Investigation of Annealing Process Effects on the Response and Stability of Sprayed Co₂SnO₄ Film under Ethanol Vapor

Abstract

The ethanol (C₂H₅OH) vapor detection properties of a layer of cubic spinel cobalt stannate as-prepared Co₂SnO₄ (denoted as CTO_as), elaborated by a low-cost spray pyrolysis method, were investigated before and after an annealing process. The response of layers under ethanol vapors of different concentrations exhibited a stable behavior in a dry atmosphere at the optimum working temperatures. The elaborated nanomaterials were characterized by transmission electron microscopy (TEM), scanning electron microscopy (SEM), X-ray diffraction (XRD), direct current measurement (DC) and impedance spectroscopy techniques (AC). After the annealing process of pure CTO_as, an n-type to p-type conductivity switch was observed and attributed to Co₃O₄ apparition in the Co₂SnO₄ structure. The layer’s ethanol response increased and became more stable after annealing. The stability was measured over a period of three months. DC and AC investigations allowed us to propose a detection mechanism involving grain boundary regions in a Co₂SnO₄/Co₃O₄ composite obtained after annealing (denoted as CTO_ann). According to its stable and receivable sensing response, the Co₂SnO₄/Co₃O₄ composite could be considered as a promising novel sensitive layer for ethanol gas sensors.

1. Introduction

Currently, the pollution crisis is one of the major global problems. In order to overcome this problem, several research groups have turned their attention towards various kinds of nanomaterials, because such materials exhibit outstanding performance in gas sensor applications for controlling the emission of polluting gases. A ternary metal oxide and composites such as NiMn₂O₄, Fe₂O₃/ZnO and Ag-SnO₂/βC₃N₄ [1,2,3] have been developed as promising materials for use in gas sensing applications. Other materials, such as metal stannates including Ba₂SnO₄, Ca₂SnO₄, Zn₂SnO₄, Cd₂SnO₄, Mg₂SnO₄, Mn₂SnO₄ and Co₂SnO₄, have also been developed by research groups [4,5,6,7,8,9,10]. These materials have been found to have interesting properties and demonstrated good optical and electrical performance, and have been widely used in the study of solar cells, super capacitors, etc. [11,12,13,14,15,16,17,18]. So far, there have been very few studies on their gas sensing performance.

In this context, special attention has been paid to the cobalt stannate Co₂SnO₄ (CTO), which has the potential for many applications in various fields, including thermally stable capacitors [19]. Nevertheless, little study has been undertaken in the field of Co₂SnO₄ synthesis, or on its gas detection properties. Thus, to the best of our knowledge, synthesis of Co₂SnO₄ by the technique of spray pyrolysis and the study of its sensing properties under ethanol vapor (EtOH) has not been yet reported. In addition, there has been no study of the sensing properties of Co₂SnO₄/Co₃O₄ composite under ethanol vapor. Among the most commonly used materials for ethanol detection are Co₃O₄ [20,21,22] and its composites, such as ZnO/Co₃O₄, Fe₂O₃@Co₃O₄, Co₃O₄–TiO₂ and Co₃O₄/Ti₃C₂T_x [23,24,25,26]. Most sensing devices for various kinds of gases are based on metallic oxide layers [27]. In the present work, the low-cost technique of spray pyrolysis followed by an annealing process was successfully used to synthetize ternary Co₂SnO₄ and also Co₂SnO₄/Co₃O₄ sensing layers for ethanol (EtOH) vapor detection. The Co₂SnO₄/Co₃O₄ composite was the resultant product from an annealing treatment, as we will show later in this work.

Ethanol vapor was chosen as the target gas in this study, due to its toxicity [28] and its application in other safety sectors. Thus, there is a necessity to develop more effective sensors for ethanol detection [29]. Therefore, many ethanol sensing methods based on physical and chemical processes have been developed [30]. In all these methods, different materials were tested in order to find an appropriate material with the best sensitivity and selectivity under ethanol vapor to be used as the active layer for an ethanol gas sensor in a given application. Such sensors could be used, for example, in food quality checks, or in the bioenergy industry in ethanol production. Ethanol is also considered as a renewable fuel, which can be produced from various materials and plants, collectively known as “biomass”. In this context, Co₂SnO₄ could be introduced as a promising new candidate for the ethanol gas sensor industries. The morphological characterizations of our prepared Co₂SnO₄ film, as well as its ethanol sensing properties and sensing mechanism, were investigated according to the annealing temperature.

