Investigation of structural, morphological, thermal, and thermoelectric properties of Zn1−xCuxAl2O4 (0.0 ≤ x ≤ 0.1)

Most promising oxide thermoelectric (TE) materials such as perovskites, layered oxide materials, Al-doped ZnO, etc, have been reported. In the present work, Zn1−x Cu x Al2O4 (0.0 ≤ x ≤ 0.1) samples were synthesized by a simple hydrothermal method. The structural, optical, morphological, and TE properties of Zn1−x Cu x Al2O4 (0.0 ≤ x ≤ 0.1) have been investigated. XRD analysis reveals that ZnAl2O4 has a single-phase cubic structure and Cu is completely dissolved in the ZnAl2O4 lattice. Thermal analysis shows that ZnAl2O4 has high thermal stability up to 1000 °C. From the UV–vis DRS analysis, the energy band gap of ZnAl2O4 decreased from 3.30 eV to 2.82 eV with increasing the content of Cu. Carrier concentration and mobility of the samples were measured by the Hall effect. The values of a carrier concentration of undoped ZnAl2O4 and Zn0.9Cu0.1Al2O4 are obtained to be 3.836 × 1013 cm−3 and 3.3 × 1016 cm−3 at 313 K and 9.6 × 1013 cm−3 for pure and 5.5 × 1016 cm−3 for Zn0.9Cu0.1Al2O4 at 673 K. TE properties of the synthesized samples have been analyzed as a function of temperature. With the optimum values of Seebeck coefficient and electrical conductivity, Zn0.9Cu0.1Al2O4 shows the highest power factor of 0.50 μW/mK2 while the pure ZnAl2O4 shows a maximum power factor of 0.19 μW/mK2 at 673 K. The Zn0.9Cu0.1Al2O4 exhibits a relatively high zT of 2.4 × 10−4 at 673 K, while pure ZnAl2O4 has a zT value of 0.4 × 10−4 at 673 K. The obtained values reveal the improvement of TE properties by increasing the Cu content in the sample.


Introduction
Alternative and sustainable energy sources are becoming more and more important in the 21 st century due to the lack of natural energy resources. There are many renewable energy sources including solar, hydropower, bioenergy, and thermal energy. Thermal energy involves in most of the industrial processes and processes happening in nature. A lot of energy is being wasted every second as heat from industry, automobiles and etc. This waste heat can be effectively recycled to reduce both the scarcity of energy sources and global warming [1,2]. Recently, thermoelectric materials (TE) play a major role in harvesting waste heat energy and convert into electrical energy. The development of high-efficiency TE material is one of the important research directions for this energy conversion [3][4][5]. In the last few decades, extensive research on thermoelectric (TE) properties of oxide materials were carried out effectively, which significantly improved their zT values. The performance of TE material is calculated using the dimensionless quantity called a figure of merit (zT) where, S-Seebeck coefficient (μV/K), σ-electrical conductivity (Sm −1 ), and κ T -total thermal conductivity (W/ mK) [6,7]. High S, σ high, and low κ T are all desirable characteristics of TE material. Doping, alloying, and nanostructuring are effective ways to accomplish it [8].
In recent decades, heavy metal-doped intermetallic compounds are developed as promising thermoelectric materials. However, the use of these intermetallic compounds is limited by their cost, less abundance in nature, high toxicity, and low operating temperature [7,8]. To overcome these difficulties, oxide-based thermoelectric materials like NaCo 2 O 4 [9], Ca 3 Co 4 O 9 [10], ZnO [11], and SrTiO 3 [12] are good candidates owing to their nontoxicity, low cost, and high thermal stability. The report by MuruguThiruvalluvan et al has enhanced the power factor of ZnO by Al doping [13]. Moreover, the Seebeck coefficient of ZnO is enhanced to 450 μV/°C at 100°C by aluminum doping [14]. Copper aluminate is another oxide material and has good thermoelectric properties. The Fe substitution in CuAlO 2 resulted in a high-power factor of 1.1 × 10 −4 W mK −2 at 1140 K [15]. C. Liu et al reported that the figure of merit of CuAlO 2 is 0.0007 at 573 K [16]. The superior electrical conductivity and ecofriendliness of copper make it a viable dopant element [17]. There are several reports on the improvement of thermoelectric properties of oxide materials by copper doping. Rehman et al reported the enhancement of the power factor of 2.3 × 10 −4 W mK −2 of ZnO by copper doping [18].
