Role of Al doping in structural, microstructural, electrical and optical characteristics of as-deposited and annealed ZnO thin films

Abstract

The role of the Al dopant (0.0 to 3.0 wt%) in modifying the structural, microstructural, electrical and optical properties of pulsed laser-deposited ZnO thin films is reported in both as-deposited (AD) and annealed (AN) films. Incorporation of Al³⁺ ions in the ZnO matrix (AZO) created localized lattice distortions in AD thin films. Breakdown in the local translational symmetry resulted in the observation of otherwise Raman inactive modes in the phonon spectra of the AZO thin films. Furthermore, Al doping enhanced the n-type character of the films, with the charge carrier density exceeding the Mott critical density of ZnO. Charge transport study at low temperatures revealed the metal–semiconductor transition. The increase in charge carrier density with Al doping concentration resulted in a blue shift in the absorption band-edge of these films. These well-characterized thin films were annealed at 800 °C in ambient air for 4 hours. Subsequent characterization revealed drastic modifications in the properties of the AZO thin films. Annealing resulted in reduced lattice distortions, thus improving the crystalline quality of the thin films. This is also supported by the enhanced intensity of E₂^High phonon mode and the disappearance of Raman inactive modes in the phonon spectra of AN thin films. AN AZO thin films showed reduced charge carrier density and increased resistivity. These radical changes in characteristics suggest that the segregation of Al³⁺ ions to grain boundaries is a consequence of annealing.

---

1 Introduction

As contemporary transparent conductive oxide (TCO) materials, aluminium-doped zinc oxide (AZO) thin films with high optical transmission and low resistivity, which are effective in a variety of applications, such as transparent thin film transistors, flat panel displays, solar cells^1–3 etc., have drawn significant attention. Al-doping in ZnO matrix is known to enhance the charge carrier density (n_e (cm⁻³)) and reduce the resistivity (ρ (Ohm cm)) of the host ZnO lattice, which is a wide band gap semiconductor.^4–9 With the increase in n_e, the optical band-gap (E_g (eV)) is also found to show a shift to higher energy (blue shift) compared to the undoped, as-deposited ZnO thin films, which can be explained on the basis of the Burstein–Moss effect and band narrowing. The transmission of these thin films with Al-doping, however, is reported to be 80–85% in the range 400 to 700 nm. These films have hence been looked upon as an alternative to the currently used indium tin oxide (ITO) thin films. Low natural reserves of indium makes it necessary to seek an alternative, which is expected to emerge in the form of AZO thin films.

Different techniques such as metal organic chemical vapour deposition (MOCVD), pulsed laser deposition (PLD), RF-sputtering, atomic layer deposition (ALD) etc. have been used to synthesize AZO thin films. The AZO thin films deposited by pulsed laser deposition technique have yielded low resistivity (∼2 × 10⁻⁴ Ω cm) and high electron mobility (∼60 cm² V⁻¹ s⁻¹), which compare very well with the existing, currently used ITO coatings.⁴ However, it has been reported that AZO thin films are sensitive to a variety of deposition parameters, which are found to alter their properties. For example, the resistivity is often found to increase significantly with an increase in substrate temperature, or after subjecting the films to post-deposition processing, like annealing.¹⁰ Vinnichenko et al.¹¹ have suggested that a narrow window exists, as far as substrate temperature is concerned, in which good quality Al-doped ZnO thin films can be grown. While such studies are being carried out, very little, or only scattered, information is available as far as variation in doping concentration and its effect on the structural, microstructural, electrical and optical properties are concerned, particularly information on the comparison of as-deposited and annealed thin films. This is essential, since the ionic size mismatch between Al³⁺ (0.53 Å) ions and Zn²⁺ (0.74 Å) ions (which are replaced in the ZnO matrix) is large. It is therefore expected to play a crucial role in dictating the above-mentioned properties of the thin films.

In the present work, a comparative investigation of structural, microstructural, electrical and optical properties of as-deposited (AD) and annealed (AN) hetro-epitaxial pulsed laserdeposited undoped ZnO and AZO thin films on c-Al₂O₃ substrate is presented. Thin films of undoped ZnO and AZO of different doping concentrations viz. 0.5, 1.0, 2.0 and 3.0 wt% were grown for this study.

2 Experimental

High purity ZnO and Al₂O₃ (99.99%, Sigma Aldrich) powders were used for the synthesis of undoped ZnO and Al-doped ZnO (AZO) pellets via the standard solid state ceramic route. The powders were used in their stoichiometric proportions for synthesizing AZO pellets with different Al concentrations of 0.5, 1.0, 2.0, and 3.0 wt%. These pellets, henceforth designated as xAZO (where x represents the doping concentration of Al viz. x = 0.0 (undoped ZnO), 0.5, 1.0, 2.0 or 3.0 wt%), were subjected to X-ray diffraction (XRD) (BRUKER D8-Advance) studies. Ni-filtered Cu K_α radiation at 1.542 Å was used for recording the XRD patterns. This recorded data was further analyzed using the technique of Rietveld refinement. The XRD and Rietveld refinement studies ensured the formation of single phase, polycrystalline pellets of x-AZO (where, x = 0.5, 1.0, 2.0 or 3.0 wt%). These well-characterized pellets were then used as targets in the PLD set up to deposit the respective thin films.

