Photoluminescence of spray pyrolysis deposited ZnO nanorods
- 8.5k Downloads
Photoluminescence of highly structured ZnO layers comprising well-shaped hexagonal rods is presented. The ZnO rods (length 500-1,000 nm, diameter 100-300 nm) were grown in air onto a preheated soda-lime glass (SGL) or ITO/SGL substrate by low-cost chemical spray pyrolysis method using zinc chloride precursor solutions and growth temperatures in the range of 450-550°C. We report the effect of the variation in deposition parameters (substrate type, growth temperature, spray rate, solvent type) on the photoluminescence properties of the spray-deposited ZnO nanorods. A dominant near band edge (NBE) emission is observed at 300 K and at 10 K. High-resolution photoluminescence measurements at 10 K reveal fine structure of the NBE band with the dominant peaks related to the bound exciton transitions. It is found that all studied technological parameters affect the excitonic photoluminescence in ZnO nanorods.
PACS: 78.55.Et, 81.15.Rs, 61.46.Km
KeywordsSpray Rate Near Band Edge Energy Dispersive Spectroscopic Chemical Spray Pyrolysis Energy Dispersive Spectroscopic Analysis
ZnO is a semiconductor material for various photonic and electrical applications. ZnO shows a unique set of physical and chemical properties, such as a wide band gap (3.37 eV), large exciton binding energy (60 meV) at room temperature, radiation hardness , piezoelectricity and photoelasticity  and surface chemistry sensitive to environment. Zinc oxide nanostructured layer comprizing nanorods, further denoted as ZnO nanorod layer (ZnONRL), is a material with large effective surface area, suitable for short-wavelength devices, such as ultraviolet (UV) light-emitting diodes (LED) [3, 4], UV nanolaser arrays , UV photodetectors , field emitters , UV protectors-filters , and chemical sensors [9, 10]. As a passive layer with light-trapping properties, ZnONRL can be used as an antireflection coating on silicon solar cells , surface-enhancing window layer in the second generation solar cells with extremely thin inorganic absorber layer (ETA)  or with dye-sensitized solar cells (DSSC) . The chemical spray pyrolysis (CSP)-deposited ZnONRL was used in ETA solar cells showing energy conversion efficiency of 4.2% .
For many of such devices, a large-scale, low-cost fabrication of high optical and crystalline quality ZnO is desirable. Low-temperature chemical synthesis methods can provide large scale and low-cost fabrication. Photoluminescence (PL) is a very sensitive and an effective method to identify the dominant recombination mechanism and defects in materials. However, according to PL study, the as-deposited ZnONRL grown via a low-temperature wet-chemical methods (chemical bath, electrodeposition, hydrothermal growth) do not show a high excitonic to visible emission intensity ratio at 300 K (room temperature), indicating a high defect concentration in these samples [15, 16, 17, 18, 19, 20, 21]. The PL properties of ZnONRL deposited by the wet-chemical methods can be improved by post-growth annealing at high temperatures of 200-850°C in forming gas or vacuum environment [15, 17, 18, 19, 21].
In this study, strong excitonic PL is observed in the as-deposited ZnONRL, grown via CSP at growth temperatures up to 550°C from zinc chloride (ZnCl2) solutions. CSP technique is a template- and catalyst-free method, allowing fast and low-cost deposition of ZnONRL. According to XRD, ZnONRL comprise c-axis (002)-oriented, hexagonal rods of pure ZnO wurtzite phase with aspect ratio up to 30 [23, 24]. The aim of this work is to study the variation in the PL response due to different growth parameters used for the CSP-deposited ZnONRL. As a result, the PL properties were found to depend on the growth temperature, substrate type, spray rate, and the solvent type. The correlation between the electrical and the photoluminescence properties of the CSP-deposited ZnONRL were reported earlier . Our previous studies on the CSP-deposited ZnONRL were focused on the development of the ZnONRL[22, 23, 24], no specific study on the PL of the CSP-deposited ZnONRL was reported until now.
