Nickel Doped Zinc Oxide Thin Films for Visible Blind Ultraviolet Photodetection Applications

The current research aims to investigate the effect of nickel doping on the structural and opto-electrical characteristics of zinc oxide thin films. Sol-gel spin coating technique has been utilized to deposit Zn1-xNixO (x = 0, 0.005, 0.010, and 0.015) films on glass substrates. X-ray diffraction (XRD) analysis confirms the formation of crystalline zinc oxide thin films with hexagonal wurtzite structure. Williamson-Hall analysis has been performed to study the individual contribution of lattice strain and crystallite size to the peak broadening in the XRD pattern. Scanning electron microscopy (SEM), Photoluminescence spectroscopy, and UV–visible spectroscopic techniques have been used to examine the surface morphology and optical properties of the deposited films. Transient photocurrent measurements have been performed on all the films under the exposure of ultraviolet (UV) light of wavelengths 365 and 254 nm with on/off cycle of 100 s, and various device key parameters such as sensitivity, responsivity, and quantum efficiency, etc have been determined. Sensitivities of the fabricated photodetectors (PDs) are found to be 5463%, 3809%, 3100%, and 831% for pristine ZnO, Zn0.995Ni0.005O, Zn0.99Ni0.01O, and Zn0.985Ni0.015O, respectively. The UV photodetection mechanism, which is based on the interaction between chemisorbed oxygen on the surface of ZnO and photo-generated holes, has been thoroughly discussed.

During the past few decades, plethora of global energy consumption has accelerated the research towards the development of sustainable and energy-efficient photodetectors. Recently, considerable attention has been paid to ultraviolet (UV) photodetectors (PDs) owing to their widespread implementation in the scientific, military, and medical sectors. [1][2][3][4] UV photodetectors are widely employed for human body protection. 5,6 Such detectors are necessary to reduce UV radiation exposure, which may harm the human beings and cause skin cancer, impair the immune system, and hasten the aging process. Although, a wide range of silicon-based UV photodetectors with high sensitivity, high signal-to-noise ratio, and fast response are available, but these detectors have some limitations, which include the need of visible and infrared light filters for spectral selectivity, a high degradation rate, and low efficiency, 7 etc. Implementation of wide bandgap semiconducting materials and the devices based on such materials have revolutionized various sectors of society, and the researchers, all over the world, are still working on these materials to enhance the performances of related devices. Because of their combined opto-electrical capabilities, semiconducting metal oxides are the potential materials for emerging optical and electrical applications. 8,9 Among them, zinc oxide (ZnO) is a wellknown metal oxide with excellent multifaceted properties like wide direct bandgap energy of 3.37 eV, existence in n-type semiconductor behavior, and exciton binding energy of 60 meV at room temperature. It is a useful material for UV photodetection applications because of its high radiation endurance, low toxicology, high transparency with electron mobility, low cost, and flexibility of fabrication via a number of preparation techniques. [10][11][12][13][14] Furthermore, pristine and doped ZnO thin films have a wide range of technological applications in light emitting diodes (LEDs), gas sensors, photocatalysts, solar cells, and other photonic devices. [15][16][17] Sol-gel 18,19 technique, chemical vapor deposition (CVD), 20 electron beam evaporation, 21 sputtering, 22 pulsed laser deposition (PLD), 23 and chemical bath methods (CBD) 24 are among the most common deposition processes used to fabricate the ZnO thin films. In fact, the photoelectric properties are greatly dependent on the fabrication process. It should also be noted that the key requirements for the fabrication of photodetectors is repeatability, adaptability, and low-cost production of materials. The electrical functionality of ZnO does not lose even in the solution state, which opens avenues to fabricate the film by simple and cost-effective chemical routes, viz., chemical bath deposition, sol-gel spin coating, etc. 25 All these processes are low-temperature and thin film can be deposited on a large scale on various substrates. We have employed a low-cost sol-gel spin coating approach to deposit the ZnO and Ni-doped ZnO thin films. Sol-gel spin coating is a unique and straightforward method of generating many types of nanostructures with control over deposition rate, temperature, growth level, and thickness, along with various other parameters. We have investigated the effect of various Ni doping concentrations (0, 0.5, 1, 1.5%) on the morphological, structural, optical, and photoelectric/or photodetection properties of nanostructured ZnO thin films.

