Piezoelectric and opto-electrical properties of silver-doped ZnO nanorods synthesized by low temperature aqueous chemical method

In this paper, we have synthesized Zn1-xAgxO (x = 0, 0.03, 0.06, and 0.09) nanorods (NRs) via the hydrothermal method at low temperature on silicon substrate. The characterization and comparison be ...


I. INTRODUCTION
The semiconductor zinc oxide (ZnO) has gained a lot of interest in the research community.ZnO is a group II-VI compound semiconductor with excellent thermal and chemical stability, has a relatively large excitonic binding energy (60 meV) and a direct wide band gap (3.37eV) at room temperature.In addition to its semiconducting properties, and due to the inherent crystal structure, ZnO also possesses strong piezoelectric properties. 1 These characteristics make ZnO a material to be suitable for applications in electronic, optoelectronic, electrochemical, and electromechanical devices such as light-emitting diodes, 2 photodetectors, 3 photodiodes, 4 gas sensors, 5 solar cells, 6 piezoelectric transducers 7 and so on.Modification of ZnO properties by impurity incorporation is currently another important issue for possible different applications.Doping in ZnO with selective elements offers an effective method to adjust their electrical, optical, magnetic and piezoelectric properties, which is crucial for their practical applications. 8For ZnO, silver (Ag) is a good candidate for adjusting its optical properties.Ag ions can act as acceptors in ZnO, existing on substitutional Zn sites or in the interstitial form. 9,10In addition, in Ag doped ZnO, the location of the acceptor level remains contentious. 9,10n general, the physical properties of ZnO are closely connected with the deposition method, deposition parameters, annealing treatments and doping.Due to these factors, the doping has been widely used to adjust the structural, electrical and optical properties of ZnO thin films. 10Ag doped ZnO thin films have drawn considerable attention in many studies, e.g. as reported in Refs.11-16.1][12][13][14][15][16] In general all these Ag doped ZnO thin film experiments were concentrated on the electrical or opto-electrical properties, but to our best knowledge, no study has been published on measuring the piezoelectric effect of Ag doped ZnO.Based on the surface roughness and the electrical resistivity, S-H Nam et al claimed that Ag doped ZnO thin films deposited by radio frequency magnetron sputtering are not appropriate for piezoelectric devices because the films have poor crystallinity and low resistivity; interestingly, they did not measured the piezoelectrical properties of their films. 17Moreover, the published papers on Ag doped ZnO nanostructures are so far very limited. 18,19Among the many possible ZnO nanostructures, well-aligned nanorods (NRs) have attracted increased interest for many applications.Un-doped well aligned ZnO NRs have been grown on many substrates using both physical as well as chemical methods. 20Recently, aqueous chemical growth (ACG) methods used to synthesize metal oxide nanostructures have become very popular among researchers. 21,222][23][24][25] It will be of great interest to investigate and control the incorporation of different impurities in ZnO nanostructures using the low temperature chemical methods.
In the present work, we report on the structural, optoelectronic and piezoelectric properties of Zn 1−x Ag x O (x = 0, 0.03, 0.06 and 0.09) NRs synthesized using the low temperature ACG approach.The structural properties were investigated using scanning electron microscopy (SEM) and powder x-ay diffraction (XRD).The chemical state of Ag in the Zn 1−x Ag x O was investigated by an X-ray photoelectron spectroscopy (XPS).The influence of the Ag doping on the optical properties of the different samples was investigated using UV-VIS spectrometer.Finally, the piezoelectric properties were investigated using direct and inverse nano-indentation measurements.

