ON THE CONTRIBUTION FACTORS OF ENHANCED PHOTOCATALYTIC ACTIVITY OF DOPED SEMICONDUCTORS: STRUCTURAL AND OPTICAL INTVESTIGATION OF Cu DOPED ZnO NANOPARTICLES

This work studied the photocatalytic activity improvement of ZnO doped with Cu ion and tried to illustrate the crucial photocatalytic enhancement factors. The structural and morphological properties of synthesized materials were characterized by X-ray diffractometry (XRD), Raman spectroscopy, and transmission electron microscopy (TEM). It showed that the crystal lattice parameters of ZnO were not affected by the doping manner. However, the crystallite size was slightly decreased in the doped sample. Furthermore, the visible photocatalytic activity of the doped sample was enhanced in comparison with that of the undoped ZnO, so that the reaction rate constants of ZnO and ZnO:Cu were 0.00528 and 0.00968 min, respectively. This work also showed that narrowing bandgap did not contribute into the enhanced photocatalytic activity rather than visible light absorption capacity and recombination rate.


INTRODUCTION
Nowadays, photocatalytic activity of photocatalysts towards the degradation of organic compounds has been widely investigated. Among many studies on semiconductors as potential photocatalysts, researches on ZnO -an excellent photocatalyst candidate -are of great interest thanks to its own characteristics as a cheap, non-toxic, and environmentally friendly material. Further studies on ZnO as a direct wide-bandgap semiconductor (Eg ~ 3.37 eV) with large exciton binding energy (60 meV at room temperature) should be carried out in advance, as compared to photocatalytic application [1]. For increasing the visible-light harvesting, bandgap engineering should be applied to decrease the ZnO bandgap and/or introduce defect states in the bandgap [2]. Furthermore, the decrease of recombination rate of electron-hole pair is an important task to be carried out with this material, especially to avoid its large exciton binding energy. Hence, previous reports showed that the photocatalytic activity of ZnO was enhanced by compositing with other semiconductors [3,4], carbon-based materials [5,6], or by doping with other metal and non-metal ions [7 -9]. These experiments were targeted to form a heterojunction of the semiconductor host and the other materials which have showed advantages in photocatalytic enhancement in many studies. However, the contribution of the enhancement factors due to the heterojunction formation, i.e. bandgap reduction, improved visible-light harvesting, and prolonged recombination rate, was not clearly illustrated.
In this work, we have proposed a simple method to enhance the photocatalytic activity of ZnO nanoparticles fabricated by sol-gel method in which Cu 2+ ion, a transition metal ion, was used for in-situ ZnO doping. UV and visible photocatalytic activities were characterized by decolorisation of methylene blue (MB). Moreover, this work has also attempted to illustrate the important factors which contribute to enhanced photocatalytic activity by characterising the optical properties from experimental photoluminescence emission and UV-vis diffuse reflectance spectra.

MATERIALS AND METHODS
The Zn1-xCuxO samples with x = 0.0 and 0.01 were prepared by sol-gel method similar to previous reports [10,11]. Typically, the zincnitratehexahydrate -Zn(NO3)2.6H2O (Sigma-Aldrich) and copper(II) nitrate trihydrate -Cu(NO 3 )2.3H2O (Sigma-Aldrich) were mixed with citric acid -C6H8O7.H2O 1M (Sigma-Aldrich) with a molar ratio of metal ions : acid = 1:3. The resulted solution was stirred at 80 o C for 1 h, then heated at 150 o C for 4 hto form a powder. After that, the powder was annealed at 400 o C for 4 h in air. The undoped and doped samples were named as ZnO and ZnO:Cu, respectively.
Structural and morphological properties were determined using a Bruker D8 Advance Xray diffractometry (XRD) with a CuKα radiation source (λ = 1.54064 Å) and a sweep range 2θ of 20 o -70 o . Optical bandgap was determined by UV-Vis absorption measurement using UV-Vis spectrophotometry (Metash, UV-S100). The photoluminescence emission characteristics were recorded at room temperature by an analog spectrophotometer (Horiba, JobinYvon) equipped with a Xe lamp as an excitation source at a wave length of 325 nm. Photocatalytic activities were investigated by UV-Vis spectrophotometry (Metash, UV-S100) by determining the degradation of methylene blue in solution under simulated ultraviolet (UV) and natural visible light irradiation.  [12]. As can be seen from the XRD of ZnO:Cu sample, there is no extra peaks related to the impurity of an external phase. Figure 1b showed the micrograph of the undoped ZnO, whereby the nanoparticles were fairly uniform in spherical morphology shape with a diameter of ~20 -25 nm. In order to study the effect of the dopant on the crystal structure of the host, lattice parameters (a and c), crystallite size and d-spacing distance should be identified. The lattice parameter could be obtained by Bragg's law:

Phase and structure
where dhkl is the spacing distance of two consecutive planes in the same direction; h, k, l are Miller indexes; n = 1; λ is the wavelength of incident X-ray 1.5406 Å. Then the lattice constants of the wurtzite structure (hexagonal) of ZnO could be calculated using the following equation [13]: The crystallite size can be calculated by using Debye-Scherrer method as follows: where D, λ, β and θ are respectively crystallite size (nm), wavelength of incident X-ray (nm), full-widthhalf-maximum-FWHM (degree), diffraction angle (degree), and 0.89 is the Debye-Scherrer effective constant. The crystallite size calculated is given in  [5,14,15]. We started at the following equation [16]: where β, θ, λ, and ε are full-widthhalf-maximum-FWHM (degree), diffraction angle (degree), wavelength of incident X-ray (nm), and microstrain, respectively. Equation (4) is a linear equation that can be written as y = ax + b; where y-axis (βcosθ) and x-axis (2sinθ) were determined by diffraction peaks of the synthesized samples. The average crystallite size and the microstrain could be estimated by a linear fitting as shown in Fig. 2. The crystallite size and microstrain were calculated and given in Table 3.  The average crystallite size of ZnO and ZnO:Cu calculated by W-H method was in good agreement with the result obtained by Derbye-Scherrer method, the crystallite size was slightly decreased in the doped sample. Furthermore, the microstrain obtained by W-H method in the doped sample was decreased showing that Cu 2+ ions could diffuse into ZnO lattice to reduce its surface energy. However, the correlation parameter (R 2 ) was low indicating that this method is not good to apply. As reported in the literature, W-H method could only have a high R 2 value when the nanocrystallite size was larger than 35 nm [5,14,15].

Photocatalytic activity experiment
The photocatalytic activity of synthesized samples was tested by measuring the decomposition of methylene blue (MB) dye under both UV and visible light irradiation [17]. A sample of 0.02 g was magnetically stirred in 250 mL of a solution containing methylene blue at a concentration of 25 ppm in dark for 60 min in order to obtain the equilibrium state of the absorption, and then the mixed solution was irradiated under UV light or solar light for photocatalytic experiments. The optical investigations were performed with UV-Vis spectrophotometry in a wavelength sweep range of 250 -800 nm.   It can be seen that the performance of ZnO:Cu samples is higher than that of pure ZnO, however the efficiencies of both samples were quite low. We then carried out the photocatalytic measurement under solar light irradiation. Figure 5 showed the photocatalytic activity of ZnO and ZnO:Cu samples under visible light irradiation on a sunny day from 8 am to 11 am [18].
It can be seen from Fig. 5a) and 5b) that the intensities of absorption spectra were decreased as the irradiation time increased up to 150 min. The intensity of the MB characteristic peak at ~ 664 nm was decreased due to the degradation of methylene blue dye [17] and the degradation efficiency of the ZnO:Cu sample was better than that of the pure ZnO. Fig. 5c) shows the photocatalytic efficiency for MB degradation of the synthesized samples over time.
After 150 min of exposure to visible light, the MB degradation efficiencies were about 54.95 % and 76.65 % for ZnO and ZnO:Cu, respectively. The doping effect on the enhanced photocatalytic activity could be from various reasons such as ZnO bandgap reduction [19], enhanced light absorption [20], and excited carriers separation [11] which all affect photoelectron-hole generation/recombination processes. The Cu 2+ ion dopant and donor states (oxygen vacancies/defects) act like a sink to electron and hole pairs, decreasing the electron-hole pair recombination that may lead to an increase in the production of high reactive radicals, OH • and 2 O •− [21] which play an important role in the photocatalytic enhancement. The overall steps of the photocatalysis process could be described as follows [22] The photocatalytic decomposition of methylene blue dye of the synthesized samples also follows the pseudo-first-order kinetic model [23 -26], which can be expressed as: Taking integral ofequation (8): where C and C0 are the methylene blue concentrations at time t = t and t = 0, respectively; K and t are the pseudo-first-order rate constant (reaction rate constant) and time, respectively; the relationship between ln(Ct/C0) and irradiation time is shown in Fig. 5d.
The obtained pseudo-first-order rate constant (K) and linear correlation coefficient (R) were summarized in Table 4. It showed that the photoreaction rate of the doped ZnO was higher than that of the pure ZnO, i.e. ~ 5.28 × 10 -3 and ~9.68 × 10 -3 min -1 for ZnO and ZnO:Cu, respectively. In order to understand the insights of the photocatalytic enhancement, optical properties of the synthesized samples were determined, which will be discussed in the next section.  Photoluminescence (PL) emission is a suitable and nondestructive technique to determine the emissions assigned to radiative recombinations of electron-hole pairs in semiconductors and the presence of impurities and point defects. In principle, the PL peak in UV range is associated to the near band-to-band emission (NBE) while the visible emission range originates from the defects, which includes zinc vacancies (VZn), interstitial zinc (Zni), interstitial oxygen (Oi) and lattice defects relating to oxygen and zinc [27]. Figure 6a showed room temperature PL spectra of the synthesized samples.

