Light Trapping-Mediated Room-Temperature Gas Sensing by Ordered ZnO Nano Structures Decorated with Plasmonic Au Nanoparticles.

An ordered array of 1D ZnO nanorods obtained by colloidal templating is shown to dramatically enhance the sensing response of NO x at room temperature by confining light and creating periodic structures. The sensitivity is measured for a concentration varying from 2 to 10 ppm (response 53% at 10 ppm) at room temperature under white light illumination with ≈225 nm hole diameter. In contrast, structures with ≈450 nm hole size show better sensing under (response 98% at 10 ppm) elevated temperatures in dark conditions, which is attributed to the increased surface chemical interactions with NO x molecules due to the porous nature and enhanced accessible surface area of ZnO nanorods. Further, the decoration of ZnO Nanorods with gold nanoparticles shows enhanced sensor performance (response 130% at 10 ppm) due to localized surface plasmon resonance under white light illumination. The findings may lead to new opportunities in the visible light-activated room-temperature NO x sensors for healthcare applications.

Introduction

Detection of nitrogen oxides (NO and NO2) is important, as toxic gases released from automobiles and thermal power plants have detrimental effects on human health, which include severe respiratory diseases and irritation of the skin.[1,2] Amongst all metal oxides such as SnO2,[3] ZrO2,[4] V2O5,[5] WO3,[6] NiO,[7] and ZnO,[8] various ZnO nanostructures have been widely utilized for sensing NO gases, as environment friendly ZnO possesses high electron mobility along with superior chemical and thermal stability. However, most of the ZnO-based NO sensors fail to operate at room temperature owing to the transfer of electrons from the conduction band to the surface-adsorbed oxygen, which in turn leads to the reduction of the conductance of the semiconductor. This bottleneck is circumvented by making the device operate at elevated temperatures (typically between 50 and 450 °C), as the intrinsic conductance increases linearly with temperature.[9,10] However, the high-temperature detection is incompatible with low power sensor circuitry and also suffers from possible detection-induced flammability and consequently, significant research efforts are being devoted toward the development of NO sensors operating at room temperature.

Light-driven gas sensing is a promising approach for room-temperature sensing due to the absence of flammability and associated hazards, as well as long-term stability because high-temperature devices are prone to degradation caused by diffusion and sintering effects. Previously, UV illumination (wavelength ≈ 200–400 nm)-driven room-temperature NO sensors have been fabricated with ZnO[11] and other metal oxides semiconductors.[12−14] However, UV-assisted NO sensing is energy consuming (solar UV radiation contains only ∼5–7% of the total energy), expensive, and also harmful for human skin and eyes. For obvious reasons, nonhazardous visible light (wavelength ≈ 400–700 nm)-activated NO sensors that can be operated using solar illumination (∼43% visible), fluorescent lamp, indoor illumination, visible light-emitting diodes (LEDs), and so forth are desirable. However, the development of room-temperature white light (UV–visible)-activated sensors is in the nascent stage. Recently, toxic heavy metal cadmium sulfide (CdS)-based thin films, nanoflakes,[15] and heterojunction (ZnO/CdS)-based visible and UV–visible activated NO sensors have been reported.[16−18] Environmentally benign metal oxide-based visible light-activated NO sensors can be fabricated by reducing the band gap using self-doping with oxygen vacancies in metal oxides such as ZnO,[19] NiO,[20] WO3,[21] and fabricating binary composites of oxygen self-doped metal oxides such as CuO1–y@ZnO1−α[22] and rGO–ZnO1–.[23] However, self-doping-based oxygen vacancies in metal oxides extend the band gap limitedly toward the visible wavelength and the fabrication process is complex and expensive, involving techniques such as arc melting at high temperature (∼3000 °C) and ultrafast cooling in the reducing atmosphere.[19,24] Lu et al. have fabricated visible light-activated NO2 sensors operating at room temperature using the 3D inverse opal-based (3D IO) In2O3–ZnO heterogeneous composite microsphere (HCM).[25] However, 3D IO-based architecture using In2O3–ZnO 3D IO HCM shows the poor extension of light response at the visible wavelength region (∼400–700 nm), and the role of porosity for retarding light response has not been studied.

