- Nano Express
- Open Access
Computational Analysis of the Optical and Charge Transport Properties of Ultrasonic Spray Pyrolysis-Grown Zinc Oxide/Graphene Hybrid Structures
© Ali and Hashim. 2016
- Received: 17 February 2016
- Accepted: 4 May 2016
- Published: 12 May 2016
We demonstrate a systematic computational analysis of the measured optical and charge transport properties of the spray pyrolysis-grown ZnO nanostructures, i.e. nanosphere clusters (NSCs), nanorods (NRs) and nanowires (NWs) for the first time. The calculated absorbance spectra based on the time-dependent density functional theory (TD-DFT) shows very close similarity with the measured behaviours under UV light. The atomic models and energy level diagrams for the grown nanostructures were developed and discussed to explain the structural defects and band gap. The induced stresses in the lattices of ZnO NSCs that formed during the pyrolysis process seem to cause the narrowing of the gap between the energy levels. ZnO NWs and NRs show homogeneous distribution of the LUMO and HOMO orbitals all over the entire heterostructure. Such distribution contributes to the reduction of the band gap down to 2.8 eV, which has been confirmed to be in a good agreement with the experimental results. ZnO NWs and NRs exhibited better emission behaviours under the UV excitation as compared to ZnO NSCs and thin film as their visible range emissions are strongly quenched. Based on the electrochemical impedance measurement, the electrical models and electrostatic potential maps were developed to calculate the electron lifetime and to explain the mobility or diffusion behaviours in the grown nanostructure, respectively.
- Molecular orbital (MO)
- Graphene oxide
- Spray pyrolysis
- Zinc oxide
- Density functional theory
Two-dimensional (2D) sheet of sp2-hybridized carbons known as graphene has attracted great attention because of its exceptional optical, electrical, chemical and mechanical properties that impose promising ability for developing new generation of functional nanomaterials for various applications [1–3]. An ideal graphene nanosheet is found to possesses more light transmittance, flexibility and conductivity than indium tin oxide or single-wall carbon nanotubes for flexible transparent conductor or electrode applications [2, 3]. Lately, various methods are reported for growing graphene in large area size. In particular, chemical vapour deposition (CVD) is considered as the most common method for preparing high-quality large-area graphene due to the excellent controllability of thicknesses of the grown layers.
Concerning the targeted applications, there have been extensive efforts to combine the unique properties of graphene with metal-oxide nanostructures, namely, zinc oxide (ZnO) nanostructures to realize a novel hybrid structure for new generation of electronic, optoelectronic and photovoltaic applications [1, 4–8]. For instance, nanorods (NRs) and nanowires (NWs) as prolonged nanostructures have privileges over other structures. The optical reflectance properties of NRs are much better than thin films, thus significantly its absorption increases, which is particularly interesting for photovoltaic and photon-induced hydrophillicity applications.
Recently, intensive works have been conducted in developing ZnO/graphene hybrid structures either by vapour-phase [9–12] or liquid-phase techniques [13–17]. Recently, we report the evolution of ZnO nanostructures grown on graphene using a low-temperature ultrasonic-assisted spray pyrolysis technique [18, 19]. The effects of pyrolysis parameters, i.e. growth temperature, precursor injection/growth time, precursor molarity and precursor flow rate, on the grown structures were investigated. The growth modelling and process optimization was carried out to explain the observed evolution of ZnO nanostructures. The responses, i.e. structure density, structure shape factor and structure size, were evaluated. The modelling and optimization of the ultrasonic spray pyrolysis parameters for the growth of ZnO nanostructures on graphene layer using the response surface methodology (RSM) method were discussed. In this article, we report the computational analysis of the measured optical and charge transport properties of the spray pyrolysis-grown ZnO nanostructures. Most of the literatures regarding ZnO/graphene hybrid structures have mainly focused on the discussions of the experimentally measured morphological, structural and optical properties without any well-established computational analysis to explain the measured optical and charged particle transport characteristics [5, 7–9, 14, 20].
