BFO and 10% La-BFO. The negative slope depicts that both samples are p-type semiconductors. The V FB estimated from Equation (3) for BFO is −0.042 V (vs. Ag/AgCl), and −0.172 V (vs. Ag/AgCl) for 10% La-BFO. This reduction in V FB value shows a downward shift in the Fermi level of the 10% La-BFO electrode, which leads to the inhibition of the recombination of photogenerated electron-hole pairs [8,37]. The presence of La in the matrix causes excitation at lower energy and a favourable pathway for the charge transfer from the CB of BFO to the CB of La, thus increasing the charge carrier concentration by hindering the recombination. These electrons, which are trapped within the lanthanum, enable holes to fully participate in the oxidation reaction. As estimated from the slope, the carrier density (N D ) values for BFO and 10% La-BFO are 6.564 × 10 22 cm −3 and 1.861 × 10 23 cm −3 , respectively. We also observed an enhancement in the carrier density of the doped electrode.
tion of photogenerated electron-hole pairs [8,37]. The presence of La in the matrix causes excitation at lower energy and a favourable pathway for the charge transfer from the CB of BFO to the CB of La, thus increasing the charge carrier concentration by hindering the recombination. These electrons, which are trapped within the lanthanum, enable holes to fully participate in the oxidation reaction. As estimated from the slope, the carrier density (ND) values for BFO and 10% La-BFO are 6.564 × 10 22 cm −3 and 1.861 × 10 23 cm −3 , respectively. We also observed an enhancement in the carrier density of the doped electrode.

Photoelectrocatalytic Degradation of Pollutants
The photoelectrocatalytic performances of the samples BFO, 5% La-BFO, 10% La-BFO and 15% La-BFO photoanodes were evaluated using Orange II dye as a typical organic pollutant. The degradation process was performed under simulated solar light as described in Section 3.3 As seen in Figure 4a, the degradation percentages of Orange II dye at BFO and 10% La-BFO photoanodes were 55.3% and 84.2%, respectively, while those of the 5% La-BFO and 15% La-BFO photoanodes were 52.6% and 57.8%, respectively, as shown in the insert. These results show that the 10% La-BFO is the best performing pho-

