Humidity-Enhanced Effects of Sputtered Au Nanoparticles on the Ethanol Sensing Performance of Thermally Evaporated ZnO Nanowires

Abstract

Mitigating the adverse effects of humidity on the sensing performance of one-dimensional (1D) metal oxide semiconductors has been a key challenge in gas sensor development. This study investigated the humidity-enhanced effect of gold (Au) nanoparticles on the ethanol response of 1D ZnO nanostructures. ZnO nanowires were grown using a thermal evaporation–condensation method with Au-catalyzed growth kinetics for the fabrication of conductometric chemical sensors. Au nanoparticles were then additionally decorated on ZnO nanostructures by sputtering with varying times from 2 to 10 s. Structural examinations demonstrated Au nanoparticles with diverse diameters (10–70 nm) widely distributed on ZnO nanowires exhibiting lengths of 0.1–1 μm and diameters of 10–30 nm. All fabricated sensors were systematically evaluated toward NO₂, NH₃, C₃H₆O, CO, and C₂H₅OH at 250–400 °C under dry and humid conditions. ZnO nanowires functionalized with Au for 5 s were found to exhibit the highest response of 93.75 to 70 ppm of C₂H₅OH, which was significantly greater than that of the pristine ZnO nanowire sensor (1.86) at 50% relative humidity (RH) and the optimum sensing temperature of 350 °C. Interestingly, the optimally Au-decorated ZnO nanowire sensor showed humidity-enhanced behaviors with significantly boosted ethanol responses at 50% RH. In addition, the optimal sensor displayed a low theoretical limit of detections (3.7 ppb) and high C₂H₅OH selectivity against interfering gases, including NO₂, NH₃, C₃H₆O, and CO. The observed ethanol response improvement and humidity-enhanced effect might be ascribed to the attributes of ohmic metal–semiconductor junctions formed at Au–ZnO interfaces and the capabilities of very fine sputtered Au nanoparticles to enhance oxygen chemisorption and suppress the formation of surface hydroxyl and hydronium species on ZnO nanowires. According to the results, Au-decorated ZnO nanowires fabricated by evaporation–condensation and sputtering could be potential candidates for use as practical C₂H₅OH sensors.

1. Introduction

Ethanol (C₂H₅OH) is a common flammable, volatile, and colorless transparent liquid with light odor under normal temperature and pressure conditions. It is a popular solvent used in various chemical, medical, and agricultural industries and a key ingredient of alcoholic beverages. − However, its vapor in air can be explosive under certain workplace environments and thus hazardous to humans. It may also cause other symptoms, including nasal, eye, and skin irritation, as well as alcohol poisoning at high exposure levels. , As a depressant, C₂H₅OH slows down bodily reactions and impairs normal brain functions, leading to various alcohol-induced problems, particularly traffic accidents caused by drunken driving. − C₂H₅OH vapor is present at a high level in the breath of intoxicated drunken drivers that have consumed an excessive quantity of alcoholic beverages. Consequently, many countries have prescribed the legal blood alcohol limits corresponding to 130–208 ppm of C₂H₅OH in human breath. , Typically, C₂H₅OH toxication can be classified into three levels, including no consumption (<44 ppm), trace drinking (44–176 ppm), and excessive drinking (>176 ppm). Utilizing sensor devices to detect C₂H₅OH in human breath is simple, rapid, and cost-effective.

One-dimensional (1D) metal oxide semiconducting (MOX) nanostructures, including nanowires, nanotubes, nanorods, and nanobelts, are particularly effective gas-sensing materials, owing to their high integration density, synthetic flexibility, low production cost, and miniaturization capabilities. − 1D MOX nanostructures made of ZnO, ,− SnO₂, − In₂O₃, TiO₂, WO₃, , CuO, , and NiO , have demonstrated decent C₂H₅OH-sensing performances. Among them, ZnO is one of the most promising n-type semiconducting metal oxides with high electron mobility, a wide band gap (3.37 eV), , large exciton binding energy (60 meV), excellent thermochemical stability, low cost, and low toxicity, making it an outstanding material for constructing functional sensing devices. ,,− However, ZnO still faces certain limitations, particularly selectivity and detection limits. Several approaches including optimizing the synthesis process, modifying surface properties, and introducing catalytic dopants have been explored to overcome these shortcomings. , Noble metal functionalization is known to be a highly effective method for improving gas-sensing performance, particularly in humid environments. The noble metals, including Ag, Ru, Mo, and Au, − can provide catalytic sites that enhance target gas adsorption and chemical sensitization, resulting in positive humidity effects on gas response. Among these, Au has been proven as a particularly effective additive for enhancing the C₂H₅OH-sensing capability of ZnO.

ZnO nanostructures with Au functionalization produced through different routes have been investigated for C₂H₅OH sensing as summarized in Table . First, the addition of 3.4 wt % Au to ZnO nanowires via a hydrothermal route enhanced the response to 100 ppm of C₂H₅OH from 3.7 to 33.6 at 380 °C. Similarly, combining 1 wt % Au with ZnO through mixed powder oxidation to form Au-loaded ZnO nanowires improved the response toward 2000 ppm of C₂H₅OH from 27 to 37 at 240 °C. Likewise, incorporating 5 wt % Au into ZnO nanowires via a thermal oxidation process boosted the response to 1000 ppm of C₂H₅OH from 23 to 170 at 320 °C. Also, Au sputtering on hydrothermally made ZnO nanowires increased the response toward 50 ppm of C₂H₅OH from 1.5 to 7 at 325 °C. Additionally, Au decoration by wet-chemical reduction on hydrothermally made ZnO nanowires improved the response to 100 ppm of C₂H₅OH from 14.7 to 22.6 at 260 °C. Besides, ion sputtering of Au on CVD-grown ZnO microwires augmented the response toward 80 ppm of C₂H₅OH from 5.5 to 6.8 at 15 °C under UV illumination. Conversely, the Au doping within ZnO nanosheets using another hydrothermal process raised the response to 60 ppm of C₂H₅OH from 0.2 to 0.35 at 300 °C. Furthermore, the Au decoration over ZnO nanoplates through an ion exchange route promoted the response toward 100 ppm of C₂H₅OH from 10 to 21 at 400 °C. Recently, Au nanoparticles were decorated on NiO/ZnO nanosheets using another hydrothermal method, elevating the response to 50 ppm of C₂H₅OH from 5 to 41 at 300 °C. Lately, Au loading by photochemical reduction on hydrothermally synthesized ZnO nanorods augmented the response toward 100 ppm of C₂H₅OH from 6.9 to 9.0 at 270 °C.