Deposition parameters and structure properties
Run (R i)
Flow rate (ml/min)
Deposition time (min)
All experimental runs were carried out under a fixed pressure of 35 × 10−4 mbar. Nitrogen gas was used for scavenging any volatile contaminant before spraying took place as well as to act as a carrier gas. Morphological and element compositional characterization was carried out using field-emission scanning electron microscopy (FESEM) equipped with energy-dispersive X-ray spectroscopy (EDS) facilities.
The optical absorption properties of the grown nanostructures were measured using UV-vis spectroscopy (Perkin Elmer/Lambda 35). Furthermore, the photoluminescence properties of the ZnO/graphene nanostructures excited at a wavelength of 335 nm at room temperature were recorded using fluorescence analyser (WITec Alpha300 M). The setup of this instrument combines the advantages of confocal and near-field optical microscopy together. Moreover, in order to investigate the possible structural defects that affect the absorption properties of the obtained structures under UV light excitation, a time-dependent density functional theory simulation (TD-DFT) was performed. The ωB97X-D functional was used to calculate the UV-vis spectrum for a variety of possible structural defects in ZnO lattice grown on single layer graphene. The Spartan 14 quantum chemistry package (Wavefunction, USA) was used to perform all calculations in this study [21, 22]. Equilibrium geometries were optimized by the B3LYP density functional method using the 6-311G** basis set; the developer of Spartan chose the Gaussian exponents for polarization functions to give the lowest energies for the modelled molecules. The polarization of the s orbitals on hydrogen atoms is crucial to accurately describe the bonding in acetylacetonate systems, particularly the hydrogen bonding. Furthermore, the 6-31G** basis set provides the p-type polarization functions for hydrogen. This can improve the total energy of the system along with the results for systems with large anions and can impose more flexibility [21, 22]. Zn-containing structures were also optimized with larger basis sets and higher levels of theory , where all correction energies were calculated using the 6-311G**, 6-311++G** and 6-311++G(2df,2pd) basis sets. The calculations involving anions and absolute acidity need to be carefully treated especially in selecting the basis sets since the excess electrons are weakly coupled to specific atoms or groups of atoms. The basis sets should provide diffuse s- and p-type functions on non-hydrogen atoms. This is usually designated by the ‘+’ sign, as in 6-311++G**. The second ‘+’ sign indicates that a diffuse function is added to hydrogen [21, 22].
Here, the first summation is over nuclei A. Z is the atomic number, and R AP is the distance between the nuclei and the point charge. The second pair of summation is over basis functions, φ i. P is the density matrix, and the integral reflects the Columbic interactions between the electrons and the point charge. r p is the distance separating the electron and the point charge. A surface for which the electrostatic potential is negative (a negative potential surface) delineates regions in a molecule which are subject to electrophilic attack. The bonds will be made to centres for which the spin density is the greatest .
The electrochemical impedance spectroscopy (EIS) measurements were performed to study the transport of charged particles through the synthesized ZnO/graphene hybrid structures. A platinum electrode was used as a current electrode (CE), and glass electrode filled with 3 mol/l KOH reference electrolyte solution was used as a reference electrode (RE). The specimen (0.5 × 1.0 cm2) was fitted onto copper sheet and connected to the working electrode (WE) clamp. The three electrodes were then plugged to Autolab potentiostat (PGSTAT128N/FRA32M, Metrohm). Zinc nitrate solution was used as an electrolyte with 0.4 M concentration. A low-amplitude sinusoidal excitation signal (voltage range of −0.015 to 0.015 V) was then introduced to the cell at frequency range from 0.1 to 105 Hz.
As presented in , by changing the flow rate, growth time, substrate temperature and molarity, diverse groups of nanostructures in terms of shape, size and density were able to be grown. As discussed in the same report, the evolution of ZnO structures was well explained by our developed modelling approach which enables a precise prediction on the structure to be grown.
The change of the energy bandgap in the as-grown ZnO nanostructures can be used to prove the rearrangement of the band structure and to provide information on the stress [19, 25]. Here, the upward shift in the energy bandgap corresponds to the occurrence of compressive stress in the crystal, whereas the narrowing of energy levels is found to be resulted from the residual tensile stress. The shift in the optical phonon mode of the ZnO nanoparticles indicates the effect of stress on the wurtzite structure. All these studies indicate that oxygen content in addition to the residual stresses can affect the optical properties of ZnO nanostructures [19, 25].