Photoelectrocatalytic Degradation of Pollutants
The photoelectrocatalytic performances of the samples BFO, 5% La-BFO, 10% La-BFO and 15% La-BFO photoanodes were evaluated using Orange II dye as a typical organic pollutant. The degradation process was performed under simulated solar light as described in Section 3.3 As seen in Figure 4a, the degradation percentages of Orange II dye at BFO and 10% La-BFO photoanodes were 55.3% and 84.2%, respectively, while those of the 5% La-BFO and 15% La-BFO photoanodes were 52.6% and 57.8%, respectively, as shown in the insert. These results show that the 10% La-BFO is the best performing photoanode for the degradation process. This improved performance can be ascribed to the narrow band gap and excellent charge separation, which reduced the recombination of electron-hole pairs. The results of the kinetic studies fit a pseudo-first order kinetic model, and Figure 4b shows lnC 0 /C t plotted against reaction time. The apparent rate constants for BFO and 10% La-BFO were 6.73 × 10 −3 min −1 and 15.4 × 10 −3 min −1 , respectively. The rate constant of 10% La-BFO shows that the abatement of Orange II was fastest with the 10% La-BFO electrode. To further show the performance of the photoelectrocatalytic process using the optimum photoanode (10% La-BFO), the degradation efficiencies obtained through photocatalytic (PC) and electrocatalytic oxidation (EC) were studied and presented in Figure 4c,d. From Figure 4c, we recorded degradation percentages of 5.7%, 60.6% and 84.2% for PC, EC and PEC, respectively. The corresponding rate constants were 0.00094 min −1 , 0.0075 min −1 and 0.0154 min −1 for PC, EC and PEC. The application of bias potential increases the charge carrier separation, and consequently improves the catalytic degradation in the presence of light (photoelectrochemical oxidation or photoelectrocatalysis) at the 10% La-BFO photoanode. This shows the advantage of PEC over PC and EC. For a better understanding of the performance of PEC in the application of bias potential and light, the degree of electrochemical enhancement (E) and the degree of process synergy (S) were calculated using Equations (4) and (5), respectively [38].
where K PEC , K PC and K EC are the apparent rate constants for the photoelectrocatalytic, photocatalytic and electrochemical degradation of Orange II dye, respectively. From Equation (4), we see that the degree of electrochemical enhancement was 0.938 (93.8%). This result suggests that the application of bias potential contributed immensely to the process. The degree of process synergy as calculated from Equation (5) was 0.451, which is greater than zero. This means that the performance of PEC degradation alone is more than the summation of the rate constants of PC and EC [39,40].
Catalysts 2021, 11, x FOR PEER REVIEW 9 of 17 through photocatalytic (PC) and electrocatalytic oxidation (EC) were studied and presented in Figure 4c,d. From Figure 4c, we recorded degradation percentages of 5.7%, 60.6% and 84.2% for PC, EC and PEC, respectively. The corresponding rate constants were 0.00094 min −1 , 0.0075 min −1 and 0.0154 min −1 for PC, EC and PEC. The application of bias potential increases the charge carrier separation, and consequently improves the catalytic degradation in the presence of light (photoelectrochemical oxidation or photoelectrocatalysis) at the 10% La-BFO photoanode. This shows the advantage of PEC over PC and EC. For a better understanding of the performance of PEC in the application of bias potential and light, the degree of electrochemical enhancement (E) and the degree of process synergy (S) were calculated using Equations (4) and (5), respectively [38].
where KPEC, KPC and KEC are the apparent rate constants for the photoelectrocatalytic, photocatalytic and electrochemical degradation of Orange II dye, respectively. From Equation (4), we see that the degree of electrochemical enhancement was 0.938 (93.8%). This result suggests that the application of bias potential contributed immensely to the process. The degree of process synergy as calculated from Equation (5) was 0.451, which is greater than zero. This means that the performance of PEC degradation alone is more than the summation of the rate constants of PC and EC [39,40]. The pH value of the solution is one of the most important aspects that affects the decolourisation of azo dyes in the oxidation process. The PEC degradation of Orange II dye at pH 4, 7 and 10 resulted in degradation percentages of 56.7, 84.2 and 43.3%, respectively (Supplementary Figure S2). This result suggests that acidic medium and neutral pH favour the decolourisation of Orange II dye. This may be attributed to the protonated hydroxyl group developed on the surface, which causes more electrostatic interaction between Orange II (anionic dye) and the semiconductor [41]. Zhang [42] reported a rapid decolourisation of acid Orange II at acidic and neutral pH using the adsorption process.
The reusability and stability study of the 10% La-BFO photoanode was evaluated by replicating Orange II degradation under visible light for up to five cycles. After each cycle, the electrode was rinsed with deionised water. As shown in Supplementary Figure S3, the degradation rates are 84.2%, 84%, 82.8%, 82.2%, and 80%. These close values indicate a good measure of stability, and suggest the reusability of the photoanode without significant losses in catalytic performance.
(a) The pH value of the solution is one of the most important aspects that affects the decolourisation of azo dyes in the oxidation process. The PEC degradation of Orange II dye at pH 4, 7 and 10 resulted in degradation percentages of 56.7, 84.2 and 43.3%, respectively (Supplementary Figure S2). This result suggests that acidic medium and neutral pH favour the decolourisation of Orange II dye. This may be attributed to the protonated hydroxyl group developed on the surface, which causes more electrostatic interaction between Orange II (anionic dye) and the semiconductor [41]. Zhang [42] reported a rapid decolourisation of acid Orange II at acidic and neutral pH using the adsorption process.
The reusability and stability study of the 10% La-BFO photoanode was evaluated by replicating Orange II degradation under visible light for up to five cycles. After each cycle, the electrode was rinsed with deionised water. As shown in Supplementary Figure S3, the degradation rates are 84.2%, 84%, 82.8%, 82.2%, and 80%. These close values indicate a good measure of stability, and suggest the reusability of the photoanode without significant losses in catalytic performance.

