Effect of Milling Time, Doping Concentration and Temperature on the Structural, Microstructural and Optical Properties of Cu Doped ZnO Nanoceramics

In this work, high energy ball milling (HEBM) technique has been employed to successfully synthesize a series of Cu-doped ZnO nanoceramics (Zn 1-x Cu x O) with Cu concentration x = 0, 0.01, 0.02, 0.03 and 0.04. The synthesized nanoceramic was analysed by XRD, FESEM, TEM, Uv-Vis-DRS. The structural, microstructural and optical properties of the synthesized sample with different duration of milling, doping concentration and temperature has been investigated. X-ray diffraction result indicates that the samples are wurtzite structure with single phase. A decrease in peak intensity was observed with increase in milling time. 10h milled Cu doped ZnO sample shows single phase. After calcination at 900 ºC in x = 0.04 few peaks related to CuO was observed indicating the solubility limit of Cu in ZnO is 3 atomic%. With increase in milling time crystallite decreases where as strain increases however after calcination crystallite increases where as strain decreases. After calcination peak intensity increases. The specific surface area of ZnO increases after Cu doping but decreases after calcination. Bond length decreases after calcination but lattice distortion increases. FESEM micrograph shows particle growth after calcination. The average particle size decreases with increase in Cu concentration(30 nm to 20 nm) whereas increases after calcination and sintering (488 nm-2706 nm).The band gap decreases with increasing in milling time also with increase in Cu concentration and after calcination. Cu doping in ZnO shows red shift indicating Cu doped ZnO samples are suitable for optoelectronic device applications.


Introduction
ZnO has outstanding properties with a direct band gap of 3.37eV, large exciton binding energy of 60 meV, transparent, nontoxic, cheap, high physical and chemical stability, semiconducting behaviour with unparallel piezoelectric, thermal, optolectronic and transport properties. Because of these peculiar properties, ZnO is the material of preference for a broad range of applications like optoelectronics, transparent electronics, spintronic and sensor devices [1,2],UV screening applications [3]. Properties of zinc oxide can be modified by incorporation of transition metals for different applications. Different dopants like Cu, Cr, Co, Mn, Ni etc. are doped into the ZnO which brings major change in the physical and chemical properties [4]. Doping by other elements such as Al, Ga, Ti, Li, 3 Eu, Ce, Y, La, Er, Au, and Pb into ZnO lattice have also reported in the literature. Despite the potential doping of these elements, Cu is selected as dopant for the present investigations as it shows many interesting applications as fluorescent sensor, acetone and ethanol gas sensor, H 2 S gas sensing, CO gas sensor, resistive random access memory, and diluted magnetic semiconductors [5]. Different synthesis methods like auto-combustion [6], ball milling [7], co precipitation [8], sol gel [9], hydrothermal [10] etc. are used for the preparation of transition metal doped ZnO nanoparticles. High energy ball milling was found to be a very efficient method for production of nanoparticles in a large scale with grain size ~100 nm [11]. The chemical reaction takes place between the constituent powder phases during ball milling which brings structural and morphological changes in the particles as well as the electrical and optical properties of materials [12].
The imperfect crystallinity of crystal develops XRD peak broadening. The properties like (a) crystallite size and (b) lattice strain can be extracted from the analysis of peak width. Because of polycrystalline aggregate crystallite size and particle size is not same [13]. Different analysis techniques like BET, SEM and TEM are used for the particle size measurement. Lattice strain arises from crystal imperfections such as lattice dislocation, grain boundary triple junction, sintering etc. [14]. High energy ball milling also brings about a huge strain in the samples [15]. Crystallite size and lattice strain can be analysed by X-ray profile analysis [16]. The characterization of nanocrystalline materials can be done with regard to (i) grain size and (ii) lattice defects [14,17]. The sintering temperature, doping of metal plays an important role for changing properties like optical, electrical etc. [18,19].
The study on band gap reduction effect of ZnO based oxides have thus paying attention of intense research [20].
In the present work we have investigated extensively the structural, morphological and band gap tuning of Cu doped ZnO nanoceramics synthesised by high energy ball milling.