The ZnONRL were deposited by pneumatic CSP method in air at growth temperatures (TG) of 480°C, 530°C, 550°C using zinc chloride (ZnCl2) precursor solutions. Two kind of substrates were used: soda-lime glass (SGL) and commercial indium tin oxide-covered glass (ITO/SGL). The growth temperature TG was controlled through the temperature of a molten tin bath used to obtain uniform heating of the substrate. Other deposition parameters were varied in the following: Three discrete values were used for the spray rate (v): 1.2, 2.2, and 6.2 ml/min. Two kinds of solvents for ZnCl2 were used: H2O and alcoholic solvent (H2O + ethanol, in ratio of 2:3 by volume), the volume of spray solution was 50 ml. The concentration of ZnCl2 in the spray solution was kept constant in the presented series ranging from 0.05 mol/L ( Effect of the substrate on PL properties of ZnONRL and Effect of the spray rate on PL properties of ZnONRL sections) to 0.1 mol/L (General properties of ZnONRL, Effect of the growth temperature on PL properties of ZnONRL, and Effect of solvent type on PL properties of ZnONRL sections). The acidity of the solution was kept at pH = 5. The studied samples were as grown. XRD measurements were performed on a Rigaku Ultima IV diffractometer with Cu Kα radiation (λ = 1.5406 Å) using the silicon strip detector D/teX Ultra.
The photoluminescence (PL) measurements of ZnONRL were made at room temperature (T = 300 K) and at T = 10 K in a closed-cycle He cryostat (Janis). He-Cd laser (325 nm) was used as an excitation source, the excitation intensity was approximately 0.5 mW/mm2. The luminescence emission in the energy region of 1.45-3.45 eV was dispersed by a computer-controlled Carl Zeiss SPM-2 monochromator (f = 0.4 m) equipped with a prism and detected by FEU-79 photomultiplier with a lock-in amplifier. High-resolution LabRam Horiba Yvon HR 800 spectrometer and CCD-detector were used in the region of 3.30-3.40 eV for a closer study of the PL of ZnONRL in UV-region. The same apparatus was used for the Raman spectroscopic studies. Zeiss EVO-MA15 apparatus was used for the scanning electron microscopic (SEM) and the energy dispersive spectroscopic (EDS) study of the ZnONRL. The apparatus was equipped with the Oxford Instruments PentaFet x3 spectrometer using the INCA Energy EDS system at accelerating voltage of 7 kV.
Results and discussion
General properties of ZnONRL
Effect of the substrate on PL properties of ZnONRL
The PL spectra of ZnONRL deposited at TG of 550°C onto the SGL and the ITO/SGL substrates, are presented as Figure 4. We can see a drastic reduction of the red band intensity at 10 K in case of ZnONRL on ITO/SGL substrate. In addition, the green PL emission band in the region of 2.2-2.8 eV is not present at T = 10 K and at T = 300 K in case of the ITO/SGL substrate used. This could be expected because strong red and green bands do not tend to co-exist [26, 33].
The broad and structureless green luminescent band is mostly reported as an intrinsic defect level due to VO[34, 35] or VZn or both, acting as a donor-deep-acceptor pair . Although controversial, the band is most likely a native complex defect of zinc vacancy involving Zni, OZn, and VO. The origin of another visible Gaussian band, the red band in region of 1.5-2.0 eV, is attributed to interstitial-type defects like Zni, Oi[15, 18] while PL bands between the red and the green band are commonly attributed to Oi[18, 38, 39, 40]. Without a comprehensive study additional to PL spectroscopy, the defect identification remains speculative .
