Optimized SILAR Growth of Vertically Aligned ZnO Nanorods for Low-Temperature Acetone Detection

Abstract

Vertically oriented morphologies of the semiconducting metal oxide (SMO) surface provide a simple and effective means of enhancing gas sensor performance. We successfully synthesized explicitly aligned ZnO nanorods using a simple automated SILAR technique to improve acetone detection. In this work, we found that vertically oriented morphologies, such as well-aligned ZnO nanorods, can significantly enhance the sensor response due to an increase in specific active area and electron mobility, allowing a faster response to changes in the gas environment. The optimal operating temperature for our ZnO nanorod-based sensors in detecting acetone gas is 260 °C. At this temperature, the sensors exhibit a 96% response with a rapid response time of just 3 s. The improved sensing performance is attributed to both electronic and chemical sensitization mechanisms, which enhance the formation of active sites and shorten electron diffusion paths.

1. Introduction

Acetone, as a volatile organic compound (VOC), is harmful and toxic and can be a serious problem for human health and safety. It can cause neurological disorders, respiratory problems, damage to internal organs, and irritation of the respiratory tract, eyes, and skin. It can cause redness, stinging, and even burning, as well as irritation of the central nervous system, resulting in headaches, dizziness, confusion, and even nausea. According to the Occupational Safety and Health Administration (OSHA), acetone is immediately hazardous at concentrations above 20,000 ppm (2%). Furthermore, acetone in human breath is a biomarker for type 1 diabetes [1,2]. The concentration of acetone from the breath of a healthy body is less than 0.76 ppm, while that of a diabetic patient is greater than 1.71 ppm [3]. Additionally, many industrial processes emit acetone vapors that must be controlled below safety thresholds typically set around several hundred ppm to prevent health hazards and environmental pollution [4]. For example, prolonged exposure to acetone concentrations exceeding 173 ppm has been shown to cause significant damage to human organs. Consequently, occupational health guidelines establish an exposure limit of 250 ppm to mitigate potential health risks [5]. Consequently, the development of gas sensors for the rapid and selective detection of acetone has attracted researchers’ interest in recent years.

Thanks to current advances in micro-nanotechnologies and the emergence of new nanostructuring processes [6], it has recently been possible to design a new generation of extremely miniature sensors housing a wide variety of highly sensitive nanostructures with novel properties for gas detection. The aim is to converge on easily reproducible sensing devices produced at advantageous manufacturing costs with optimum sensing performance [6,7,8]. Optimizing the cost of these devices requires affordable materials combined with simple, controllable manufacturing techniques that are ideally suited to industrial-scale mass production.

While commercially available MOS-based sensors are widely acknowledged for their low unit cost and ease of integration, it is important to note that the cost-effectiveness of the final device does not necessarily reflect the complexity or capital expenditure of the fabrication technique used. Physical deposition methods, such as sputtering or PLD, though enabling high-quality films, often entail significant infrastructure investment due to their reliance on vacuum systems, high temperatures, and cleanroom conditions. In contrast, solution-based methods, such as SILAR, sol–gel, or hydrothermal synthesis, offer cost-effective, scalable, and low-temperature alternatives that are particularly advantageous in resource-limited contexts. However, these chemical methods must also demonstrate sufficient reproducibility and stability to ensure commercial viability [9,10,11,12,13]. To address this issue, we chose zinc oxide (ZnO) as an affordable, non-toxic semiconductor material with a high surface-to-volume ratio, unidirectional electron transport, enhanced gas diffusion, and high sensitivity to various gases and vapors. To produce and integrate ZnO thin films on detection micro platforms, we chose to use the “Successive Ionic Layer Adsorption and Reaction (SILAR)” technology in this work. This method, automated at our laboratory scale, bridges this gap by combining the economic and procedural simplicity of chemical methods with improved control and reproducibility, making it a promising candidate for the scalable fabrication of high-performance gas sensors.

