Enhanced photocatalytic performance of CuFeO2-ZnO heterostructures for methylene blue degradation under sunlight

One of the strategies to overcome the drawbacks of fast charge recombination of a photocatalyst is to develop semiconductor heterostructures. Herein, we report a two-step precipitation-hydrothermal process to create CuFeO2-ZnO heterostructures with different weight percentages of CuFeO2 (0.5, 1, 5, and 10%). Though X-ray diffraction detected the presence of CuFeO2 on ZnO above 5%, Raman spectroscopy could reveal the presence of CuFeO2 phase as low as 0.5 wt%. For all of the compositions, the bandgap of ZnO did not vary (3.15 eV) on forming heterostructures with CuFeO2. The oxidation of methylene blue under sunlight was used to determine the photocatalytic performance of the heterostructures. In comparison to pure ZnO and CuFeO2, CuFeO2-ZnO heterostructures exhibited a better photocatalytic efficiency. Overall, 5 wt% CuFeO2 on ZnO showed 100% degradation with a rate constant of 0.272 ± 0.002 min−1, which is 16 times faster than ZnO. Time-resolved photoluminescence analysis indicated a higher lifespan of charge carriers in the 5wt% CuFeO2-ZnO heterostructure (32.3 ns) than that of CuFeO2 (0.85 ns) and ZnO (27.6 ns). The Mott–Schottky flat band potentials of ZnO and CuFeO2 was determined to be -0.82 and 1.17 eV, respectively, revealing the presence of Type I heterostructures. The heterostructures also showed outstanding recyclability, with a degradation rate of 97% even after four cycles. The current study shows the significance of forming p-type CuFeO2 and n-type ZnO heterostructures for enhanced photocatalysis.