1. Characteristics of Au/ZnO-Based Ethanol Sensors Made by Different Techniques.

sensing material	method	temp. (C)	conc. (ppm)	response	refs
ZnO nanowires	hydrothermal	380 (dry air)	100	3.7
3.40 wt % Au/ZnO nanowires				33.6
ZnO nanowires	hydrothermal (ZnO) and sputtering (Au)	325 (dry air)	50	1.5
3.70 atom % Au-sputtered ZnO nanowires				7
ZnO nanowires	mixed powder oxidation	240 (dry air)	2000	27
1 wt % Au/ZnO nanowires				37
ZnO nanowires	hydrothermal process (ZnO) and wet-chemical reduction (Au)	260 (dry air)	100	14.7
Au-decorated ZnO nanowires				22.6
ZnO nanowires	thermal oxidation	320 (dry air)	1000	23
5 wt % Au/ZnO nanowires				170
ZnO nanoplates	ion exchange	400 (dry air)	100	10
1% Au-decorated ZnO nanoplates				21
ZnO microwires	CVD (ZnO) and ion sputtering (Au)	15 (dry air)	80	5.5
Au-sputtered ZnO microwires				6.8
ZnO nanosheets	hydrothermal	300 (dry air)	60	0.2
9 atom % Au-doped ZnO nanosheets				0.35
ZnO nanosheets	hydrothermal	300 (dry air)	50	5
1.2 wt % Au-decorated NiO/ZnO nanosheets				41.3
ZnO nanorods	hydrothermal (ZnO) and photochemical deposition (Au)	270 (dry air)	100	6.9
1.79 atom % Au-decorated ZnO nanorods				9.0
ZnO nanowires	evaporation–condensation	350 (50% RH)	500	2
ZnO nanowires	evaporation–condensation (ZnO) and sputtering (Au)	350 (50% RH)	70	1.9	this work
			70	94
3.25 wt % Au-sputtered ZnO nanowires (5 s)			5	15

^aOperated under UV irradiation.

^bDefined in Section .

These reports demonstrate that Au functionalization on various ZnO nanostructures provides different enhancement effects on C₂H₅OH-sensing responses within the C₂H₅OH concentration range of 50–2000 ppm at diverse working temperatures ranging from 15 to 380 °C, highly depending on the production methods, structural features, and Au content. − ,,− The review results motivate an open research gap of investigating Au-ZnO nanostructures made via another distinctive method, which may offer superior C₂H₅OH-sensing performance. Moreover, these studies have mostly evaluated the ethanol-sensing performance of Au-modified ZnO sensors under dry conditions or at a single relative humidity (RH) value, leaving additional research gaps in their humidity-dependent effects and related sensing mechanisms.

To address the research gaps, this study investigated the ethanol-sensing properties with varying humidity levels of Au-decorated ZnO nanowires made by a distinctive fabrication method based on the combination of thermal evaporation–condensation and sputtering for the first time. Thermal evaporation–condensation via a vapor–liquid–solid (VLS) mechanism is an effective approach for the synthesis of one-dimensional nanostructures with key advantages over other processes listed in Table , including high yield, high structural quality, simplicity, and high reproducibility. Previously, ZnO nanowires and nanorods were successfully synthesized via VLS mechanisms using different metal catalysts and studied for gas-sensing applications. The study reported important roles of catalysts in ZnO nanostructures and gas-sensing properties. Particularly, ZnO nanowires grown using the Au catalyst provided good selectivity to hydrogen at 350 °C. In this work, the ZnO nanowires made via this method were modified by additionally sputtering the Au catalyst with varying sputtering times to obtain decorating Au nanoparticles and clusters of different sizes. The fabricated conductometric sensors were systematically investigated and optimized for C₂H₅OH sensing at low concentrations under dry and humid conditions to evaluate the humidity effects for practical ethanol sensor applications in humid environments. Key contributions including an unusual humidity-enhanced effect of sputtered Au nanoparticles on the ethanol response of ZnO nanowires and the proposed mechanisms responsible for the observed results would be elaborately discussed.

2. Experiments

2.1. Formation of Pristine and Au-Decorated ZnO Nanowires

ZnO nanowires were grown using the Au catalyst for nanowire nucleation. In the first step, 2 × 2 × 0.25 mm alumina substrates were cleansed by soaking in acetone for 15 min followed by drying with an air flow to remove dust and organic contaminants. A metallic Au catalyst layer was then deposited on cleaned substrates by radiofrequency (RF) magnetron sputtering with the plasma power of 70 W under 7 cm³/min an argon (Ar) flow at 5 × 10^–3 mbar. Afterward, a thermal evaporation–condensation process was employed to grow ZnO nanowires on top of the Au layer. In brief, ZnO powder (99.9% purity, Sigma-Aldrich) and Au-catalyzed alumina substrates were put upstream and downstream in a horizontal tube furnace, respectively. The tubular reactor was heated to 1200 °C under a 100 cm³/min Ar flow at a constant pressure of 10 mbar. The growth was undertaken at this temperature for 15 min before cooling to room temperature. To obtain Au-decorated ZnO nanowires, Au nanoparticles were sputtered over ZnO nanowires on the substrates with varying times from 2 to 10 s using an RF plasma power of 50 W under a 5 cm³/min Ar flow at 5 × 10^–3 mbar.

2.2. Fabrication of Gas-Sensing Devices

Gas-sensing devices were fabricated by depositing platinum (Pt) electrodes via direct-current (DC) magnetron sputtering on pristine and Au-decorated ZnO nanowire films, followed by the deposition of a Pt heating layer on the other side of alumina substrates. Initially, a 50 nm titanium–tungsten (TiW) adhesion layer was sputtered through a 200 μm-wide interdigitated shadow mask onto the ZnO nanowire film (the front side of the substrate) using 70 W Ar plasma at 5.3 × 10^–3 mbar and 300 °C. Next, a 1 μm-thick Pt electrode layer was deposited over the adhesive layer with the same sputtering parameters as those of the TiW layer. Subsequently, a 50 nm TiW adhesion layer and a 900 nm-thick Pt heating layer were sputtered through a heater shadow mask on the backside of the alumina substrate using the same sputtering conditions as those of the front-side layers. The fabricated devices were then mounted by bonding with Au wires to 4-pin TO packages as demonstrated in Figure . Lastly, the electrodes and the sensing materials were stabilized by aging at 400 °C for 48 h in ambient air.

1.