Zinc-based spinel oxide materials are having high thermal stability and are thereby suitable for thermoelectric applications at high temperatures. ZnAl 2 O 4 is one of the promising materials which crystallizes in cubic, spinel-type AB 2 O 4 structure where Zn 2+ ions occupy the tetrahedral site (A) and Al 3+ ions occupy the octahedral sites (B). ZnAl 2 O 4 and other spinel oxides are suitable to host lattices for different activator ions or dopants. The most fascinating ions for ZnAl 2 O 4 doping are transition metal ions [19]. The band gap of metal ion doped ZnAl 2 O 4 can be tuned by doping at either tetrahedral Zn 2+ ion site or octahedral Al 3+ ion site. By doping lanthanum and transition metal ions, the band gap of ZnAl 2 O 4 can be reduced [19]. ZnAl 2 O 4 is mostly studied for catalytic properties [20][21][22]. In 2014, Dwivedi et al theoretically investigates the TE properties of ZnAl 2 O 4 and obtains the figure of merit (zT) value of 0.77 at 300 K [23]. However, the thermoelectric properties of ZnAl 2 O 4 have not been experimentally studied so far. Therefore, the purpose of the present work is to experimentally investigate the structural, optical, and electrical transport properties of pure and Cu-doped ZnAl 2 O 4 . There are many methods to synthesis ZnAl 2 O 4 such as solid-state reaction, co-precipitation, sol-gel, and hydrothermal [24][25][26][27][28]. In the present work, zinc aluminate is synthesized by hydrothermal method owing to its simplicity and cost-effectiveness.
In this work, Cu doped ZnAl 2 O 4 through a hydrothermal method. Structural characterization was done using XRD and the morphology of the undoped ZnAl 2 O 4 and Cu-doped ZnAl 2 O 4 samples. TG-DTA, Seebeck coefficient, electrical resistivity, Hall effect measurement, and thermal conductivity were used to analyses the thermal and thermoelectric properties of pure and Cu-doped ZnAl 2 O 4 .

Experimental section
Pure and Cu-doped ZnAl 2 O 4 samples were synthesized by hydrothermal method. For the synthesis, zinc acetate, copper acetate, and aluminium nitrate (99%, Sigma Aldrich) were used as source materials without further purification. For the synthesis of ZnAl 2 O 4 , Zinc acetate and aluminium nitrate precursors were taken in the mole ratio of 1:2 and dissolved in 50 ml of distilled water under magnetic stirring. After dissolving precursors, Ammonia solution was slowly added until the pH reached to 10. The final solution was transferred into Teflon lined autoclave and kept in a furnace at 180°C for 12h. After 12 h, it was allowed to cool to room temperature naturally. Then the resultant solution was centrifuged and washed with distilled water and ethanol. The obtained powder was dried at 80°C for 12 h. The as-prepared powders were calcined at 900°C for 3 h for getting a phase pure spinel structure.
For synthesizing copper-doped zinc aluminate, zinc acetate (0.18 mol) and copper acetate (0.02 mol) were dissolved in 50 ml of distilled water under magnetic stirring. 0.4M of Aluminium nitrate was added to the above solution. After the dissolving of precursors, the ammonia solution was added for adjusting the pH to 10. The resultant solution was transferred to an autoclave and put in a furnace at 180°C for 12 h. Then the autoclave was allowed to cool naturally. The end solution was centrifuged and washed. The resultant powder was dried at 80°C for 12 h. The final powder was calcined at 900°C for 3 h.

Characterization techniques
The prepared zinc aluminate sample was analyzed by x-ray diffractometer (BRUKER USA D8 Advance, Davinci) to analyze the phase confirmation and crystal structure. The instrument was configured with Cu K α (λ=1.5406Å) radiation in the 2θ range between 20 and 80 degrees at a 40 kV and 30 mA current. The crystallite size of the samples was calculated by using Scherrer's formula. From the high-intensity peak value and planes, the lattice parameters were calculated. TG-DTA measurements were done for as-prepared samples to know the thermal stability and weight loss during the heating treatment. NETZSCH STA 2500 instrument was used for the thermal analysis. TG-DTA was carried out from room temperature to 1000°C with a ramping rate of 20°C min −1 under a Nitrogen atmosphere. For the analysis of TG-DTA, 6.805 mg of the as-prepared material was put into the alumina crucible. The optical properties and band gap values of the samples were analyzed by Diffuse Reflectance Spectra (DRS) recorded by UV-vis spectrophotometer (Perkin-Elmer-650) attached to the Diffuse Reflectance system. Dopant incorporation and modes of vibrations were analyzed by Raman spectrometer (LabSpec 6, HORIBA SCIENTIFIC). A 532 nm line from an Nd:YAG laser was used for off-resonance excitation with less than 4 mW power at the sample. The instrument was calibrated to the same accuracy using silicon standards.