Introduction
The growing negative impact of industrial waste on humans and the atmosphere has prompted researchers to focus on renewable energy and the environment. To receive clean water, toxic-free removal of these large amounts of hazardous industrial effluents is necessary. Photocatalysis technology is being investigated as one of the significant methodologies to remove environmental contamination. Several studies on TiO 2 [1], ZnO [2], ZnS [3], CdS [4],CdSe, V 2 O 5 [5], WO 3 [6], Fe 2 O 3 [7] and CuO [8] have proven their usage for photocatalytic removal of organic azo dyes. Among the vario us photocatalysts, the n-type semiconductor zinc oxide (ZnO) has garnered much attention due to its low cost, ease of morphological modification, high redox potential, oxidation resistance and non-toxic nature. ZnO has wide bandgap energy (* 3.2 eV) and the absorption of UV photon generates electron and hole which degrade the effluents. The use of readily available solar energy is economically advantageous and a viable process for large scale photocatalysis. However, UV region forms only 4% of the total sunlight radiation which lowers the process efficiency. In addition, faster recombination of electron and hole in ZnO reduces the free carrier concentration which further hinders the photocatalytic performance [9].
Various efforts have been made to lower the recombination of electron and hole. Recently, p-type delafossite materials such as CuAlO 2 , CuFeO 2 have received considerable attention as photocathodes for photoelectrochemical water splitting/ hydrogen evolution reaction and as transparent conducting oxides [10,11]. In particular, CuFeO 2 delafossite has drawn great attention due to its low-cost constituents and abundance. CuFeO 2 has been used in various applications such as photocatalytic degradation and PEC water splitting [12]. In addition, the absorption of CuFeO 2 in the visible region with a narrow bandgap energy of * 1.5 eV makes it efficient to harvest solar spectrum. However, the main challenge in CuFeO 2 is the meagre transport properties due to higher electron-hole recombination rate.
To promote the absorption of these semiconductors in the visible region and inhibit the electron-hole recombination, researchers adopted several strategies such as introducing dopants [13][14][15], manoeuvring the size and shape of the photocatalytic materials [16,17], facet engineering [18,19] and formation of heterojunctions [20,21]. Among these, contriving the photocatalyst materials by forming heterojunctions with other semiconductor oxides forms an interfacial electric field which effectively separate the chargecarriers thereby enhancing the photocatalytic activity. Meng et al. observed an improvement of photocatalytic CO 2 reduction on the formation of Ni(OH) 2 / TiO 2 interface [22]. AgI/BiVO 4 has shown 43% higher photocatalytic activity compared to BiVO 4 [23].  [26]. Among many heterojunctions viz. metal-semiconductor, semiconductor-semiconductor (p-p type, p-n type, n-n type), p-n heterojunctions have shown to be favouring effective charge separation, longer lifetime and rapid charge transfer. CuO-ZnO p-n type heterostructure has shown better performance for gas sensing and photocatalytic degradation [27][28][29]. Shaheer et al. reported that reduced GO supported TiO 2 -In 0.5 WO 3 showing a 12 fold improvement in comparison to bare TiO 2 [30]. Similarly, ZnO-CdS, ZnO-ZnS, ZnO-CuO, ZnO-GaN heterostructures have shown increased performance in comparison to pristine ZnO. ZnOSe/ZnO/boron doped ZnO exhibited reduced electron-hole recombination rate compared to pristine ZnO. ZnO/NiFe 2 O 4 nanocomposite exhibited enhanced degradation efficiency compared to ZnO [31]. Further to enhance the visible light-harvesting, heterojunction formation with narrow bandgap semiconductors has found to be advantageous.
The present study reports the construction of heterostructure between a p-type narrow bandgap (CuFeO 2 ) and n-type wide bandgap semiconductor (ZnO) for improved charge carrier transport and lifetime transport and test for methylene blue degradation and the results have been correlated to the physical-chemical properties of the material. Thus, this paper would provide a scientific outlook to further design novel heterostructures for enhanced solar harvesting.
2 Experimental procedure 2.1 Synthesis of pure ZnO Zinc oxide (ZnO) was synthesised by precipitation method using zinc nitrate hexahydrate (Zn(NO 3 ) 2-6H 2 O, Himedia, 99%) and sodium hydroxide (NaOH, Himedia, 99%) as precursors. 14 g of zinc nitrate was dissolved in 250 mL of distilled water. To this, 3 M NaOH was added to maintain the pH of 11. The resultant solution was stirred for 12 h to form a precipitate. Then the precipitate was centrifuged three times using distilled water and ethanol and dried at 80°C overnight. The dried powder was calcined at a temperature of 500°C for 4 h.

Synthesis of pure CuFeO 2
CuFeO 2 (CFO) was synthesised by hydrothermal method. An aqueous solution of copper (Cu(NO 3 ) 2-6H 2 O, Himedia, 99%) and iron nitrates (Fe(NO 3 ) 3-9H 2 O, Himedia,99%) was prepared by dissolving in stoichiometric ratio. To this, 3 M NaOH (Himedia, 99%) was added to achieve a pH of 14. The solution was stirred for 30 min, then 1 mL of hydrazine hydrate (Merck, 80% solution in water) was added to the mixture. Then the solution was decanted to a Teflon vessel which was later autoclaved at 180°C for 24 h. After this process, the sample was washed 3 times, two times with double distilled water and once with ethanol. The resultant nanopowder was collected after drying at 80°C overnight. Hereafter, CuFeO 2 nanopowders are represented as CFO in this manuscript for discussion.