Radiation from an excimer laser (KrF, wavelength λ = 248 nm, pulse duration t_p = 20 ns, pulse repetition rate = 5 Hz) was focused onto the target mounted in the deposition chamber. The polar angle between the incident radiation and the normal to the target surface was 45°. The energy density of the incident radiation on the surface of the target was maintained at 2 J cm⁻². Prior to deposition, the c-Al₂O₃ substrates were ultrasonically degreased sequentially, in tri-chloroethylene, acetone and methanol for five minutes each. The cleaned substrate was mounted on a heater and placed parallel to the target at a distance of 6 cm. The deposition chamber was evacuated to a base pressure of 1 × 10⁻⁶ Torr. High purity (99.995%) oxygen was introduced into the chamber and maintained at a constant pressure of 1 × 10⁻⁴ Torr during the deposition. The c-Al₂O₃ substrate was heated to 600 °C and maintained at that temperature during the deposition process. The target was continuously rotated at a speed of 10 rpm during deposition to avoid texturing/pitting of the surface. All the depositions were carried out over 30 minutes. After deposition, the substrates with the deposits were slowly cooled (2 °C min⁻¹) to room temperature (RT) under the same oxygen pressure before removing them for investigation. The AD-xAZO thin films were subjected to structural (XRD), microstructural (micro-Raman spectroscopy), atomic force microscopy (AFM), electrical (RT and low temperature resistivity and charge carrier concentration) and optical (UV-visible spectroscopy) studies. The well-characterized films were annealed at a higher temperature of 800 °C^12,13 in ambient air for 4 hours and then slowly cooled as before to RT. Henceforth, the as-deposited (AD) and annealed (AN), undoped ZnO thin films are designated as AD-ZnO and AN-ZnO thin films, respectively. In the case of thin films deposited using the pellets (targets) xAZO, the AD and AN thin films are represented as AD-xAZO and AN-xAZO thin films, respectively.

The thickness of all the films was estimated by Talystep (Rank Taylor Hobson Limited, England). The structural and microstructural investigations of these thin films were done using the XRD, micro-Raman spectroscopy (Horiba Jobin Yvon Labram HR800) and AFM (JEOL, JSPM-5200). The phonon spectra of all the AD and AN thin films were recorded in back scattering geometry between 90 and 900 cm⁻¹. A continuous wave Ar⁺ laser (λ = 488 nm, output power of 10 mW) in tandem with a charge coupled device was used as an excitation source and detector, respectively. The recorded spectra were deconvoluted using Lorentz curves to comprehend the various phonon modes exhibited by the AD and AN thin films. The surface microstructure/surface morphology of the AD and AN AZO thin films were studied using the AFM in the non-contact mode. These thin films were subjected to electrical characterization, wherein the low temperature (300–50 K) resistivity (ρ–T) measurements were recorded using a computerized four probe technique in the van der Pauw geometry. The charge carrier concentration (n_e) was estimated from the Hall measurements. In order to minimize the contact potential, indium pads were deposited on the thin films using a low power ultrasonic gun. The transmission spectra of the AD and AN thin films were recorded using a UV-visible spectrophotometer (JASCO UV-vis-NIR Spectrophotometer Model V-670) in the range 200 to 700 nm.

Additional characterization of the AD-xAZO thin films by X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) is presented as ESI.†

3 Results and discussion

The thickness of the AD-xAZO thin films determined by Talystep was found to be 300 ± 10 nm.

A Structural characterization

(a) AD-xAZO thin films. Fig. 1(a) shows the θ–2θ XRD pattern of the AD-3.0AZO thin film grown on c-Al₂O₃ substrate. Only two Bragg reflections corresponding to the (0002) and (0004) planes of wurtzite ZnO (JCPDS card no. 36-1451) were observed in the range of 20 to 80° besides the (0006) plane of c-Al₂O₃ substrate. No other Bragg reflections, which could indicate the presence of any other crystallographic orientations of ZnO, or phases of either Al₂O₃ or Zn–Al–O complexes, were detected. This indicates that single phase, highly c-axis oriented AD-3.0AZO thin films were grown successfully on the c-Al₂O₃ substrate. Similar XRD patterns were recorded for thin films with different Al doping concentration, i.e. AD-xAZO thin films (x = 0.5, 1.0, 2.0 wt%), as well as undoped ZnO thin films i.e. AD-xAZO thin films (x = 0.0), and hence are not shown here. However, the peak position (2θ value) corresponding to the (0002) plane was observed to shift continuously to higher 2θ value, as is evident from Fig. 1(b).

Fig. 1 (a) XRD pattern of AD-3.0AZO thin film and (b) the variation in (0002) peak position for different Al doping concentrations.