Photographs of (a) fabricated sensing device mounted by bonding with Au wires to a 4-pin TO package and (b) sensor chip layout illustrating the components and functional elements on each side of the chip.

2.3. Conductometric Gas Sensor Evaluation

The fabricated gas sensors were tested by monitoring their electrical conductivity variations inside a 1 L stainless-steel chamber placed within a climatic enclosure (MTC 120, Angelantoni, Italy) maintained at 20 °C. Dry air was passed through a Drechsel bottle held in a thermostatic tub at 25 °C to control the humidity level. A condensation vessel kept at 20 °C was used to remove excess moisture and supply a saturated vapor flow. A sensor response was evaluated toward different gases with a fixed total flow rate of 200 cm³/min at 50% RH and 20 °C. All target gas sources of NO₂, NH₃, C₃H₆O, CO, and C₂H₅OH were certified gas cylinders balanced in dry synthetic air (SIAD S.p.A., Bergamo, Italy). The source concentrations of CO, C₃H₆O, NH₃, and NO₂ were 2000, 200, 100, and 10 ppm, respectively. For C₂H₅OH, 100 and 200 ppm cylinders were used to generate low (<70 ppm) and high (70 ppm) concentrations, respectively.

A constant voltage at 1 V was applied to the sensors and their currents were recorded using a dedicated picoammeter (Keithley, 6485). These sensors were heated to a working temperature in the range of 250–400 °C by controlling the currents of their Pt heaters and stabilized for 10 h before exposing to a target gas at a designated concentration for 20 min, after which synthetic air was resumed for 25 min to return to their original states. The sensor response was computed based on the change in electrical conductance according to the definitions

$$\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{e}=\frac{G_{\mathrm{g}\mathrm{a}\mathrm{s}}-G_{\mathrm{a}\mathrm{i}\mathrm{r}}}{G_{\mathrm{a}\mathrm{i}\mathrm{r}}}\mathrm{for}\mathrm{reducing}\mathrm{gas}$$ 1

$$\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{e}=\frac{G_{\mathrm{a}\mathrm{i}\mathrm{r}}-G_{\mathrm{g}\mathrm{a}\mathrm{s}}}{G_{\mathrm{g}\mathrm{a}\mathrm{s}}}\mathrm{for}\mathrm{oxidizing}\mathrm{gas}$$ 2

where G _gas and G _air represent the steady-state sensor conductance values measured under a target gas and air, respectively. In addition, the response and recovery times (t _res & t _rec) were evaluated from the times required to attain 90% of the stable states of G _gas and G _air after applying gas sample and resuming air flow, respectively. ,

3. Results and Discussion

3.1. Structural Properties of Pristine and Au-Decorated ZnO Nanowires

The crystalline and phase components of the Au-ZnO nanostructures were assessed using X-ray diffraction (XRD, Phillips X-pert) with Cu Kα radiation (λ = 0.15418 nm) as reported in Figure . All XRD profiles exhibited well-defined peaks corresponding to (100), (002), (101), (102), (110), and (103) planes of a hexagonal ZnO structure (JCPDS No. 36–1451) together with the (200) plane of a cubic Au phase (JCPDS No. 04–0784) and other strong peaks of a rhombohedral Al₂O₃ phase (JCPDS No. 46–1212) from the substrate. The minor Au phase appeared in the pristine sample could be associated with the Au catalyst predeposited on the Al₂O₃ substrate for the nucleation of ZnO nanowires. The Al₂O₃ substrate peaks were dominant, and the minor Au peak could still be detected from very small catalyst nanoparticles located at the bases of ZnO nanowires since X-rays penetrated deeply into the bulk. The XRD results confirmed the coexistence of all expected phases with no additional impurity.

2.

XRD patterns of pristine ZnO nanowires and Au-sputtered ZnO nanowires on Al₂O₃ substrates prepared with sputtering times of 2, 5, and 10 s.

The compositional features of ZnO-based nanostructures were examined using scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (SEM/EDS, JSM-6335F, JEOL). Figure a illustrates a typical top-view SEM image of the pristine ZnO material, displaying highly dense nanowires with diameters of 10–30 nm and lengths of >1 μm. Figure b–d shows Au-decorated ZnO nanowires sputtered for 2, 5, and 10 s, respectively. In the high-resolution SEM image provided in Figure b, the Au nanoparticles were not clearly observed, presumably because of their ultrasmall sizes at the shortest sputtering time of 2 s. Distinctively, the inset images of Figure c,d reveal Au nanoparticles with sizes in the ranges of ∼10–20 nm and ∼30–70 nm for 5 and 10 s sputtering times, respectively. It should be noted that the particle size below 10 nm could not be reliably estimated due to the SEM instrument’s resolution limit. Figure e comparatively presents the cross-sectional SEM images of pristine ZnO and 5 s Au-sputtered Zn layers on Al₂O₃ substrates. Noticeably, the pristine ZnO nanowires (left image in Figure e) oriented slantingly with no apparent catalyst nanoparticles at their tips, suggesting a root-growth VLS mechanism. According to the established VLS growth kinetic models for quasi-one-dimensional metal oxides, ,, ZnO vapor continuously dissolved into liquid Au nanoparticulate droplets (formed at a high growth temperature) that strongly adhered on an Al₂O₃ substrate, causing ZnO supersaturation and outward precipitation of solid ZnO nanowires. However, Au catalyst nanoparticles were not apparently seen at the bases because they were most likely enclosed by ZnO nanowires. Nevertheless, the presence of Au at the ZnO bases could be experimentally indicated by the XRD result of the pristine ZnO sample, which clearly showed a minor Au phase detectable via X-ray penetration into the layer (Figure ). It also implied that Au nanocatalysts were not significantly consumed and did not form an alloy with the Al₂O₃ substrate during the root growth process. This growth mode occurred due to strong adhesion of Au droplets on the Al₂O₃ surface owing to their compatible wetting and surface tension properties under the growth conditions. , The ZnO nanowires grown via this mechanism usually arranged in nonvertical or tilted directions as seen in Figure e due to random precipitation out of Au nanodroplets, resulting in the outward growth of ZnO crystals with diverse crystallographic orientations as identified by the XRD data (Figure ). ,

3.

Top-view SEM pictures of (a) pristine ZnO nanowires and Au-decorated ZnO nanowires with sputtering times of (b) 2, (c) 5, and (d) 10 s, including the corresponding high-resolution SEM images (insets). (e) Cross-sectional SEM images comparing the pristine ZnO nanowires with 5 s Au-sputtered ZnO nanowires on Al₂O₃ substrates and (f) EDS data containing O, Zn, and Au elemental maps along with the corresponding EDS spectrum of 5 s Au-sputtered ZnO nanowires.