Characterization
The phase analysis was done using X-ray diffractometer (Rigaku, Ultima IV, Japan) with Cu K a radiation (k = 1.5406 Å ) at a scan rate of 0.02°per second in the 2h range between 20 and 80 o . The Raman spectra were obtained by Renishaw, RM 2000 (UK) with a laser excitation wavelength of 785 nm and 0.3 mW power. The nature of chemical bonding and the formation of heterostructure was identified using Thermo Nicolet 6700 (USA) Fourier Transformed Infrared (FTIR) spectrometer ranging from 800 to 400 cm -1 using KBr as background. The morphological characterization was done using a Scanning Electron Microscope (Hitachi, Model-S-34000 N, Japan). The optical analysis was carried out using UV-Visible spectrophotometer (Shimadzu, UV-3600, Japan). Time-resolved photoluminescence (TRPL) analysis was done using fluorescence lifetime spectrometer (Jobin Yvon, FLUOROLOG-FL3-11, USA). The electrochemical analysis was done using PAR-STAT 4000A potentiostat (USA). Mott Schottky and electrochemical impedance analysis of the heterostructures were carried out using a three-electrode cell consisting of Ag/AgCl as reference electrode and platinum as counter electrode.

Photocatalytic acivity
The photocatalytic activity of xCF-ZnO heterostructures was evaluated by the photodegradation of Methylene Blue trihydrate (Merck) under sunlight. 2 mg of the xCF-ZnO heterostructures was added to 10 mL of methylene blue dye solution having a concentration of 10 mg/L (2.67 9 10 -5 M, 10 ppm). The combined solution of the respective photocatalyst and dye was stirred in dark for 30 min to achieve the adsorption-desorption equilibrium. Then, the reactant solution was exposed to sunlight for 60 min. The intensity of the sunlight was noted using a lux meter (85,000 lx). The latitude and longitude coordinates were 12°0 0 56.967 00 N and 79°51 0 8.978 00 E. To monitor the degradation process, 1 mL of the exposed solution was collected at an interval of every 10 min, subsequently centrifuged and was analysed by UV-Visible spectrophotometer.

Phase analysis
Phase identification of the samples was done by X-ray diffraction analysis. Figure 1 shows the XRD pattern of pristine ZnO, CuFeO 2 and xCF-ZnO In 0.5CF-ZnO and 1CF-ZnO, XRD peaks other than ZnO was not observed. But in 5 and 10 CF-ZnO, additional peaks corresponding to CuFeO 2 were observed at 35.3°and 40.7°(denoted by arrow) in addition to the diffraction peaks of ZnO. This substantiates the presence of both ZnO and CuFeO 2 in the synthesized material. The peaks other than (012) and (104) were not distinct due to the low scattering ability of CFO on ZnO. The absence of peaks specific to CuFeO 2 in xCF-ZnO (x \ 5) arise from the lower concentration of CuFeO 2 which might be lower than the detection limit of XRD. No additional peaks were observed that implies the purity of the synthesised material. The calculated lattice parameter of ZnO was found to be a, b = 0.3348 nm, c = 0.569 nm and CFO was a = b = 0.304 nm, c = 1.702 nm. No peak shift was observed in xCF-ZnO and the lattice parameter of ZnO remained constant substantiating the successful formation of heterostructure without doping.
The crystallite size of the synthesised sample was computed by Debye-Scherer's formula. The mean crystallite size of ZnO was calculated to be 32 nm, while that of 0.5,1,5 and 10%CF-ZnO was calculated to be 34, 32, 35 and 34 nm, respectively. It is clear that the mean crystallite size of ZnO did not vary since ZnO was first synthesised using precipitation method and then hydrothermally treated to form a p-n heterostructure. This denotes that the hydrothermal condition did not affect the previously synthesised ZnO. The crystallite size of CFO was calculated to be 41 ± 3 nm, while that of 10CF-ZnO was calculated to be 18 ± 2 nm. The decreased crystallite size in the composite is due to the hindrance in the nucleation of the host structure (CuFeO 2 ) in the presence of foreign ZnO.