In this work, we have selectively fabricated organized ZnO 1D structures to enhance gas sensing performance at room temperature by trapping low-intensity (400 μW/cm2) white light. Hexagonal close-packed monolayer array of polystyrene (PS) colloids with different particle diameters (dp) has been used as a template to grow 1D ZnO nanorods hydrothermally on radio frequency (RF)-sputtered ZnO thin films. Our experiment shows that ZnO nanorods were grown with dp = 600 nm colloids (structure identified as Z450) show superior response at an elevated temperature (200 °C). In contrast, the best response at room temperature under low-intensity white light is observed when the nanorods are grown using dp = 300 nm colloids (Z225). While the maximum accessible area in the Z450 structure is responsible for sensing at elevated temperatures in dark, enhanced light trapping in the Z225 structure is responsible for sensing at room temperature under low-intensity white light. Further, we used solid-state dewetting-mediated uniform deposition of Au nanoparticles for extending the absorption wavelength further toward the visible region to achieve rapid and higher response for Z225 samples that shows a response time of 110 s and a decay time of 100 s. The study shows that the resonant plasmonic effect of Au nanoparticles leading to strong absorption at the visible band, along with the scattering and local field enhancement[26−28] can be effectively utilized to overcome the limitation of ZnO-based sensors operating at room temperature.

Results and Discussion

Figure shows the field-emission scanning electron microscopy (FESEM) images of vertically aligned ZnO nanorods grown without the template as the control sample in Figure a and with UVO-mediated colloidal transfer-printed PS templates in Figure b–d on ZnO/indium tin oxide (ITO)/glass substrates. The average diameter and length of the nanorods are found to be ∼80 nm and ∼1.0 μm, respectively, after a growth time of 4.5 h. ZnO nanorods exhibit a hexagonal regular structure all over the substrate, as shown in Figure a–d. In order to obtain well-oriented ZnO nanorod samples, three types of PS colloids with diameters dp = 300, 600, and 800 nm were used, The growth of well-oriented porous ZnO nanorods solely depends on the hexagonal close-packed arrangement of monolayer colloidal crystals, as well as the infiltration of the nutrient solution in the interior interstices of the colloid templates. Figure b–d shows the images of ZnO nanorods prepared over the contour of ordered periodic spherical holes of an average diameter of ∼225, ∼450, and ∼640 nm, respectively and consequently, these samples are designated hereafter as Z225, Z450, and Z640, respectively. ZnO nanorods grown with colloidal templates reveal aligned nanostructures with the porous surface up to a height that matches the dp of the colloids used that is 300 nm for Z225, 600 nm for Z450, and 800 nm for Z640 samples. Beyond this height, it becomes smooth similar to generic hexagonal growth of ZnO nanorods. In case of a colloidal template, the growth of ZnO nanorods occurs along the contours of the spheres, which differs from the conventional nanosphere lithography mask-based vapor phase deposition techniques where the growth takes place over the interstices of the mask only.[29,30] The average diameter of ZnO nanorods grown using the colloidal templates is found to be ∼45 ± 5, ∼80 ± 4, and ∼80 ± 5 nm for Z225, Z450, and Z640 samples, respectively. The diameter of ZnO nanorods (∼80 nm) in Z450 (Figure c) and Z640 (Figure d) samples are identical to the control sample (∼80 nm, Figure a), which may be attributed to the unperturbed colloidal contour oriented growth of ZnO nanorods in Z450 and Z640 samples as adequate space in adjacent colloids were available for growth. Following the growth of nanorods, the average hole radius in all samples is found to be approximately 25% lower than that of the initial PS sphere diameter. Near identical percentage shrinkage of the spheres irrespective of the dp of the colloids is attributed to the same rate of decomposition and vaporization of PS colloids during the hydrothermal process. In order to quantify the rate of decomposition and vaporization of PS colloids as well as the subsequent growth of ZnO nanorods, we have specifically prepared ZnO nanorods using 600 nm (Z450) PS colloidal templates at different growth times and is shown in Figure .