Here, Abs is the absorbance and t is the ZnO array thickness. The Tauc plot was generated by plotting (αhν) 2 vs. (hν) as shown in Fig. 3. The band gap was estimated by the extrapolation of the linear part of the curve at (αhν)2 equals to zero. It can be seen in Fig. 3a for the NW-based heterostructure that the direct band gap was found to be 3.40 eV, while for the NR-based heterostructure, the band gap was reduced to 2.90 eV as shown in Fig. 3b. Furthermore, the band gap for the NSC-based structure was found to reduce to 3.78 eV as shown in Fig. 3c, which could be attributed to the narrowing of the gaps between the energy levels caused by the red shift effect in correlation with the existence of lattice defects, i.e., oxygen vacancies. The estimated band gaps were confirmed to be in good agreement with the published results [8, 19].
The reason of mismatch is attributed to the ability of the DFT calculations to take into account the spinning condition of electrons at excited states for complex oxides. This is mainly because of the redundancy of iteration in the density matrix caused by the spinning direction vector. Here, the percentage of similarity is calculated as the ratio of the area under both calculated curve (red) and measured curve (blue). On the other hand, the average peak position matching ratio between TD-DFT calculations and experiments is found to be around 97.5 %. Thus, the presented atomic models resulted from this simulation could be considered valid for explaining the structural defects that lead to the shown absorbance behaviour under UV light.
As shown in Fig. 4a, the ZnO NW-based structure contains excessive amount of O atoms (red spheres) in its lattice as compared to the perfect wurtzite structure . It can be seen that the side Zn atom (green spheres) at the ZnO branch is covalently bonded to three extra O atoms with single bond (denoted by dashed circle). Besides, the top edge Zn atoms of the branch are singly bonded to three O atoms per each (denoted by dashed oval). Meanwhile, by considering the ZnO NR-based structures, it involves less content of O atoms as compared to the ZnO NW structure as shown in Fig. 4b. However, by comparing it to the perfect ZnO wurtzite, NR structure still contains three more O atoms (denoted by dashed oval) as can be seen in Fig. 4b. Finally, the ZnO NSC arrays show lack of O atoms inside the ZnO lattice as shown in Fig. 4c. It was reported in literature that O defects have great influences on the optical properties of the ZnO structures , and controlling these defects can lead to remarkable enhancement in the optoelectronic properties of the material especially for the photocatalytic ability [27, 28].
It is found that S1 is the state where the charges travel from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) as a result of excitation. As can be seen in Fig. 5, the LUMO orbitals are localized at the graphene matrix while the HOMO orbitals are mainly localized at the free end of the ZnO branch. The single bonded O atoms at the side of the branch also have LUMO orbitals. Thus, the distance to be travelled by electrons seems to be long and that is the main cause of the large band gap.
Concerning S2, it is the state where the electrons need to travel from HOMO-1 to LUMO after being subjected to an excitation that exceeds 3.9 eV. It is shown in Fig. 5 that the HOMO-1 orbitals are localized on the ZnO branch rather than the graphene matrix; this makes the distance travelled by electrons becomes shorter. However being HOMO-1 electrons make it higher in energy than HOMO. Similar scenario repeats with S3 and S4; however, in this case, the HOMO-2 orbitals are distributed much better all over the structure which enhances the transfer of excited atoms. Such enhancement in orbital distribution at HOMO-1 and HOMO-2 is the main cause of the narrowing of the gaps between energy levels which is caused by process-induced stresses and lead to a gradual spectrum as shown in Fig. 4c.
It can be seen in Fig. 6 that HOMO-1 and HOMO-2 are delocalized from the graphene matrix. This can be observed from the red and blue areas that exist only at graphene matrix in atomic model, corresponding to energy levels of LUMO, LUMO + 1 and HOMO, while disappear in other energy levels in the graphene matrix. Due to the proper distribution of the LUMO and LUMO + 1 all over the heterostructure provided by the graphene layer, it allows the transfer of charges during the excitation states of S3, S4 and S5 with energy barrier less than 4 eV. Thus, it can be said that the localization of LUMO at graphene matrix lead to a better charge transfer from graphene side to ZnO end which results in the reduction of the band gap and the increase of the lifetime of the transporting charge. Consequently, the charge and hole recombination rate is expected to be enhanced.