After sputtering Au for 5 s (right image in Figure e), small Au nanoparticles could be observed near the nanowire tips as marked by the arrows. The thickness values of pristine ZnO and ZnO-Au 5 s films were directly estimated based on the scale bars of cross-sectional images in Figure e to be similarly around 1.0–1.3 μm. The elemental distributions and composition of the Au-decorated ZnO nanowires were determined by EDS mapping analysis as illustrated in Figure f. The EDS maps of O, Zn, and Au confirmed the uniform distributions of these elements over the selected region. The quantified elemental compositions from the EDS spectrum were 18.35 wt % (48.53 atom %) O, 78.40 wt % (50.77 atom %) Zn, and 3.25 wt % (0.70 atom %) Au, signifying an appreciable Au loading amount on ZnO nanowires. Table reports the quantified elemental compositions of all samples based on EDS data. The observed Zn:O atomic ratios were consistent with the expected ZnO stoichiometry, confirming the formation of the ZnO phase in all samples. Furthermore, the Au content increased accordingly with increasing sputtering duration, reaching 2.64 wt % (0.54 atom %), 3.25 wt % (0.70 atom %), and 6.30 wt % (1.44 atom %), for sputtering times of 2, 5, and 10 s, respectively.

2. Elemental Analysis Results from the EDS Spectra of Pristine and Au-Sputtered ZnO Nanowires with Sputtering Times of 2, 5, and 10 s.

	O		Zn		Au
samples	wt %	atom %	wt %	atom %	wt %	atom %	Zn/O ratio
ZnO	18.70	48.45	81.30	51.55			1.06
ZnO-Au 2 s	20.72	52.21	76.64	47.25	2.64	0.54	0.90
ZnO-Au 5 s	18.35	48.53	78.40	50.77	3.25	0.70	1.05
ZnO-Au 10 s	15.88	44.81	77.82	53.75	6.30	1.44	1.20

X-ray photoelectron spectroscopy (XPS, AXIS Ultra DLD, Kratos) was employed to determine the chemical compositions and elemental oxidation states of the pristine and 5 s Au-sputtered ZnO nanowires as shown in Figure . The broad-scan spectra (Figure a) confirmed the presence of all anticipated elements and carbon (C) from typical surface organic contaminants. From the inset table of Figure a, the pristine ZnO nanowires contained Zn and O elements with the Zn:O atomic ratio of 1:1.08 in good agreement with the expected ZnO stoichiometry (1:1). However, the determination of elemental composition for 5 s Au-sputtered ZnO nanowires was not feasible due to the spectral overlap between the Au 4f and Zn 3p core-level peaks that prevented accurate deconvolution and quantitative analysis. For the Zn 2p region (Figure b), the pristine and Au-sputtered ZnO samples displayed the Zn 2p_1/2:Zn 2p_3/2 doublet pairs positioned at 1044.5:1021.5 eV and 1044.9:1021.9 eV, respectively. These binding energies similarly corresponded to the Zn²⁺ state of ZnO. ,− Concerning the O 1s level (Figure c), both samples similarly comprised three Gaussian components. The deconvoluted peaks of pristine ZnO nanowires appeared at 530.2, 531.0, and 532.1 eV when those of 5 s Au-sputtered ZnO nanowires were located at 530.7, 531.9, and 532.6 eV. The major O 1s peak at 530.2–530.7 eV could be assigned to oxygen ions (O^2–) in the wurtzite ZnO lattice, , while the minor ones at 531.0–531.9 and 532.1–532.6 eV could be related to chemisorbed oxygen species and hydroxyl species (OH) on ZnO surfaces, respectively. Regarding the Au element (Figure d), the overlapped Au 4f and Zn 3p core levels of 5 s Au-sputtered ZnO nanowires could be deconvoluted into Zn 3p_1/2:Zn 3p_3/2 peaks at 91.6:88.7 eV and Au 4f_5/2:Au 4f_7/2 peaks at 87.7:84.0 eV. , The Zn 3p peak positions confirmed the presence of the Zn²⁺ state, while the Au 4f peak locations closely corresponded to a metallic Au⁰ state. ,− The findings verified the formation of metallic Au nanoparticles on ZnO nanowires.

4.

XPS spectra of pristine ZnO nanowires and 5 s Au-sputtered ZnO nanowires on Al₂O₃ substrates: (a) survey scan and high-resolution core-level scans of (b) Zn 2p, (c) O 1s, and (d) Au 4f-Zn 3p.

3.2. Gas-Sensing Characteristics

The selectivity behaviors of all ZnO-based sensors were tested among environmental gases at relevant concentrations, comprising C₂H₅OH (70 ppm), C₂H₆O (70 ppm), NH₃ (35 ppm), NO₂ (3 ppm), and CO (700 ppm) at 350 °C and 50% RH as reported in Figure . The pristine ZnO sensor exhibited relatively high NO₂ response compared with those of other gases, indicating a limited selectivity toward NO₂. Upon sputtering Au for 2–5 s, the C₂H₅OH responses rose greatly, and the C₃H₆O responses increased moderately, while the responses to NO₂, CO, and NH₃ changed minorly. As the sputtering time further increased to 10 s, the responses to all gases decreased remarkably to very low levels, indicating the adverse effects of the excessive Au amount on gas-sensing performance. Hence, the optimal Au sputtering time of 5 s provided the highest C₂H₅OH selectivity at 350 °C and 50% RH.

5.

Selectivity behaviors of pristine ZnO nanowires and Au-sputtered ZnO nanowires prepared with sputtering times of 2, 5, and 10 s among various gases including CO (700 ppm), C₂H₅OH (70 ppm), C₃H₆O (70 ppm), NH₃ (35 ppm), and NO₂ (3 ppm) at 350 °C and 50% RH.

Figure demonstrates the responses of all ZnO-based sensors at different working temperatures (250–400 °C) toward 70 ppm of C₂H₅OH under 50% RH. The C₂H₅OH response of the pristine ZnO sensor increased gradually with increasing temperature from 250 to 400 °C. Distinctively, the responses of 2–10 s Au-sputtered ZnO sensors rose significantly with rising temperature from 250 to 350 °C before declining at 400 °C, resulting in an optimum sensing temperature of 350 °C. Notably, the 5 s Au-sputtered ZnO sensor offered the maximum response of ∼94 toward 70 ppm of C₂H₅OH at 350 °C and 50% RH. Consequently, the Au-decorated ZnO sensor with the optimum Au sputtering time of 5 s gave high response enhancement of nearly 50 times compared with the pristine ZnO sensor (response = 1.86 to 70 ppm of C₂H₅OH) at 50% RH and 350 °C.