Figure 1

FESEM images of (a) nanorods on flat rf-sputtered ZnO thin films (control sample), (b) 300 nm (Z225), (c) 600 nm (Z450), and (d) 800 nm (Z640) PS colloid-templated growth of ZnO nanorods on rf-sputtered ZnO thin films.

Figure 2

FESEM images of (a) 1.5, (b) 2.5, (c) 3.5, and (d) 4.5 h growth of Z450 samples; (e) percentage shrinkage of colloidal spheres and ZnO nanorods length as a function of time during hydrothermal growth.

Figure shows the images of ZnO nanorods grown using the 600 nm PS colloid template at different growth times. It is observed that the growth of ZnO is initiated from the contour of PS colloids, and it follows voids of the colloids with increasing growth time. The percentage shrinkage of diameter and length of the nanorods has been estimated using ImageJ (version 1.50 e) software, which is shown in Figure e. The colloid diameter is shrunk from 600 to 450 nm, which is nearly 25% after 4.5 h of ZnO nanorods growth as shown in statistical estimation in Figure e. It is also confirmed from Figure b–e that the diameter shrinkage rate after 4.5 h of hydrothermal growth using dp = 300 nm and 800 nm is nearly equal to dp = 600 nm sample. This observation confirms that following the removal of the colloidal template, hole diameters can be controlled using the initial colloids diameter, nutrient concentration, and time of growth of ZnO nanorods in a particular hydrothermal growth condition.

Figure a,b shows the images of Au-decorated Z225 structures prepared using the thermal solid-state dewetting method.[31] The deposition of Au NPs on ZnO NRs by thermal evaporation has been used to ensure the purity of the metals as well as the control of the size distribution and the density of NPs by controlling initial metal film thickness and annealing temperature. FESEM images reveal the formation of Au NPs with a diameter range of ∼10–50 nm, which is anchored on the surface of ZnO NRs. The low magnification image in Figure a confirms the high density of Au NPs deposited on the surface of the nanorods, while the high magnification image of Figure b reveals the deposition of Au NPs uniformly on the side surface as well as the top of the nanorods.

Figure 3

FESEM images of Au nanoparticle-decorated Z225/Au NP sensors for (a) low and (b) high magnification.

In order to investigate the optical properties of as-prepared ordered ZnO structures, structure-dependent transmission and photoluminescence (PL) spectra were studied at room temperature. Figure a illustrates the spectra of ZnO nanorods grown on colloidal templates and flat ZnO seed layer surfaces (control), revealing a broad transmission dip in the UV range due to strong absorption. The ordered samples show a significant decrement in transmission in comparison to control ZnO nanorods grown on the flat surface. It is observed from the spectra that the Z225 sample exhibits negligible transmission at the absorption edge of ZnO in comparison to other samples, which is likely due to the enhanced light trapping within the structures. Upon decreasing the colloidal diameter down to the subwavelength range, it becomes possible to confine the light within the structures leading to increase in the absorption of ZnO.[32,33]

Figure 4

(a) Transmission spectra of colloidal template-assisted grown samples and control samples. (b) PL spectra of different diameter ZnO 1D nanostructures using colloid templates along with the control sample. (c) Z225/Au NP samples annealed at different temperatures.