The morphological and elemental analysis was carried out for the prepared samples using High-Resolution Scanning Electron Microscope (HR-SEM) (Thermocientific, Apreo S). For HR-SEM analysis, the synthesized material was evenly coated on carbon tape. To analyze the lattice fringes and Selected Area Electron Diffraction (SAED) pattern of the prepared sample, HR-TEM analysis was carried out using High-Resolution Transmission Electron Microscope (HR-TEM)(JEOL Japan, JEM-2100 Plus). For HR-TEM examination, the material was dispersed in ethanol by sonicating for 5 min before being loaded on a 500-mesh copper grid. Electrical transport properties and Seebeck Coefficient of the prepared samples were measured using the Hall effect measurement system (Excel Instruments, India) and Hot Stage Seebeck Measurement system (Marine India, India). The thermal conductivity was measured between the temperature ranges of 300 to 673 K using the LFA-467 Hyper Laser Flash instrument under Argon atmosphere. For specific heat capacity measurement in the Laser Flash method, copper was used as reference material. The circular shape pellets having (13 × 1 mm) dimension were used for Hall Effect and thermal conductivity measurements. The rectangular shape pellets having (16 × 5 × 3 mm) dimensions were used for Seebeck coefficient measurements.   Rietveld analysis is a powerful tool to determine both structural and microstructural parameters [29]. The profile of x-ray diffraction data is fitted using Rietveld analysis performed at room temperature using GSAS software [30]. The refined XRD graphs (observed, calculated, and difference profiles) are shown in figure 1(a). The cell parameters and R-factor values from the refinement are shown in table 1. The low values of χ 2 and R-factor show that there is a good fit between observed and calculated profiles. The lattice parameters of pure and doped samples are contracted from 8.086 Å to 8.053 Å with increasing Cu concentration may be due to the difference in ionic radii between Zn and Cu (0.65 Å (Zn), 0.57 Å (Cu)) [31]. The intensity change and peak shift of high intensity diffraction peak (311) is shown in figure 1(b). The intensity variation is ascribed to the change in the peak broadening. The intensity of the high intense peak (311) is decreased as Cu concentration increases which originate from the increase in FWHM values (table 1). Therefore, the Cu concentration affects the intensity of ZAO. The slight shift in the peak position of ZCA can be attributed to the small difference between the ionic radii of Zn 2+ and Cu 2+ . Variation of cell parameters and crystallite size with respect to dopant concentration is shown in figure 1(c). The refined 3D crystal structure of ZAO and ZCA4 were derived from VESTA software and shown in The dislocation density (δ) and micro-strain (ε) were calculated by the following equations (3) and (4).  Table 1 shows the structural parameters calculated from XRD and Rietveld Refinement. The low values of the reliability factors R wp and R p confirm the goodness of fit. From table 1, the decrease in the lattice parameter values are du to the difference in ionic radius of Zn 2+ and Cu 2+ and peak broadening causes the decrease in crystallite size. As listed in table 1, the dislocation density and micro-strain are increased as increasing Cu concentration. This rise in microstrain and dislocation density is a sign of more lattice defects.  figure 3. Three stages of weight loss have been noted in both ZAO and ZCA4. The first stage of weight loss is observed at a lower temperature (<200) which may be due to the evaporation of surface water molecules in the samples. This agrees with the endothermic peaks at low temperatures (<200°C) in the DTA curve. The second stage of weight loss is very broad and it is observed from 200 to 400°C. This weight loss may be ascribed to the evaporation of structural water, and the decomposition of residual nitrates and other organic residuals. This is also shown in the DTA curve of the broad endothermic peak at 280°C. The third stage of weight loss occurs between 400°C to 700°C and it may be due to the decomposition of complexes and other residues. This weight loss is also revealed in DTA as an endothermic peak between 430°C to 550°C. After 700°C, the minimum amount of weight loss is noted and it reveals that both ZAO and ZCA4 have high thermal stability up to 1000°C.

UV-vis DRS analysis of pure and Cu-doped ZAO sample
The optical property of ZnAl 2 O 4 was analyzed by UV-vis-DRS spectroscopy. The absorbance spectra of the samples are calculated from the diffuse reflectance of the samples using the Kubulka-Munk formula [34].
where F(R) is the Kubelka-Munk function, α -absorption coefficient, and R -reflectance. The absorption spectra of the samples are shown in figure 4. From figure 4, it is observed that all the samples have strong absorption at the UV region. This can be related to the transition of charge from oxygen to Cu-  doped ZnAl 2 O 4 also exhibit this absorption with absorption extending towards the visible region with increasing the concentration of Cu. The broad absorption hump arising between 550 to 1100 nm. This could be a result of Cu 2+ ions going through a d-d electronic transition [31,34]. The absorption peaks up to 1100 nm in Cu 2+ , which has a 3d 9 -electron configuration with a free electron in the d orbital, are caused by the transition of this one electron.