Experimental Details
Zinc acetate di-hydrate (Hi-media) and Diethanolamine (Himedia) were used as precursors and sol stabilizer, respectively. Nickel acetate (Sigma Aldrich) was used as the source of nickel atoms. All the chemicals used were of analytical grade and were utilized without any additional purification. Equimolar solutions (0.4 M) of zinc acetate and Diethanolamine (DEA) were prepared by dispersing a suitable quantity of Zn(CH 3 COO) 2 .2H 2 O and DEA in 30 ml 2-propanol. The whole process of thin film deposition and device fabrication is shown in schematic Fig. 1.
The solution was stirred on a magnetic stirrer for 1 h before being left to age for 24 h. Similarly, three more solutions were prepared using 0.5, 1, and 1.5 molar percentage of nickel acetate. Glass substrates were degreased by ultra-sonication in detergent, hydrochloric acid, and acetone, alternatively. Finally, the glass substrates were washed with de-ionized water and dried in a hot air oven. Aged sol was spin-coated over a glass substrate at the rotation speed of 1500 rpm, and after each coating film was dried at 120°C. Five coats of each sol were deposited on different glass substrates. All the deposited films were annealed at 550°C in a tubular furnace in the ambient environment. For photocurrent measurement, devices were fabricated by depositing two parallel aluminum electrodes by the thermal evaporation technique.The chamber was kept at a vacuum of 10 −5 torr during the deposition. z E-mail: sajjandahiya1@gmail.com; rawalishpal@kmc.du.ac.in The phase composition and structure of annealed films were investigated using an X-ray diffraction pattern recorded on a Rigaku MiniFlex-600 diffractometer equipped with monochromatic Cu-K α radiation (=1.54056 Å) at a scan rate of 1°min −1 and a step size of 0.02°. The morphology of the deposited films was studied using a high-resolution scanning electron microscope (ZEISS model Evo 18) at a magnification of 50kX and accelerating potential of 20 kV. The elemental compositions were quantified using an energy dispersive X-ray spectroscope (EDAX make Oxford instruments equipped with X-act-10mm 2 SDD Detector) fitted with the afore mentioned SEM setup. The optical properties of the deposited films were investigated by taking UV-vis spectra with a Shimadzu UV-Vis spectrometer model UV 3600 plus. Photoluminescence spectra of the deposited films have been recorded over Horiba instruments made sophisticated fluorescence spectrometer (FL3C-21). All electrical measurements were done in MSM (metal-semiconductor-metal) geometry. Ultraviolet (UV) photodetection measurements were accomplished on Keithley's 617 electrometer interfaced with the Lab-View program through GPIB cable. Monochromatic UV sources of wavelengths 365 nm and 254 nm were used to illuminate the devices during these measurements. Figure 2 depicts the X-ray diffraction patterns of pristine and nickel-doped zinc oxide thin films. The formation of a hexagonal wurtzite-structured ZnO thin film is confirmed by the perfect similarity of the XRD spectra with JCPDS card # 80-0075. [26][27][28] All of the observable peaks can be correlated to the wurtzite hexagonal structure shown in Fig. 2. The absence of any extra peak in the XRD pattern of doped ZnO film confirms the incorporation of nickel atoms in the ZnO crystal lattice. It is obvious that the intensity of the peaks in the Ni-doped thin films is lower than that of the undoped (pure) ZnO film. This drop in the intensity of the diffraction peaks of the Ni-doped thin film shows that Ni doping has severely reduced the crystalline character of the ZnO thin films. The observed shift in XRD peaks can be explained in terms of small ionic radius of Ni 2+ as compared to Zn 2+ .

Results and Discussion
The following equations have been used to determine unit-cell parameters a, c, and volume in accordance with the wurtzite hexagonal structure: 29 Where d hkl is inter-planer spacing for respective miller indices, a and c are lattice parameters of the hexagonal crystal lattice. All the calculated parameters are approximately equal to as given in JCPDS card # 80-0075.The ionic radius of Ni 2+ is 69 pm, while that of Zn 2+ is 74 pm. Table I clearly shows that the value of lattice parameters continues to decrease due to the small ionic radius of Ni 2+ ions.