PL spectra of the control and ordered ZnO nanorod samples are shown in Figure b. The spectra display two prominent peaks: one in the UV region attributed to the band edge (BE) emission of ZnO and other in the visible region owing to the intrinsic defects of ZnO. BE emission is found to be enhanced in ordered ZnO samples because the ratio of UV to defect emission intensity gets enhanced for Z225 (∼0.87), Z450 (∼0.59), and for Z640 (∼0.69) samples, as compared to the controlled one (∼0.36). A large enhancement in the Z225 structure is primarily attributed to the increased charge generation and recombination within the nanorods due to light trapping. It is also observed that the position of BE is red-shifted for ordered ZnO samples as compared to the control structure because of the changed dimension of nanorods. The visible emission can be attributed to the multiple point defects presents in ZnO such as Zn and O vacancies (VZn and VO), Zn and O interstitials (Zni and Oi), and antisites (OZn and ZnO), similar to earlier observation by Gogurla et al.[27] Four emission peaks are associated with single-ionized oxygen vacancies (VO+) at 523 nm, doubly ionized oxygen vacancies (VO++) at 585 nm, oxygen interstitial (Oi) at 657 nm, and an infrared tail at 759 nm due to transition from complex defects.

This phenomenon along with PL enhancement at BE suggests that optical path length can significantly be enhanced by subwavelength dimensional nanosphere-assisted growth of nanorods. Z450 samples show relatively low emission compared to other ordered samples due to the low density of nanorods and increased hole diameters, which results in high transmission and low absorption. Based on this observation Z225, Z450, and control samples are chosen to be explored for further measurement as light trapping-based sensors.

Figure c shows the absorption spectra of different temperature-annealed (150, 250, 350, and 450 °C) Au thin films over Z225 structures. The spectra display absorption in the UV and visible regions. The visible band prominent in 350 and 450 °C is attributed to the localized surface plasmon (LSPR) effect of Au nanoparticles after solid-state dewetting of Au thin films.[31] The LSPR band is more prominent in the Z225/Au NP samples annealed at 350 and 450 °C and the same is less prominent and red-shifted for samples with unannealed and annealed at 150 and 250 °C temperatures, respectively. The absorption characteristics in the wavelength range of 350–450 nm remain unaltered for Z225/Au NP and Z225 samples, which suggest that no further evolution of ZnO nanostructures following Au NP incorporation.

Gas Sensing Performance Evaluation

The testing of samples for gas sensing has been carried out under both dark and illumination at room and elevated temperatures. Temperature-dependent and light-induced gas sensing at room temperature have been carried upon exposure to several gases such as NO and other volatile organic compounds by measuring sensor resistance. In order to reduce the energy consumption, the effect of both temperature and white light together has not been investigated for sensing. The sensors were initially exposed to dry air for 20 min to stabilize the baseline resistance. The sensor response (S) of NO is calculated from the relation[27] where RNO is the resistance of the film in the presence of NO gas and Rair is the film resistance in the presence of dry air.

