SrZnO nanostructures grown on templated <0001> Al₂O₃ substrates by pulsed laser deposition

ZnO is a remarkable semiconductor with a wide and direct bandgap of 3.37 eV. It has been comprehensively studied due to its excellent properties which can be explored in a variety of optoelectronics applications, such as in dye-sensitized solar cells (DSSCs),^1,2 chemical and biological sensors,^3,4 energy-storage devices and inverted-polymer solar cells (IPSCs).^5,6 ZnO is inherently an n-type material and surface modifications and nanostructure designing via doping/alloying of ZnO with transition metals (such as Al, Cu, Co, Cd, Mn, Ba, Fe, Cr, and Ca, etc.) has been widely explored to produce ZnO materials with high-reflectance, high conductivity, good magnetic properties, as well to tailor ZnO with other materials for optical bandgap manipulation and photocatalytic property enhancement.^5,7–13

Recently, Sr-doped ZnO nanostructures have received wider attentions for possible various applications (such as in solar cells, gas sensors, MEMS, love wave filter, and wastewater treatment), and that at some appropriate doping levels, ZnO doped with Sr has shown an improved crystallinity, luminescence, photocatalytic and anti-bacterial activities in addition to enhanced UV emissions.^{10–12,14–21}

Alkaline-earth Strontium (Sr) atoms have been doped within the ZnO wurtzite structure using a variety of techniques to form different SrZnO nanostructures and the results were remarkable.^{10,11,15,18,19,21–24} Yousefi et al. produced nanoparticles by doping ZnO with Sr up to 6% via sol-gel method and noted an enhancement in the visible-light photocatalytic activities of the polycrystalline samples.¹⁵ Zheng et al. produce highly porous SrZnO (Sr=5%) structures with good photocatalytic performance for the degradation of Rhodamine B solution.¹⁸ Up to 3% Sr doping on ZnO, Srivastava et al. produced nanoparticles using a sol-gel spin coating technique and observed an improvement in the bandgap of ZnO, which is higher compared to Mg and Ca dopings.²² Similarly, using the same sol-gel technique and doping up to 7%, Xu et al. produced porous nanostructures and observed a blue shift with improvement in the crystalline quality and UV emission performance.¹⁹ Li et al. grew hexagonal columnar SrZnO crystallites using microwave hydrothermal method to dope ZnO with Sr and reported an improvement of the photocatalytic properties of the doped crystallites.²³ Using reflux method, hexagonal Sr-doped ZnO nanorods were grown with red shift in the bandgap enhancing its white-like luminescence.²⁴ Recently, by co-precipitation method, Kumaran and Muhaleedharan reported a much higher photocatalytic activity of Sr²⁺ doped ZnO with a molar ratio of 1:2 compared to the undoped ZnO.¹¹ On the other hand, Harish et al. using one-pot hydrothermal approach, synthesized ZnO/SrO nanostructures consisting of ZnO nanorods and SrO nanoparticles with photocatalytic degradation (up to 9 times under UV).¹⁰ Further, T. Das et al. doped ZnO with Sr (up to 3%) using solid state reaction with resulting modifications in the structure, bandgap, conductivity, and photocatalytic activity of the SrZnO nanostructures.¹² Very recently, using spray-pyrolysis technique, Raghavendra et al. reported an enhancement of up to 250 % in the UV emission at RT due to improvement in surface morphology at higher concentration of Sr atoms in the ZnO wurtzite structures resulting from oxygen vacancies induced by Sr doping.²¹ 

However, in all these reported growths of SrZnO, there seem to be a consensus that the applicability of Sr-doped ZnO is dependent on the kind and shapes of the formed nanostructures and on the synthesis method used. Further, those reports were too few (and even conflicting) to fully understand the effects and mechanism of Sr-incorporation into the lattice of ZnO structure resulting in different structures. The formation of SrZnO nanostructures appears to be dependent on the process/approach used in the synthesis, that is, most reported results were somehow technique-dependent.²¹ Moreover, most Sr-dopings on ZnO as cited previously were done below 7%.