In addition, nanostructured ZnO thin films are also available in a variety of forms, such as nanorods, nanowires [11,12,14,15], nanospheres [16,17], nanograins [18], and nanosheets, which can form plate- or flake-like shapes [19,20]. Beyond these commonly investigated forms, other complex architectures, such as nanoflowers, nanotubes, and nanoporous films, have been demonstrated to significantly influence gas sensing properties due to their enhanced surface area and porosity [21,22,23]. Notably, nanoporous ZnO layers synthesized via spark ablation techniques exhibit a highly porous network that facilitates gas diffusion and increases the density of active sites, thereby improving sensor response and response time. These diverse nanostructures, as comprehensively reviewed by Mani et al. [21], provide critical insights into tailoring ZnO morphologies to optimize gas sensor performance. However, vertically oriented shapes remain the most recommended because of their larger surface area, enabling them to make increased and complete contact with certain gases and better electron mobility [24,25].

Furthermore, studies on acetone gas detection have explored various sensor morphologies, each with specific metrological characteristics and performance limitations. These include poor response even at high concentrations, long response times, and high operating temperatures (over 300 °C). On the other hand, for the approaches based on nanorod morphology, whether for pristine or doped ZnO (using ZnO nanorods as the main material and introducing nanoparticles of different metals), shapes with non-aligned and non-homogeneous random orientation are observed, affecting sensor response by increasing current losses as well as the signal. For example, Salehi et al. [26] constructed a hybrid gas sensor based on electrospun ZnO nanofibers and reduced graphene oxide (rGO) to detect acetone. The fabricated sensor showed an increase in response from 230 to 400% for a 200 ppm acetone concentration. Wang et al. [27] synthesized ZnO flowers using a thermal decomposition method in organic solvents, followed by gold deposition on ZnO. The gold nanoparticle-modified ZnO gas sensor showed superior performance in acetone detection. Yang et al. [28] fabricated ZnO nanorods by a simple precipitation reaction and then modified them with Au nanoparticles. The gold nanoparticle-modified ZnO gas sensor showed superior performance at 300 °C, a response of ~350%, and a response/recovery of ~15/20 s. Huang et al. [29] fabricated ZnO nanosticks using a simple hydrothermal method, which was loaded in a further step with Au, Pd, and Au-Pd. The ZnO gas sensor decorated with Au-Pd nanoparticles showed better acetone detection performance, achieving ~246% response at 250 °C, with ~20/25 s response/recovery times. Al-Hadeethi et al. [30] synthesized Ag-doped ZnO nanoneedles by a facile hydrothermal method. However, the fabricated acetone sensor reached ~220% response at 370 °C and ~25/35 s response/recovery times. Y. Shen et al. [31] synthesized Au-Pd/ZnO nanorods using a simple hydrothermal technique. The synthesized ZnO nanorods showed a response of 220% under a 225 °C operating temperature and an acetone concentration of 50 ppm with ~8/12 s response/recovery. Meng et al. [32] synthesize hollow ZnO/ZnFe₂O₄ microspheres with a sea urchin-like shell–core structure by doping ZnO. The ZnO/ZnFe₂O₄-6 sensor exhibited the best response to 100 ppm acetone gas at an optimal operating temperature of 250 °C, which was four times higher than that of pure ZnFe₂O₄ sensors, and had ~400% response at 250 °C for 100 ppm acetone with ~12/18 s response/recovery. Wu et al. [33] presented a bilayer sensor combining ZnO nanorods on ZnFe₂O₄, which showed enhanced acetone detection at an optimal temperature around 250 °C, with a response of approximately 400% and response/recovery times near 12 and 18 s. Li et al. [34] developed Pt-decorated ZnO nanorods yielding highly efficient acetone sensing performance with rapid response times (~6 s) at operating temperatures near 200 °C.

Furthermore, we acknowledge the important theoretical findings of Hua et al. [35], who demonstrated that, under conditions of constant grain size and carrier concentration, spherical morphologies exhibit enhanced transducer functions compared to columnar or plate-like structures. This highlights the critical influence of morphological shape on sensor performance. In parallel, the superiority of aligned nanorod morphologies for acetone detection is well established; vertically aligned ZnO nanorods offer an increased effective surface area and promote efficient electron transport pathways, thereby facilitating stronger interactions with gas molecules. This leads to enhanced response, faster response, and quicker recovery times. Additionally, combining aligned nanorods with noble metal nanoparticle decoration forms heterojunctions that significantly boost sensor selectivity and lower operating temperature requirements. These morphological features are key factors behind the superior performance of recent sensors utilizing aligned ZnO nanorods, as demonstrated in the works of Wu et al. and Li et al. [33,34].