Williamson-Hall size-strain analysis.-The Williamson-Hall (W-H) technique may be used to analyze an XRD pattern to calculate the crystallite size and lattice strain. However, crystallite size can also be calculated by Scherrer's method but it is less  accurate as compared to the W-H method because it does not eliminate the effect of peak widening caused by lattice strain. The W-H approach individually addresses peak broadening caused by crystallite size (β s ) and lattice strain (β d ), providing a value for both that is substantially more accurate. 30 Here, β hkl is the full-width at half maximum (FWHM), k is constant having a value of 0.94, λ is the wavelength of X-ray, which is 1.54056 Å for Cu-K α , ε is lattice strain and D is crystallite size. Now, straight line fitting of 4sinθ vs βcosθ curve, may be used to directly determine crystallite size (D) and lattice strain (ε) as shown in Fig. 3.
SEM micrograms have been used to perform topographical examination of the deposited films. Figures 4a-4d shows SEM micrograph of the films deposited with 0, 0.5, 1 and 1.5% molar concentration of Zn 2+ ions. All the films have been gold coated before the SEM measurement to avoid charging of the surface and to stimulate the emission of secondary electrons.SEM micrographs of all the deposited films confirm the formation of the dense and uniform films. Insets of the Figs. 4a-4d show the EDS spectra of the pristine and nickel-doped films. EDS spectra confirm the doping of the ZnO films with Ni 2+ ions. However, the presence of the Si, Ca, Mg, Na, and Al elements in the EDS spectra could be attributed to the soda lime glass substrate. The appearance of gold in the EDS spectra of all the samples is due to the coatings of gold applied before the measurements.
Optical properties.-The optical characteristics of pristine and Ni-doped ZnO thin films become increasingly prominent when particle size is pushed to the nanoscale. UV-visible spectroscopy has been used to comprehensively evaluate the optical properties of pure and Ni-doped ZnO thin films. Since different dopants can produce different crystal defects in the ZnO crystal structure, the band gap of these compound semiconductor nanoparticles varies accordingly. Figure 5 depicts the UV-Visible transmission spectra of the pure and Ni-doped ZnO thin films deposited over the glass substrates. The transmission spectra of the deposited films have been recorded over the wavelength range from 300 to 800 nm. The deposited films are found to be almost transparent to the radiations in the 400 to 800 nm wavelength range with a transmittance greater than 96%. When compared to pristine ZnO, the absorption edge of Ni-doped ZnO thin films exhibits substantial blue shift. Because of the inclusion of Ni ions, the observed blue shift suggests a change in the band structure in Ni-doped ZnO films. Bandgap value for the films having Ni content 0, 0.5, 1 and 1.5% is found to be 3.41, 3.45, 3.46 and 3.48 eV, respectively. The observed trend in bandgaps with an increase in Ni content is in good agreement with many published reports. [31][32][33][34] Band gaps are observed to increase with an increase in the nickel concentration as shown in the insets of Fig. 5. The size effect is unlikely to be responsible for the observed blue shift because the diameters of the produced nanostructures are higher than the excitonic Bohr radius in ZnO. The band structure modification might be caused by the sp-d exchange interaction between ZnO band electrons and localized d-electrons associated with doped Ni 2+ cations. The widening of the optical energy band gap and the blueshift of the absorption edge can also be attributed to an increase in carrier concentration and are explained in theory by the Moss-Burstein band filling phenomena, which is commonly observed in ntype semiconductors. 35,36 The optical properties of the deposited films have been further investigated by Photoluminescence spectra in the wavelength range 400-550 nm, as shown in Fig. 6. Each PL spectrum of the pristine as well as the Ni-doped ZnO films consist of two sharp peaks around 413 and 437 nm and a broad peak around 464 nm. Strong UV peak around 413 nm apparently originates from recombination of free excitons. 37 The main feature is an evident decrease of the UV peak intensity with Nickel doping, which indicates a decrease in the recombination rate of the photogenerated carriers, and this may result from the doped Ni 2+ ions that act as traps to capture the photogenerated electrons or holes. Most likely, the low intensity visible emissions are linked with the vacancies in the singly ionised oxygen compounds that emerge from photogenerated recombination. It appears that additional defects in the nanoparticles sample may be generated as a result of the inclusion of dopants into the ZnO lattice. 38    To check the repeatability of fabricated devices five cycles of transient photoresponse have been recorded and it has been found that all the devices have a regenerative response, which is the main requirement for a photodetector. The charge carriers transit in zinc oxide are controlled by oxygen molecules adsorbed on the film surface, which withdraw electrons from the metal oxide semiconductor's conduction band to generate oxygen ions. This causes a drop in charge carrier concentration and is responsible for poor conductivity in metal oxides. The surface-adsorbed oxygen molecules at the grain boundaries produce a depletion zone that is susceptible to upward band bending or the creation of a potential barrier. The extent of band bending depends upon the width of depletion layer at the grain boundaries.