Photoelectrocatalytic Degradation of Cocktail Pollutants and Pharmaceuticals
The 10% La-BFO was further used to simultaneously degrade 5 ppm each of Orange II, Congo red and methylene blue. This route serves as a guide to elucidate the performance of the photoanode in the degradation of myriad pollutants. There was considerable degradation of the pollutants as a result of the peak intensity reduction, as illustrated by the UV-vis spectrophotometer shown in Supplementary Figure S4. The degradation percentage of the dye mixture was calculated to be 44.8, 43.9 and 56% for Orange II, Congo red and methylene blue, respectively. The individual degradations of Congo red and methylene blue were calculated to be 76% and 93.7%, respectively. As expected, the rate of degradation in the cocktail was lower as compared to the individual dyes, owing to the presence of (i) higher concentrations of organic molecules, (ii) the increased complexity of the molecules that must be oxidised, and (iii) the matrix effect created by the presence of other organics in the solution. Furthermore, we applied the photoanode in the degradation of sulfamethoxazole and acetaminophen (data not shown) with degradation values of 36.8% and 54%, respectively. These results showed the capability of the photoanode in a competitive environment of numerous recalcitrant pollutants.

Trapping Experiment and Proposed Photoelectrocatalytic Activity Mechanism
A scavenger experiment was carried out to show the role of active species in the PEC degradation of Orange II and other pollutants used in this study. From the result shown in Figure S5, 0.5 mM of ethylenediaminetetraacetate (EDTA) salt and 0.2 mM of t-butanol (t-BuOH) was used to suppress the effects of hole and hydroxyl radical, respectively. The degradation percentage was reduced to 20% and 10% upon the addition of EDTA and t-BuOH, respectively. This shows that hydroxyl radical (·OH) and hole (h + ) played synergistic roles in degrading the pollutants in this study. Table S1 shows the percentage degradation of Orange II dye in pH, stability and scavenger studies. Furthermore, to understand the separation of the photogenerated electron-holes in the composite over the pristine BFO, it is imperative to determine the band gap edge energy of the valence band and conduction band. The potentials at the point zero charge of the conduction band (CB) and valence band (VB) are estimated using Mulliken electronegativity in Equations (6) and (7) [43].
E VB = E g + E CB (7) where E CB and E VB are the band potentials, X is the electronegativity of the semiconductor, and x is the electronegativity of the elements using the formula [X = [x(A) a x(B) b x(C) c ] 1/a+b+c ; therefore, X is the electronegativity of BFO, calculated to be 5.89. E e is the energy of the free electron on the hydrogen scale (4.5 eV), and E g is the band gap of the semiconductor obtained from Equation (1). Via the above equation, E VB and E CB for BFO were calculated to be 2.59 eV and 0.19 eV, respectively. As illustrated in the scheme in Figure 5, when solar light irradiates on the La-BFO electrode, it produces electrons e − and holes h + (Equation (8)). Since the conduction band edge of BFO (0.19 eV) vs. NHE is more positive than the standard redox potential E 0 (O 2 /O 2 − ) (−0.33 eV vs. NHE), the electrons at the CB of BFO cannot reduce O 2 to O 2 ·− . Thus, the superoxide radical is not the active species. Furthermore, the edge of the CB potential of BFO is more negative than the standard redox potential E 0 (O 2 /H 2 O 2 ) (0.685 eV vs. NHE), implying that oxygen adsorbed on the surface of the 10% La-BFO composite could react with electrons to form H 2 O 2 , which will further react with electron to form hydroxyl radicals (·OH) (Equations (9) and (10)). Hydroxyl radical has strong oxidising potential, enabling it to participate in photocatalytic reactions. The VB edge potential of BFO (2.59 eV vs. NHE) is more positive than the standard redox potential E 0 (OH/OH − ) (1.99 eV vs. NHE), suggesting that the accumulated holes on the VB of the composite can oxidise OH ads − to form ·OH [44,45]. Via this mechanism, the hole can directly attack the dye to mineralise it, or can adsorb water to form a hydroxyl radical (Equations (11) and (12)). The doping of metals expands the wavelength of a semiconductor to absorb more visible light by creating localised electronic states, which results in a shift of the band gap and also alters the band edge by shifting the Fermi level in the doped system [24,46]. The doping of a rare earth element (La) into bismuth ferrite (BFO) in a similar way showed higher absorbance in the visible region, giving rise to a smaller band gap. According to the Mott-Schottky plot, BFO exhibited a p-type semiconductor, which means that the localised energy level is above the VB edge. The possible mechanisms that led to the suppression of photogenerated electron-hole pairs could have involved the lanthanum acting as an electron-trapping agent and an impurity sinking site for the photogenerated electron-hole pairs, which suppressed the recombination of the photogenerated electron-hole pairs. Alternatively, with the application of bias potential, the electrons could have been dragged to the counter electrode and reacted with the adsorbed oxygen to form hydroxyl radicals. This makes PEC an excellent method for the degradation of emerging pollutants, and is also one of its advantages over photocatalysis. These possible mechanisms will enable the holes to participate in directly attacking the pollutants or be adsorbed with H 2 O to form hydroxyl radicals (·OH), which, according to the scavengers study, are the dominant oxidising species. The mechanism of reaction is summarised in Equations (8)