Finally, it can be understood from the spectra shown in Fig. 7 that ZnO NWs and NRs exhibit better emission behaviour under UV excitation as compared to ZnO NSCs and thin film where their visible range emissions are strongly quenched.
Charged Particle Transport Properties
By comparing the Nyquist plot (Fig. 8 a) to the Bode plot (Fig. 8 b), it can be understood that at low frequency range, the electrons transfer via Randles-like behaviour at the graphene/electrolyte interface as represented by a constant phase element of the Randles circuit component Y 0 as shown in the equivalent circuit (Fig. 8c). This conclusion is attributed to the semi-circle that appears in the Nyquist plot just prior to the linear part [30, 31]. The calculated lifetime (τ) of an electron through the graphene layer is found to be τ c1 = 1/2πf c1 = 3.18 ms , where f c is the frequency of the charged particle. At middle range of frequency, the electrons tend to transfer at the ZnO/electrolyte interface via semi-infinite length diffusion (represented by capacitance C 1 and resistance R 1 as shown in Fig. 8 c [23, 24]). The calculated lifetime of electron through the ZnO structure is found to be 265 ms at f c2 [30, 31]. The above performance could be more explained by paying attention to the electrochemistry of the measurements, i.e. investigating the reactions in which the charged particles (ions and/or electrons) cross the interface between different phases of matter, such as the interface between electrodes and electrolyte.
The diffusion through the sample and the electrode starts after the IHP plane is formed. At this stage, the charged particle transport is mainly controlled by electrostatic attraction, transport channel geometry and charge carriers concentration [32–34].
As can be seen in Fig. 10, the blue areas dominate the map and are located at the ZnO branches, while the population at the graphene layer is found to be minor. This difference in population can explain why the electron lifetime at the ZnO side of 265 ms can drop to be as low as 3.18 ms. It seems to show that the crowded ZnO channels have less mobile electrons than the graphene layer. Thus, the rate of diffusion towards the graphene is slow; however, after reaching the less crowded graphene surface, the charged particles can move much faster . Furthermore, the red and purple arrows shown in Fig. 10 indicate the long and narrow channels suitable for a finite length diffusion at the ZnO side which explain the Randles-like diffusion shown in the Nyquist plot of Fig. 8 a. While, the black arrows show wider and shorter channels suitable for semi-infinite length diffusion at graphene matrix.
However, as compared to the potential map of the ZnO NWs/graphene sample, the blue channels are getting smaller. On the other hand, the red iso-surfaces are getting bigger and thus the existence of negatively charged particles at the graphene layer is almost prohibited. This change in channel size clarify the reduction of electron life time at the ZnO side to be 159 ms and the increase of the life time up to 7.95 ms at graphene matrix as compared to the ZnO NWs/graphene. In fact, the reduction of the iso-surfaces of the high electrostatic potential (blue iso-surfaces at electrostatic potential map) indicates the reduction of the number of diffusing particles at the ZnO branch, thus increasing the mobility. The rate of diffusion towards the graphene is increasing, however after reaching the less crowded graphene surface, the charged particles can move much faster. This resulted from the localization of LUMO at graphene matrix which leads to a better charge transfer from graphene side to ZnO end. This results to the reduction of the band gap and the increase of the life time of the transporting charge as discussed in Fig. 6. Furthermore, the green arrows shown in Fig. 12 indicate the long and narrow channels suitable for a semi-infinite length diffusion at the ZnO side which explain the semi-infinite diffusion as shown by the Nyquist plot in Fig. 11a. While, the black arrows show the few separated spots available for finite length diffusion at graphene matrix.
At middle range of frequency, the electrons tend to transfer at the ZnO NSCs/electrolyte interface via finite length diffusion (represented by O element shown in Fig. 13c). The calculated lifetime of electron through the NSCs was found to be τ c1 = 1/2πf c1 = 19.89ms. At low range of frequency, electrons transfer through the graphene via hyperbolic tangent diffusion (represented by T-element shown in Fig. 13c.