The inset figure shows the energy band gap of materials. The band gaps are estimated by extrapolating the intercept of the tangent line at (αhν) 2 = 0, using the Tauc formula a n n )where A is constant, n is 2 for the direct band gap, and ½ for the indirect band gap respectively. ZnAl 2 O 4 is a direct band gap semiconductor [35]. The energy gap of all the samples is found to be in the range of 3.30-2.82eV and this value indicates that all the prepared samples are semiconducting materials. From the figure, it is evident that the band gap values of undoped and Cu-doped samples decreased from 3.30 eV to 2.82 eV. This decrease may be ascribed to the induced defect levels of Cu-doping [36]. The energy band gap's narrowing with Cu 2+ doping may be the result of sub bands forming between the bands and merging with the conduction band. The red shift observed in the Cu doped ZnAl 2 O 4 is accounted by numerous variables, including crystallite size, structural parameters, carrier concentrations, the presence of impurities, and lattice strain, have an impact on the band gap value. The values of obtained energy gap are agrees with the previously reported literature [33,34]. Among all of the samples' band gap values, the ZCA4 sample has an optimum band gap value which is suitable for a significant improvement in thermoelectric characteristics. Consequently, additional morphological and thermoelectric properties were conducted on ZAO and ZCA4 samples.

Raman spectroscopy ofpure and Cu-doped ZAO sample
Raman spectroscopy is a powerful technique to know the dopant incorporation, and lattice defects, and also Raman spectroscopy is used to find the crystal quality and secondary phases. From the group theory, ZnAl 2 O 4 (Gahnite) should have five Raman active modes [37] with the following representation Figure 5 shows the Raman spectra of the prepared samples. The high intensity and strong peak are seen at 662 cm −1 and this is attributed to the high-frequency T 2g mode, while the next high intense peak is observed at 422 cm −1 due to E g mode of vibration. The high frequency T 2g mode is the main Raman mode for the ZnAl 2 O 4 crystal structure.
The previous report showed that the phonon modes with low frequencies are owing to motions of Zn ions and the high-frequency modes are due to motions of O and Al ions [38,39]. In this material, the high-frequency mode at 662 cm −1 is caused by the motion of O atoms in the AlO 6 octahedral [40]. The disorder or defect of material is typically identified by some asymmetry in the peak positions [37]. From figure 5, it is evidently noted that the peaks corresponding to T 2g and E g modes for ZCA4 are asymmetric. This asymmetry in the peak positions indicates the existence of imperfections in the material.  Figure 6 shows the HR-SEM image, Elemental mapping, and Energy Dispersive x-ray Analysis (EDX) spectrum of Cu-doped ZnAl 2 O 4 . From the figure 6(a)-(e) it is understood that all the samples are highly aggregated and   Figure 7 shows the HR-TEM images and Selected Area Electron Diffraction (SAED)patterns of undoped and Cu-doped zinc aluminate. From figures 7(a) and 8(a), it is obvious that the morphology of both ZAO and ZCA4 is in a spherical shape. Figures 7(b) and 8(b) shows the size distribution histogram of ZAO and ZCA4 and the average particle size of ZAO and ZCA4 is calculated to be 25 nm and 15 nm, respectively. Figures 7(d) and 8(d) show the HRTEM images of ZAO and ZCA4 and the insets clearly shows the lattice fringes with thed values of 0.24 nm and 0.27 nm for ZAO and ZCA4 whichcorresponding to (311) and (220) crystal planes, respectively. This clearly indicates that ZAO and ZCA4 are highly crystallized along (311) planes and it is closely agreed with XRD results (table 1). The SAED patterns of ZAO and ZCA4 are displayed in figures 7 (c) and 8(c). SAED pattern consists of many regular concentric rings with different radii caused by the polycrystalline nature of the samples. The diffraction rings from the center to the outside corresponding to the crystal planes (220), (311), (422), (511) and (440) respectively. The SAED and HR-TEM images are well agreed with XRD results. Figure 9(a) shows the temperature-dependent Seebeck coefficient (S) of the ZAO and ZCA4 samples. The negative sign of S for all samples shows the n-type nature of the sample. The Seebeck coefficient increased with temperature for all the samples. This result could be explained by the phonon traction effect, in which a phonon flows from the high-temperature to the low-temperature end of the semiconductor, collides with the carrier, and then produces the seam stream with the phonons, boosting the Seebeck coefficient. S is decreased when the concentration of Cu increases. Since the number of free charge carriers is increased when increasing the dopant concentration, the Seebeck coefficient is decreased according to Mott's relation. The value of S is −437 μVK −1 for ZAO and −353 μVK −1 for ZCA4 at 673 K.