Table 1 shows the peak position corresponding to the (0002) plane and the c-axis value for different AD-xAZO thin films. It is evident from the table (and Fig. 1(b)) that for undoped and lower doping concentrations, i.e. 0.0, 0.5 and 1.0 wt%, the peak position shifts towards higher 2θ values approaching the standard bulk value of ZnO i.e. 34.42° (JCPDS card no. 36-1451). The observed shift was found to increase with the increase in Al doping concentration. This trend in shifting of the peak position to higher 2θ value persisted even for higher doping concentration, i.e. x = 2.0 and 3.0 wt%, going well beyond the standard value of 34.42°. Shifting of the peak to higher 2θ values indicates the reduction in the interplanar distance “d”, which is in accordance with Bragg’s law and is thought to be due to ionic size mismatch between Zn²⁺ (0.74 Å) and Al³⁺ ions (0.53 Å). The substitution of smaller Al³⁺ ions for relatively larger Zn²⁺ ions at tetrahedral sites causes the decrease in the “d” value.

Table 1 The Bragg angle and c-parameter of both AD-xAZO and AN-xAZO thin films

xAZO	AD-2θ positions (degree)	AD-c-axis parameter (Å)	AN-2θ positions (degree)	AN-c-axis parameter (Å)
0.0	34.27	5.234	34.36	5.220
0.5	34.30	5.229	34.44	5.209
1.0	34.38	5.218	34.46	5.206
2.0	34.56	5.191	34.47	5.204
3.0	34.70	5.171	34.47	5.204

---

The in-plane stress in the basal plane of AD-xAZO thin films is estimated using the biaxial strain model.^14,15 The c-axis parameters mentioned in Table 1 were used for these calculations. The variation in the in-plane stress of these films as a function of doping concentration is shown in Fig. 2. It can be seen that films with doping concentration x = 0.0, 0.5 and 1.0 wt% show negative in-plane stress values (compressive stress), while those with doping concentration of x = 2.0 and 3.0 wt% show positive in-plane stress values (extensive stress). In the doping range studied here, the in-plane stress is hence found to vary systematically, from compressive to extensive, when the Al concentration increased from x = 0.0 to 3.0 wt%.

Fig. 2 Variation in in-plane stress and FWHM of AD-xAZO thin films.

Fig. 2 also shows the variation in the full width at half maximum (FWHM) of the peak corresponding to the (0002) plane. From these values and the Scherrer formula, the crystallite sizes in these films were estimated. As the Al doping concentration increased, the crystallite size was ascertained to decrease up to a doping concentration of 1.0 wt% (i.e. AD-1.0AZO). Beyond this doping concentration, i.e. for x = 2.0 and 3.0AZO thin films, the crystallite size increased monotonically.

(b) AN-xAZO thin films. Fig. 3(a) shows the XRD pattern of the AN-3.0AZO thin film. Again, only two Bragg reflections corresponding to the (0002) and (0004) planes of wurtzite ZnO were clearly observed and no noticeable modification due to annealing was seen. The (0006) reflection of the c-Al₂O₃ substrate can also be seen. No other peaks, either corresponding to any other crystallographic orientation of ZnO, or any other phases, such as Al₂O₃ etc., were observed as a consequence of annealing, indicating that features like single phase, highly c-axis oriented growth of Al-doped wurtzite ZnO thin films were retained after annealing. Similar results were obtained in the XRD patterns of the other AN-xAZO thin films (where x = 0.0, 0.5, 1.0 and 2.0 wt%) and hence are not shown here.

Fig. 3 (a) XRD of AN-3.0AZO thin film, (b) variation in (0002) peak positions of the annealed AZO thin films for different Al doping concentrations.

However, the shift in the peak position (i.e. 2θ value) of the diffracted X-ray radiation from the (0002) plane showed interesting behaviour. Fig. 3(b) shows a slow scan of the (0002) peak for AN-xAZO (where, x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films. In the case of AN-0.0AZO (2θ = 34.36°) thin films, it is observed to shift to a higher 2θ value (i.e. towards the standard bulk value, 2θ = 34.42°), compared to its AD counterpart. Similar behaviour is also exhibited by the AN-xAZO (where x = 0.5 and 1.0 wt%) thin films. The peak position of the (0002) plane for these films also shifts to higher 2θ value (viz. 2θ = 34.44° and 34.46°, respectively) when compared to their AD counterparts, i.e. AD-xAZO (where x = 0.5 and 1.0 wt%) thin films. However, interestingly, in the case of AN thin films with higher doping concentrations, i.e. AN-xAZO (where x = 2.0 and 3.0 wt%), the shift in the peak position is observed to be to a lower 2θ value (viz. 2θ = 34.47° in both cases) compared to its AD counterpart, AD-xAZO (where x = 2.0 and 3.0 wt%) thin films. These 2θ positions of the (0002) plane and the corresponding c-axis parameters of all the AN-xAZO thin films are also indicated in Table 1. It is hence evident that, on annealing, the AD-0.0AZO (undoped ZnO thin film) and the AD-xAZO thin films, irrespective of the doping concentration, show a shift in the peak positions of the diffracted X-ray radiation “2θ value” from the (0002) plane towards the standard value. The shift in peak position of AN-0.0% AZO thin film towards the standard value is possibly due to the improvement in the oxygen stoichiometry. Incorporation of oxygen during annealing relaxes the compressive strain in the AN-0.0AZO crystallites. The inter-planar distance “d” approaches the standard value, thus resulting in shifting of the peak position of the (0002) Bragg plane to a higher 2θ value. The shift in the (0002) peak position of the AN-xAZO (where x = 0.5, 1.0, 2.0 and 3.0 wt%) thin films towards the standard value can be attributed, together with improvement in the oxygen stoichiometry, to microstructural changes in the thin films as a consequence of annealing. It is highly probable that the Al ions have segregated from the substitutional sites (Zn²⁺ sites) in the ZnO matrix or the interstitial positions to the grain boundaries, thus reducing the localized lattice distortion caused by their smaller ionic size. This has further improved the crystalline quality of the films. Secondly, the FWHM of the peaks corresponding to the (0002) plane of all the AN-xAZO thin films were also found to be the same.