2. Experimental Procedure

2.1. Co₂SnO₄ Film Deposition

Co₂SnO₄ (CTO) films were obtained using 0.02 mole of cobalt (II) chloride dihexahydrate (CoCl₂.6H₂O) mixed with 0.01 M of chloride dihydrate (SnCl₂.2H₂O) in 100 mL of distilled water at ambient temperature. All the used reagents were commercially available, and no purification was needed. Then, 2.0 M of NaOH solution, used as mineralizer, was added under magnetic stirring for 15 min; a blue precipitate was obtained instantaneously. The final molar ratio of [Sn]:[Co] was fixed at 1:2. The mixture was then transferred to Teflon-lined stainless steel autoclave and kept at 240 °C for 48 h. The resulting precipitates were collected by centrifugation and washed with distilled water, and finally dried in an electric oven at 120 °C for 12 h. The obtained Co₂SnO₄ precursor was dissolved in absolute alcohol and sprayed on glass substrates at a growth temperature of 500 °C.

During the spray deposition, nitrogen (N₂) was used as a carrier gas. In addition, to ensure the best uniformity of the deposited layers, the horizontal sweep of the nozzle (0.5 mm diameter) was optimized. The flow rate of the precursor mixture was taken constantly at 5 mL.min⁻¹. After deposition, the elaborated films were allowed to cool to room temperature.

Finally, two Co₂SnO₄ samples were prepared. One of them (denoted as CTO_ann) was calcined at 800 °C for 8 h with a heating rate of 10 °C/min under standard atmospheric conditions, after which the sample was allowed to cool to room temperature. The aim of this step was the investigation of the annealing effects on the Co₂SnO₄ layer’s sensitivity toward ethanol vapor, which is probably controlled by the phase transition in the as-prepared layer. The choice of temperature and calcination duration were in order to ensure the synthesis of a stable sensitive phase, as the working temperatures under the ethanol vapor would be varied. The second sample, which remained as prepared, was denoted as CTO_as. The morphological results are reported in Figure 1.

We noted that the surface morphologies of CTO_ann (Figure 1b) were more granular and porous compared with the as-prepared sample (CTO_as) presented in Figure 1a. These morphological differences could be beneficial for ethanol detection by CTO_ann.

2.2. Characterization Techniques

The layers’ X-ray diffraction (XRD) characterizations were performed using a Philips X’Pert MPD XRD diffractometer with monochromatic radiation Cu-Kα (1.51418 Å), in the range of 20–80°, with a scanning rate of 0.05°.s⁻¹. SEM investigations were undertaken using a JSM-6510 scanning electron microscope from JEOL, Ltd., with accelerating voltage (0.5 kV to 30.0 kV) at low vacuum (10 Pa–150 Pa) and magnification up to 600,000×. The prepared samples were also characterized by a Tecnai F20 (200 kV) transmission electron microscope (TEM) from FEI. The samples’ DC and AC gas sensor electrical signals were measured using a source/pico-ammeter HP4140B and Solartron 1250 impedance analyzer from L.M.M.A laboratory, respectively, at a low-bias voltage of 50 mV.

2.3. Experimental Setup for Ethanol Detection

Figure 2 presents the experimental setup used for ethanol detection in which the sensors were exposed to a flow rate of 1 L.min⁻¹ of dry air used as a carrier gas for ethanol vapor coming from a flask placed in a heating bath. The ethanol concentrations were controlled by two mass flows (d₁, d₂) and given by the following equation [31].

$$\left[{\mathrm{C}}_{\mathrm{EtOH}}\right]\left(\mathrm{ppm}\right)=\left(\frac{{\mathrm{xd}}_1}{\left(1+\mathrm{x}\right){\mathrm{d}}_1+{\mathrm{d}}_2}\right)\times {10}^6$$

(1)

The vapor molar fraction x at fixed T_vap was given by:

$$\mathrm{x}=\frac{{\mathrm{P}}_{\mathrm{vap}}}{{\mathrm{P}}_{\mathrm{atm}}}$$

(2)

P_atm and P_vap were the atmospheric and partial vapor pressure, respectively, in the test chamber [32]. For all ethanol concentrations, the heated bath circulator temperature T_vap was fixed at 30 °C in order to fix the vapor pressure P_vap which was deduced from the database curve of ethanol pressure from the CRC Handbook of Chemistry and Physics [33].