Preparation of Bi1−xLaxFeO3 Perovskite
In a typical experiment, Bi1-xLaxFeO3 (mole fraction x = 0.00, 0.05, 0.10, 0.15) was synthesised by the hydrothermal/calcination method with modification [47]. Quantities of 0.1 M of bismuth, lanthanum and iron were weighed and dissolved in DI in a beaker with continuous stirring for 60 min. A 40 mL volume of 8 M KOH (mineraliser) was added into the mixture, and the pH was adjusted to 8 using 5M nitric acid. The mixture was covered and stirred at room temperature for 72 h. Thereafter, the formed mixture was transferred into a sealed a stainless-steel autoclave (Teflon-lined) and was heated at 200 °C for 24 h, after which it was allowed to cool to room temperature. The brownish product was collected by centrifugation, washed with deionised water and absolute ethanol several times, and then dried in the oven following 1 h of calcination at 500 °C.

Structural, Morphology and Optical Characterisation of the Prepared Electrodes
The PANalytical X'Pert PRO X-ray diffractometer (XRD, Rigagu Ultima IV, Tokyo, Japan) was used to study the degree of crystallinity and the structure. The diffraction patterns were collected from the 2θ degree angle of 10-80° at the step size of 0.0170° using Cu radiation with a wavelength 0.154 nm. The generator was operated at a voltage of 40

Preparation of Bi 1−x La x FeO 3 Perovskite
In a typical experiment, Bi 1-x La x FeO 3 (mole fraction x = 0.00, 0.05, 0.10, 0.15) was synthesised by the hydrothermal/calcination method with modification [47]. Quantities of 0.1 M of bismuth, lanthanum and iron were weighed and dissolved in DI in a beaker with continuous stirring for 60 min. A 40 mL volume of 8 M KOH (mineraliser) was added into the mixture, and the pH was adjusted to 8 using 5 M nitric acid. The mixture was covered and stirred at room temperature for 72 h. Thereafter, the formed mixture was transferred into a sealed a stainless-steel autoclave (Teflon-lined) and was heated at 200 • C for 24 h, after which it was allowed to cool to room temperature. The brownish product was collected by centrifugation, washed with deionised water and absolute ethanol several times, and then dried in the oven following 1 h of calcination at 500 • C.

