Fabrication and characterization of ultraviolet photosensors from ZnO nanowires prepared using chemical bath deposition method

One-dimensional nanostructures, such as nanorods, nanowires (NWs), nanotubes, and nanobelts, exhibit multifunctional properties, as well as possess unique features that are significantly different from those of their bulk counterparts.^1–3 For example, the properties of devices fabricated from nanoscale materials, such as carbon nanotube-based field-effect transistors,⁴ tin oxide-based chemical sensors,⁵ and zinc oxide (ZnO)-based optoelectronic devices,⁶ demonstrate significantly enhanced property/efficiency compared with conventional devices fabricated using their macrostructure counterparts. Although a wide variety of nanoscale materials are being investigated for application for a specific function, interest in growth and characterization of ZnO NWs has significantly increased.^7–13 One primary reason is that ZnO is a direct band semiconductor that demonstrates a band gap of $∼$3.37 eV and is potentially useful in various applications related to dye-sensitized solar cells,¹⁴ nanosensors,¹⁵ and nanogenerators,¹⁶ as well as in energy conversion and storage.¹⁷ 

In most of the aforementioned studies, ZnO NWs were mainly prepared using techniques such as vapor–liquid–solid (VLS)¹⁸ or chemical vapor deposition,¹⁹ both of which are high-temperature processes. Several investigators have also utilized the chemical bath deposition (CBD) technique to synthesize ZnO nanostructures. The CBD method is an inexpensive method that can be performed at moderate temperatures (∼95 °C). In fact, synthesis of desired nanostructures via CBD requires deposition of a suitable seed layer on specific substrates, which act as a nucleation layer prior to the growth process. However, the current state of the art technique uses expensive and/or cumbersome methods to deposit the seed layers needed for CBD. These methods include mixing ZnO with polyvinyl alcohol as a complex seed layer²⁰ or using other expensive techniques such as RF magnetron sputtering.²¹ These complicated and expensive processes can become a serious bottleneck in technology development, which requires cheap methods that lead to scale up production of nanomaterials. Therefore, significant attention has been devoted to discover simple ways to obtain efficient seeding materials for ZnO NW growth. For example, a layer of ZnO nanoparticles (NPs) prepared using zinc acetate dihydrate dissolved in ethanol serves as a good seed layer on a variety of substrates (e.g., silicon and glass).^22–24 Such a finding indicates that simple and highly efficient routes to develop and deposit seed layers for CBD processes are needed to bolster nanomaterial synthesis and future nanotechnology-based applications.

This paper presents a simple route to synthesize ZnO NWs by using commercially available ZnO NPs, which are spin coated over substrates to serve as seed layer. The NWs grown using this method demonstrate crystalline structure and are similar to ZnO NWs grown by using other methods in terms of physical and optical properties.

ZnO NPs (diameter, $∼$20 nm) were purchased from mkNANO. Hexamethylenetetramine (HMT, C₆H₁₂N₄) and zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O) were purchased from Fisher Scientific Company. These materials were used without any further purification and/or other treatments.

ZnO NPs were mixed with 2-propanol (IPA) (1.5 g/20 ml) and were dispersed by sonicating the solution for 30 min in a bath sonicator. A few droplets of the solution were deposited on the surface of a thoroughly cleaned silicon dioxide/silicon (SiO₂/Si) substrate (⁠$2×2 cm2$⁠) placed on a spin coater and then spun at 3000 rpm for 30 s to achieve uniform distribution of the NPs on the substrate. After spin coating, the SiO₂/Si substrate was annealed on a hot plate at 200 °C for 10 min to allow solvent evaporation and to improve adhesion of the NPs on the surface. Once the deposition of the seed layer was completed, a simple chemical bath with an equimolar ratio of HMT and Zn(NO₃)·6H₂O was prepared following the method described by Vayssieres et al.²⁵ with few modifications. The SiO₂/Si substrate was fixed vertically in a flask by using a Teflon holder. The whole assembly was placed in an oven for 16 h at 95 °C. Once the CBD process was completed, the SiO₂/Si substrate was removed from the oven and was left to dry under ambient conditions prior to any characterization. Figure 1 shows the schematic of the synthesis procedure.