Surface analysis
The vibrational properties of the heterostructures were examined by Raman spectroscopy. Figure 2 depicts the Raman spectra of xCF-ZnO system. As per group theory, the optical modes present in a wurtzite ZnO are given in Eq. (1) where A 1 and E 1 are active modes which are further divided into longitudinal optical (A 1 LO and E 1 LO) and transverse optical (A 1 TO and E 1 TO) modes. In this, A 1 , E 1 and E 2 represents Raman-active modes while B 2 are Raman inactive modes. The prominent peak at 436 cm -1 corresponds to the E 2 mode of vibration in ZnO lattice which represents the parallel vibration of paired Zn and O atom in the same direction. Besides the E 2 band, the small intense peaks at 381 and 585 cm -1 is accredited to A 1 (TO) mode and A 1 (LO)/E 1 (LO), respectively. The peak at 332 is due to the multi-phonon scattering modes that can be assigned to the optical phonon overtone with A 1 symmetry [32]. CuFeO 2 delafossite with rhombohedral structure contains twelve vibrational modes at the zone centre, represented as where subscript g represents Raman active modes, subscript u denotes the infrared active modes. CFO exhibits A 1g and E g mode of vibrations at 680 and 382 cm -1 , respectively. In xCF-ZnO system, the peak at 680 cm -1 emerges with the increase in concentration from 0.5 to 10CF-ZnO, which is characteristic to the vibration of rhombohedral CuFeO 2 . Further, no shift in the representative peaks of CFO and ZnO was observed denoting the absence of Cu or Fe in ZnO lattice.

Interfacial analysis
FTIR spectra of ZnO, CFO and xCF-ZnO heterostructures are depicted in Fig. 3. From the spectra, metal-oxygen bonding has been identified. In ZnO, the vibration band from 500 to 400 cm -1 is characteristic to stretching vibration between Zn and O. In CFO, two strong absorption bands were observed at 436 and 651 cm -1 . The first one corresponds to both Cu-O stretching vibration and O-Cu-O asymmetric vibration while the latter corresponds to the stretching vibration of Fe 3? -O in FeO6 octahedra of the delafossite structure [33]. Further, it can be noticed that the formation of CF-ZnO heterostructure narrows down the vibrational band corresponding to Zn-O. With the increase in CFO percentage, Cu-O vibrational band at 651 cm -1 became stronger, depicting the formation of delafossite structure. In addition, the peak shifted to lower wavenumber with the increase in CFO percentage which denotes the increase in interfacial interaction thereby confirming the heterostructure formation [34].

Morphological study
The surface morphology of the heterostructure was inspected by scanning electron microscope. Figure 4a, b depicts the SEM images of 0.5CF-ZnO and 5CF-ZnO. The heterostructure exhibits spherical morphology. The average particle size was acquired by calculating the particle size of around 100 particles. Inset represents the histogram of particle size distribution. The histogram exhibited lognormal distribution with a mean particle size of 113 and  119 nm for 0.5CF-ZnO and 5CF-ZnO, respectively, which indicates a minimal variation in particle size. The samples exhibited separate particles projecting the homogeneity of the samples. Further, no change in the morphology of ZnO occurred upon forming CF-ZnO heterostructure. Figure 4c depicts the EDX spectra of 5CF-ZnO with the elemental weight composition in the inset table. This confirms the stoichiometry of the prepared samples.

Optical analysis
The optical absorption of the heterostructures was inspected to find the bandgap of the samples. The absorption peak between 200 and 240 nm in CFO is due to the charge transfer excitation from the valence band to the conduction band [35]. In comparison to ZnO, it can be observed that with an increase in wt% of CFO in ZnO, the optical absorption in visible region increases. This indicates the enhanced solar light harvesting property upon forming ZnO-CFO heterostructures compared to pure ZnO which is essential for enhancing the photocatalytic performance. Both ZnO and CFO being indirect semiconductors, the bandgap energy was obtained using Eq. 3 where, a, h, m, A, and E g are the absorption coefficient, Planck constant, frequency, constant and bandgap, respectively. Extrapolating the plot of [a hm] 2 vs. the energy to x-axis provides the bandgap. The bandgap values were found to be 3.15, 3.19, 3.17, 3.12, 3.11 and 1.5 eV for ZnO, 0.5, 1, 5, 10 CF-ZnO and CFO, respectively. ZnO with bandgap 3.15 eV absorbs visible light scarcely. But CFO nanoparticles strongly absorb visible light due to their narrow bandgap. Thus, combining these two materials enhances the visible light absorption without substantial change in bandgap.