Most techniques on Sr-doping onto ZnO resulted in nanostructures in powder form. Although the resulting materials possess good properties, their applicability in optoelectronics applications is limited, if not, difficult and very challenging. One of the main technical challenges in the integration of nanostructures as an active sensing nanomaterial onto the sensor structure is the growth of nanostructures directly into the substrates.²⁵ Normally, powder nanostructures are screen-printed manually on a substrate which suffers from adhesion problem resulting in poor quality contact. To fully realize the applicability of Sr-doped ZnO, like in the case for sensor applications, the nanomaterials have to be deposited epitaxially by some methods (like CVD and PVD) in the form of a film into a particular substrates. Here, a laser ablation process, also known as pulsed laser deposition (PLD), has been used to grow oxides, carbides, nitrides, and magnetic/superconducting films, and lately it has been optimized to grow nanostructures of ZnO and ZnO alloys.²⁶ PLD has the unique capability to transfer the exact stoichiometry of a target with complex material composition to a substrate making PLD the preferred physical vapor process to deposit high-quality thin films and nanostructures. PLD-grown nanostructures, therefore, offers a very good alternative to sensor production by enhancing nanostructure-substrate contact and providing uniform material composition at the same time.

In this study, we have optimized the parameters of PLD to design and synthesize pointed leaf-like- and pitted olive-like nanostructures of SrZnO alloys directly onto <0001> Al₂O₃ substrates using 5, 10, and 25% SrO-doped ZnO as target materials. This technique uses a template approach, where a thin film of the same material is templated/grown first on the Al₂O₃ substrate at a lower temperature and pressure.²⁷ To this date, there is only one report on synthesis of SrZnO nanostructures by PLD technique carried out previously by our group, and where ZnO is doped with Sr up to 25%,²⁸ although ZnO has been reportedly doped with other transition metals up to 35%.⁷ It is therefore interesting to understand the mechanism of SrZnO alloying at such high percentages. The formed SrZnO nanostructures were characterized by UV/Vis spectrometry, photoluminescence (PL), Transmission Electron Microscopy (TEM), Scanning Electron Microscopy (SEM), and X-ray diffraction (XRD).

ZnO nanostructures have been grown by PLD technique without any metal catalyst.^29,30 In this work, the targets used for ablation were commercially-available targets from Process Materials, Inc. (USA) with 99.9% purity. Three 2-inch targets consisting of alloyed ZnO/SrO with 95:5, 90:10, and 75:25 mol % were used to grow nanostructures onto <0001> oriented Al₂O₃ substrates.

The SrZnO nanostructures were grown using the Neocera Pioneer 180 PLD system, which was evacuated to a base pressure of ∼ 10^-8 Torr prior to growth. A KrF-type Laser (Compex 205, Lambda-Physik, 248 nm) was focused at 45^o to the normal surface of the target at a pulse rate of 10 Hz. The fluence of the incident pulsed beam is estimated to be at ∼6-8 J/cm². The substrate-to-target distance was varied from 5 to 9 cm to follow the shape of the plume. The <0001> Al₂O₃ substrates were cleaned ultrasonically using ethanol, deionized water, and acetone. Prior to the deposition, the substrates were degassed at ∼800 ^oC to remove any native oxide and other contaminants on the sample surface. The substrates were all heated at the back-side by a resistive coil and monitored by a thermocouple attached to the heater assembly. Before deposition, the targets were ablated for a few minutes to ensure fresh and contamination-free ZnO and SrZnO surfaces.

As texture, thickness, and crystal size of the seed layer affect the quality of the grown nanostructures, the templating/seeding method has been used to improve/promote ZnO adhesion on the substrate and was at the same time used as a starting layer for the growth of epitaxial structures.³¹ The templating/seeding technique in PLD also reduces stress/strain at the interface between the substrate and the film due to lattice mismatch. In this work, a thin SrZnO buffer film was first grown at 300^oC, 10 mTorr O₂ background. This serves as a template prior to the growth of the SrZnO alloy nanostructures at 650^oC at less than 1 Torr O₂ background. The formed nanostructures were annealed for 30 min in a 1-Torr Ar atmosphere, before cooling down to room temperature at a rate of 20 ^o/min under the same environment.

The morphological features of the grown SrZnO nanostructures were characterized by High-Resolution Transmission Electron Microscopy (JEOL-2100 HRTEM, Japan) operated at 200 kV, Scanning Electron Microscopy (JEOL-7600F SEM), and X-ray Diffraction (Xpert PRO PANanalytical XRD), while their electronic properties were observed by UV/Vis spectroscopy (Shimadzu UV-2550) and Photoluminescence (Horiba iHR 320 PL).