To further investigate the structural properties of the as-prepared ZnO NWs, a sample for TEM was prepared by placing ZnO NWs on SiO₂/Si in IPA solution. The mixture was sonicated for 60 min, which yields a well-dispersed ZnO NWs in IPA. A few drops of the solution were placed on a TEM grid. The prepared solution was also used to evaluate the optical band gap of ZnO NWs.

To characterize the photoelectric transport of these NWs, a few NWs were transferred onto pre-patterned gold (Au) electrodes through drop casting (3–4 drops) ZnO NWs suspended in IPA. The distance between the electrodes is ∼750 nm, which is comparable to the length of the ZnO NWs and allows a few NWs to form a bridge between the two electrodes. Prior to measuring the electrical transport, a fabricated device (device 1) was annealed at 200 °C under high vacuum for 2 h to eliminate solvent from the surface of the NWs. Annealing also generally improves the quality of the contacts. Another device (device 2) was fabricated with ZnO nanowires, grown directly on commercially available, large area (mm × mm), inter-digitated gold electrodes by CBD (16 h, 95 °C) using the seeding method previously described. This allowed a huge number of wires to cross the gold inter digits creating a large area device.

A scanning electron microscopy (SEM) system (FEI Quanta FEG 450) was used to determine the surface morphology of the as-grown ZnO NWs. The nanostructure diameter and length were further investigated using a TEM system (Hitachi H-7650). X-ray diffraction (XRD) measurements were also performed at room temperature to reveal the crystal structure of ZnO NWs by using a GBC mini-material analyzer that employs Cu Kα radiation and Bragg–Brentano geometry. Ultraviolet-Visible Spectrophotometer (UV-Vis) system (HACH-DR/4000) was used to determine the optical band gap of ZnO NWs. In addition, optical microscopy at 50–1000× magnification was performed using a Nikon L150 microscope equipped with CFI60 optics to reveal detailed information on the fabricated devices prior to opto-electro measurements. A Renishaw InVia confocal Raman microscope equipped with 50× objective lens was used at 514 nm excitation to provide a laser spot size of 2 μm diameter for Raman spectroscopic characterization. The laser power used was maintained at 1.5 mW. Electrical transport was measured using Kietheley 2400 series current/voltage source meters.

The surface morphology of the grown ZnO NWs was investigated from the SEM and TEM images. Figures 2(a) and 2(b) show the SEM images (top view) of the ZnO NWs grown on SiO₂/Si through CBD. The SEM images revealed that the synthesized ZnO NWs with tapered tips are densely assembled over a large area. Although the NWs were tilted with respect to the substrate surface, extensive SEM investigation revealed that they were continuously distributed over the entire substrate. Figures 2(c)–2(e) present the results of TEM imaging, which revealed that the average length of the ZnO NWs is ∼1–1.3 μm. Furthermore, the TEM image shows that these NWs have different crystal orientations in areas indicated by arrows in Figure 2(c). The average diameter of NW distribution (Figure 2(f)) is 40–50 nm as shown in various microscopic images (TEM and SEM).

The optical band gap of these NWs was estimated using UV-Vis spectroscopy. Figure 3(a) presents the absorption spectroscopy of ZnO NWs in the UV-Vis spectrum region ranging from $∼320$ nm to $∼600$ nm. The optical band gap of ZnO NWs can be determined from the UV-Vis spectrum by using the following equation:²⁶ 