6.

Sensor responses of pristine and 2 s to 10 s Au-sputtered ZnO nanowires to 70 ppm of C₂H₅OH at 50% RH and 250–400 °C.

The C₂H₅OH-sensing characteristics of pristine and Au-sputtered ZnO nanowires were further evaluated for practical applications by exposing to various C₂H₅OH concentrations (5–70 ppm) in dry air (0% RH) and humid air (30–50% RH at 20 °C) at the best sensing temperature of 350 °C as demonstrated in Figure . In dry air (Figure a), the baseline conductance tended to increase upon Au sputtering. The effect might be attributed to the formation of ohmic interfaces between Au nanoparticles and ZnO nanowires (metal–semiconductor: M–S junctions), to be additionally discussed in the next section. Upon C₂H₅OH injection, the conductance of each sensor increased sharply, affirming an n-type semiconducting character toward a reducing gas. After reapplying dry air, the conductance gradually returned to its baseline, signifying good reversibility. Importantly, the conductance change due to C₂H₅OH exposure of the ZnO nanowire-based sensor was substantially enhanced with increasing Au sputtering time up to 5 s, followed by a significant reduction at a longer sputtering time (10 s).

7.

Conductance profiles of pristine and 2 s to 10 s Au-sputtered ZnO sensors while exposing to 5–70 ppm of C₂H₅OH and related sensor response versus C₂H₅OH concentration in (a, b) dry air (0% RH), humid air with (c, d) 30% RH, and (e, f) 50% RH, (g) conductance profiles of the 5 s Au-sputtered ZnO sensor to 70 ppm of C₂H₅OH with 0–50% RH, and (h) response to 70 ppm of C₂H₅OH of all sensors with 0–50% RH at 350 °C.

The sensor response from the conductance change calculated at each C₂H₅OH concentration as presented in Figure b rose correspondingly to the highest value as the Au deposition time increased up to 5 s at 350 °C. At 70 ppm of C₂H₅OH, the 5 s Au-sputtered ZnO sensor exhibited the superior response of 8.6 with t _res and t _rec of 520 and 1200 s compared with the pristine ZnO sensor displaying a response of 1.86 with a t _res and t _rec of 280 and 640 s, respectively. These long response and recovery times were in line with the typical values reported in the literature for this dynamic-type setups, in which sensors operated under a constant flow of gas. They arise from the combination of several phenomena, including reaction kinetics at oxide surfaces, gas diffusion through the pores and structural defects of materials, and gas flow through the setup pipelines to fill and purge the test chamber. In our case, the latter effect would play an important role because our chamber was designed to accommodate many sensors with a volume of 1 L for comparative studies. At the employed gas flow rate of 200 cm³/min, a minimum time of 300 s was needed to completely fill the chamber. Considering the working temperature of our sensors (350 °C), response, and recovery times measured with small-volume setups and high flow rates, suitable for fast filling/purging within a few seconds, should lie in the range of 10–100 s. Diffusion of point defects through nanowires would instead be expected on the order of 2–5 s for our nanowires (T = 350 °C, diameter between 10 and 30 nm). Based on the information, we could roughly estimate the intrinsic response and recovery times of our sensors, i.e., values not affected by the characteristic times of the setup, as not exceeding 100 s, with either reaction kinetics of gas diffusion through pores or defects, being the limiting phenomena.

Concerning the gas concentration dependency, the sensor response (y) could correlate well with C₂H₅OH concentration (x) according to a power relation, y = ax ^b , as listed inside Figure b, where a and b represent the gas sensitivity factor and the exponential constant linked to the surface reactions between dominant oxygen species and analyte molecules, respectively. ,, This relationship could be used to estimate the theoretical limit of detection (LOD) of each sensor at a minimal response threshold, y _min, of 0.1, set based on the three-time signal-to-noise criterion (3 × 0.033). The LOD (x) of the optimal Au-ZnO sensor was calculated to be 0.024 ppm (from the inset of Figure b: y = 0.8012x ^0.5567 = 0.1), which was acceptably low for many ethanol-sensing applications.

The C₂H₅OH-sensing characteristics of all sensors measured under humid conditions (30 and 50% RH at 20 °C) are displayed in Figure c–f. Evidently, the conductance changes of Au-sputtered ZnO sensors (2, 5, and 10 s) were remarkably high compared with their corresponding changes under the dry conditions. The effect might be related to the additional electrons generated through the reducing reaction between water molecules and preadsorbed O^– species on Au/ZnO surfaces. − The increased carrier concentration in Au could promote the Au catalytic activity, thereby enhancing the gas response. The response enhancement was interestingly dependent on the RH value. At the low humidity level of 30% RH (Figure c–d), the 2 s Au-sputtered ZnO sensor demonstrated the optimal ethanol response of 72.2 to 70 ppm of C₂H₅OH at 350 °C, with t _res = 599 s, t _rec = 1283 s, and LOD (x) = 16 ppb (calculated from the inset of Figure d: y = 0.25535x ^0.7831 = 0.1). Conversely, the 5 s Au-sputtered ZnO sensor offered a superior response of 93.8 to 70 ppm of C₂H₅OH, with t _res = 550 s, t _rec = 1220 s, and a very low LOD (x) of 3.7 ppb (computed from the inset of Figure f: y = 4.8731x ^0.6947 = 0.1) at the higher humidity level of 50% RH (Figure e).

The impact of humidity on the response toward 70 ppm of C₂H₅OH of the optimal Au-sputtered ZnO sensor (ZnO-Au 5 s) at 350 °C is presented in Figure g. Notably, the ZnO-Au 5 s sensor exhibited lower baseline conductance but larger change in conductance upon exposure to ethanol at 50% RH than those at 0% and 30% RH. This behavior indicated possible additional oxidation of Au-ZnO surfaces by H₂O molecules. The corresponding ethanol responses of the ZnO-Au 5 s sensor were compared with those of other sensors at 0–50% RH as displayed in Figure f. Apparently, the ZnO-Au 5 s sensor displayed the highest ethanol responses at 50% RH when the ZnO-Au 2 s and ZnO-Au10s sensors showed the highest responses at 30% RH. In the case of pristine ZnO sensors, the ethanol responses were low and slightly increased from 1.9 to 2.4 with increasing humidity level from 0 to 30% RH before declining to almost the same value at 50% RH (inset). It should be noted that our experimental constraints involving present test conditions and instrumentations prevented measurements at higher humidity levels (>50% RH). The limitation disallowed the evaluation of sensor performance for breath analysis under the actual breath conditions (>80% RH). Future studies involving measurements at high humidity levels should be conducted to verify their suitability in these important applications. The behaviors of baseline conductance and the remarkable ethanol response enhancement under humid conditions of the Au-sputtered ZnO nanowires would be elaborated in the subsequent mechanistic section.