Figure a,b shows the semilogarithmic plots of the I–V characteristics of all ordered ZnO devices along with the control one under dark and illuminated conditions. It is observed that the Z225 and Z450 structures exhibit significantly low dark currents in comparison to the control nanorod device (Figure a,b). The reduction in dark current for Z225 and Z450 devices is mainly due to the rough nature of nanorods, which increases the adsorption of oxygen molecules on the surface. The current increases for all devices under light illumination, indicating the photoconductive response. Interestingly, it is found that the photocurrent enhances for Z225 and Z450 devices over the control one. The photo-to-dark current ratio (Iph/Id) is found to be 6 × 103, 0.75 × 103, and 0.36 × 103 at −5 V for Z225, Z450, and control devices, respectively. An enhancement in the Z225 device is observed under both the forward and reverse bias conditions in the presence of light as compared to Z450 and control samples. Enhancement in a photo-to-dark (Iph/Id) current ratio of 2 and 16 is observed for Z450 and Z225 samples over the control device, respectively. This enhanced photoresponse for Z450 and Z225 devices is owing to the higher collection rate of captured and photogenerated electrons in ZnO as compared to the control sample. The enhanced photoresponse of the Z225 device in comparison to the Z450 one is attributed to the improved light confinement in the structures, as observed in the transmittance spectra. It is known that the photogenerated electrons and excess electrons from the adsorbed oxygen mainly contribute to the photoconductivity of ZnO. The porous nature of ZnO nanorods in ordered structures can lead to the higher adsorption of oxygen molecules resulting in excess electrons in the devices. In addition, an enhanced light trapping in the Z225 device further boosts its photoresponse due to the subwavelength structures. Moreover, the photo-to-dark ratio is nearly the same under 325 nm LASER and white light illumination, which signifies that Z225, Z450, and control samples are well responsive under white light. Therefore, photogenerated and excess electrons can be effectively used to interact with NO molecules for operating the sensors at room temperature under white light illumination. The dynamic response and recovery of the Z225 device is further enhanced by the decoration of Au nanoparticles on its surface. The incorporation of Au nanoparticles results in a combinatorial effect of nanoparticle-induced scattering of incoming white light, catalytic activation for the dissociation of oxygen molecules, and LSPR-induced UV–visible absorption in Z225 structures. The improved photoabsorption signifies an increase in the number of interacting photogenerated electrons toward NO molecules. Figure c shows the semilogarithmic plot of the I–V characteristics of Z225/Au NP and Z225 sensors under dark and 100 mW/cm2 white light illumination. The photo-to-dark current ratio of Z225/Au NP (Iph/Id ≈ 5.8 × 104) is observed to be enhanced 10 times in comparison to Z225 sensors (Iph/Id ≈ 6 × 103). A slight decrease in dark current in Z225/Au NP sensors might be attributed to the increment of low conductivity depletion region, as Au NPs with a higher work function (∼5.1 eV) withdraw electrons from the Z225 surface having a relatively low work function of ∼4.1 eV. In the presence of white light, electrons are excited to the surface plasmon band of Au NPs and under the decay process, high-energy LSPR induced generated free electrons are transferred to the conduction band of Z225. Additionally, photogenerated holes migrate toward Au nanoparticles due to the interactive Coulombic attraction leading to trapping at the Au NP and Z225 Schottky barrier interface, which results in the interaction of a large number of electrons with NO molecules under white light illumination.

Figure 5

Semilog plot of the current–voltage characteristics of Z225, Z450, and control devices in dark and under (a) 100 mW/cm2 white light and (b) 48 mW/cm2 325 nm laser illumination. (c) Current–voltage characteristics of Z225/Au NP under 100 mW/cm2 white light illumination.

Figure a1–a3 shows the dynamic response–recovery curves of template-assisted grown ZnO nanorods and control sensors for different concentrations of NO exposure in dry air, under white light illumination with a power of 400 μw/cm2. The response also shows a gradual increase with the concentration of NO. As seen in figure, the sensors show a significant change in the response even at a low part-per-million concentration levels for all devices. Moreover, the Z225 sensor shows 4 times increment in sensing response in comparison to the control device upon exposure to 10 ppm NO. Figure b shows the response as a function of NO concentration for all devices, indicating a gradual increase with the concentration of NO. This observation suggests that the Z225 sensor has an improved sensing response due to the enhanced light trapping effect. We know that the oxygen species (O2–, O–, and O2–) adsorb on the ZnO surface by capturing free electrons and create a low conductive depletion layer under dry air conditions. In the presence of white light, the electron–hole pairs are generated in ZnO nanorods (Figure c), and the photogenerated holes migrate to the surface to decrease the width of the depletion layer by desorbing the oxygen molecules. This results in a decrement of the resistance for sensors under illumination. Here, the photogenerated and excess electrons from adsorbed oxygen (Figure d) mainly contribute to the photocurrent. The electronegative NO molecules which have a tendency to withdraw electrons from the n-type semiconductor surface easily interacts with the photogenerated electrons at the ZnO semiconductor surface leading to the decrease of conductance and increase of resistance of the sensors (Figure e).

Figure 6

Dynamic sensing response–recovery curves of the sensors (a1) Z225 (a2) Z450, and (a3) control samples at different concentrations of NO in dry air under illumination. (b) Response characteristics under NO exposure with different concentrations of dry air under white flash light illumination. Schematic representation of interacting gases toward ZnO under light illumination, (c) ambient condition, (d) under dry air, and (e) NO gas in dry air.