When exposed to UV light, the photogenerated electron-hole pairs are either adsorbed by the oxygen molecules or gathered at the electrodes. While electrons are responsible for increased photocurrent through a reduction in potential barrier height and the size of the space-charged region at the grains and grain boundaries, photogenerated holes interact with oxygen ions on the film surface. Thus, the degree of band bending or change in potential barrier height in UV light controls the photo-response in the metal oxide semiconductor films. [40][41][42][43] Parameters of Photodetector's Performance.-There are some parameters which are commonly used to assess a photodetector's performance. Here, we have calculated and discussed some of them.
Photoresponse factor (S) or sensitivity of a photodetector is the percentage change in current on UV exposure relative to dark current   Under the illumination of UV light (λ = 365 nm), the sensitivity of the deposited films was found to be 5463%, 3809%, 3100%, and 831% for pristine ZnO, Zn 0.995 Ni 0.005 O, Zn 0.99 Ni 0.01 O, and Zn 0.985 Ni 0.015 O, respectively. Thus, the photodetection measurements indicate that the sensitivity of ZnO thin films decreases with an increase in nickel content in the thin film. However, the measured value of sensitivity is much higher than the reported by Chu et al. 44 They have reported 71.45, 238.75, and 393.04% sensitivity for different Nickel contents (0, 0.4, and 0.8%) in ZnO thin films. The decrease in sensitivity could be due to a increase in the band gap of ZnO on increasing nickel percentage as reported by Owoeye et al.. 26 The observed sensitivity value of the present thin films towards the UV radiations is comparable or better than the previously reported literature values and the observed sensitivity values for the ZnO based devices by different research groups are presented in tabular form in Table II. Responsivity is another parameter used to measure the device's sensitivity toward the intensity of incident light. It is the ratio of obtained output current to the incident power density. Mathematically, it can be calculated by the following relation:   As soon as the device is exposed to UV light, it takes a while to attain its maximum photocurrent value; similar behavior is observed on switching off the light. The transient photocurrent of a photodetector can be expressed by the following equation Here, t o is the time constant and gives us an idea about the temporal response of the device. Transient photoresponse of the fabricated devices was also recorded under the illumination of UV light of wavelength 254 nm and shown in Fig. 8. Single cycle of photocurrent under the exposure of UV light (365 nm) of intensity 0.979 mW cm −2 was fitted exponentially (Fig. 9) using Eq. 8. Obtained values of rise and decay time are given in Table III. It is found that the prepared photodetectors have a comparatively long decay time when compared to their rising time. The direct band-to-band transfer of photo-generated charge carriers may be the primary cause of the short rise time. The large decay time may be due to the time taken by surface valances to re-adsorb the oxygen molecule.

Conclusions
In summary, the sol-gel spin coating process has been used to deposit pristine ZnO and Ni (0.5, 1, and 1.5 wt%) doped ZnO thin films on glass substrates. Crystalline forms with hexagonal wurtzite phase of the deposited films have been confirmed using XRD analysis. Crystallite sizes have been calculated using Williamson-Hall approach, which are found to be 14.83,