Photocatalytic activity
The photocatalytic activity of xCF-ZnO heterostructures was evaluated by the photodegradation of methylene blue under sunlight. 2 mg of the xCF-ZnO heterostructures was added to 10 mL of methylene blue dye solution having a concentration of 10 mg/L. The combined solution of the respective photocatalyst and dye was stirred in dark for 30 min to achieve the adsorption-desorption equilibrium. Then, the reactant solution was exposed to sunlight for 60 min. The intensity of the sunlight was noted using a lux meter (85,000 lx). The latitude and longitude coordinates were 12°0 0 56.967 00 N and 79°51 0 8.978 00 E. 1 mL of the exposed solution was collected at an interval of every 10 min, subsequently centrifuged and analyzed by UV-Vis spectrophotometer. Figure 6 shows the UV-Vis absorption spectra for MB degradation by the prepared heterostructures at an interval of 10 min. The degradation efficiency of the respective photocatalysts was evaluated from the UV-Vis spectra using Eq. 4, where; C 0 is the initial concentration at 0 min and final concentration of the dye at 60 min. Figure 7a represents the bar diagram of the degradation efficiency of pure ZnO, CFO and CF-ZnO heterostructures. Pristine ZnO showed only 38% degradation of methylene blue at 30 min while xCF-ZnO heterostructures showed enhanced degradation efficiency. 5CF-ZnO exhibited 100% photodegradation in 30 min. The increased photocatalytic ability of xCF-ZnO can be attributed to the combined effect of CFO and ZnO. The reaction kinetics for methylene blue degradation was quantified to compare the photocatalytic activity. The reaction exhibited pseudo-firstorder kinetics and the rate constant is calculated using Eq. 5.
where, k is the rate constant and C is the concentration of the dye at different time intervals.  The recyclability of the heterostructure was evaluated by sequentially retrieving the sample by centrifuging and drying at 60°C for 1 h after every degradation experiment. 5CF-ZnO was tested for recyclability and Fig. S1 shows the UV-Vis absorption spectra of 5CF-ZnO for the degradation of methylene blue from 2nd to 4th consecutive cycles. Figure 7d shows the degradation efficiency for 4 Fig. 6 Absorption spectra for the degradation of methylene blue by sunlight for ZnO, xCF-ZnO, and CFO consecutive cycles. The degradation was found to be lowered only by 3% which might be due to the weight loss of samples during the recovery process. Thus, the prepared heterostructure showed excellent stable performance. The phase stability of the best photocatalyst was analysed by XRD. Fig S2 represents XRD pattern of 5CF-ZnO after 3 cycles. It clearly depicts the stability of the photocatalyst without any chemical or phase change. Table 2 displays the comparative data on the photocatalytic activity of various heterostructures reported in the literature. In spite of the lower catalyst concentration used, the rate constant for the degradation of methylene blue was found to be higher even at comparatively larger MB

Time-resolved photoluminescence (TRPL) analysis
To proclaim the constructive effect of heterostructure formation on the photocatalytic performance, the lifetime of the excited charge carriers was measured by TRPL study. The samples were excited at 325 nm and the photoluminescence decay was monitored at 400 nm. Figure 8 depicts the TRPL decay curves for xCF-ZnO heterostructures. ZnO exhibits a multiexponential decay process and was fitted using bi-exponential model as represented in Eq. 6 to obtain two decay times s 1 and s 2 .
where, A 1 and A 2 represent the amplitude. s 1 signifies the time constant for short decay process generated due to quasi-free excitons and s 2 signifies the time for long decay process which is characteristic to localised exciton recombination [41]. The average lifetime (s avg ) of the charge carriers were calculated from the fitted values using Eq. 7 and are represented in Table 3.
From Table 3, it could be observed that long lifetime increased from 29.6 ns in ZnO 33.4 ns in 5CF-ZnO. The lifetime increased with increase in weight percentage of CuFeO 2 in ZnO upto 5% farther on decreased in 10CF-ZnO. 5CF-ZnO exhibited a