Materials And Method
Nanocrystalline Zn1-xCuxO (x = 0, 0.01, 0.02, 0.03 and 0.04) was synthesized by High Energy Ball Milling method (HEBM, PM400Retsch, Germany). Stochiometric amount of ZnO and CuO were taken according to their atomic ratio in tungsten carbide vials with tungsten carbide balls in ball to powder 4 ratio 3:1 and were milled for 10h with an altering stop of 30 min after each 1hr of milling to prevent wear at a speed of 300 rpm. The synthesized powders were calcinated at 900 ºC at a heating rate of 2 ºC/min. The structural parameters of 10h milled calcined powder was carried out by using XRD (D8 Advance, Bruker) at room temperature. The morphologies of the ball milled powder, the pellet sample after calcinations at 900 ºC for 2h and also for the 900 ºC sintered pellet was investigated by Field Emission Scanning Electron Microscopy (FESEM, Carl Zeiss NTS Ltd, UK). The Transmission Electron Microscopy (TEM) of the synthesized sample was carried out by the JEOL 2010 TEM with 200 kV acceleration voltages. The UV-Vis absorption spectrum of the samples was recorded via the use of Lambda 750 UV/Vis/NIR Perkin Elmer spectrophotometer. The Rietveld refinement of the XRD patterns was performed by using the "FullProf" program.  Figure 2 (a, c, e, g and i) and their enlarged view are shown in Figure 2 (b, d, f, h and j) respectively. The highest intensity peak is along (101) plane and the shifting of this plane for different duration of milling was shown in Figure 3 (a-e).The XRD pattern of 0h (unmilled) sample shows distinct peaks belonging to the starting oxides of CuO and ZnO. The strongest peak at 2θ≈38.74º correspond to the principal peak of CuO (marked as Δ, Figure   2(d, f, h and j)) and peak belonging to ZnO phase (marked as *, Figure 2(c, e, g and i)). The formation of Zn 1-x Cu x O (x = 0-0.04) was observed with 4h of milling itself and further increase in milling time for 10h, the crystallite size reduces to 20nm. The XRD pattern of 10h milled Zn 1-x Cu x O samples clearly shows the crystalline behaviour. The peak shows that crystal structure of Zn 1-x Cu x O nano powders are similar to JCPDS data card No. 36-1451 [21][22][23][24][25][26]. A decrease in peak intensity and peak 5 broadening observed due to ball milling which indicates the reduction in size of the sample. This expansion of peak confirmed the decrease in size and the improvement of strain in the prepared samples.It is clear from the XRD pattern of 10h milled sample that there is no secondary peak related to CuO or Cu related secondary and impurity phase which confirms the purity as well as formation of single phase of the sample [26][27][28] The peak shifting for different duration of milling was observed from (Figure 3) shows the shifting towards the lower 2θ angle.This may be due to various kinds of strain arises in the synthesized samples due to milling and also due to introduction of Cu into ZnO lattice.

The effect of milling time on crystallite size (D) and strain (ε)
The variation of crystallite size and strain for different duration of milling time was shown in Figure   4(a-e). The Williamson-Hall method was used to find out the average crystallite size and strain.
From the Figure 4, it was found that with increasing milling time the crystallite size reduces where as the strain enhances i.e. a inverse relationship exist between crystallite size and lattice strain with different duration of milling. The considerable increment in the lattice strain may be due to decrease of size of the powders during milling which led to reduction in crystallite size and increase in lattice strain [29,30]. The crystallite size of the samples ranges from ~101 nm to ~11nm where as the strain in the range of 0.16 x 10 -3 to 5.02 x 10 -3 for ZnO and CuO sample. After Copper doping the crystallite size of ZnO reduces where as strain decreases.  The Rietveld refinement XRD of 10h ball milled and calcination at 900ºC was shown in Figure 7(a-e).

Effect of calcination temperature
The X-Ray Diffraction pattern of the 10h milled calcined samples also shows that the samples with Cu 6 concentration up to 3 atomic% (x = 0.03) maintains their structure intact. Figure 5(b) for x = 0.04 few peaks at 2θ = 35.61° and 38.74° (marked by the symbol Δ) related to low intense peak of CuO were observed. Figure 7(e) agreed with the experimental result. It shows that after 10h of milling, the Cu atoms were diffused into the ZnO and substituted Zn. But after calcination at 900 ºC, sample up to x = 0.03 substituted Cu 2+ ion retain the Zn 2+ ion positions in the ZnO, giving single phase Zn 1-x Cu x O material, but above x = 0.03, CuO precipitates, generating CuO peaks in the XRD pattern. This is because after calcination Cu ion diffused in ZnO crystal get separated from ZnO and produce a secondary phase. Similar observation was also reported by Y. Liu et al. [31].
For each sample the FWHM of the XRD peak decrease and the intensity increase as compared to uncalcined sample indicating the decrease in the strain during calcination and the crystallite size have grown larger.