3. Results and Discussion

3.1. Characterization of Elaborated Films

The XRD layer pattern of the as-prepared sample (CTO_as) confirmed the synthesis of pure Co₂SnO₄ with a spinel phase, as shown in Figure 3. All diffraction peaks could be indexed in the JCPDS 29-0514 card for the Co₂SnO₄ spinel cubic single phase. After annealing, the obtained sample (CTO_ann) was identified as a mixture of spinel Co₂SnO₄ phase and spinel Co₃O₄ pure phases; the obtained Co₂SnO₄/Co₃O₄ composite was indexed by both standard cards: JCPDS no. 29-0514 for the Co₂SnO₄ spinel phase and JCPDS 78-1970 for the Co₃O₄ spinel phase.

The crystallite sizes of the prepared samples were calculated using the following Scherrer formula (Equation (3)):

$${\mathrm{D}}_{\mathrm{hkl}}=\frac{0.9\mathsf{\lambda }}{\mathsf{\beta }.\mathrm{Cos}\mathsf{\theta }}$$

(3)

where λ is the wavelength of the implemented X-ray diffraction, β is the full width at half maximum (FWHM) of the (hkl) diffraction plane in radians and h is the corresponding incident angle.

The average crystallite sizes were found to be about 80 nm for the CTO_as pure Co₂SnO₄ (311) phase and 50 nm and 30 nm for the Co₂SnO₄ (311) and Co₃O₄ (111) directions, respectively. In the polycrystalline CTO_ann layer, no other phase formation of impurities was observed in the diffractogram. Figure 4a,b show the high-magnification SEM images of the samples obtained before and after the annealing treatment at 800 °C.

We note that the high-magnification SEM image shows a random combination of compact grain agglomerations with average sizes ranging from about 50 nm to 150 nm in size. After annealing, the film shows a large number of individual grains with average sizes ranging from about 20 nm to 50 nm in size, leading to an increase in porosity and offering a large active surface area available for oxygen molecule adsorption.

High surface area is the main factor in a sensor’s sensitive layer which enables it to achieve higher sensitivity.

The polycrystalline nature of the annealed film offered a greater available surface area for atom fraction and oxygen molecule adsorption on the surfaces of grains and at interfaces between grains (grain boundaries), suggesting that the microstructure of the films is suitable for ethanol vapor sensing applications. The variations in the sizes of particles seen in SEM images were comparable to the grain sizes estimated by XRD measurements.

In addition, the phase transition from Co₂SnO₄ (CTO_as) to Co₂SnO₄/Co₃O₄ composite (CTO_ann) was confirmed by TEM images (Figure 4). Figure 4a shows the Co₂SnO₄ (311) monocrystal structure of the as-prepared film with an interplanar spacing of 0.26 nm, whereas Figure 4b shows the multicrystal structure of the annealed film with high crystallinity and purity. The interplanar spacing for CTO_as and CTO_ann is calculated as 0.26 nm and 0.46 nm, respectively, matching well with the interplanar spacing of Co₂SnO₄ (311) and Co₃O₄ (111), and confirming the elaboration of the Co₂SnO₄/Co₃O₄ composite.

3.2. Co₂SnO₄ and Co₂SnO₄/Co₃O₄ Sensing Results and Discussion under Ethanol

The samples’ ethanol sensing tests were performed in the concentration range of ethanol vapor at 100–1000 ppm; the gas response “S” is defined as follows:

$$\mathrm{S}=\frac{{\mathrm{I}}_{\mathrm{gas}}-{\mathrm{I}}_0}{{\mathrm{I}}_0}$$

(4)

for the CTO_as sample, which presented the conductivity behavior of an n-type semiconductor; and

$$\mathrm{S}=\frac{{\mathrm{I}}_{\mathrm{gas}}-{\mathrm{I}}_0}{{\mathrm{I}}_{\mathrm{gas}}}$$

(5)

for the CTO_ann sample, which showed the conductivity behavior of a p-type semiconductor.

I_gas and I₀ are the current intensities of layers under ethanol gas (vapor) and dry air, respectively. Firstly, in order to determine the optimal working temperatures, the concentration of ethanol was fixed at 500 ppm and the temperature was varied from 100 °C to 250 °C. The obtained optimal working temperatures were found to be 150 °C for CTO_as and 200 °C for CTO_ann, as shown in Figure 5a,b.