B Microstructural characterization

(a) AD-xAZO thin films. Raman spectroscopy is a powerful, non-destructive technique used for studying the crystalline quality, dopant incorporation, induced defects and disorder in doped semiconductor thin films. Wurtzite ZnO belongs to the C_6v (P6₃mc) space group. At the Γ point of the Brillouin zone, the different vibrational modes predicted by group theory are A₁ + 2B₁ + E₁ + 2E₂.^16–18 The modes A₁ and E₁ are symmetric and polar, and are split into longitudinal and transverse optical (LO and TO) components. The B₁ modes are both infrared and Raman inactive modes. The E₂ mode is also symmetric but non-polar and exhibits two frequencies viz. E₂^High and E₂^Low.¹⁸ The Raman spectra were recorded in the range 90 to 900 cm⁻¹, deconvoluted using the Lorentz curves and analyzed.

Fig. 4(a) shows the typical Raman spectra of the c-Al₂O₃ substrate and the AD-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films deposited on c-Al₂O₃ substrate. Different phonon modes of the c-Al₂O₃ substrate were observed at ∼377, 416, 429, 448, 576 and 750 cm⁻¹. These peaks are indicated in Fig. 3(a) by * (asterix), so as to distinguish them from the other peaks arising due to AD-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films. The phonon modes of AD-0.0AZO thin film appear at ∼99, 332 and 437 cm⁻¹. These were earlier attributed to the E₂^Low, two phonons from the K-M-Σ around 160 cm⁻¹ (also called as acoustic overtone) and the E₂^High modes, respectively.¹⁹ The A₁(LO) mode of AD-0.0AZO thin film, which is expected at around 568 cm⁻¹, is also usually observed in the back scattered geometry in which the present Raman spectra has been recorded. However, this mode of ZnO is subsumed by the strong phonon mode of c-Al₂O₃ appearing at 576 cm⁻¹. Excluding this mode, all the other modes arising due to the sapphire substrate and AD-0.0AZO are observed to be distinctly different. The presence of the E₂^High mode at ∼437 cm⁻¹ (standard value of E₂^High mode appears at 438 cm⁻¹) indicates that the wurtzite crystal structure is preserved in our thin films up to a doping concentration of 3.0 wt%, i.e. AD-3.0AZO thin films.¹⁵ The intensity and shift in peak position of the E₂^High mode in ZnO can be associated with the crystalline quality and the stress produced in the films, respectively. With the increase in Al doping concentration in ZnO matrix, the intensity of the E₂^High phonon mode of AD-xAZO is observed to decrease. This reduction in intensity indicated the deterioration in the crystalline quality, which supports our earlier results obtained using XRD studies.

Fig. 4 Micro-Raman spectra of (a) c-Al₂O₃ substrate, and AD-xAZO (x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films and (b) c-Al₂O₃ substrate, and AN-xAZO (x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films.

Micro-Raman spectra of the AD-xAZO (where x = 0.5, 1.0, 2.0 and 3.0 wt%) thin films showed additional vibration modes at ∼276, 465, 511, 570 and 643 cm⁻¹, which are attributed to the B₁^Low, phonon density of states (PDOS), 2B₁^Low, B₁^High and TA + LO, respectively.^18,20 The B₁ modes (i.e. B₁^Low, B₁^High and 2B₁^Low) are Raman inactive modes and were not observed in undoped ZnO thin films deposited on c-Al₂O₃ substrates. These Raman inactive modes, as discussed by Bundesmann et al.,¹⁸ arise due to defects created in the host lattice, which in the present study is manifested in the form of breaking of local translational symmetry due to Al doping. It is interesting to note that the intensity of B₁ modes (i.e. B₁^Low, 2B₁^Low and B₁^High) increases with the increase in Al doping concentration. This can be explained on the basis of enhanced carrier concentration due to Al doping (see Section D of this paper for enhanced carrier concentration). Al doping increases the charge carrier concentration in ZnO crystallites. Between two neighbouring crystallites with different n_e, a depletion region is formed. The electric field developed across these crystallites aids in enhancing the scattering and hence the increase in intensity.²¹ The increase in intensity is thus due to the electric field induced Raman scattering. Furthermore, it is interesting to note that an additional phonon mode is observed at ∼663 cm⁻¹, which appears exclusively in cases of AD-xAZO (where, x = 2.0 and 3.0 wt%) thin films. This mode has been earlier assigned to the TA + LO combination.²²