On the other side, the rate of diffusion towards the graphene is totally prohibited due to the isolation of the charged particles that reach the sites of the ZnO branch by the dominating red areas. This can be attributed to the excited state of S1 as discussed in Fig. 5, where the charges travel from the HOMO orbital that localized at the free end of the ZnO branch to the LUMO orbital that localized at the graphene matrix. Thus, the distance to be travelled by electrons seems to be long and that is the main cause of the large band gap. In other words, the role of the graphene layer is eliminated in some way. Furthermore, the green arrows shown in Fig. 14 indicate the localized isolated channels suitable for a finite length diffusion at the ZnO side which explain the finite length diffusion band shown in the Nyquist plot in Fig. 13a. The blue arrows show few separated spots available for the Warburg infinite length diffusion at the ZnO terminal. Finally, the yellow arrows indicate the possible channel for a hyperbolic diffusion of negatively charged particles.
In this work, the computational analysis of the measured optical and charge transport properties of the spray pyrolysis-grown ZnO nanostructures was developed. The induced stresses in the lattices of ZnO NSCs that formed during the pyrolysis process seem to cause the narrowing of the gap between the energy levels. On the other hand, ZnO NWs and NRs show homogeneous distribution of the LUMO and HOMO orbitals all over the entire heterostructure. Such distribution contributes to the reduction of the band gap down to 2.8 eV, which has been confirmed to be in a good agreement with the experimental results. It was found that LUMO orbitals are distributed all over the ZnO branch in the NW and NR structures as a result of the excessive O atoms. Furthermore, ZnO NWs and NRs exhibited better emission behaviour under the UV excitation as compared to ZnO NSCs and thin film as their visible range emissions are strongly quenched. The electrostatic potential density map for the ZnO NWs show crowded ZnO channels with less mobile electrons than the graphene layer where the rate of diffusion towards the graphene was found to be small. However, after reaching the less crowded graphene surface, the charged particles move much faster. The electrostatic potential density map for the ZnO NRs indicates the reduction of the number of diffusing particles at the ZnO branch, thus increasing the mobility. The rate of diffusion towards the graphene is increasing; however, after reaching the less crowded graphene surface, the charged particles can move much faster. Finally, for the ZnO NSCs, the number of diffusing particles at the ZnO branch was drastically reduced, thus the mobility is dramatically increased. On the other side, the rate of diffusion towards the graphene is totally prohibited due to the isolation of the charged particles reach cites of the ZnO branch by the dominating red areas. In other words, the role of the graphene layer is eliminated in some way.
A.A. Amgad thanks the Malaysia-Japan International Institute of Technology for providing a scholarship. This work was funded by the Malaysia-Japan International Institute of Technology; Universiti Teknologi Malaysia; the Malaysian Ministry of Science, Technology and Innovation; and the Malaysian Ministry of Education through various research Grants.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Muszynski R, Seger B, Kamat P, Phys J (2008) Chem C 112:5263–5266Google Scholar