After optimal working temperature deduction, a dynamic measurement under gas was made in order to deduce the response (τ_res) and the recovery (τ_rec) times, reported in the last table. The chemical reactions which occurred between the ethanol vapor and the samples’ sensitive layers when the samples were exposed to gas were also analyzed, and the most probable detection mechanism was deduced. For these measurements, each film was exposed to six different concentrations of ethanol, namely 100, 200, 300, 400, 500 and 1000 ppm, at its optimum temperature. The duration of each cycle was about 10 min under ethanol vapor. In the overall measurements, dry air was used as a carrier gas.

CTO_ann film showed an improvement in ethanol sensitivity compared with the as-prepared sensor CTO_as with a switch from n- to p-type conductivity; this result could be attributed to gas interaction with the p-type Co₃O₄ phase formation after the annealing process. In addition, we noted an increase in grain boundaries in the CTO_ann sensitive layer which became more granular, with small grains, therefore increasing the density of grain boundaries, which are the favored oxygen adsorption regions. The main reason for the observed grain size decrease is the porosity decrease in the grains with the increase in annealing temperature (i.e., the grain become more compact) as reported in the literature [34]. For working temperatures below 200 °C, the oxygen molecular form ${\mathrm{O}}_2^{-}$ dominates, and above this temperature the ionic forms O⁻ and O²⁻ are the predominant adsorbed oxygen species [35]. Thus, the possible ethanol interaction mechanisms with the CTO_as film at the optimum temperature of 150 °C could be presented by the following reactions for temperatures below 200 °C:

$${\mathrm{O}}_{2\left(\mathrm{gas}\right)}^{}+{\mathrm{e}}_{\left(\mathrm{bulk}\right)}^{-}\to {\mathrm{O}}_{2\left(\mathrm{surface}\right)}^{-}$$

(6)

$${\mathrm{C}}_2{\mathrm{H}}_5{\mathrm{OH}}_{\left(gas\right)}+3{\mathrm{O}}_{2\left(\mathrm{surface}\right)}^{-}\leftrightarrow 3{\mathrm{H}}_2{\mathrm{O}}_{\left(gas\right)}{}^{}+2{\mathrm{CO}}_{2\left(gas\right)}^{}+3{\mathrm{e}}_{\left(bulk\right)}^{-}{}^{}$$

(7)

The resulting CO₂ and the pre-adsorbed OH⁻ will react together as follows:

$${\mathrm{CO}}_{2\left(\mathrm{gas}\right)}^{}+{\mathrm{OH}}_{\left(\mathrm{surface}\right)}^{-}\leftrightarrow {\mathrm{HCO}}_{3\left(\mathrm{surface}\right)}^{-}$$

(8)

$${\mathrm{HCO}}_{3\left(\mathrm{surface}\right)}^{-}+{\mathrm{OH}}_{\left(\mathrm{surface}\right)}^{-}\leftrightarrow {\mathrm{CO}}_{3\left(bulk\right)}^{2-}+{\mathrm{H}}_2{\mathrm{O}}_{\left(\mathrm{gas}/\mathrm{surface}\right)}$$

(9)

Equation (6) shows that the oxygen molecule is absorbed on the layer’s surface without dissociating, by trapping an electron from the bulk, and it becomes negatively charged. After that, the $\left({\mathrm{O}}_2^{-}\right)$adsorbed oxygen species will react with ethanol vapor (C₂H₅OH) to produce water vapor and carbon dioxide (CO₂) and an electron (e⁻), according to Equation (7). The carbon dioxide (CO₂) will react with preadsorbed hydroxide (OH⁻) according to Equation (8), to provide a reaction product, a bicarbonate ion $\left({\mathrm{HCO}}_3^{-}\right)$, which in turn will react with preadsorbed hydroxide OH⁻ to produce a carbonate anion $\left({\mathrm{CO}}_3^{2-}\right)$ and water as a final product (Equation (9)).