(b) AN-xAZO thin films. Fig. 4(b) shows the Raman spectra of c-Al₂O₃ substrate and AN-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films. The most striking features in these spectra are (a) the absence of Raman inactive modes in the AN-xAZO thin films (which were observed in the AD-xAZO thin films) and (b) the E₂^High mode appearing very conspicuously in the spectra of all the AN-xAZO thin films. The conspicuous presence of E₂^High mode indicates that the crystalline quality of the films has been enhanced due to annealing. This is further supported by the absence of Raman inactive modes in the AN-xAZO thin films, which indicates the restoration of local translational symmetry in the crystallites of the AN-xAZO thin films. This is possible only if the Al ions, with lower ionic size, have segregated to the grain boundaries during annealing.

C Microstructures/surface morphological characterization

(a) AD-xAZO thin films. Changes in the morphology/microstructure due to changes in the level of Al doping in the AD-xAZO thin films were studied using AFM in the non-contact mode. Fig. 5(i) and (ii) show 2- and 3-dimensional images of the AD-xAZO thin films, respectively. Fig. 5(i)(a–e) shows 2-dimensional images of the AD-xAZO thin films, where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%. Significant structural changes are seen in the AFM images. In the case of AD-xAZO (x = 0.0 wt%) thin films, small circular/oblong islands of the order of 65–70 nm are seen growing on relatively larger regions. In the case of AD-xAZO (x = 0.5 wt%) thin films, holes of roughly the order 90 to 100 nm are observed. With a further increase in Al concentration, AD-xAZO (x = 1.0 wt%), the surface was seen to be covered with larger islands. In the case of AD-xAZO (x = 2.0 wt%), structures resembling self-assembled structures with secondary growth were seen. With a further increase in Al doping, the microstructures showed uniform, unstructured growth of nano-islands, which are smaller than those observed in the case of undoped AD-xAZO (x = 0.0 wt%) thin films, along with regions corresponding to small pockets of shallow deposition.

Fig. 5 (i) 2D-AFM images of AD-xAZO where, x = (a) 0.0, (b) 0.5, (c) 1.0, (d) 2.0 and (e) 3.0 wt% thin films. (ii) 3D-AFM images of AD-xAZO, where x = (a) 0.0, (b) 0.5, (c) 1.0, (d) 2.0 and (e) 3.0 wt% thin films.

Fig. 5(ii)(a–e) shows 3-dimensional images of AD-xAZO thin films, where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%. In all these images, nano-protrusions clustered together were seen. However, the width and the density were found to vary with Al doping. In the case of AD-xAZO (x = 0.0 wt%) thin films, the nano-protrusions appeared to be columnar with lower relative density. In the case of AD-xAZO (x = 0.5 wt%) thin films, the holes surrounded by the nano-protrusions appeared to be similar in shape to volcanoes. In the case of AD-xAZO (x = 1.0 wt%) thin films, the nano-protrusions were faceted and had broader bases, similar in shape to pyramids. The AD-xAZO (x = 2.0 wt%) thin films showed the development of self-assembled structures comprising of these nano-protrusions, while the AD-xAZO (x = 3.0 wt%) showed surface structures similar to that of AD-xAZO (x = 0.0 wt%) thin films. However, the nano-protrusions in the case of AD-xAZO (x = 3.0 wt%), were much finer. The RMS (Root Mean Square) roughness of these films was found to decrease (∼39, 33, 25, 22 and 17 nm) consistently with increasing Al doping concentration (i.e. AD-xAZO (x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%).

(b) AN-xAZO thin films. On annealing, the surface morphology of the films underwent significant changes. Fig. 6(i) and (ii) show 2- and 3-dimensional images of the AN-xAZO thin films, respectively. Fig. 6(i)(a–e) shows the 2-dimensional images of the AN-xAZO thin films, where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%. From these figures, it is clear that, on annealing, the crystallite sizes have grown to approximately 400 nm. Coalescence of smaller crystallites leading to larger crystallites has possibly occurred during the annealing process. Secondary growth of ZnO on the larger crystallites is also evident from these images.

Fig. 6 (i) 2D-AFM images of AN-xAZO where, x = (a) 0.0, (b) 0.5, (c) 1.0, (d) 2.0 and (e) 3.0 wt% thin films. (ii) 3D-AFM images of AN-xAZO where, x = (a) 0.0, (b) 0.5, (c) 1.0, (d) 2.0 and (e) 3.0 wt% thin films.

Fig. 6(ii)(a–e) shows 3-dimensional images of the AN-xAZO thin films, where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt% respectively. From all these images, smoother, self-assembled structures can be clearly observed. The surface roughness, however, has increased with the increase in Al doping concentration.