- Kim YJ, Hadiyawarman A, Yoon A, Kim M, Yi GC, Liu C (2011) Nanotechnology 22:245603View ArticleGoogle Scholar
- Kim YJ, Yoo H, Lee CH, Park JB, Baek H, Kim M, Yi GC (2012) Adv Mater 24:5565View ArticleGoogle Scholar
- Mali KS, Greenwood J, Adisoejoso J, Phillipson R, De Feyter S (2015) Nanoscale 7:1566View ArticleGoogle Scholar
- Song WT, Xie J, Liu SY, Zheng YX, Cao GS, Zhu TJ, Zhao XB (2012) Int J Electrochem Sci 7:2164Google Scholar
- Bai H, Li C, Shi G (2011) Adv Mater 23:1089–1115View ArticleGoogle Scholar
- Florianoa EA, Sambranoc JR, Scalvia LV, Geraldo V (2010) Mater Res 13:437View ArticleGoogle Scholar
- Hwang JO, Lee DH, Kim JY, Han TH, Kim BH, Park M, No K, Kim SO (2011) J Mater Chem 21:3432View ArticleGoogle Scholar
- Ahmad NF, Rusli NI, Mahmood MR, Yasui K, Hashim AM (2014) Nanoscale Res Lett 9:83View ArticleGoogle Scholar
- Biroju RK, Tilak N, Rajender G, Dhara S, Giri PK (2015) Nanotechnology 26:145601View ArticleGoogle Scholar
- Li Y, Zhao X, Zhang P, Ning J, Li J, Su Z, Wei G (2015) J Mater Chem C 3:4126View ArticleGoogle Scholar
- Sun H, Liu Y, Yu Y, Ahmad M, Nan D, Zhu J (2014) Electrochim Acta 118:1View ArticleGoogle Scholar
- Aziz NSA, Mahmood MR, Yasui K, Hashim AM (2014) Nanoscale Res Lett 9:95View ArticleGoogle Scholar
- Aziz NSA, Nishiyama T, Rusli NI, Mahmood MR, Yasui K, Hashim AM (2014) Nanoscale Res Lett 9:337View ArticleGoogle Scholar
- Ghosh D, Giri S, Moniruzzaman M, Basu T, Mandal M, Das CK (2014) Dalton Trans 43:11067View ArticleGoogle Scholar
- Mendoza-Sánchez B, Coelho J, Pokle A, Nicolosi V (2015) Electrochim Acta 174:696View ArticleGoogle Scholar
- Xie R, Fan G, Ma Q, Yang L, Li F (2014) J Mater Chem A 2:7880View ArticleGoogle Scholar
- Ali A, Hashim A (2015) Nanoscale Res Lett 10:452View ArticleGoogle Scholar
- Yingying LV, Yu L, Huang H, Feng Y, Chen D, Xie X (2012) Nanotechnology 23:065402View ArticleGoogle Scholar
- Faia PM, Furtado CS, Ferreira AJ (2005) Sensor Actuat B-Chem 107:353View ArticleGoogle Scholar
- Xu P, Tang Q, Zhou ZJ (2013) Nanotechnology 24:305401View ArticleGoogle Scholar
- Ali A, Hashim A (2015) Nanoscale Res Lett 10:299View ArticleGoogle Scholar
- Hehre WJ (2003) Theoretical models. In: Pople JA (ed) A Guide to Molecular Mechanics and Quantum Chemical Calculations. Wavefunction. Inc, USA, pp 385–445Google Scholar
- Wei-Qing L, Zhong-Guan L, Dong-Xing K, Lin-Hua H, Song-Yuan D (2013) Electrochim Acta 88:395View ArticleGoogle Scholar
- Sendi RK, Mahmud S (2013) J Phys Sci 24:15Google Scholar
- Wang T, Jiao Z, Chen T, Li Y, Ren W, Lin S, Lu G, Ye J, Bi Y (2013) Nanoscale 5:7552View ArticleGoogle Scholar
- Morkoç H, Özgur Ü (2009) Zinc oxide: fundamentals, materials and device technology. WILEY-VCH Verlag, Weinheim, p 1, ch. 1View ArticleGoogle Scholar
- Ivill M, Pearton SJ, Rawal S, Leu L, Sadik P, Das R, Hebard AF, Chisholm M, Budai JD, Norton DP (2008) New J Phys 10:065002View ArticleGoogle Scholar
- Guisheng L, Dieqing Z, Jimmy CY (2009) Phys Chem Chem Phys 11:3775View ArticleGoogle Scholar
- Archana PS, Naveen Kumar E, Vijila C, Ramakrishna S, Yusoff MM, Jose R (2013) Dalton Trans 42:1024View ArticleGoogle Scholar
- Wang X, Liu Y, Zhou X, Li B, Wang H, Zhao W, Huang H, Liang C, Yu X, Liu Z, Shen H (2012) J Mater Chem 22:17531View ArticleGoogle Scholar
- Zhang LL, Zhao XS (2009) Chem Soc Rev 38:2520View ArticleGoogle Scholar
- Liu R, Duay J, Lee SB (2011) Chem commun 47:1384View ArticleGoogle Scholar
- Sparreboom W, van den Berg A, Eijkel JCT (2010) New J Phys 12:015004View ArticleGoogle Scholar
- Molins S, Trebotich D, Steefel CI, Shen C (2012) Water Resour Res 48:1View ArticleGoogle Scholar