In this study, we employed an automated and optimized SILAR technique to synthesize explicitly aligned pure ZnO nanorods. SILAR is simpler, faster, more cost-effective, and compatible with low-temperature deposition on diverse substrates, while enabling precise control over nanorod alignment and density. Thus, we exploit the effect of an explicitly well-aligned nanorod morphology, characterized by distinct properties in terms of the surface-to-volume ratio and electron mobility, to improve acetone gas sensing capabilities. Gas detection investigations on ZnO nanorods revealed a higher response to acetone compared to other VOCs, such as ethanol. In addition, the sensors feature a faster response process, a wide detection range, and a maximum response, as well as an excellent response and long-term stability.

The main contributions of this study include (i) the implementation of a fully automated SILAR process enabling highly reproducible and uniform ZnO nanorod growth; (ii) systematic optimization of chemical and mechanical parameters to achieve dense, vertically aligned undoped ZnO nanorods with enhanced sensing properties; and (iii) demonstration of competitive acetone sensing performance at 260 °C with excellent repeatability and stability, highlighting the potential of process engineering as an effective route to high-performance sensors without doping.

2. Experimental Section

2.1. Synthesis of ZnO Nanorods

Zinc chloride (ZnCl₂, 97%), ammonia solution (25%), and acetone were purchased from Sigma-Aldrich. All chemicals were used without further purification. All aqueous solutions were prepared using distilled water having a resistance of less than 2 MΩ∙cm. ZnO thin films were deposited on glass and alumina substrates using an automated SILAR process. This process involved successive substrate immersion in three beakers according to the protocol described in Figure 1. Prior to film deposition, the substrates were subjected to a cleaning process described in previously published work [36,37].

Briefly, the cationic Zn²⁺ solution was prepared by mixing zinc chloride (ZnCl₂-2H₂O) with ammonia (NH₃) quantities, keeping the ZnCl₂ salt concentration constant at 0.07 M and the pH at 10.5, adjusted with ammonia. In order to obtain a homogeneous and transparent solution, the mixture was continuously magnetically stirred at room temperature for 1 h [38,39,40,41,42]. The substrate was then immersed in the precursor solution to enhance the adsorption of the aqueous zinc ammonium solution. For all experiments, the withdrawal speed was set at 30 cm/min and the number of growth cycles at 40. The [Zn(NH₃)₄]²⁺-enriched substrate was then immersed in hot water, followed by a rinse solution containing distilled water at room temperature to remove excess ions adsorbed to its surface.

In the last step, the samples were subjected to a postdeposition thermal treatment at 400 °C for 1 h, with a heating and cooling ramp of 5 °C per minute, under ambient air to remove the solvents and allow the precursors to oxidize and crystallize into ZnO nanocrystals.

2.2. Characterization

The study of zinc oxide film formation was carried out by scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) using JEOL model JSM-6700F equipment (JEOL, Tokyo, Japan). X-ray diffraction (XRD) patterns were acquired through a Bruker D8 diffractometer (Bruker, Karlsruhe, Germany) under Cu Kα irradiation (λ = 1.54060 Å). XRD data were recorded in the 10–90° range with a step of 0.02°. In addition, the chemical composition of the samples was determined by X-ray photoelectron spectroscopy (XPS) using an ESCALAB Xi+ (Thermo SCIENTIFIC Surface Analysis, Waltham, MA, USA) equipped with a multi-channel hemispherical electron analyzer (dual X-ray source) operating with Al K radiation (h = 1486.2 eV). A C 1s peak at 284.8 eV was used as a reference.

2.3. Manufacture and Measurement of the Gas Sensor

A commercial ceramic alumina micro-transducer was used to manufacture the gas sensor, with interdigitated gold electrodes on the front face, 200 µm thick and separated by a 200 µm gap. On the rear side of the alumina substrate, there is a heating element screen-printed for the temperature control of the ZnO film. The gas response of the ZnO samples was assessed by measuring the change in electrical resistance.