Effect of calcination temperature on structural parameters
The structural parameters of 10h milled Zn 1-x Cu x O samples obtained from the Rietveld refinement of the X-ray diffraction pattern using fullprof software [32] was shown in Figure 8(a-f), where as Figure   9(a-f) shows the structural parameter of calcined (900ºC) samples. The change in lattice parameter 'a' and 'c' with different Copper concentration were given in Figure 8(a). The lattice constant of ZnO was found to be increased with Cu doping and it was changing regularly up to x = 0.02. The observed variation in lattice parameter related to the variation in ionic radii of Zn 2+ (0.74Å) and Cu 2+ (0.73Å) ion [33,34]. The lattice parameter of the calcined sample (Figure 9(a)) shows that it reduces with Cu doping and it enhances linearly with Cu concentration. This also proved indirectly that the Cu 2+ was substituted into ZnO lattice [35].The c/a ratio shows the hexagonal structure of ZnO is not effected by Cu doping as well as with changing concentration. The volume of the unit cell was calculated by using the equation [36].
The volume increases slowly with Cu doping (Figure 8(b)), whereas after calcination, the volume 7 decreases (Figure 9(b)). The constant (c/a) ratio shows the hexagonal structure of pure ZnO remain undisturbed after Cu doping or calcination [27].
The crystallite size of ZnO reduces with Cu doping (Figure 8(c)) and also with increase in Cu concentration up to (x = 0.03) but for x = 0.04 it slightly increases, whereas the strain increases linearly. The crystallite size of Cu doped ZnO increases significantly after calcination where as the lattice strain decreases with increase in crystallite size (Figure 9(c)) in comparison to uncalcined sample because of lattice reduction and defect concentration [37]. It is well known fact that the effect of temperature plays a important role in the crystallite size of ZnO nano phase [27].
The variation of specific surface area and u-parameter with Cu concentration for uncalcined sample was shown in Figure 8(d) and for calcined sample in Figure 9(d).
The specific surface area (S) is calculated by using the formula [38] The Oxygen positional parameter (u) is calculated by using the relation [39] The specific surface area of ZnO increases after Cu doping for all composition except x = 0.03, where it decreases. Which may be due to increase in particle size as observed from FESEM ( Figure 12).
However a dramatic decrease in specific surface area was observed after calcination and hence increases the crystallite size. This value of specific surface area of uncalcined sample indicates the higher reactivity of sample [39]. The oxygen position parameter (u) explains the relative dislocation between Zn & O sub lattice along the c-axis. There is a strong relationship lies between the c/a ratio and u parameter. The c/a ratio decreases with distortion of tetrahedral angle. Here an increment in u value was observed with Cu doping as well as with calcination.
The bond length (L) is analyzed by using the relation 8 where a and c are lattice parameter and u is the oxygen positional parameter [26,40].
The crystal lattice distortion degree is calculated by using the relation [41] The effect of doping on bond length Zn-O (L) and crystal lattice distortion degree (R) was shown in

Crystal Structure Super Cell Model
The crystal structure super cell model of ZnO and Cu doped ZnO was created using VESTA software.
In order to achieve realistic experimental dopant concentration, we used periodic 3x3x3 supercell of ZnO and Cu doped ZnO sample using DFT theory, which are drawn from XRD result [26]. The structure of ideal ZnO has a hexagonal wurtzite. Figure 10 Figure 11(c and d). As above 3% Cu precipitates, so Cu is not replacing Zn from 9 its crystal lattice and is placed in the interstitial position. Figure 11