Figure 1 (a), (b), and (c) show the SEM images of the different SrZnO nanostructures grown on <0001> oriented single crystal Al₂O₃ substrate using ZnO targets with 5, 10, and 25% SrO contents, respectively. The seed-layer surface can also be seen in each graph.

Fig. 1(a) and 1(b) show pointed leaf-like nanostructures for 5 and 10% doping, respectively, while, Fig. 1(c) shows the pitted olive-like nanostructures for the 25% doping. EDX analyses were also performed to determine the elemental composition of each structure. Figure 1(d) shows a representative EDX spectrum of the SrZnO structure, where all structures confirm the presence of Zn and Sr elements. The Cu and C signals come from the carbon-coated copper grid sample holder. No other elements were found in the EDX spectrum; thus confirming the purity of the structures. The uniformity of Sr distribution in a single nanostructure was further investigated using EDX line mapping in scanning TEM mode. As shown in the inset of Fig. 1(c), where a pitted olive-like nanostructure is scratched off from the 25% sample and analyzed by STEM-EDX, it can also be noted that Sr is uniformly distributed throughout the pitted olive-like nanostructure; hence the structure is mainly attributed neither to ZnO nor to SrO structures alone, but to that of the SrZnO alloy.

The growth of ZnO depends on process, preparations, and other growth conditions and the underlying mechanism is not clearly well-understood. However, the idealized growth routine of ZnO is proposed to follow the velocity-to-direction condition given by: V_<0001>>V_<01-10>>V_<000-1>, where the <0001> direction shows the fastest growth, thus resulting normally in ZnO tipped/pointed nanorods grown along the <0001> direction.³² Hence, adding dopants with much larger ionic radii would tend to alter the usual growth direction of ZnO. Sr ions have much larger radii compared to Zn; as such, the much slower growing faces of ZnO could then somehow compete with the (0001)-face resulting in pointed-tip ZnO structures at about 10% doping. At 25% doping, the (01-10) faces could have a growth rate much faster than the (0001); thus resulting in the pitted olive-like ZnO structures which could be due to the reduction of strain and surface energy ultimately dissolving the sharp edges of the structures.¹² 

XRD measurements were carried out to study the crystallinity of the samples. For comparison, ZnO film from the 99.99% pure ZnO target was also grown on Al₂O₃ using the same growth conditions with that of the 3 SrZnO samples. Figure 2 shows the XRD patterns of the SrZnO structures for 5% SrO (red), 10% SrO (blue), and 25% SrO (green). All diffraction peaks for doped samples show sharp and intense peak at ∼34.5^o. In the absence of all other peaks, all samples have shown the growth of single-phase ZnO that is highly-oriented along the <0002> direction for the doped samples.

After aligning the substrate peaks, all (0002) ZnO peaks align well in all XRD spectra. No other peaks indicating the presence of SrO is observed. Interestingly, XRD did not seem to exhibit structural distortion on the ZnO crystal structure even at 25% doping, although Sr could have substituted Zn in the ZnO structure or be inserted in the interstitials. The introduction of the larger Sr²⁺ atoms into the ZnO structure could introduce lattice defects due to larger ionic lattice mismatch. A slight change in the lattice parameter was noted in the work of Das at al.,¹² where Sr atoms could have substituted into the ZnO structures without changing the wurzite structure of the ZnO, although a transition to rock-salt like structure was also proposed for Sr-doped ZnO films grown by PLD at 750 ^oC without templating/seeding.²⁸ Other works also reported polycrystalline Sr-doped ZnO with increased lattice stress and strain.¹⁵ In our samples, the Sr-doping up to 25% by PLD technique did not seem to alter the hexagonally-oriented wurtzite structure of ZnO, although SEM have already shown a completely different structures. Templating/seeding could have reduced the stress/strain between the interface of the film and the substrates resulting in well-crystallized nanostructures. The non-distortion of the ZnO structure could also denote that Sr is introduced into the Zn vacancies or into the interstitials, although this assertion remains to be investigated.