II. EXPERIMENTAL PROCEDURE
At first, silicon substrate coated with silver was cleaned by sonication in acetone, deionized water, and isopropanol, respectively.Then, the substrate preparation technique developed by Green et al 26 was used to improve the quality of the grown nanorods.For the growth of the Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09), an equimolar concentration (0.075 M) of hexamethylenetetramine (HMT), and a mixture of zinc nitrate hexahydrate and silver nitrate solutions were prepared and mixed together.The different Ag concentrations were obtained by mixing different volume ratios of the zinc nitrate hexahydrate and silver nitrate.Next, the prepared solution was poured in a beaker and the pre-treated substrates were immersed in the solution with the growth side facing downward.After that, the beaker was sealed and heated in a laboratory oven at 90 o C for 6 hours.Then, the growth beaker was allowed to cool down to room temperature.Finally, after the growth process, the samples were rinsed with deionized water and dried with flowing nitrogen in order to remove the residual salts.
A Schottky contact on the Zn 1−x Ag x O NRs grown on silicon substrate (x = 0, 0.03, 0.06 and 0.09), was achieved using gold (Au) contact on top of the nanorods.The crystal structure of the pure and Ag-doped ZnO NRs arrays grown hydrothermally were investigated by Powder x-ray diffraction (XRD) Philips PW 1729 diffractometer equipped with Cu-Kα radiation (λ = 1.5418Å).The surface morphology and physical parameters were measured using Field-emission scanning electron microscopy (FE-SEM) Gemini LEO 1550.The influence of the Ag doping on the optical properties of the different samples was investigated using a PerkinElmer Lambda 800/900 spectrometer part no.BV900ND0.The direct and converse piezoelectric properties tests were performed by a nanoindentation technique using a Hysitron IT-950-Triboindenter at room temperature, as described in details in Refs.27 and 28.  were measured using the same acquisition parameters and hence are comparable.Three diffraction peaks are well consistent with the hexagonal phase of diffraction peaks of ZnO, and in agreement with the JCPDS Card No. 36-1451 file.In addition to the (111) Ag diffraction peak (from the substrate), the x-ray diffraction spectra of the different Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09), samples show diffraction peaks corresponding to the ZnO (002), (100) and (101) planes.The (002) reflection peak is intense and sharper in nature, as compared to other peaks, indicating a preferential c-axis growth orientation all the NRs.Nevertheless its intensity is decreasing with increasing the Ag concentration.Increasing the concentration of Ag leads to decrease the crystal quality of the ZnO NRs, in agreement with previous results reported by Xu et al. 10 002) peak.This shift indicates that the Ag ions replaced the Zn sites in the ZnO NRs crystal matrix.Additionally and because the radius of Ag 2+ ion (1.22 Å) is a greater than that of Zn 2+ ion (0.72 Å), the increase of the number of Ag 2+ ion the Zn ions lattice sites contraction the lattice parameter. 13The values of full width half maximum (FWHM) and (002) intensity for the different Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09) samples are shown in Figure 1(c).It can also be seen that FWHM is increased by the increasing the Ag dopant concentration.The FWHM change with the Ag incorporation demonstrates that the crystallinity of the NRs decreases with increasing the Ag concentration, indicating that large amount of Ag atoms may inhibit the c-axis preferential growth of the ZnO NRs. 15 SEM images of the different Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09) are shown in Figure 2. It is clear that all the NRs grown samples have a hexagonal structure with uniform, well aligned c-axis oriented nature.The average diameter is between 100-150 nm and the approximate length is 1µm.As can be seen from these SEM images, the addition of the Ag does not affect the morphology: neither the size nor the spatial distributions of the NRs have been altered by the presence of the Ag in the ZnO matrix.

B. Electronic structure characterization
X-ray photoelectron spectroscopy was used to investigate the charge state and chemical composition of Ag in the Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09).Figure 3 shows the XPS spectra of the Ag 3d peaks (Ag3d 5/2 and Ag3d 3/2 ) for the pure ZnO (pink), Zn 0.97 Ag 0.03 O (red), Zn 0.94 Ag 0.06 O (blue) and Zn 0.91 Ag 0.09 O NRs (black).For the three doped samples, the XPS signal from Ag 3d photoemission was substantially detected, for instance the peak of Ag 3d in the Zn 0.91 Ag 0.09 O NRs sample appears at around 368.12 eV and 374.17 eV for Ag 3d 5/2 and Ag 3d 3/2 , respectively.Therefore, these spectra clearly confirm the incorporation of Ag into the ZnO crystal lattice.0][31][32] Besides the appearance of Ag in the doped film, there is a shift of Ag 3d peak upon the different doping levels.From Zn 0.97 Ag 0.03 O NRs to Zn 0.91 Ag 0.09 O NRs, the peak position of Ag3d 5/2 downshifts to the lower binding about 0.1 eV, as marked by the dashed line in Figure 3, where all the intensity were normalized to show the change of peak position.Such behaviors could indicate that there is more metallic properties of Ag in the higher doped sample than lower doped sample.