Electrochemical impedance spectroscopy
To explore the charge transfer dynamics of the heterostructures for the enhanced photocatalytic activity, the electrochemical measurements were carried out. Figure 9 shows the Nyquist plot of xCF-ZnO the samples. Mott-Schottky analysis was carried out to probe the alignment of band structure in CFO and ZnO. The relative positions of the valence band and the conduction band of ZnO, CuFeO 2 were evaluated from the flat band potential (V fb ) and the bandgap values. Figure 10 depicts the Mott-Schottky (M-S) curves of CFO and ZnO. A positive/negative slope indicates that the CFO/ZnO is an n-type/p-type and affirms that electrons/holes are the majority of the charge carriers. V fb was obtained by extrapolating the plot of 1/C 2 vs potential to X-axis. The flat band potential of ZnO and CuFeO 2 was calculated to be -0.82 and 1.17 eV vs Ag/AgCl, respectively. It is well known from literature, that the difference between the flatband potential and the bottom/top edge of the conduction/valence band is negligible for an-type/ptype material [42]. Hence, we assume the value of flat band potential as the conduction/ valence band edge for n and p-type semiconductor, respectively. The band structure alignment between ZnO and CFO was identified using flat-band potentials and bandgap energy using Eq. 8.

Mechanism of photocatalytic activity
According to the calculated values of energy bands, the photocatalytic degradation mechanism of CF-ZnO heterostructures was proposed as shown in Fig. 11. The energy levels of CFO lies within the energy levels of ZnO forming a Type I (Straddling) pn heterostructure. After the contact formation, the energy levels of CFO shifts upwards while the energy levels of ZnO shifts downwards until the Fermi levels of CFO and ZnO forms an equilibrium. This results in the conduction band edge of CFO higher than ZnO. As a result, an inner electric field forms between the p-CFO and n-ZnO with minority charge carriers. When exposed to sunlight, CFO and ZnO absorbs photons and generate electron-hole pairs. The excited electrons migrate from the conduction band of CFO to the surface of ZnO across the interface. Simultaneously, the holes are transferred from the valence band of ZnO to CFO due to the more negative VB potential of CFO than that of the ZnO. This transfer of electron and holes in the opposite direction increases the lifetime of photo-generated charge carriers in the heterostructures. The oxygen dissolved in the aqueous solution adsorb on the surface of the photocatalyst and react with these excited electrons and generate active free radical Á O 2 . Simultaneously, the holes react with H 2 O to produce Á OH-free radicals. This effectuates the degradation of methylene blue which are adsorbed on the surface of the photocatalyst by the free radicals. Thus, the formation of

Conclusion
The goal of this research was to create a new p-n CuFeO 2 -ZnO heterostructure using a two-step precipitation-hydrothermal process. The weight ratios of p-CuFeO 2 and n-ZnO are important in controlling the photocatalytic properties. For the photodegradation of methylene blue, 5CF-ZnO exhibited 100% degradation in 30 min with a 16-fold increase in rate constant (0.272 min -1 ) when compared to ZnO (0.017 min -1 ). The CuFeO 2 -ZnO weight ratios in the constructed heterostructures were kept at stoichiometry and confirmed via EDX analysis. The increase in charge carrier lifespan in the heterostructure was discovered by the TRPL investigation. The creation of type I p-n heterostructure was revealed by Mott-Schottky analysis of flat band showed that the proposed engineering of CuFeO 2 -ZnO based p-n heterostructure improved photocatalytic methylene blue degradation by increasing visible light absorption, lowering charge transfer resistance, and reducing charge carrier recombination. Overall, this strategy of creating p-n heterostructure would pave the way for developing other efficient solar harvesting devices.