The gas sensing properties of the prepared ZnO sensors were investigated using a gas testing system described in a previously published article [43,44]. The sensor’s response to VOC vapors was monitored by measuring the evolution of its electrical resistance as a function of time using a voltage divider circuit connected to an acquisition board programmed under the LabVIEW environment. The sensor response, R (%), was defined according to Formula (1), where R_a and R_g are the sensor electrical resistance in air and VOC vapors, respectively [45]:

$$\mathit{R}\left(\mathbf{\%}\right)={\displaystyle \frac{{\mathit{R}}_{\mathit{a}}-{\mathit{R}}_{\mathit{g}}}{{\mathit{R}}_{\mathit{a}}}}\times 100$$

(1)

where R_a and R_g are the sensor resistances in air and in the presence of the target gas, respectively.

Sensitivity refers to the sensor’s quantitative ability to detect variations in analyte concentration within its linear operating range. It is defined as the slope of the response-versus-concentration curve, representing the proportional change in the sensor signal per unit increase in analyte concentration [46]. The response and recovery time were defined as the time required to achieve 90% of the total change in sensor resistance upon exposure to and removal of the target gas, respectively [47]. The VOC vapors were generated using the static liquid–gas distribution method, which involves preparing the desired gas concentration by allowing the gas to equilibrate with a liquid in a closed system. Once the sensor resistance has reached a stable relative value, the target gas is injected. The concentration is controlled based on the amount of liquid and the system conditions, taking into account the test chamber volume of 1.8 L. Throughout the measurement cycle, pure air was used as the carrier gas and circulated through the gas sensor chamber at a constant flow rate of 50 sccm. During the sensor response test, the relative humidity value in the test box was set at 30%.

3. Results and Discussion

3.1. Structural Study

Figure 2 shows the XRD diffractograms of the ZnO nanostructured films fabricated by the automated SILAR technique. The samples exhibited sharp characteristic peaks of the wurtzite ZnO hexagonal structure [48]. The crystallite sizes (D) of the ZnO films were inferred using the well-known Debye–Scherrer equation:

$$\mathit{D}={\displaystyle \frac{\mathbf{0.9}\mathit{\lambda }}{\mathit{\beta \; cos}\left(\mathit{ϴ}\right)}}$$

(2)

Since the Scherer equation does not consider lattice microstructures, such as grain boundaries, stacking errors, and inherent distortion resulting from defects in nanocrystals, it relies solely on the effect of crystal size on XRD peak broadening. However, to calculate crystallite size and intrinsic strain more accurately and conveniently, the Williamson–Hall approach is strongly recommended. As both size and strain contribute to XRD peak broadening, the Williamson–Hall method is employed, through a combination of small crystal size and strain partial, to distinguish between the two effects, which leads to a separation of the broadening into reflections rather than the Scherrer 1/cosθ connection. The calculation of crystallite size using the Williamson–Hall method is performed by determining the intercept of the curve β_Tcos(θ) as a function of 4sin(θ), according to the following size- and strain-related influence equations:

$${\mathit{\beta }}_{\mathit{T}}={\mathit{\beta }}_{\mathit{D}}+{\mathit{\beta }}_{\mathit{\epsilon }}$$

(3)

where ${\mathit{\beta }}_{\mathit{T}}$ is the total broadening, ${\mathit{\beta }}_{\mathit{D}}$ is size broadening, and ${\mathit{\beta }}_{\mathit{\epsilon }}$ is strain broadening.

$${\mathit{\beta }}_{\mathit{\epsilon }}=4\mathit{\epsilon }\mathit{t}\mathit{a}\mathit{n} \left(\mathit{\theta }\right)$$

(4)

$${\mathit{\beta }}_{\mathit{D}}={\displaystyle \frac{\mathit{k}\mathit{\lambda }}{\mathit{D}\mathit{c}\mathit{o}\mathit{s} \left(\mathit{\theta }\right)}}$$

(5)

then:

$${\mathsf{\beta }}_{\mathbf{T}}={\displaystyle \frac{\mathbf{k}\mathsf{\lambda }}{\mathbf{D}\mathbf{c}\mathbf{o}\mathbf{s}\left(\mathsf{\theta }\right)}}+4\mathit{\epsilon }\mathbf{t}\mathbf{a}\mathit{n}\left(\mathit{\theta }\right)$$

(6)

Multiplying both sides of Equation (6) by cos(θ), we find

$$\mathit{c}\mathit{o}\mathit{s}\left(\mathit{\theta }\right){\mathit{\beta }}_{\mathit{T}}=\mathit{\epsilon }\left(4\mathit{sin}\left(\mathit{\theta }\right)\right)+{\displaystyle \frac{\mathit{K}\mathit{\lambda }}{\mathit{D}}}$$

(7)

So, by determining the slope and intercept, we can deduce the lattice strain and crystallite size, respectively.