Microstructural Characterization
Microstructural analysis of the ball milled sample has been carried out by FESEM, EDX and TEM. Figure 12(i-iii) shows the FESEM micrograph of all ball milled (uncalcined and calcined) Zn 1-x Cu x O samples. The surface morphology of ZnO changes with doping and temperature as observed from FESEM micrograph. The majority of the particles have spherical in shape and the normal particle size is in the nanometric regime. The bigger sized aggregation were noticed with increasing Cu concentration (Figure 12(i)). This nano powder calcined at 900ºC and compacted to prepare the pellet. The similar pattern of increasing particle size was also observed in case of unsintered Cu doped pellets (Figure 12(ii)). This can be due to the agglomeration of particle after calcination. It shows that the calcined/sintered particles are polycrystalline in nature. It has been observed that the relative density of all the samples increases after sintering at 900ºC. However, the relative densities of unsintered samples are less (~92%). After sintering the density of all sample increases to (97-99%). The particle size distribution of all samples was shown in Figure 14 and the adj. R-Square is 0.767. From Figure, it was clear that the average particle size of uncalcined powder is in the range of 18 nm to 37nm (Figure 14(i)). Whereas after calcination the particle size increases from 209 nm to 915 nm (Figure 14(ii)). The particle size of uncalcined Cu doped ZnO sample was also calculated by TEM ( Figure 13). Again after sintering the particle size increases and is within 488 nm to 2708 nm (Figure 14(iii)). It was also found that, the average particle size enhances with increase in copper concentration.

The chemical composition of pure ZnO and Cu doped ZnO (Zn 1-x Cu x O) with different concentration of
Cu was analyzed by energy dispersive X-ray spectroscopy (EDX) and was depicted in Figure 15

Optical Properties
Studies of optical properties have special importance for optoelectronics applications. The optical properties has been studied by using. UV-Vis Diffuse Reflectance Spectroscopy (DRS).
As ZnO is a direct band gap material [39] and the energy band gap (E g ) can be calculated by direct transition between conduction band and valence band [42]. It was determined by plotting the square of the Kubelka-Munk function F(R) 2 Vs hν and extrapolating the linear part of the curve to F(R) 2 = 0.DRS technique can extract the band gap values of powder semiconductor without any ambiguity.

Effect of milling time on optical energy band gap (E g )
The band gap energy (E g ) of ZnO and copper doped ZnO was estimated by using Kubelka-Munk method [43][44][45][46]. It promotes the transformation of the measure diffuse reflectance and the extraction of E g value with best accuracy.
In case of a semiconductor material the band gap energy varies with the crystal defect developed within the material due to doping [47]. In order to see the effect of milling time and size effect on band gap energy (E g ), the reflectant spectra of 2% Cu doped ZnO (x = 0.02) was studied with different milling time (2,4,8 and 10) and was shown in Figure 16 and their corresponding E g value was shown in Figure 17.
The effect of particle size was clearly observed in the bad gap. The gradual red shift was attributed to size reduction and increase in strain during the milling process. The estimated band gap energy (E g ) for 2h milled sample was found to be 3.11eV which decreases for 4h and 8h (3.02eV and 2.98eV) which may be because of reduction in crystallite size with increase in milling time, but for 10h milled sample it was found to be 3.09 eV (Figure 17). This investigation is in good agreement with some of the recent reports [48][49][50]. Thus mechanical milling can be considered as an important tool to control the optical properties of ZnO for optoelectronic device application. Figure 18 shows the reflectant spectra of 10h milled calcined Zn 1-x Cu x O nanoceramics. The sharp absorption edge at 372 nm in pure ZnO shifts slightly towards longer wavelength edge for the Cu doped samples may be attributed to increase in grain size after calcination.

Effect of Cu concentration on optical energy band gap
It was clearly noticed from Figure 19 that the band gap reduces with doping. A number of researcher also reported that there is a reduction in band gap of ZnO band gap by the introduction of doping element [51][52][53][54]. Again there is regular decrease in band gap of ZnO (red shift) with enhancement in Cu concentration up to 3% Cu doping (x = 0.03) and for x = 0.04, it slightly increases which supports XRD result. Diouri et al. also reported the similar type of red shift in energy band gap for transition metal doped II-VI semiconductor [55,56] which is due to the p-d spin exchange interactions. A reduction in the band gap for Cu doped ZnO nanoparticles was also observed by Elilarassi et al. [57,58].

Effect of calcination temperature on optical energy band gap
This decrease in band gap of ZnO is mainly due to the strong p-d mixing of O and Cu. This is supported by the observation of Bylsma et al. [59].

Compliance with Ethical Standards:
The authors declare that they have no conflict of interest.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.