D Electrical characterization

(a) AD-xAZO thin films. Incorporation of Al³⁺ ions is known to modify the electrical character of the ZnO matrix. Two possible contributions can be considered, firstly, that Al³⁺ acts as an effective donor, as it replaces Zn²⁺ ions in the ZnO matrix⁴ and, secondly, that the interstitial Al ions will modify the sub-lattice of oxygen in terms of oxygen vacancies, which further leads to an increase in charge carrier density (n_e). Al doping is hence expected to change the electrical properties viz. charge carrier density (n_e), RT resistivity (ρ_RT) and temperature dependent resistivity (ρ_T) of the as-deposited ZnO thin films.

The resistivity of pulsed laser-deposited AZO thin films depends on various factors, such as thickness of the films,²³ structural defects²⁴ etc. The lowest resistivity of 0.85 × 10⁻⁴ Ω cm with a charge carrier concentration of 1.54 × 10²¹ cm⁻³ was reported by H. Agura et al.²⁴ Judicious modifications in pulsed laser deposition parameters (energy density) and application of intense external magnetic field resulted in this achievement. This lowest observed resistivity was attributed to good crystal growth occurring directly from the substrate–thin film interface. Under normal PLD conditions, Minami et al.²⁵ have indicated that the charge carrier concentration obtained in pulsed laser-deposited AZO thin films was 1.31 × 10²¹ cm⁻³, while the resistivity obtained was 1.3 × 10⁻⁴ Ω cm. Özgür et al.² have also indicated that the charge carrier concentration obtained in AZO thin films was >10²¹ cm⁻³, while the resistivity of such films was below 2 × 10⁻⁴ Ω cm.

Fig. 7 shows the typical variation in n_e of the AD-xAZO thin films as a function of Al doping concentration. It is observed to increase from 1 × 10¹⁷ cm⁻³ in the case of AD-0.0AZO thin films to 1 × 10²¹ cm⁻³ for AD-0.5AZO thin films. Increasing the Al doping concentration beyond 0.5 wt% in ZnO resulted in a decrease in n_e. It could be reduced to 7 × 10²⁰ cm⁻³ for AD-1.0AZO, 5 × 10²⁰ cm⁻³ for AD-2.0AZO and 5 × 10¹⁹ cm⁻³ for AD-3.0AZO thin films. The n_e in the AD-3.0AZO thin film was found to be approximately two orders of magnitude lower than that of the AD-0.5AZO thin film, but is still higher by two orders of magnitude than that of the AD-0.0AZO thin film. The decrease in n_e with an increase in Al doping concentration is commensurate with an increase in the resistivity of the films, as shown in Fig. 7. Compared to the undoped AD-0.0AZO thin films, the AD-0.5AZO thin film shows a drastic drop in the room temperature resistivity (ρ_RT). The ρ_RT was observed to increase marginally but consistently as the Al doping concentration increased from 0.5 to 3.0 wt%. This is probably due to segregation of some Al ions to grain boundaries, which then act as trap centres for charge carriers, resulting in decreased charge carrier concentration, and hence increased resistivity. It is essential to note that, even though the resistivity of 1.0, 2.0 and 3.0AZO thin films shows slight increase, it is still significantly lower than the resistivity of AD-xAZO thin films.

Fig. 7 Plot of charge carrier concentration and RT resistivity of the AD-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films as a function of x.

Furthermore, in the case of AD-xAZO (where x = 0.5, 1.0, 2.0 and 3.0 wt%) thin films, it is seen that n_e has exceeded the Mott critical density of undoped ZnO, i.e. 4 × 10¹⁸ cm⁻³.⁸ Variation in the resistivity of these degenerate semiconductor films in the low temperature range of 300 to 50 K has hence been investigated. Fig. 8(a–e) show the variation in the resistivity of AD-xAZO (where x = 0.5, 1.0, 2.0 and 3.0 wt%) thin films as a function of temperature. As is evident from Fig. 8(a), on reducing the temperature from 300 K to 50 K, the resistivity of AD-0.0AZO thin film is observed to increase, demonstrating the typical semiconducting nature of the film. The lowest resistivity value of this film is observed to be 50 mΩ cm at 300 K. However, for all AD-xAZO thin films, the resistivity is observed to decrease with a decrease in temperature (metallic behaviour) until approximately 150 K. With a further decrease in temperature, the resistivity either remained the same (as in the case of AD-0.5AZO thin films) or showed a slight increase (as in the cases of AD-1.0AZO, AD-2.0AZO and AD-3.0AZO thin films). The lowest resistivity value obtained amongst these films was 8.7 × 10⁻⁴ Ω cm at 125 K for AD-0.5AZO. This value is comparable to the lowest values obtained by other researchers.^26,27 It is also interesting to note from Fig. 8(b–e) that the lowest resistivity value in the AD-xAZO thin films increases with an increase in Al doping concentration. In degenerate semiconductors, charge transport is due to weakly localized electrons, which has been discussed in detail in our earlier communication.²⁸

Fig. 8 Temperature dependent ρ of AD-xAZO thin films, where x = (a) 0.0, (b) 0.5, (c) 1.0, (d) 2.0 and (e) 3.0 wt%.