The response values achieved at low concentrations under humid conditions and the response enhancement due to Au loading were higher than those of other recently developed Au/ZnO-based C₂H₅OH gas sensors as summarized in Table . For instance, Au-decorated ZnO nanostructures synthesized via hydrothermal, photochemical, and other wet-chemical routes provided responses ranging from 0.35 to 41.3 toward 50–100 ppm of C₂H₅OH at elevated temperatures (200–380 °C). ,,,,, For other fabrication techniques, including thermal oxidation, mixed powder oxidation, and ion exchange, the resulting sensors required much higher C₂H₅OH concentrations (1000–2000 ppm) to achieve comparable or higher responses. ,, Specially, Au-sputtered CVD-grown ZnO microwires showed a modest response of approximately 5 to 80 ppm of C₂H₅OH at room temperature (15 °C) but required UV irradiation.

The comparative performance results indicated that the sensors based on Au-sputtered ZnO nanowires could be promising for practical ethanol sensing applications. The production costs of the sensors were then estimated to assess their economic feasibility for commercialization. Based on the present laboratory-scale production, the material and operating costs of thermal evaporation–condensation, sputtering, and sensor packaging processes were tentatively estimated to be in the ranges of 12–15, 4–5, and 8–10 Euro per sensor, respectively. ,− The use of Au sputtering insignificantly increased the unit cost because of very short sputtering times. Thus, the total cost would be approximately in the range of 24–30 Euro per sensor. In the future industrial-scale production using large-batch fabrication facilities, the unit cost would expectedly drop considerably to 8–10 Euro, suggesting their commercial viability. Compared with standard commercial metal oxide gas sensors having the current price in the range of 10–20 Euro per sensor, this sensor technology could be competitive owing to the employed low-cost techniques compatible with large-scale production while delivering superior ethanol-sensing performance.

3.3. Mechanistic Discussions of Au-Decorated ZnO Nanowires

From the electrical data (Figure ), Au sputtering with the optimal time of 5 s significantly promoted the baseline conductance and ethanol response of ZnO nanowires. The attributes could be ascribed to two mechanistic reasons: (1) the creation of metal–semiconductor (M–S) junctions at Au/ZnO interfaces and (2) the catalytic activity of Au nanoparticles distributed across the ZnO nanowire surfaces. The XPS analysis confirmed the presence of metallic Au (Au⁰) nanoparticles (Φ_Au ∼ 5.1 eV) decorated on nanowires exhibiting the predominant Zn²⁺ oxidation state of n-type ZnO (E _g ∼ 3.4 eV and Φ_ZnO ∼ 5.3 eV , ), which could be represented with the energy band diagram as depicted in Figure a. Upon the junction formation, the band bending due to the work function disparity (Φ_Au < Φ_ZnO) could cause electrons in the conduction band of ZnO to accumulate near the Au–ZnO interface, establishing an ohmic M–S contact as presented in Figure b. The ohmic nature of the junction was presently inferred from the work function assumptions and from the observed trend of increase in the baseline conductance of ZnO nanowire sensors after Au decoration. Thus, additional experimental data, including current–voltage characteristics of pristine ZnO nanowires in contact with Au electrodes, would be needed to verify the inference in the future.

8.

Energy band diagrams of metallic Au and ZnO nanowires (a) before and (b) after the Au/ZnO nanowire junction under thermal equilibrium.

The mechanistic models of the C₂H₅OH-sensing process for ZnO nanowire-based sensors under dry and humid conditions were proposed as displayed in Figure . In air under dry conditions (Figure a), the conductivity of pristine n-type ZnO nanowires would be determined by ionization of intrinsic defects (primarily oxygen vacancies) and chemisorption of oxygen species (O^– and O^2–), whose dominant form could depend on the working temperature (T) according to ,,,

$${\mathrm{O}}_{(\mathrm{ads})}+{{\mathrm{e}}^{-}}_{(\mathrm{ZnO}\mathrm{CB})}\to {{\mathrm{O}}^{-}}_{(\mathrm{ads})}100\mathrm{°C}<\mathrm{T}<300\mathrm{°C}$$ 3

$${{\mathrm{O}}^{-}}_{(\mathrm{ads})}+{{\mathrm{e}}^{-}}_{(\mathrm{ZnO}\mathrm{CB})}\to {{\mathrm{O}}^{2-}}_{(\mathrm{ads})}\mathrm{T}>300\mathrm{°C}$$ 4

These oxygen species would capture free electrons from the ZnO conduction band, generating electron depletion layers on the nanowire surfaces. This could cause a decrease in sensor conductance, resulting in a low baseline conductance of ZnO as observed in Figure .

9.

Mechanistic models of the C₂H₅OH-sensing process: (a, b) pristine ZnO nanowires and (c, d) optimal Au-sputtered ZnO nanowires (ZnO-Au 5 s) under dry conditions and zoomed-in surface models under humid conditions for (e–h) pristine ZnO nanowires and (i–l) optimal Au-sputtered ZnO nanowire sensors.

With Au decoration (Figure c), the ohmic M–S junctions of the Au–ZnO interface could form, leading to an increase in the baseline conductance of the Au-decorated ZnO sensor. Meanwhile, the catalytic activity of Au would exert an opposite influence through the spillover effect, in which catalytic Au nanoparticles dissociated atmospheric O₂ molecules into O atoms. This effect would enhance the formation of O^– or O^2– species by taking electrons from ZnO according to Reactions and . These oxygen ions could migrate (“spill over”) from the Au sites onto ZnO surfaces, thereby further extracting electrons and partly lowering the baseline conductance in air. The reactions would lead to the induction of electron-depleted regions on the ZnO surfaces and at the Au–ZnO interfaces. However, the depletion widths at the interfaces would be thinner than those on the surfaces due to the higher concentration of electrons supplied by the ohmic M–S junctions. The recession of electron depletion regions due to the ohmic M–S junction effect could be responsible for the significant increase in conductance observed upon Au sputtering (Figure a). When the Au sputtering time was long, many Au nanoparticles tended to become poorly dispersed and agglomerated. Such aggregation adversely affected their size-dependent electrical properties, particularly the ability to modulate the widths of carrier depletion layers, since larger nanoparticles would be inherently less effective compared with smaller ones. The SEM image in Figure d supported this assumption, presenting the tendency of Au nanoparticle size increase with increasing sputtering time from 5 to 10 s.