Optical properties
The optical band gap of the Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09) samples was determined using Tauc method.From this method, the (αhν) 2 plot versus hν for ZnO is as shown in  Figure 4, and according to the equation: 18,33,34 (αhν where α is the optical absorption coefficient of the material, hν is the photon energy, A is a constant coefficient, Eg is the optical band gap, and the exponent r depends on the nature of the nature of the transition of the material.Band gap narrowing upon doping is a well-known general phenomenon in semiconductors, not just in ZnO.Shallow level donor impurities create energy levels in the band gap near the conduction band edge and shallow acceptor impurities create energy levels near the valence band edge.With increase in the amount of doping, the density of states of these dopants increase and form a continuum of states 8,9 just like in the bands, and effectively the band gap decreases.In this work, the values of optical band gap for the Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09) samples are shown in Figure 4, calculated by the extrapolation method.The values obtained were 3.30 eV (black curve), 3.26 eV (red curve), 3.22 eV (orange curve) and 3.17 eV (blue curve), for the different Zn 1−x Ag x O NRs (x = 0, 0.03, 0.06 and 0.09) samples, respectively.As it can be seen, the optical band gap decreased by increasing the amount of the Ag doping.This might be an indication that the Ag has substituted the Zn in the lattice.This is consistent with the observation from the x-ray results discussed above.So we can conclude that the optical band gap of Ag doped ZnO nanostructures is strongly dependent on the lattice sites of Ag in the ZnO.Additionally, the interaction of Ag states with the ZnO host states resulted in creating energy levels in the band gap that leads to reduce the optical band gap.

Piezoelectric properties
a. Direct piezoelectric effect.A Schematic diagram of nanoindentation instrument used for measurement of converse piezoelectric under the applied voltage shown in Figure 5(a).shows the relation between the maximum applied force and the generated piezo potential at the point of maximum applied force for the four different samples.Among different reported results, A. Khan et al investigated ZnO nanowires (NWs) grown on conductive fabric.The generated output potential was up to 13 mV at 1000 µN applied load. 35M. Hussain et al obtained up to 3 mV at 400 µN applied load for ZnO NRs sample synthesized on FTO. 36Here from our samples, it's observed that when the applied load was increased up to 160 µN, the resulting piezoelectric potential consistently increased for the four samples.In addition, while piezoelectric potential values of up to 7 mV were generated for pure ZnO NRs, output values of 2, 1.9 and 1. b. Converse piezoelectric effect.The converse piezoelectric effect is measured by applying a DC voltage in the range 0 to -40 V while there is a relatively low applied force of 15 µN applied to the samples to allow the tip to be always in physical contact with the NRs; as shown in Figure 6(a).
In order to evaluate the performance of any piezoelectric material clamped to a substrate, the most important parameter to calculate is the effective piezoelectric coefficient d eff 33 .In the case of NRs, this coefficient is directly related to the change of the longitudinal elongation ∆l when the NRs are subject to a change of the applied voltage ∆V in their c axis direction: d eff 33 = ∆l/∆V. 27,28,37he converse piezoelectric effect d eff 33 , of technological importance for the design of devices, can be related to the "true" piezoelectric coefficient d 33 of bulk material by the following relationship: 27,28,37 where s 11 , s 12 , and s 13 are the mechanical compliances of the piezoelectric NRs.
From the results are shown in Figure 6(b)-6(e) we can estimate the value of d eff 33 to be ∼ 130 pm/V for the pure ZnO NRs sample.A piezoelectric coefficient value of 33.2 pmV −1 from ZnO NWs grown on conductive fabric substrate was reported by A. Khan et al. 35 Additionally, the value of the d eff 33 for the Zn 0.97 Ag 0.03 O, Zn 0.94 Ag 0.06 O and Zn 0.91 Ag 0.09 O NRs were about 8.75, 6.25 and 4 pm/V, respectively.Figure 7 illustrate the substantial decrease of the piezoelectric response with the addition of Ag.
Because of the non-central symmetric feature in the ZnO wurtzite structure, the cations and anions are tetrahedrally coordinated and the centers of the positive ions and negative ions overlap with each other under strain-free conditions.When an external stress is applied, the centre of the cations and anions is displaced, and this produces a non-zero dipole moment.A constructive sum of these dipole moments results in a macroscopic potential, which is the origin of piezoelectricity. 38n some materials, like AlN, the doping with one foreign element can increase the piezoelectric response, as demonstrated in Sc-doped AlN thin films. 28On the other hand, in ZnO thin films, the crystalline quality decreased considerably with increasing Ag atom%. 17By introducing a very small doping amount, the effect will be such that; this dopant addition distorts the unit cell of the ZnO crystal.If the doping level is further increased, there will be a high deformation of the crystalline structure, leading to loss of the symmetry and decreasing thereafter the piezo response.Our optical and structural results are consistent and similar to those reported in Ref. 17, nevertheless, we have not observed any change of morphology of Ag doped ZnO nanorods.