Structural, Morphology and Optical Characterisation of the Prepared Electrodes
The PANalytical X'Pert PRO X-ray diffractometer (XRD, Rigagu Ultima IV, Tokyo, Japan) was used to study the degree of crystallinity and the structure. The diffraction patterns were collected from the 2θ degree angle of 10-80 • at the step size of 0.0170 • using Cu radiation with a wavelength 0.154 nm. The generator was operated at a voltage of 40 kV and a current of 40 mA. The FTIR (Alpha Bruker, Germany) measurement ranged from 400 to 4000 cm −1 . Electron micrographs were collected with field emission scanning electron microscopy (JEOL JSM-7500F, JEOL Ltd., Tokyo, Japan). UV-vis diffuse reflectance spectroscopy (DRS) measurements were taken with a Cary 60 UV-vis spectrophotometer (Agilent Technologies, Selangor, Malaysia).

Preparation of Bi 1−x La x FeO 3 Photoanodes for Electrochemical and Photoelectrochemical Experiments
In preparing the photoanodes, 30 mg of the prepared samples were each dissolved in isopropanol and nafion solution in a ratio of 3:1; thereafter, the paste was drop-cast on FTO glass (50 mm × 13 mm × 2.2 mm, surface resistivity of ∼7 Ω/sq). The electrodes were dried in the oven for 2 h at 70 • C and were used as working electrodes. The electrochemical analyses were performed using Autolab PGSTAT204 (Metrohm, Utrecht, The Netherlands) potentiostat/galvanostat with a three-electrode system. The prepared FTO-BFO, FTO-5 wt. % La-BFO, FTO-10 wt. %-BFO and FTO-15 wt. % BFO photoanodes served as the working electrodes, while platinum foil and Ag/AgCl (3.0 M NaCl) were employed as counter and reference electrodes, respectively. A 100 W xenon lamp (solar simulator) was employed as the light source for the photoelectrocatalytic degradation. The photoanodes were facing the incident light at a distance of 10 cm. Chronoamperometry (photocurrent response) was carried out in 0.1 M Na 2 SO 4 solution. Impedance spectroscopy was carried out in a 5 mM solution of K 3 [Fe(CN) 6 ] 3−/4− (prepared in 0.1 M KCl solution). Potential scans were performed to obtain data for Mott-Schottky plots without the application of light at room temperature in a pH 7 solution. For the PEC degradation experiments, 5 mg L −1 of Orange II was prepared in a 0.1 M solution of Na 2 SO 4 as a supporting electrolyte in a 100 mL quartz glass for 120 min. Electrochemical degradation experiments were carried out in the absence of light with the application of only bias potential, while photocatalysis was carried out in the presence of light and working electrodes only. Aliquots were taken at different time interval to evaluate the concentration decay and degradation pattern of the Orange II using a UV-vis spectrophotometer. The effect of pH on the PEC removal efficiency was studied. In addition, the optimum photoanode 10% La-BFO was used to degrade methylene blue, Congo red, sulfamethoxazole and acetaminophen.

Conclusions
We successfully synthesised multiferroic Bi 1-x La x FeO 3 (x = 0.05, 0.10, and 0.15) via the hydrothermal method with a low calcination temperature. The formation of La-doped BFO perovskite was confirmed by the XRD, SEM/EDS and FTIR results. The doping of this composite improved the photocurrent response to about three times higher than the pure BFO. The doping of BFO with La 3+ caused a reduction in band gap and an extension of wavelength into the visible light region. When the 10% La-BFO electrode was applied for the photoelectrocatalytic degradation of Orange II dye, a degradation efficiency of 84.2% (higher than that of BFO alone) was recorded within 120 min. This photoanode also showed a decent degradation ability in the presence of a cocktail of PEC dyes and pharmaceutical pollutants, suggesting its robustness.

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