where $α$ refers to the absorption coefficient, $h$ is the Planck's constant, $ν$ is the frequency, $L$ is a constant, and $Eg$ is the optical band gap. The nature of the optical band gap of ZnO NWs can be anticipated from the value of the exponent $r$⁠. For materials with a direct band gap (e.g., ZnO), r = ½.²⁷ Thus, the extrapolation of the plot of (⁠$αhν$⁠)² (on y-axis) as a function of $hν$ (x-axis) can provide a good estimate of the band gap of the material under investigation. The inset in Figure 3(a) (also known as the Tauc plot) shows that the tangent line (linked to $α=0$⁠) meets the horizontal axis at a value of 3.61 eV, which corresponds to the optical band gap of ZnO NWs.²⁸ The obtained $Eg$ is larger than the bandgap of bulk ZnO with a significant shift of 240 meV. Many theoretical models, such as quantum confinement effect,²⁷ color centers,²⁹ free-exciton collision,³⁰ surface states,³¹ and Burstein–Moss (BM) effect,³² were developed to understand the cause of increasing band gap in NWs (also called blue shift of $Eg$⁠). Among these theories, the quantum confinement effect has often been associated with the shift of $Eg$ when the geometric dimensions of a material have the same order as the Bohr exciton radius (⁠$rB=2.34 nm for bulk ZnO)$⁠. However, microscopy analysis indicates that the diameters of ZnO NWs in the current study are larger than the value of exciton Bohr radius; thus, the increasing band gap value is not caused by the quantum confinement effect. In fact, many reports have shown that the band gap shift in ZnO NWs occurs at diameters far beyond the quantum confinement regime.^31–34 One possible reason for this shift is the BM effect, which is an important phenomenon in n-type semiconductors (e.g., ZnO). In the BM effect, when the carrier concentration is sufficiently high, the Fermi level (FL) moves into the conduction band, leading to an increased band gap.^32,34–36 In particular, a long duration of CBD growth (⁠$∼$16 h) causes oxygen ion consumption in the bath, leading to non-stoichiometric distribution of Zn and O atoms in the NWs structure.³⁴ The loss of oxygen in ZnO lattice induces the increase in carrier concentration, causing FL to move into the conduction band.^32,34–36 Therefore, the increasing optical band gap of the synthesized ZnO NWs presented in this study can be attributed to the BM effect.

The crystalline structure of the grown ZnO NWs was verified through XRD measurements. Figure 3(b) (top) shows the XRD pattern of the NWs. The diffraction peaks related to the (100), (002), (101), (102), (110), (103), and (201) planes were easily identifiable from the XRD pattern, which is similar to the pattern obtained from the as-received ZnO NPs (Figure 3(b), bottom). These peaks correspond to ZnO crystals that exhibit a hexagonal Wurtzite structure when compared with the standard data.³⁷ The relative intensity ratio between the (002) and (101) diffraction peaks is typically used to determine the orientation of ZnO NWs.³⁸ The NWs in this study are apparently aligned along the c-axis polar direction according to the intensity ratio obtained from ZnO NWs.³⁷ These results are consistent with the reports on ZnO NWs that exhibit a Wurtzite structure and are obtained using a similar growth method.³⁹ 

According to the group theory, the Wurzite ZnO belongs to the $C6v4$ space group. Wurzite ZnO has eight sets of optical phonon modes at the Γ point of the Brillouin zone. These modes can be classified as A₁ + E₁ + 2E₂ (Raman active modes), 2B₁ (Raman silent mode), and A₁ + E₁ (infrared (IR) active modes).^40,41 The multi-phonon process (2E₂ modes) splits into two Raman active modes, namely, E₂ (high) and E₂ (low), which are associated with the vibration of oxygen and heavy zinc atoms, respectively.⁴² Furthermore, the IR active modes (A₁ and E₁) separate into two compounds: transverse optical (TO) and longitudinal optical (LO) phonons. Figure 4 shows the Raman spectrum of the as-grown ZnO NWs on SiO₂/Si. The peaks were assigned to the following vibration modes of ZnO: 2E₂ (333.23 cm⁻¹), A₁ (TO) (381.56 cm⁻¹), E₂ (high) (438.6 cm⁻¹), and E1 (LO) (580.2 cm⁻¹). The existence of 2E₂ peak at $∼$333 cm⁻¹ indicates that the grown ZnO NWs possibly exhibit a strong crystalline order.⁴¹ In addition, the dominant peak intensity of E₂ (high) line at 438 cm⁻¹ can be associated with the hexagonal Wurzite ZnO NWs, which are oriented along the c-axis.⁴¹ The presence of E₁ (LO) generally indicates the presence of impurities and/or defects, such as oxygen vacancies, which typically occur during growth.⁴³ The peak located at $∼$520.2 cm⁻¹ belongs to the Si substrate on which these NWs are grown.