All recorded peaks were indexed according to JCPDS data (89-1379) with space group P63mc. As shown in

Table 1. Crystallographic parameters of ZnO samples calculated by the Scherrer and Williamson methods.

		Scherrer Method					Williamson Method
2θ (°)	hkl	d (Å)	FWHM (β)	Lattice Parameters (Å)	Crystallite Size D (Å)	Dislocation Density δ (×10+14) Lines m−2	βcos(ϴ(rad))	4sin(ϴ(rad))	Crystallite Size D (Å)
31.82	1 0 0	0.281	0.29923	a = 3.244 c = 5.18107			0.00599	1.0966
34.46	0 0 2	0.2601	0.29788		260.4	14.7	0.00434	1.1862	371.58
36.29	1 0 1	0.2474	0.3561				0.00581	1.2467

, the crystallographic parameters of the synthesized ZnO films show a crystallite size of 260.4 Å with a dislocation density of 14.7 × 10⁺¹⁴ lines m⁻². The results of the Williamson–Hall method gave an order crystallite size of 371.58 Å.

3.2. Compositional and Morphological Analysis

Figure 3a shows SEM micrographs of the ZnO samples. The identified morphology is a one-dimensional nanorod-like structure with order lengths of 1.4 µm. Such one-dimensional structures, particularly vertically oriented ones, are recommended morphologies for achieving a better compromise between the surface-to-volume ratio and electron mobility. The morphology of the well-aligned nanorods was carefully analyzed using Image J software (v. 1.54p), and the results revealed an average diameter of around 176 nm (Figure 3b).

For gas sensor applications, the aim is to maximize the surface-to-volume contact ratio between a material and gas while increasing electron mobility. The results of the elemental mapping of the treated ZnO samples, obtained by EDX (Figure 3c), show the presence of the main elements, namely, oxygen (O) and zinc (Zn).

XPS is essential for understanding the film’s chemical state, electronic structure, and composition. To this end, Figure 4a shows the XPS survey spectra of ZnO, with Zn, O, and C as the main constituents. Figure 4b shows the high-resolution XPS spectrum of the Zn 2p region, with two bands at 1045.0 and 1021.5 eV, corresponding to Zn 2p1/2 and Zn 2p3/2, respectively. The gap between the Zn 2p3/2 and Zn 2p1/2 electronic states was calculated as 23.0 eV. This value is generated by spin–orbit splitting and is identical to that previously reported elsewhere [49,50]. XPS spectra of the central O1s level were carried out to better understand oxygen-related defect states in ZnO thin films.

As shown in Figure 4c, the O1s XPS peak can be decomposed into two Gaussian components at ∼531.4 and 530 eV, the band at 530.0 eV can be attributed to oxygen vacancies present in the ZnO lattice and chemisorbed oxygen species on the surface (OC), and the band at 531.4 eV can be indexed to lattice-bound O₂-oxygen species (OL) in the Wurtzite structure of hexagonal ZnO [51].

3.3. Gas Sensor Properties

The sensing performance of semiconductor-based gas sensors is strongly influenced by operating temperature, as it governs the kinetics of gas adsorption and surface reactions. In this study, we evaluate the response of the pure ZnO-based sensor toward acetone and ethanol vapors at various operating temperatures to identify the optimal temperature that yields the highest response under a fixed gas concentration [52,53,54]. Figure 5a shows the evolution of the response of the ZnO-based sensor under ethanol and acetone vapors as a function of temperature. During the gas tests, the concentration of both vapors was kept constant at 1%. Thus, the optimum temperature for ethanol detection is observed at 220 °C, while that for acetone is at 260 °C. Under these conditions, the sensor exhibits a 70% and 96% response for ethanol and acetone, respectively. As shown in Figure 5b, the dependence of sensing properties on temperature was characterized by measuring the change in electrical resistance as a function of temperature regulated by a heating element. It was observed that the electrical resistance of the sensors decreases with increasing temperature.