(b) AN-xAZO thin films. Fig. 9 shows the variation in RT electrical resistivity (ρ_RT, Ω cm) of AN-xAZO (where, x = 0.0, 0.5, 1.0, 2.0, and 3.0 wt%) thin films. The resistivity of the AN-xAZO (where, x = 0.0 wt%) thin films is around 5.0 MΩ cm. The resistivity of the AN-xAZO thin films is observed to decrease as the Al doping concentration increased from 0.5 wt% (4.0 MΩ cm) to 2.0 wt% (0.9 MΩ cm). With further increase in Al doping concentration to 3.0 wt% (AN-3.0AZO), the resistivity increases to 2.0 MΩ cm. The variation in the charge carrier concentration (n_e) of the annealed thin films is also shown in Fig. 9. As the Al doping concentration increased from 0.0 to 2.0 wt% in AN-xAZO the n_e increased from 4 × 10¹⁴ to 7 × 10¹⁶ cm⁻³. With further increase in the doping concentration i.e. AN-3.0AZO the charge carrier concentration decreased to 3 × 10¹⁶ cm⁻³. However, it is essential to note that the AN-xAZO thin films show significantly lower charge carrier concentration when compared to their AD counterparts.

Fig. 9 Plot of charge carrier concentration and room temperature resistivity of AN-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films as a function of x.

E Optical characterization

(a) AD-xAZO thin films. The AD-xAZO thin films were found to be highly transparent in the range 400 to 700 nm, as is evident from Fig. 10(a). The average transmittance in this wavelength range is ∼90%. All the AD-xAZO thin films showed a sharp drop in transmission around the absorption edge. With an increase in Al doping concentration, the absorption edge shifts towards higher energy. The RT band gap of these films (estimated from Tauc plot) changes from 3.36 eV for AD-0.0AZO to 3.75 eV for AD-3.0AZO thin films, i.e. a blue shift is observed. This blue shift in the band gap energy arises due to an increase in charge carrier density as a result of varying doping concentration of Al³⁺ ions (Burstein–Moss effect).^29,30

Fig. 10 Transmission spectra of (a) AD-xAZO and (b) AN-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films in the UV-visible region.

(b) AN-xAZO thin films. Fig. 10(b) shows the transmission spectra of AN-xAZO (where x = 0.0, 0.5, 1.0, 2.0 and 3.0 wt%) thin films in the range 300 to 700 nm. It is interesting to note that the transmission of these films in the range 400 to 700 nm has not been affected by annealing. However, the absorption edge of each of these spectra viz. AN-xAZO (where x = 0.5, 1.0, 2.0 and 3.0 wt%) has shifted and coincided with the AN-xAZO (where x = 0.0 wt%) thin film at λ ∼375 nm. This also happens to be the absorption edge of undoped AD-xAZO (where x = 0.0 wt%) thin film. This clearly indicates that, on annealing, all these spectra show a red shift, i.e. the charge carrier concentration is reduced on annealing.

On comparing the results of structural, microstructural, electrical and optical studies of AD-xAZO and AN-xAZO thin films, it is strongly evident that, on annealing the AD-xAZO thin films at 800 °C for 4 h in ambient air, the dopant viz. Al ions have segregated to the grain boundaries. The segregated Al ions probably form an amorphic network of Al₂O₃, which forms a barrier at the grain boundaries prohibiting inter-grain charge transport and thus enhancing the resistivity of the AN-xAZO thin films. Further work on understanding this barrier is in progress and will be communicated elsewhere.

4 Conclusions

In conclusion, it can be said that the structural, microstructural, electrical and optical properties of ZnO thin films were significantly modified due to incorporation of Al in the ZnO matrix at tetrahedral Zn sites. Annealing drastically modified the aforementioned properties of the AD-xAZO thin films.

The analysis correlating the structural, microstructural, electrical and optical properties of the AD and AN-xAZO thin films is as follows: as-deposited, undoped ZnO and all the as-deposited AZO thin films are textured along the (0002) orientation. The presence of strain (along the c-axis) in the AD films suggests the incorporation of Al³⁺ ion at the Zn²⁺ tetrahedral sites in ZnO matrix. Incorporation of Al³⁺ ion has enhanced n_e which leads to (a) formation of degenerate semiconductor, (b) the degenerate semiconductor exhibiting metal to semiconductor transition, with the lowest resistivity (ρ = 8.7 × 10⁻⁴ Ω cm) being observed in the case of AD-0.5AZO thin film and (c) the observation of a blue shift in the absorption edge of AD-xAZO thin films, commensurate with the Burstein–Moss effect. After annealing at 800 °C, the AZO thin films are structurally unaltered and highly textured along the (0002) orientation. Annealing has significantly relaxed the in-plane strain compared to their AD counterparts. However, no shift in the absorption edge is seen for the AN thin films. The optical transmission of these films (AD and AN) is ∼90% in the range of 400 to 700 nm, which is excellent for TCO application of AZO thin films. The absence of Raman inactive modes in AN-AZO thin films (which were observed in the case of AD-AZO thin films), suggests that the local tetrahedral symmetry has been restored. The incorporation of oxygen has resulted in reduction of n_e and enhanced ρ. All these factors indicate that dopant Al has segregated to the grain boundaries.