At room temperature (T = 300 K), for the ZnONRL grown at 550°C onto ITO/SGL, no green emission is detected, see Figure 4 inset. Therefore, we conclude that ZnONRL with a lower content of defects can be deposited using the ITO instead of SGL substrates, however, such a growth mechanism of the ZnONRL onto the ITO remains unclear. It may be speculated that a more homogenious lateral heat distribution may be responsible for the significant improvement of the NBE to red band intensity ratio (INBE/IRED) in the case of using the ITO/SGL substrate for deposition of ZnONR. Another speculation may be that ITO acts as a barrier, restraining the diffusion of some elements, e.g., sodium from the SGL glass during high-temperature spray process as it has been recorded by X-ray photoelectron spectroscopy for the CSP-deposited CuInS2 films . In such a case, borosilicate glass could be a suitable replacement for the SGL substrate. Whatever the reason might be, ITO-covered substrates are preferred in order to achieve ZnONRL with high crystal quality during the CSP process. This observation is of significant importance due to the fact that for many electronic applications including solar cells, a conductive and transparent material, such as ITO, is required as the substrate material.
Effect of the growth temperature on PL properties of ZnONRL
In connection with the results in the Effect of the substrate on PL properties of ZnONRL section, higher TG of 550°C and the ITO/SGL substrates are preferred for the deposition of ZnONRL in order to achieve high optical quality by the CSP method. The best samples were grown at 550°C onto ITO/SGL substrate, showing NBE emission only. Although, no significant reduction in FWHM is observed while using a higher growth temperature of 550°C, an apparent shift of the main NBE peak at approximately 3.360 eV towards higher energy could be an indicator of additional transitions. The high ratio of INBE to IVISIBLE emissions alone, although widely reported as an indication of good  or even excellent optical quality [7, 44], may not be enough to characterize the purity of ZnONRL[15, 43]. However, the resolvable fine structure of the excitonic PL emission could be taken as an indication of a relatively high crystal quality of the CSP deposited ZnONRL.
Effect of the spray rate on PL properties of ZnONRL
The rise in the intensity of the peaks could be correlated to a longer deposition period, hence, a lower growth rate. In addition, the FWHM of the main peak decreases from 4.0 meV down to 2.2 meV, accompanied by a small shift from 3.360 eV to 3.358 eV in the exciton peak position. A shift in the position of a defect-bound exciton is a clear indication of a change in the type of the dominant impurity in the ZnONRL, while a smaller FWHM is a sign of increased crystal quality. For comparison, peak widths of 6-9 meV and 4.7 meV are reported for ZnONRL grown by hydrothermal method with post-deposition annealing  and thermal evaporation and vapor-phase transport , respectively. A FWHM of 1 meV is characteristic of the excitonic peak at 3.360 eV reported for a single crystal MOCVD-deposited ZnONRL, measured at 3.4 K . The PL intensity in a log-scale reveals a free exciton peak at 3.378 eV appearing only at a spray rate of 1.2 ml/min (see Figure 6 inset), providing confirmation that a low defect density is obtained in ZnONRL. Thus, in addition to high growth temperatures, low spray rates are preferred for CSP deposition of ZnONRL in case aqueous solutions are used.
Effect of solvent type on PL properties of ZnONRL
In case the alcoholic solvent is used, the excitonic band intensity increases approximately 30 times. The PL intensity ratio INBE/IRED increases from 34 in case of aqueous up to 294 in case alcoholic solution is used. This observation could be explained taking into account a similar observation from the effect of the TG (Effect of the growth temperature on PL properties of ZnONRL section). The use of an alcoholic solution results in a smaller spray droplet size, leading to a more homogeneous distribution of the sprayed solution over the substrate, compared to an aqueous solution. In addition to a smaller droplet size, the high volatility of the alcoholic solvent is very likely to cause a lower rate of the substrate cooling. Thus, a higher value of the effective growth temperature is expected in the reaction zone of pyrolysis, resulting in a somewhat higher growth temperature than the stated TG. This can be discussed in the following.