Figure 6a shows a cross-sensitivity diagram for the detection of acetone and ethanol at 260 °C. At this temperature, the ZnO-based sensor shows a more pronounced affinity toward acetone compared with ethanol. The cross-sensitivity of the gas sensor was calculated as the ratio of the sensor’s response to the target gas compared to the total response of the sensor to all gases, multiplied by 100 to express it as a percentage. However, it is important to note that at equivalent concentration levels, the response values for acetone and ethanol remain relatively close, indicating a limitation in selectivity. This overlap may result from similar physicochemical interactions between the ZnO surface and both volatile organic compounds. Further optimization strategies, such as doping or surface functionalization, could be considered to improve molecular discrimination in future work. Figure 6b shows the sensor’s dynamic response to three successive injections of 0.4% acetone. This curve indicates that the sensor’s response to acetone is highly reproducible after three successive injections of 0.4% acetone. This curve indicates an excellent reproducibility of the sensor’s response to acetone.

Figure 7 shows the dynamic responses of the ZnO-based sensor at 260 °C to different acetone concentrations. The sensor response behaves as a rapid drop in sensor resistance following injection of this compound, compared with its baseline resistance in the latter’s absence. After removal of the acetone, the sensor gradually recovers its baseline resistance as the vapor desorbs from the surface of the sensitive layer, as shown in Figure 7. As an example, the response and recovery times displayed for the 1% acetone concentration were 3 s and 64 min, respectively. As depicted in Figure 8, the ZnO-based sensor demonstrates a nearly 0.958 linear response for acetone concentrations up to 1%, with a 0.9995 linearity observed for concentrations up to 0.6%. This linear region of the response curve is particularly important for accurate sensor calibration and dependable performance in practical applications [55]. Beyond 1%, the sensor begins to saturate.

The pure zinc oxide sensor for acetone detection developed in this work achieved a response of 96% at an activation temperature of 260 °C, with a fast response time of approximately 3 s. These performances surpass those reported in the literature for pure zinc oxide, particularly in terms of lower activation temperature and faster response time, thereby enhancing its efficiency for gas detection. Generally, the resistance of the ZnO-based sensor is affected by the thickness of the electronic depletion layer on the nanostructure’s surface, affecting charge carrier transport and overall sensing performance [56,57,58]. Compared to other published works, as reported in Table 2, where the responses of pure ZnO rarely exceed 8% at similar or higher temperatures [26,59], and response times often remain longer than 10 s, our result demonstrates significant improvement. Furthermore, doped or decorated ZnO nanostructures show varying performances, with responses ranging from ~12% to over 150%, and response times between 3 and 80 s, depending on the synthesis methods and dopants used. This highlights the crucial role of material engineering and nanostructure alignment in achieving optimal sensing characteristics.

Table 2. Acetone ZnO-based sensor response of various materials reported in the literature.

Material	T (°C)	Response (%)	Response Time (s)	Ref
ZnO	260	8	---	[59]
ZnO nanorods	219	12.9	13	[28]
Ag/ZnO nanoneedles	370	30.23	10	[30]
ZnO pure	400	2.3	--	[26]
ZnO indium	200	19.3	---	[60]
ZnO nanorods	320	50	60	[61]
Al-ZnO nanorods	250	50	12	[62]
ZnO nanorods (aligned)	280	150	9
ZnO nanoparticles (NPs)	300	11.54	25	[63]
2% Ni-doped ZnO	RT	~78	33	[64]
0.5% Fe-ZnO	365	105.7	3	[65]
Hierarchical ZnO + Au	300	76	6	[66]
5 wt% Ni- ZnO NFs	260	132.14	80	[67]
Au-Pd ZnO nanorods	225	14	7	[31]
ZnO nanorods	260	96	3	This work

Table 2. Acetone ZnO-based sensor response of various materials reported in the literature.

Material	T (°C)	Response (%)	Response Time (s)	Ref
ZnO	260	8	---	[59]
ZnO nanorods	219	12.9	13	[28]
Ag/ZnO nanoneedles	370	30.23	10	[30]
ZnO pure	400	2.3	--	[26]
ZnO indium	200	19.3	---	[60]
ZnO nanorods	320	50	60	[61]
Al-ZnO nanorods	250	50	12	[62]
ZnO nanorods (aligned)	280	150	9
ZnO nanoparticles (NPs)	300	11.54	25	[63]
2% Ni-doped ZnO	RT	~78	33	[64]
0.5% Fe-ZnO	365	105.7	3	[65]
Hierarchical ZnO + Au	300	76	6	[66]
5 wt% Ni- ZnO NFs	260	132.14	80	[67]
Au-Pd ZnO nanorods	225	14	7	[31]
ZnO nanorods	260	96	3	This work