When exposed to C₂H₅OH under dry conditions, the C₂H₅OH molecules would directly adsorb onto ZnO nanowire surfaces and react with prechemisorbed oxygen ions according to ,,

$${\mathrm{C}}_2{\mathrm{H}}_5{\mathrm{OH}}_{(\mathrm{g})}\to {\mathrm{C}}_2{\mathrm{H}}_5{\mathrm{OH}}_{(\mathrm{ads})}$$ 5

$${\mathrm{C}}_2{\mathrm{H}}_5{\mathrm{OH}}_{(\mathrm{ads})}+6{{\mathrm{O}}^{-}}_{(\mathrm{ads})}\to 2{\mathrm{CO}}_2+3{\mathrm{H}}_2\mathrm{O}+6{\mathrm{e}}^{-}$$ 6

$${\mathrm{C}}_2{\mathrm{H}}_5{\mathrm{OH}}_{(\mathrm{ads})}+6{{\mathrm{O}}^{2-}}_{(\mathrm{ads})}\to 2{\mathrm{CO}}_2+3{\mathrm{H}}_2\mathrm{O}+12{\mathrm{e}}^{-}$$ 7

The reactions could occur only at a limited number of active sites with oxygen species on pristine ZnO nanowires, resulting in a low ethanol response (Figure b). With Au loading (Figure d), Au catalytic nanoparticles could significantly enhance the adsorption of C₂H₅OH molecules on ZnO surfaces (Reaction ()), thereby promoting C₂H₅OH oxidation (Reactions and ) and increasing C₂H₅OH response. The detailed reaction processes could be elaborated through the following steps. First, C₂H₅OH molecules could selectively adsorb at Au nanoparticulate surfaces, owing to their high catalytic activity. ,,, These adsorbed molecules would then diffuse and spill over onto ZnO nanowire surfaces. Subsequently, consecutive oxidation reactions with O^– or O^2– species would produce H₂O and CO₂ while returning electrons to the conduction band. ,,, Electron liberation would reduce the widths of depletion layers (Figure d), causing an increase in conductance as observed previously in Figure a. Additionally, the oxidation of C₂H₅OH could be further facilitated by the oxygen spillover effect, which provided extra preabsorbed oxygen species. This effect could be maximized at the optimum sensing temperature of 350 °C when the adsorption of O^– or O^2– dominated over their desorption. Particularly, the optimal Au sputtering time of 5 s induced a prominent increase in conductance relative to the case of the unloaded ZnO sensor as shown in Figure a. At this sputtering time, the pronounced catalytic activity of Au nanoparticles could result in more active sites for C₂H₅OH adsorption and reaction with surface oxygen species, leading to the optimal C₂H₅OH-sensing performance of Au-ZnO nanowires.

Under humid conditions at a low humidity level (30% RH), the pristine ZnO sensor exhibited slightly higher ethanol response than that under the dry conditions at the optimal working temperature of 350 °C as seen earlier in Figure h (inset). This slight enhancement could be attributed to the reducing reaction between H₂O molecules and preadsorbed oxygen species (O^– or O^2–) as described by ,−

$${\mathrm{H}}_2{\mathrm{O}}_{(\mathrm{g})}\to {\mathrm{H}}^{+}+{\mathrm{OH}}^{-}$$ 8

$${\mathrm{H}}^{+}+{{\mathrm{O}}^{-}}_{(\mathrm{ads})}\to {\mathrm{OH}}_{(\mathrm{ads})\mathrm{Zn}}$$ 9

$${\mathrm{H}}_2{\mathrm{O}}_{(\mathrm{g})}+{{\mathrm{O}}^{-}}_{(\mathrm{ads})}\to 2{\mathrm{OH}}_{(\mathrm{ads})\mathrm{Zn}}+{\mathrm{e}}^{-}$$ 10

$${\mathrm{H}}_2{\mathrm{O}}_{(\mathrm{g})}+{{\mathrm{O}}^{2-}}_{(\mathrm{ads})}\to 2{\mathrm{OH}}_{(\mathrm{ads})\mathrm{Zn}}+2{\mathrm{e}}^{-}$$ 11

In the reactions, H₂O molecules dissociated into hydroxyl groups (OH^–) and protons (H⁺). The hydroxyl groups would then bond with surface zinc cations, releasing additional electrons back into the conduction band of the ZnO (Figure e). These extra electrons would facilitate the creation of oxygen species (O^– and O^2–) to react with ethanol molecules via Reactions –, causing reduced widths of depletion layers (Figure f) and slightly enhanced ethanol response of the pristine ZnO sensor (Figure d).

At a high humidity level of 50% RH, water molecules could undergo a two-step adsorption process. Initially, water molecules would chemisorb onto active surface sites with oxygen species (O^– and O^2–). Next, continuous water adsorption could lead to the physisorption of additional water molecules onto the initial chemisorbed layer, forming multilayers interconnected via hydrogen bonds (Figure g). As a result, protons could move freely within these physisorbed water multilayers via the Grotthuss mechanism (H₂O + H₃O⁺ → H₃O⁺ + H₂O). ,,− This phenomenon could result in the degraded ethanol response of the pristine ZnO sensor with increasing humidity level from 30 to 50% RH, as observed previously in Figure h (inset). The degradation would occur because the active sites became heavily occupied by water molecules, thereby reducing the probability of interactions with ethanol molecules on the ZnO surface as shown in Figure h.