Due to the fact that the PL spectrum A in Figure 7 coincides with spectrum A in Figure 5, we can see that the use of an alcoholic solution instead of an aqueous solution increases the NBE band intensity (approximately × 30) in ZnONRL deposited at 480°C (Figure 7). Similar effect has been observed while increasing the TG from 480°C up to 550°C (Figure 5). In Effect of the growth temperature on PL properties of ZnONRL section, we tentatively attributed this effect to an increased surface-to-volume ratio of the ZnONRL. Similarly, the ZnONRL grown at 480°C from alcoholic solution (see SEM image in Figure 3c) shows a significantly increased surface-to-volume ratio compared to ZnONRL grown at 480°C from aqueous solution (see Figure 3b). Therefore, one possible cause of a higher NBE intensity of spectrum B compared to spectrum A in Figure 7, could be a higher surface-to-volume ratio, in addition to improved crystal quality.
Although the density of rods and the aspect ratio is known to increase when deposited from alcoholic solutions , the effect of the temperature cannot be overlooked either in the sense of an increased NBE intensity. As a result, it can be concluded that a lower TG or a higher spray rate is acceptable for the CSP deposition of the ZnONRL in case an alcoholic solution is used, compared to the TG and spray rates used for aqueous solutions (Effect of the growth temperature on PL properties of ZnONRL and Effect of the spray rate on PL properties of ZnONRL sections). The use of an alcoholic instead of aqueous solution could be applied in order to achieve a higher surface-to-volume ratio at a similar TG. It must be pointed out, however, that the PL of the ZnONRL deposited at a low spray rate (1.2 ml/min) from aqueous solution was superior to all considered samples in sense of narrow NBE peaks (Effect of the spray rate on PL properties of ZnONRL section).
We report photoluminescence properties in as-deposited ZnO nanorod layers (ZnONRL), grown via fast and low-cost pneumatic CSP method in air. The results indicate that a good crystal quality can be achieved, while ZnONRL with the best optical quality were deposited at a growth temperature of 550°C using ITO/SGL substrate. No green or red emission was detected for these samples, measured at 10 K. In addition, green emission is absent at 300 K for the best samples. The prevalence of the bound exciton transitions suggests a low defect density in our best ZnONRL. Low spray rates (1.2 ml/min) are recommended in case aqueous solution is sprayed, otherwise alcoholic solution is preferred for the deposition of ZnONRL. Lower substrate temperature is acceptable in case alcoholic solution is used, in order to achieve comparable crystal quality, compared to the use of aqueous solution. At T = 10 K, fitted excitonic peak widths (FWHM) are 2-5 meV. For comparison, peak widths of 6-9 meV are reported for ZnONRL grown by hydrothermal method with post-deposition annealing . These results indicate that the quality of the spray-deposited ZnONRL is superior compared to other solution-based methods in the sense of photoluminescence properties. The results of this study imply that the spray-deposited ZnONRL is sufficient for the use as a transparent window layer in an ETA solar cell and for applications to other low-cost photonic devices at room temperature.
The study is financially supported by: Estonian Ministry of Education and Research (SF0140092s08) and Estonian Science Foundation grants: ETF8509, ETF6954, and ETF8282.
- 29.Meyer BK, Alves H, Hofmann DM, Kriegseis W, Forster D, Bertram F, Christen J, Hoffmann A, Strassburg M, Dworzak M, Haboeck U, Rodina AV: Bound exciton and donor-acceptor pair recombinations in ZnO. Physica Status Solidi B-Basic Research 2004, 241: 231–260. 10.1002/pssb.200301962CrossRefGoogle Scholar
- 46.Dedova T, Krunks M, Mere A, Klauson J, Volobujeva O: Preparation of shape and size-controlled zinc oxide nanostructures by chemical spray pyrolysis technique. In Zinc Oxide and Related Materials; Warrendale. Edited by: Christen J, Jagadish C, Look DC, Yao T, Bertram F. Materials Research Society; 2007:359–365.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.