Further, the intensity of the XRD peaks decreases for 10% compared to 5% which suggest slight deterioration in the crystallinity or slight deformation of the wurtzite structure. However, it could be noted that the intensity of 25% is higher than that of the 10%, which could signify a different growth mechanism for the 25% sample. No other characteristic peak was noted in all patterns, except for the 10% doping, where a peak at 56.5^o appear which could be attributed to ZnO (110). The decrease in intensity for 10% is accompanied with a decrease in the PL signal, which will be discussed in the later part of this paper. Since Sr²⁺(1.10Å) is much larger than Zn²⁺(0.74Å), the lattice distortion is expected as doping increases. However the peak position of the ZnO (0002) did not indicate such distortion. More studies are therefore required to further analyze and understand the underlying mechanism for such behavior.

Figure 3 shows the UV/Vis spectra of the 3 SrZnO samples along with the undoped one obtained at Transmittance Mode. The undoped ZnO sample shows a sharp drop in transmittance at around 375 nm. To get the change in bandgap due to doping in the samples, the derivatives (dT/dλ) are obtained from the transmittance spectra, where the derivatives shifting to the lower wavelengths denote an increase in the bandgap and vice versa (see inset in Figure 3 ). It can be noted that there is a slight decrease in the bandgap for 10% doping probably due to the increased distortion in the structure, changes in grain sizes and/or additional defects creating additional states at the conduction band edge. The reduction of bandgap was also noted by Das et al.^12,21 However, a noticeable increase in bandgap is well resolved for the 25% alloy, which could be attributed to the Burstein-Moss effect, where the increased number of carrier concentration due to the creation of donor levels at the interstitials brought by the Sr²⁺ ions result in raising the Fermi level above the conduction band edge, thus increasing the measured bandgap.^19,21

Figure 4 shows PL spectra of all samples. Each spectrum is characterized by a sharp emission peak at around 375 nm and a broad yellow-green peak centered at around 540 nm. The UV peaks at 375 nm are attributed to band-to-band recombination in all samples. The green emission is attributed to deep-level trap states, e.g. interstitials, vacancies located in the middle of the bandgap.^33–35 The undoped ZnO shows a sharper UV peak with a much lower blue-green emission, denoting a very lower concentration of defects. The 5% doping indicates an increasing concentration of defect sites as depicted by a lower near band edge (NBE) peak in the UV range and a higher peak in the blue-green region. At 10 % doping, the PL signature deteriorates as NBE emission decreases, although the deep trap PL peak is also reduced at the same time. Increasing the doping to 25 % has apparently led to decreasing defects as indicated in the blue-green region and to an enhancement in the NBE emission, which could be attributed to passivation of oxygen vacancies by Sr-atom substitution.²¹ However, based on other reported studies, the enhancement of PL by Sr doping is not well explained and needs further studies to determine the effect of Sr doping in ZnO and its PL characteristic.

The PLD parameters have been optimized to design and synthesize Strontium-alloyed pointed leaf-like and pitted olive-like alloyed SrZnO nanostructures using a two-step templating/seeding approach in the 5-25% doping level. Prior to the growth of the nanostructures, a thin SrZnO seed layer/template was first deposited on Al₂O₃ substrates at substrate temperature of ∼300^oC and background oxygen pressure of 10 mTorr. XRD results showed that the SrZnO nanostructures grown by ablating ZnO targets with 5, 10, and 25 % SrO are single-crystal hexagonal nanostructures with <0002> preferential orientation without changes in the lattice structure. At higher SrO concentration, the morphology shifted to spherical due the dissolution of the sharp edges brought by reduction of strain and surface energy. The samples, however, showed different behaviours in their optical bandgaps and PL characteristics.

To date, this is the only reported work on the design and synthesis of pointed leaf-like- and pitted olive-like SrZnO nanostructures grown by pulsed laser ablation by using a 2-step templating/seeding approach, under high doping levels, as most Sr-doped ZnO studies were done below 7% doping.

Although the mechanism on the changes in the ZnO optical properties via Sr doping remains unclear, the alloyed PLD-grown SrZnO is succesfully grown on a substrate; thus flooring the ground for these very promising nanomaterials towards potential applications in biosensors, love-wave filters, solar cells, and ultrasonic oscillators.

The authors would like to extend their sincere appreciation to the Deanship of Scientific Research at King Saud University for funding this work through Research Group No. RG–1435–002.