The sensing mechanism can be stated using the following reactions

Upon switching off the NO gas, the resistance of the sensor decreases due to the abundance of free electrons generated under illumination. The enhanced sensing response for the Z225 sensor might be attributed to the increased photoelectron generation due to improved light confinement in the structures. The observed results suggest that the light confinement in subwavelength ordered nature of the ZnO sensors can be effectively used to enhance the gas sensing performance. The mechanism involved in this light-induced sensing is quite different than that for high-temperature sensing.[34] Apart from the light-induced room-temperature sensing, the devices were also tested at high temperature in dark conditions to investigate the effect of porosity of the sensing material in the presence of gas. The optimum temperature has been found to be 200 °C.

Figure a1–a3 shows the dynamic sensing response–recovery curves of template-assisted grown ZnO nanorods and control devices under different concentrations of NO exposure in dry air at 200 °C. The timing pulse of 10 min “ON” and 10 min “OFF” is retained based on the steady-state response of the sensors at 2 ppm NO exposure at dry air conditions (Figure S2 of the online Supporting Information) during the dynamic response–recovery measurements. The response of all devices shows a gradual increase upon exposure to NO, as shown in Figure b. The response is found to be enhanced by a factor of 1.21 for Z225 (∼77.7%) and 1.53 for Z450 (∼98%), as compared to the control one (∼64%) upon exposure to 10 ppm NO gas at 200 °C. The enhancement for ordered ZnO sensors can be well understood from the porous nature of the nanostructures in comparison to control samples. Because of the porosity and enhanced surface area of nanorods, a large number of NO molecules are adsorbed on the surface, which results in a higher sensing response for Z225 and Z450 sensors than the control sample. Furthermore, the decoration of Au nanoparticles using solid-state dewetting creates a higher density of electrons in the ordered semiconductor surface under visible light illumination, which results in rapid and improved gas sensing response at room temperature.

Figure 7

Dynamic sensing response–recovery curves of the sensors (a1) Z225, (a2) Z450, and (a3) control at different concentrations of NO in dry air at 200 °C. (b) (%) response characteristics under NO exposure with different concentrations of dry air.

Figure a shows the dynamic response–recovery curves of Au nanoparticle-decorated nanorod sensors under different concentrations of NO exposure in dry air, under white light illumination with a power of 400 μW/cm2. The decoration of Au NPs on light trapping Z225 samples (Z225/Au NP) exhibits a rapid and higher response (130% at 10 ppm) under illumination in comparison to the control Z225 structure.

Figure 8

Dynamic sensing response–recovery curves of (a) Z225/Au NP sensors at different concentrations of NO in dry air under illumination; (b) selectivity of Z225/Au NP to 10 ppm vapors of different gases under white light illumination schematic representation of s Z225/Au NP under (c) light illumination and (d) under NO gas in dry air.

Figure b shows the response of Z225/Au NP sensors to six different gases under the same concentration of 10 ppm under white light illumination. It is observed that sensors show a higher response to NO and lower response to DCB, acetone, toluene, ethanol, and CO. The response of NO gas is almost 42.5 times higher in comparison to DCB vapor for Z225/Au NP sensors (Figure b) and 4.8 times to DCB for Z225 sensors (Figure S2 of the online Supporting Information) in the presence of white light. This observation suggests that ordered ZnO nanostructures decorated with solid-state dewetted Au nanoparticles can be used for highly selective NO sensors under low-intensity white light conditions. The lowest detection limit of NO using Z225/Au NP sensors can be estimated using criteria for minimum percentage response detectable to be 1.[35] The estimated limit of detection has been found to be approximately 550 ppb. The calculated detection limit suggests the potential of low concentration NO gas detection using Z225/Au NP sensors at room temperature under low-intensity white light illumination.