Figure 4 shows that the value of E₂ (high) peak for the synthesized ZnO NWs demonstrated a blueshift of 1.6 cm⁻¹ compared with the 437 cm⁻¹ for bulk ZnO.⁴⁴ The shift of E₂ (high) mode for ZnO NWs toward a high frequency is typically caused by the biaxial compressive stress within the c-axis^40,41 and can be estimated by using the following model:⁴⁵ 

where $Δω$ refers to the shift of E₂ (high) mode and $σ$ represents the value of stress. The blueshift in the ZnO NWs synthesized in this study is $∼$0.365 $GPa$⁠. The calculated value is less than those reported for ZnO film deposited on Si (0.9 < $σ$ < 9.99 $GPa$⁠).⁴¹ The previously reported smaller stress values for NWs grown using the VLS technique can be understood from the perspective of stress relaxation effect.^40,44,46 Similar processes are also possibly responsible for the reduced biaxial compressive stress in these NWs.

Current-voltage (IV) measurements were carried on at ambient condition to investigate the electrical properties of the fabricated UV photosensor (device 1, Figure 5(a)) in dark and under UV illumination. Figure 5(b) shows the IVs in both cases, which indicates non-linear behavior because of the Schottky barrier formed at the interface between Au and ZnO NWs. A significant increase in the source drain current was observed upon exposing the device to UV radiation (Figure 5(b)). Before UV light was applied, the measured current (I_dark) was observed to be very low. However, after exposure of the device to UV light, the current increased sharply (I_light), which implied a pronounced increase in conductance. The effective barrier height (EBH, $∅b)$ can be determined from the I–V characteristics by including series resistance $(Rs)$ as a factor that affects the electrical characteristics. The ideality factor (⁠$n)$⁠, which measures the deviation of the device from the ideal diode, can be determined using Cheung's function:⁴⁷ 

where $q$ is the electron charge, $k$ is the Boltzmann constant, $T$ is temperature in Kelvin, $A$ is the effective diode area, and $A*$ is the effective Richardson constant (⁠$32 A cm−2 K−2$ for bulk ZnO). By using Equation (3), and by plotting $∂V∂(lnI)$ as a function of $I$ (not shown here), the value of $n$ can be extracted from the intercept. This value can be used in Equation (4) to calculate $M(I)$⁠, which can then be plotted as a function of $I$ (not shown here) to determine the EBH from the intercept by using Equation (5). Table I summarizes the values of both parameters (⁠$∅b$ and $n$⁠) in the dark and under UV illumination in the low voltage region.

Under dark conditions, the value of the EBH was $∼$0.55 eV, which is comparable to the reported values for Au/ZnO Schottky diode based on ZnO NWs fabricated on different substrates.⁴⁸ EBH decreases to $∼$ 0.46 eV after UV exposure because of the large number of photogenerated carrier pairs, which leads to an increase in the majority carriers. Although the photogenerated carriers ionize the interface states, which results in an increase in barrier height, the large number of photogenerated electrons reduces the EBH.^49–51 

To further analyze the photoconductive properties, the chemical states of NW surface with and without UV radiation can be clarified based on the previous findings.^52,53 Under dark conditions, the surface of ZnO NWs absorbs oxygen from the atmosphere and forms a depletion layer, thereby producing negatively charged ions.^52,53 As a result, the absorbed oxygen molecules on the surface of the ZnO NWs trap some of the free electrons, whereas the mobility of the remaining electrons decrease because of the depletion layers created on the surface

The water vapor molecules absorbed by the surface of ZnO NWs also enhances the depletion layer. The water vapor molecules capture not only free electrons but also free holes, thereby further lowering the conductivity NWs.⁵³ Consequently, H₂O molecules significantly affect the conductivity more than O₂ molecules in ZnO NWs

Electron-hole pairs are generated (⁠$hv→h++e−)$ after applying UV light with a wavelength λ of $∼365 nm$⁠. The photogenerated holes are separated from the electrons by strong local electric fields,⁵⁴ which reduce the electron-hole recombination process, thereby increasing the carrier lifetime. Consequently, the conductivity increases because of the increase in carrier density. The holes then relocate to the surface and suppress the depletion region by discharging the adsorbed oxygen ions, thereby forming photo-desorbed oxygen.