IV. CONCLUSION
In summary, silver (Ag) doped ZnO NRs have been successfully synthesized on silicon substrate.Structural characterization (x-ray and SEM) indicated that the morphology of the NRs was preserved while growing ZnO NRs samples up to 9% atom fraction of Ag doping concentration.The increase of Ag concentration result in creating an energy donor levels in the band gap which leads to reduce the optical band gap.XPS data clearly shows the incorporation of Ag into the ZnO crystal lattice.The direct and converse piezoelectric properties of highly c-axis oriented ZnO and Ag doped ZnO NRs grown by low temperature ACG method on silicon substrate were measured and analyzed by the nanoindentation technique.The value of the piezoelectric coefficient d eff 33 was found to decrease from 130 pm/V to 8.75 pm/V for the pure ZnO NRs and Zn 0.97 Ag 0.03 O, respectively.Upon further increase of the Ag fraction the piezoelectric coefficient has slightly decreased further.When the nanoindenter is used to measure the direct piezoelectric effect, the piezoelectric potential generated values from the pure sample of ZnO NRs and for the Zn 0.97 Ag 0.03 O were decreased from 7 mV to 2 mV, respectively.These results indicated that, even preserving the crystallinity and electrical resistivity of the Ag doped ZnO NRs, the material is not suitable for piezoelectric device applications.
FIG. 1.(a) XRD patterns all the Zn 1−x Ag x O NRs grown on silicon substrate (x value is as indicated).(b) The XRD patterns of the (002) diffraction peaks.(c) The FWHM and the 002 peak intensity as a function of doping concentration.

FIG. 2 .
FIG. 2. SEM image of all the silver doped Zn 1−x Ag x O NRs grown on silicon substrate (x value is as indicated).
Figure 1(b) shows the XRD patterns of the (002) diffraction peaks for all samples.In this Figure there is also a clear slight shift of the angular position of the (

077163- 5 Nour
FIG. 3. XPS spectrum of Ag 3d peaks for the Zn 1−x Ag x O NRs grown on silicon substrate (x value is as indicated).All XPS peaks were normalized.The dashed line indicates the change of the Ag peak position with doping.

FIG. 5 .
FIG. 5. (a) Schematic diagram of nanoindentation instrument used for measurement of direct piezopotential.(b) Generated piezoelectric potential as a function of applied load.

FIG. 6 .
FIG. 6.(a) Schematic diagram of nanoindentation instrument used for measurement of converse piezoelectric under the applied voltage.Penetration depth as a function of time (6 < t < 10) of: (b) ZnONRs (c-e) Zn 1−x Ag x O NRs grown on silicon substrate (x value is as indicated) for switch from V = 0 to V = −40V.
4 mV were generated form the Zn 0.97 Ag 0.03 O, Zn 0.94 Ag 0.06 O and Zn 0.91 Ag 0.09 O NRs, respectively.It is clear that the addition of Ag dopant decreases substantially the generated piezoelectric voltage.Missing points in the Figure for samples with Zn 0.94 Ag 0.06 O and Zn 0.91 Ag 0.09 O NRs were for points that did not generate voltage, probably due to a short circuit.