With Au loading, the negative effects of humidity on gas-sensing performance could be mitigated via three potential mechanisms: (i) promoting oxygen adsorption and chemical sensitization, (ii) inhibiting hydroxyl formation and protecting the active sites on the ZnO surface, and (iii) suppressing the Grotthuss process. − According to the ethanol-sensing results under humid conditions of the optimal sensors (ZnO-Au 5 s) (Figure c–h), the ethanol response was enhanced with increasing humidity, particularly at 50% RH. Typically, water vapor would compete with oxygen molecules for adsorption on active sites, leading to the formation of hydroxyl species (OH^–). This process could reduce the amount of chemisorbed oxygen species (O^– or O^2–) on ZnO surfaces, ,, thereby diminishing the sensor effectiveness at high humidity levels. However, the loaded Au nanoparticles would provide highly effective adsorption sites for oxygen rather than water vapor and inhibit hydroxyl group formation on ZnO surfaces, ,,,, giving additional supply of reactive oxygen species under humid conditions. Specifically, the catalytic properties of Au could promote the dissociation of adsorbed oxygen atoms into reactive oxygen species, which subsequently migrated via the spillover effect onto ZnO surfaces (Figure i). After exposing to ethanol, ethanol molecules would react with oxygen species, which also directly adsorbed on Au catalytic nanoparticles, leading to a greater reduction of the depletion regions at the Au–ZnO interfaces (Figure j) compared with those under dry conditions (Figure d). As a result, the presence of the Au additive could moderately enhance the ethanol response of the ZnO nanowires under a weakly humid condition (30% RH) as evident in Figure c,d.

Under a highly humid condition (50% RH), the sensor conductance change would still be mainly caused by reducing reactions with ethanol rather than water molecules owing to the structural attributes of the optimal Au-sputtered ZnO nanowires. Under such a humid environment, water molecules would normally form multilayers, inducing the Grotthuss mechanism, as previously described for the pristine ZnO sensor. The presence of Au nanoparticles distributed over the ZnO nanowire surfaces (Figure c) could mitigate this problem by increasing nanoscale roughness and changing the surface energy, largely preventing the formation of continuous water layers. ,, As a result, the proton hopping pathways in the Grotthuss mechanism would be limited. Furthermore, the sensor could benefit from extra electrons induced by humidity (Reactions and ), leading to enhanced oxygen chemisorption at the Au-ZnO interfaces with a limited influence of the Grotthuss mechanism. The effects combined with Au’s ability to induce and supply oxygen species via the spillover effect (Figure k) could result in a significantly pronounced ethanol response compared with that at a low humidity level (30% RH), as formerly seen in Figure . Notably, the ZnO-Au 5 s sensor exhibited the lowest baseline conductance at 50% RH, possibly corresponding to the highest density of surface oxygen species induced by the catalytic oxygen chemisorption of Au under humid conditions. Consequently, a high density of oxygen species available for reaction with ethanol molecules on the optimal Au-sputtered ZnO nanowires could be attained even under highly humid conditions (Figure l), resulting in a remarkably enhanced ethanol response (Figure e–h).

Distinctively, the 2 s Au-sputtered ZnO nanowire sensors (ZnO-Au 2 s) displayed the best ethanol response at 30% RH before degrading considerably as the humidity level rose from 30 to 50% RH (Figure h). At low humidity (30% RH), the moderate presence of water molecules could enhance the ethanol sensing performance by providing additional electrons to the ZnO conduction band through reducing reactions (Reactions –). These extra electrons could facilitate the creation of oxygen species (O^– and O^2–) reacting with ethanol. For the ZnO-Au 2 s sensor, the highly dispersed, ultrasmall Au nanoparticles might provide a more optimal catalytic surface for the oxygen and ethanol spillover effects at this low humidity level owing to the presence of low-density water molecules compared with the ZnO-Au 5 s sensor, leading to the highest ethanol response. At high humidity (50% RH), the density of ultrasmall Au nanoparticles on the ZnO-Au 2 s surfaces might be insufficient to limit the formation of physisorbed multilayers of water, which induced the Grotthuss proton hopping pathways. As a result, the Au-ZnO surfaces would be highly covered with water molecules, significantly diminishing the reducing reaction between ethanol molecules and preadsorbed oxygen species. For the ZnO-Au 5 s sensor, Au nanoparticles were moderately sized and well-distributed to effectively enhance spillover oxygen chemisorption while limiting the interference from the Grotthuss mechanism under highly humid conditions (Figure l). In the case of the ZnO-Au10s sensor, aggregation of Au into relatively large nanoparticles on ZnO nanowires (Figure d) also led to substantially reduced ethanol responses under humid conditions, while the presence of Au nanoparticles still resulted in higher ethanol responses than those under dry conditions (Figure h), which could also be attributed to the suppression of the Grotthuss mechanism by Au nanoparticles.

Lastly, it should be remarked that the validity of the proposed mechanisms would need to be verified with direct spectroscopic evidence of surface species, particularly in situ diffuse reflectance infrared Fourier transform spectra (DRIFTS) and temperature-programmed desorption (TPD) data, under operating conditions.

4. Conclusions

In summary, ZnO nanowires were synthesized via the thermal evaporation–condensation method and subsequently decorated by Au nanoparticles with different sputtering times of 2–10 s to optimize the nanostructure and ethanol-sensing performance. Among the fabricated sensors, the Au-decorated ZnO nanowires prepared with the sputtering time of 5 s exhibited the highest response of 93.75 to 70 ppm of C₂H₅OH, which was markedly superior to that of the pristine ZnO nanowire sensor (1.86) at the optimum sensing temperature of 350 °C and 50% RH. Besides, the optimally Au-decorated ZnO nanowire sensor displayed humidity-enhanced behaviors with significantly enhanced ethanol responses at 50% RH. Additionally, these nanowires displayed high C₂H₅OH selectivity relative to CO, C₃H₆O, NH₃, and NO₂ at their relevant concentrations. The humidity-enhanced mechanism might be ascribed to the ability of very finely sputtered Au nanoparticles to enhance oxygen chemisorption and suppress the formation of surface hydroxyl and hydronium species on ZnO nanowires. The observed ethanol response improvement and humidity-enhanced effect were proposed based on the formation of ohmic metal–semiconductor junctions formed at Au–ZnO interfaces and the distinctive capabilities of very finely sputtered Au nanoparticles on the enhancement of oxygen chemisorption and suppression of the Grotthuss mechanism. The significance of this study was the finding of unusual humidity-enhanced effects from sputtered Au nanoparticles on thermally evaporated ZnO nanowires, which resulted in a highly effective ethanol sensor under humid conditions for practical ethanol-sensing applications.

S.S.: investigation, visualization, and writingoriginal draft. A.W.: investigation and writingreview and editing. G.S.: writingreview and editing. A.P.: conceptualization, investigation, methodology, supervision, and writingreview and editing. D.Z.: investigation, methodology, supervision, and writingreview and editing. E.C.: conceptualization, investigation, methodology, supervision, and writingreview and editing. C.L.: conceptualization, investigation, methodology, supervision, and writingreview and editing. M.P.: investigation, methodology, visualization, writingoriginal draft, and writingreview and editing.

The authors declare no competing financial interest.