Au nanoparticle decoration on the surface of Z225 forms the localized metal–semiconductor Schottky junction followed by surface band bending as shown in band diagram Figure c. Under white light illumination, LSPR-induced hot electrons are spatially transferred toward the Z225 surface, resulting in a large number of oxygen species (O2–, O–, and O2–) to be adsorbed on the ZnO surface by capturing free electrons, which enhances low conductive depletion layer width under dry air conditions. Here, the photogenerated hot electrons from Au along with ZnO and excess electrons from adsorbed oxygen mainly contribute to the photocurrent, so the electronegative NO molecules can interact with the photogenerated electrons (Figure d). This increases the resistance in the sensor as shown in Figure a. The resistance again decreases in the sensor when we switch off the NO gas. The enhanced sensing response for the Z225/Au NP sensor under white light illumination might be attributed to the increased photoelectron generation due to nanoparticle-induced scattering of incoming white light, which catalytically activate the dissociation of oxygen molecules and LSPR induced UV–visible absorption in the light-trapping Z225 structures.

The performance of the ordered Au-decorated ZnO sensors is comparable or superior to those previously reported visible light-sensitive room-temperature NO sensors, as shown in Table .

Table 1

Comparison of Sensing Properties of NO Gas Sensors Operating at Room Temperature

sensor type	concentration (ppm)	illumination condition (wavelength/power)	response (ΔR/R)%	reference number
MOS2	5	UV (∼365 nm/1.2 mW/cm2)	18	(36)
ZnO–Ag	5	UV (∼470 nm/75 mW/cm2)	400	(37)
CdS	5	green LED (∼530 nm/21 W/m2)	89	(38)
ZnO–PbS	6	NIR (∼850 nm/1 mW/cm2)	320	(39)
ZnO/Au NP	6	white light (∼400μ W/cm–2)	78	present work

Conclusions

We have fabricated colloidal sphere-assisted selectively arranged 1D ZnO nanorods of different diameters on ITO/glass substrates. The performances of ordered 1D ZnO nanorod structures as NO sensors under light illumination at room temperature and higher temperature under dark conditions have been investigated. The sensor fabricated using the template of PS colloids with a diameter of 300 nm (Z225) has shown improved gas sensing response (53% at 10 ppm) in comparison to other sensors operated at room temperature under white light illumination. This observation suggests that the light trapping mechanism and selectively ordered structure can be effectively used to design low power sensors operating at room temperature. On the other hand, ZnO nanorods grown on the template of PS colloids with a diameter of 600 nm (Z450) have shown improved gas sensing response (98% at 10 ppm) in comparison to other sensors operated at an optimum temperature of 200 °C in dark conditions. This observation signifies that the enhanced accessible area due to the low density of nanorods as well as higher surface to volume ratio of the 1D structure results in superior sensing response in Z450 samples. Further, the decoration of plasmonic Au NPs on light-trapping Z225 samples (Z225/Au NP) lead to a rapid and enhanced response (130% at 10 ppm) under visible light illumination.

Experimental Section

Preparation of ZnO Seed Layer on ITO Substrate

In order to prepare the ZnO seed layer, stripped ITO coated glass (Sigma-Aldrich, 15–25 Ω/sq) of 0.5 cm × 0.5 cm pieces was taken as the substrate. ITO substrates were cleaned using sequential ultrasonication and boiling in solvents. Acetone and the mixed solvent containing water, ammonium hydroxide solution, and hydrogen peroxide in the 5:1:1 ratio was used as solvents for cleaning. Cleaned substrates were dried in ambient conditions using N2 stream. A ∼30 nm thin film of the ZnO seed layer was deposited on the substrates using RF magnetron sputtering from a 3 in. diameter ZnO target. The deposition took place at a substrate temperature of 200 °C in the Ar and O2 mixture ratio of 2:1. An RF power of 25 W and working pressure of 0.02 mbar were maintained during 30 min of deposition time.