Optical properties a) Photoluminescence (PL) emission
For both samples, a broad intense deep-level (DL) emission appears in a range of 400 -750 nm. In general, the PL emission intensity of the ZnO:Cu was lower than that of the pure ZnO indicating that the radiative recombination rate was decreased in the doped sample. Hence the lifetimes of carriers, photoelectron and hole, were prolonger when introducing Cu 2+ ions into the ZnO host, thus these reactive centers could have more time to decompose methylene blue to enhance the photocatalytic activity.

b) UV-vis diffuse reflectance spectroscopy (DRS)
UV-vis DRS of the synthesized samples shown in Fig. 6b indicatesthat the ZnO:Cu sample absorbed visible light better than the undoped ZnO sample. Hence, the increase of the light absorption capacity in the ZnO:Cu sample could be a good reason for enhancing the photocatalytic activity. Furthermore, the DRS of the doped sample shifted to the longer wavelength compared to the pure ZnO which can lead to the bandgap reduction, promoting the doped sample to absorb more visible light and thus, to improve the visible photocatalytic ability [18]. For further investigation of the bandgap value of the synthesized samples, the extrapolated method was applied to determine the bandgap value through Tauc's plot for the direct ZnO bandgap by the following equation: where n is type of optical transition, α and A are absorption coefficient and optical transitiondependent constant, hυ andEg are the energy of the photon and the band gap energy, respectively. In this case, ZnO is a direct wide-bandgap semiconductor, therefore the value of the optical transition n is ½ (plotted as a(hυ) 2 versus E) [28]. Figures 6c and 6d showed the bandgap values of ZnO and ZnO:Cu, respectively, whereby the bandgap estimated for pure ZnO (3.17 eV) isalmost the same as for ZnO:Cu (3.18 eV), which contradicts previous reports of the doped ZnO studies [29,30]. We then used the first derivative method to obtain the bandgap of the synthesized powders [31] as given in Fig. 7. The result showed that the bandgaps obtained by the first derivative method were similar to Tauc's plot method, i.e. 3.15 eV and 3.13 eV for ZnO and ZnO:Cu samples, respectively. It could be stated that the doping manner did not change the bandgap value of ZnO. Many ZnO photocatalytic studies have concluded that the reduction of bandgap in the doped ZnO nanostructure results in improved photocatalytic activity. In this case, however, it can be seen that the bandgap reduction was insignificant and not the matter of enhanced photocatalytic activity. The enhancement of the photocatalytic activity in the doped sample could be from improving light harvesting and lowering recombination rate in the doped sample. Furthermore, their values were remained in UV-region.

CONCLUSIONS
This work has studied the photocatalytic improvement of ZnO doped with Cu 2+ ions in the degradation of methylene blue in both UV and visible range. The ZnO nanoparticle and its Cudoped material were synthesized by a facile sol-gel method giving a homogeneous particle size of ~ 20 nm which was in agreement with Scherrer-Debye and W-H calculations. Furthermore, we have showed that the bandgap narrowing obtained from doping was not a major factor in the photocatalytic improvement. The increase of visible-light absorption and the decrease in recombination rate are attributed to the photocatalytic enhancement.