To determine the effect of illumination power on the current, device 1 was exposed to different UV illumination intensities. For this purpose, the device was illuminated under two different UV illumination power densities (⁠$50$ and $250 μW/cm2$⁠). Figure 5(c) shows that the current increases with increasing UV illumination power density.⁵⁵ To estimate the rise and decay time constants, one cycle of UV illumination under a power density of $250 μW/cm2$ was fitted (see Figure 5(d)) using bi-exponential⁵⁶ expression of the form.

where $τON,τOFF1$⁠, and $τOFF2$ represent the relaxation time constants, whereas $I0$⁠, $A, B,$ and $C$ are positive constants. By applying a 10 V bias, the rising relaxation time $τON$ was $1.7 s$⁠, whereas the decaying relaxation times were estimated as follows: $τOFF1=1.08 s$ and $τOFF2=12.7 s$⁠. The fast current rise was related to the carrier generation in ZnO NWs due to UV illumination. The decay time constants show two different mechanisms. The decay time was initially very rapid with a time constant $τOFF1$⁠, which is represented by the bi-exponential equation as shown in the inset in Figure 5(d). This rapid decay was mainly caused by the direct recombination of holes–electrons. The initial rapid decay was followed by a slow time decay $τOFF2$⁠, which corresponds to re-absorption of oxygen and water vapor molecules on the NW surface, creating a depletion layer and leading to an increase in resistance. The actual rise time can be estimated as t_rise = 4.42 s, which is the time required for the current to reach 90% of its highest value. Similarly, the decay time (t_decay = 13.15 s) is evaluated as the time required for the current to drop by 10% of its peak value.

To evaluate the photoresponse performance of the UV photosensors, the photoresponse factor (S) is defined as the ratio of the effective photocurrent (⁠$Iph=Ilight−Idark$⁠) to dark current (⁠$Idark$⁠)⁵⁴ 

Photoresponse factor is one of the parameters considered as a decisive factor for high performance of photodetectors.⁵⁷ 

The obtained photocurrent from device 1 is $∼$245 nA, which is expected due to the small area connection of a few NWs. However, the photoresponse factor obtained from this device is $∼$40, which is significantly larger than previous reports for ZnO NWs-based UV photosensors.^54,58–60 The calculated values of the rise and decay times for this device are comparable with the past reports.^54,58–60

To further show the photodetection abilities of ZnO NWs, device 2 was measured in dark and under UV illumination with a power density of $250 μW/cm2$ in ambient condition (not shown here). This device showed photocurrent of $∼$360 $μA$ under 50 mV excitation due to the large number of NWs participating in photoconduction process. This photocurrent value is higher than the reported data.^54,58–60 Device 2 also showed a high photoresponse factor of $∼$120 along with an actual rise and decay time of 30 s and 8.1 s, respectively. Table II shows the collected data from device 1 and 2 as compared to results found in the literature. The overall performances of these devices indicate the possibility of using ZnO NWs to develop photodetectors that demonstrate improved performance.

A simple seed layer method to simplify the chemical bath deposition technique was investigated and successfully implemented for ZnO NW growth. The as-grown NWs demonstrated a good crystalline structure, which is similar to that of ZnO NWs prepared by using other methods or treatments. ZnO NWs fabricated using this method were photoactive and demonstrated substantial photocurrent generation under UV illumination. The UV photosensor devices fabricated using ZnO NWs also obtained a high photoresponse factor. These findings suggest the possibility of growing and integrating robust and optically active ZnO nanowires on large-area substrates to perform specific functions in electronics and in opto-electronic applications. The present seeding route was applied only to grow ZnO nanostructures, but this technique can be possibly extended in equivalent seeding methods to synthesize a wide variety of materials needed for diverse applications.

A.S.A is grateful for the financial support provided by the Higher Committee for Education Development of Iraq. S.T. also acknowledges the funding provided by the U.S. National Science Foundation through Grant No. NSF-PIRE OISE-0968405. N.P. and L.B. are supported by the U.S. Army Research Office MURI Grant No. W911NF-11-1-0362. V.C. acknowledges the support from The Brazilian National Council for Scientific and Technological Development (CNPq) - (249070/2013-8).