Acknowledgements

One of the authors, KPA, would like to acknowledge the financial support extended by BCUD, Savitribai Phule Pune University, Pune for carrying out this work. SDS would like to thank UGC for providing scholarship. The authors also would like to thank Dr V. G. Sathe, DAE-UGC CSR IUC, Indore for recording Raman spectra and Dr K. R. Patil, NCL, Pune for recording the XPS.

References

1. C. Klingshirn, ChemPhysChem, 2007, 8, 782.
2. Ü. Özgür, Proc. IEEE, 2010, 98, 1255.
3. C. Klingshirn, Phys. Status Solidi B, 2007, 244, 3027.
4. C. F. Klingshirn, B. K. Meyer, A. Waag, A. Hoffmann and J. Geurts, Zinc Oxide, From Fundamental Properties Towards Novel Applications, Springer series in Material Science, 2010, p. 120.
5. T. Minami, H. Nanto and S. Takata, Jpn. J. Appl. Phys., 1984, 23, L280–L282.
6. A. Sarkar, S. Ghosh, S. Chaudhuri and A. K. Pal, Thin Solid Films, 1991, 204, 255.
7. J. Steinhauser, S. Faÿ, N. Oliveira, E. Vallat-Sauvain and C. Ballif, Appl. Phys. Lett., 2007, 90, 142107.
8. V. Bhosle, A. Tiwari and J. Narayan, J. Appl. Phys., 2006, 100, 033713.
9. A. K. Das, P. Misra, R. S. Ajimsha, A. Bose, S. C. Joshi, D. M. Phase and L. M. Kukreja, J. Appl. Phys., 2012, 112, 103706.
10. J. F. Chang, W. C. Lin and M. H. Hon, Appl. Surf. Sci., 2001, 183, 18.
11. M. Vinnichenko, R. Gago, S. Cornelius, N. Shevchenko, A. Rogozin, A. Kolitsch, F. Munnik and W. Möller, Appl. Phys. Lett., 2010, 96, 141907.
12. L. M. Kukreja, P. Misra, J. Fallert, D. M. Phase and H. Kalt, J. Appl. Phys., 2012, 112, 013525.
13. H. S. Kang, J. S. Kang, S. S. Pang, E. S. Shim and S. Y. Lee, Mater. Sci. Eng., B, 2003, 102, 313.
14. D. Singh and Y. P. Varshni, Phys. Rev. B, 1981, 24, 4340.
15. S. Maniv, W. D. Westwood and E. Colombini, J. Vac. Sci. Technol., 1962, 20, 162.
16. R. Loudon, Adv. Phys., 1964, 13, 423.
17. C. A. Arguello, D. L. Rousseau and S. P. S. Proto, Phys. Rev. B, 1969, 181, 1351.
18. C. Bundesmann, N. Ashkenov, M. Schubert, D. Spemann, T. Butz, E. M. Kaidashev, M. Lorenz and M. Grundmann, Appl. Phys. Lett., 2003, 83, 1974.
19. J. M. Calleja and M. Cardona, Phys. Rev. B, 1977, 16, 3753.
20. F. J. Manjon, et al., J. Appl. Phys., 2005, 97, 053516.
21. M. Tzolov, N. Tzenov, D. Dimova-Malinovska, M. Kalitzova, C. Pizzuto, G. Vitali, G. Zollo and I. Ivanov, Thin Solid Films, 2000, 379, 28.
22. J. Serrano, F. J. Manjon, A. H. Romero, F. Widulle, R. Lauck and M. Cardona, Phys. Rev. B, 2004, 69, 094306.
23. B.-Z. Dong, G.-J. Fang, J.-F. Wang, W.-J. Guan and X.-Z. Zhao, J. Appl. Phys., 2007, 101, 033713.
24. H. Agura, A. Suzuki, T. Matsushita, T. Aoki and M. Okuda, Thin Solid Films, 2003, 445, 263.
25. T. Minami, Semicond. Sci. Technol., 2005, 20, S35 and the references therein.
26. O. Bamiduro, H. Mustafa, R. Mundle, R. B. Konda and A. K. Pradhan, Appl. Phys. Lett., 2007, 90, 252108.
27. C. Fournier, O. Bamiduro, H. Mustafa, R. Mundle, R. B. Konda, F. Williams and A. K. Pradhan, Semicond. Sci. Technol., 2008, 23, 085019.
28. S. D. Shinde, A. V. Deshmukh, S. K. Date, V. G. Sathe and K. P. Adhi, Thin Solid Films, 2011, 520, 1212.
29. E. Burstein, Phys. Rev. B, 1954, 93, 632.
30. T. S. Moss, Proc. Phys. Soc., London, Sect. B, 1954, 67, 775.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra14799e

---