Generating Hexagonal Close Pack on ZnO Seed Layer/ITO Substrate

Hexagonal close-packed (HCP) monolayer colloids were generated from monodispersed colloidal water suspension (concentration of about ∼10% wt) of PS spheres (Scheme , step 1). Spheres with a mean diameter of dp = 300 nm (Z225), dp = 600 nm (Z450), and dp = 800 nm (Z640) were used for differently ordered HCP samples of monolayer colloidal crystals. The array was placed on the ZnO seed layer-coated ITO glass by UVO—mediated colloidal transfer printing, reported elsewhere.[40−42]

Scheme 1

Schematic Diagram Indicating Fabrication Steps of PS Colloids Template-Assisted Growth of Au-Decorated Ordered ZnO Nanorods on rf-Sputtered ZnO Thin Films

Creating Ordered ZnO Nanorod Arrays

The growth mechanism of ZnO nanorods on the RF-sputtered ZnO thin film is shown schematically in step 2, Scheme . In brief, the ZnO nanorods were prepared by placing monolayer colloidal crystals on ZnO/ITO/glass substrates in a hydrothermal chamber containing 60 mL millipore aqueous solution of 1:1 0.025 M of zinc nitrate hexahydrate and hexamethylenetetramine kept at 90 °C for 4.5 h. Prior to this, seed layer-coated glass substrates were preheated for 2 min in 120 °C, which is above the glass transition temperature of PS colloids. Heating above the glass transition temperature led to necking between the adjacent particles and restricted the dislodging of colloids during solution growth of ZnO nanorods. Following ZnO nanorod growth, the PS spheres were removed by washing in the 1-chloropentane solvent followed by annealing at 200 °C (step 3, Scheme ).

Decorating Au Nanoparticles on Ordered ZnO Nanorod Arrays

Au nanoparticles were decorated at room temperature on the white light-activated best performing Z225 sensor using thermal evaporation followed by solid-state dewetting of the as-deposited film under high-temperature annealing (step 4, Scheme ). In detail, 3 mg Au was resistively evaporated from a tungsten boat at a base pressure of 4 × 10–6 Torr for 3 min on unannealed ZnO nanorods/ITO substrates. Au thin film-deposited samples were placed in a quartz tube and annealed in a preheated oven at a predetermined temperature of 400 °C till 30 min and cooled down to room temperature for another 30 min of time.

Characterization

The morphology of ordered ZnO nanorods and control (nanorods without template) samples were investigated using FESEM (JSM7610F, JEOL, Japan). UV–visible spectroscopy was carried out with deuterium- and halogen-combined excitation source with the spectra being recorded using an AvaSpec-3648 spectrometer. PL measurements at room temperature were performed using a He–Cd laser with a wavelength of 325 nm as the excitation source having an output power of 45 mW and a TRIAX 320 monochromator fitted with a cooled Hamamatsu R928 photomultiplier detector. ImageJ (version 1.50 e) was used to statistically analyze the data. The current–voltage (I–V) and photoresponse characteristics of the heterojunction devices were studied at room temperature using a Keithley semiconductor parameter analyzer (4200 SCS), along with a He–Cd laser (325 nm, 30 mW) and a broadband solar simulator (AM 1.5, 100 mW/cm2) for UV and visible range illumination, respectively.

Fabricating Gas Sensors and Optoelectronic Device

The current–voltage (I–V) characteristics and gas sensor response were measured using a lateral photoconductor device configuration for investigating the potentiality of these sensors under dark and illumination conditions.

Gas Sensing Measurement Setup

Gas sensing properties of the samples were measured using a custom-made gas sensor setup with a sealed and transparent quartz window stainless steel closed container as a chamber (Figure S1 of the online Supporting Information). The device was kept inside the chamber with the probes connected to the devices from ITO as electrical contacts. The in situ heater inside the chamber was used for high-temperature sensing measurements. The light-induced sensing properties were measured using a white light of power 400 μW/cm2. Digital mass flow controllers from Alicat were used to control the flow rates of test and carrier gases at an appropriate proportion. The resistance of devices was measured using an Agilent 349721 LXI data acquisition (DAQ) unit fitted with 34901A 20 channel multiplexer switches and a digital ohm meter. BenchLink Data Logger Pro software was used for sensor response data acquisition.