Thermoelectric properties of pure and Cu-doped ZAO sample
The temperature dependence of electrical resistivity of ZAO and ZCA4 are shown in figure 9(b). The figure shows that the resistivity decreased with temperature for both samples, which confirms their semiconducting nature. Mostly, in zinc aluminate, the electrical conduction is mostly caused by the hopping mechanism [41]. Decreasing of ρ with temperature is mainly due to the thermally activated charge carriers. The ρ value reaches the value of 1.2 Ωm at 673 K for ZAO and this value is decreased to 0.24 Ωm at 673 K for ZCA4. This is due to the substitution of Cu 2+ ions substitute at the octahedral sites of Zn 2+ ions, which contributes for decrease in electrical resistivity [42]. Thus, the electrical resistivity is decreased when Cu 2+ ions substitute the Zn 2+ ions.
The power factor s = PF S 2 is calculated and shown in figure 9(c). Despite a decrease in the Seebeck coefficient, the power factor rises with both temperature and Cu concentration. This is caused by the fall in resistivity with an increase in Cu concentration. The PF value for ZAO and ZCA4 are 0.19 μWm −1 K −2 and 0.50 μWm −1 K −2 , respectively at 673 K.
The carrier concentration (n) and mobility (μ)of the pure and Cu-doped ZAO as a function of temperature are shown in figures 10(a) and (b) respectively. The Carrier concentration increased from 3.836 × 10 13 cm −3 to  3.3 × 10 16 cm −3 with increasing the Cu concentration and this is due to alloying effect. Resistivity decreased with increasing Cu concentration because the dopant increases the carrier concentration, which in turn bands the gap reduced and thus increases the conductivity. The following equation (7) describes how resistivity and carrier concentration change as doping concentration increases.
Where ρ is the resistivity, n is carrier concentration, e is the charge of the carriers and μ is the mobility of the charge carriers. From figures 10(a), (b), it is noted that the carrier concentration and mobility are weak dependents of temperature and it can be ascribed to the prepared material is degenerately doped semiconductor.
The Mott's relation is described in equation (8) governs how carrier concentration and Seebeck coefficient change with doping concentration.
Where k b is Boltzman constant, e is the charge of the carriers, h is Planks constant, m ⁎ is the effective mass, n is the concentration of the charge carriers and T is the absolute temperature. From this relation, it is noted that  Seebeck coefficient has an inverse relationship with the carrier concentration. This statement is proved by the graph of change of Seebeck coefficient and carrier concentration with dopant concentration that is shown in figures 9(a) and 10(a) & (b). Figure 11(a)-(c) shows the κ T , κ e , and κ L of ZAO and ZCA4 samples. The total thermal (κ T )conductivity is decreased from 2.1 Wm −1 K −1 to 1.4 Wm −1 K −1 for ZCA4 at 673 K. This drastic reduction in thermal conductivity is due to the induced phonon scattering by the metal dopants that produce the defects which confirmed by Raman analysis (figure 5). When transition metal ions substituted at Zn 2+ sites, defects, and interfaces are created and this defect will strengthen the phonon scattering, which reduces the κ L [43]. κ T consists of electronic (κ e ) and lattices thermal conductivity (κ L ), i.e., k k k = + . From the values of k T and PF, the figure of merit is calculated for entire samples. The obtained values of zT are 4.0 × 10 -5 and 2.4 × 10 -4 at 673 K for ZAO and ZCA4 respectively. The figure of merit value is closely comparable to that of similar oxide materials [16, 43, and 44] and this is displayed in table 2. Figure 12 demonstrates the variation of zT of the materials as a function of temperature. The result indicates the possibility of future improvement in TE properties of zinc aluminate by introducing the Cu-dopants and by optimizing the doping composition. The zT value is much lower than the theoretical value [23], becauseof the low values of electrical conductivity of the prepared material. This low electrical conductivity value may be caused by the combined effect of nano structuring, alloying etc., When the number of grains of nano structured material were make into pellet, electron trapping is high rather than electron transfer. Therefore, the electrical conductivity is decreased. The figure of merit value of the prepared material could be enhanced by doping and optimizing the composition of the dopants.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.