3.4. Gas Sensing Mechanism

In the presence of air (carrier gas), oxygen molecules are absorbed on the surface of ZnO nanorods (n-type). They are then ionized to oxygen species (O₂⁻, O⁻, O²⁻) by capturing electrons from ZnO, forming an electron depletion layer on the surface of the nanostructure. This process can be described by Equation (8). Under these conditions, the sensor acquires a given base resistance [58,68].

$${{\mathrm{O}}_2}_{\left(\mathrm{g}\mathrm{a}\mathrm{z}\right)}\leftrightarrow {{\mathrm{O}}_2}_{\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)}$$

(8)

$${{\mathrm{O}}_2}_{\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)}+{\mathrm{e}}^{-}\leftrightarrow {{O_2}^{-}}_{\left(ads\right) } \mathrm{T} <100°\mathrm{C}$$

(9)

$${{O_2}^{-}}_{\left(ads\right) }+{\mathrm{e}}^{-}\leftrightarrow {{2O}^{-}}_{\left(ads\right) } 100 °\mathrm{C} < \mathrm{T} < 300 °\mathrm{C}$$

(10)

$${{2\mathrm{O}}^{-}}_{\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)}+{\mathrm{e}}^{-}\leftrightarrow {O^{2-}}_{\left(ads\right) } \mathrm{T} >300°\mathrm{C}$$

(11)

On the other hand, when the surface of ZnO, activated at 260 °C, interacts with a reducing atmosphere, such as acetone, the oxygen species adsorbed on the film surface (O⁻ being the dominant oxygen species at the temperature of 260 °C) react with the acetone molecules, following the reaction described below [69,70,71]:

$${{{\mathbf{C}\mathbf{H}}_3\mathbf{C}\mathbf{O}\mathbf{C}\mathbf{H}}_3}_{\left(\mathbf{g}\mathbf{a}\mathbf{s}\right)}+{{8\mathbf{O}}^{-}}_{\left(\mathbf{a}\mathbf{d}\mathbf{s}\right)}\to {3\mathbf{C}\mathbf{O}}_2+{3\mathbf{H}}_2\mathbf{O}+{8\mathbf{e}}^{-}$$

(12)

This process releases electrons to the ZnO nanorods, and the electron depletion layer becomes thinner, decreasing sensor resistance. The higher the gas concentration, the greater the decrease (until saturation is reached). The high response observed in our nanorod-based sensors is partly because electrons are transported along the axial direction, minimizing the current losses that occur between grain boundaries, as is the case in nanogranular film-based sensors.

4. Conclusions

The gas sensing experiments, conducted by exposing the fabricated sensors to reducing atmospheres of acetone and ethanol, confirmed the gas reactivity of the nanostructured films produced using our protocol. Notably, the ZnO nanorod-based sensor demonstrated enhanced selectivity toward acetone at an optimal operating temperature of 260 °C, showing a highly reproducible and quasi-linear response of 99.94% as a function of concentration. These results highlight the critical role of nanostructure morphology and specific surface area in optimizing gas sensing performance. The detection exhibited an increased response (sensitivity) and a fast response time at relatively low temperatures, which is directly linked to the morphology of the obtained nanorods, characterized by vertically oriented nanostructures and a high specific surface area. This vertical alignment enhances the exposure of active sites, facilitating the adsorption of gas molecules and improving the sensor’s response. Moreover, this orientation optimizes charge transport by reducing carrier recombination, which improves the sensor’s reactivity and stability across various conditions. These findings underline the importance of nanostructure morphology and surface area in optimizing performance. While positive aspects, such as high response, fast response, and low operating temperature, were demonstrated, limitations remain, especially regarding recovery time and stability under varying environmental conditions. To further enhance sensor efficiency, particularly in terms of recovery time, environmental control through the regulation of humidity and airflow plays a crucial role by accelerating the desorption of adsorbed gas molecules. Combining this strategy with advanced approaches, such as nanostructured material engineering and pulsed thermal activation, could lead to the development of high-performance gas sensors with improved responsiveness and stability under diverse operating conditions.