The resulting negative charges (carbonate anion and electron) will move to the CTO_as bulk which, according to the XRD pattern and TEM image already presented in Figure 3 and Figure 4, respectively, present only a pure phase of Co₂SnO₄ and thus increase the carrier charge density. This explains the current increases under ethanol, with the film showing n-type conductivity as already illustrated in Figure 5c at 150 °C. Conversely, for CTO_ann film, which presents a phase mixture of Co₂SnO₄/Co₃O₄, we obtained an opposite behavior under ethanol, as also shown in Figure 5c at 200 °C, showing the p-type conductivity which is dominated by the Co₃O₄ species, and which explains the n- to p-type switch observed for annealed film (CTO_ann) under ethanol vapor. The probable main ethanol reaction at the optimum temperature of 200 °C for CTO_ann is the oxidizing of ethanol given by the following reactions [36]:

$$\frac{1}{2}{\mathrm{O}}_{2\left(\mathrm{gas}\right)}+{\mathrm{e}}_{\left(\mathrm{bulk}\right)}^{-}\to {\mathrm{O}}_{\left(\mathrm{surface}\right)}^{-}$$

(10)

$${\mathrm{C}}_2{\mathrm{H}}_5{\mathrm{OH}}_{\left(\mathrm{gas}\right)}+{\mathrm{O}}_{\left(\mathrm{surface}\right)}^{-}\to {\mathrm{CH}}_3{\mathrm{CHO}}_{\left(\mathrm{surface}\right)}+{\mathrm{H}}_2{\mathrm{O}}_{\left(\mathrm{gas}\right)}+{\mathrm{e}}_{\left(\mathrm{bulk}\right)}^{-}$$

(11)

Equation (10) shows that the oxygen molecule is absorbed on the layer’s surface with dissociation by trapping an electron from the bulk and becoming negatively charged. Following this, the (O⁻) adsorbed oxygen species will react with ethanol vapor (C₂H₅OH) to give water vapor, an acetaldehyde molecule (CH₃COH) and an electron (e⁻), according to Equation (11).

The electron produced by the previous reaction (Equation (11)) is injected into the bulk, which induces a decrease in the hole concentrations, causing a decrease in current intensity. In this context, Barsan and Weimar show that the preadsorbed ionized form of oxygen in oxide film is produced by the following reaction [35]:

$$\frac{\delta }{2}{O}_2^{gas}+\alpha \overline{e}+V\to O_{\delta S}^{-\theta }$$

(12)

where V is an oxygen-unoccupied site, $O_{\delta V}^{-\theta }$is the adsorbed oxygen form; θ = 1 for single ionized species, δ = 1 for atomic species. We can infer from this theory that the response S is:

$$S=A[C_{EtOH}{]}^{\frac{\delta }{\theta +1}}=A[C_{EtOH}{]}^{\gamma }\mathrm{where}\mathsf{\gamma }=\delta /\left(\theta +1\right)$$

(13)

The experimental response evolution versus ethanol concentration presented in Figure 5d is in good accordance with Barsan and Weimar’s theory. From experimental data, we deduce the fitting of the following γ values:

γ ≈ 0.5 at 200 °C, thus θ = 1 and δ = 1, implying that the layer is porous with small grains

γ ≈ 1 at 150 °C, thus θ = 1 and δ = 2, implying that the layer is compact with large grains

These results confirm our previous proposed mechanisms for ethanol detection by the CTO_as and CTO_ann layers and are also in good accordance with the morphological investigation reported in Figure 1.

3.2.1. Detection Mechanism under Ethanol

The ethanol response improvement of CTO_ann is the result of porosity as well as an increase in grain boundaries. The grain boundary regions are the favored oxygen adsorption regions, as already mentioned above. Consequently, when the proportion of grain boundaries increases in the film due to a decrease in grain size, the oxidation reaction rate became more pronounced for this film wherein the current intensity variation increases due to the increases in charge transfer, and thus the response variation increases. In order to confirm the advanced effects of grains and grain boundaries in the proposed interaction mechanism between ethanol and films, an AC investigation of elaborated films, presented in Figure 6, was performed at the optimum temperatures under 500 ppm of ethanol vapor.

The AC analysis evidenced that, under ethanol, the CTO_as film impedance could be modelled by a single parallel RC circuit giving one semicircle in a Nyquist plot, attributed to the grain regions. In this case, the CTO_as ethanol response is controlled only by the grain regions, whereas for CTO_ann, we note the apparition of a second semicircle showing that the response is mainly controlled by the grain boundary regions.

The RC modeling results are reproduced in

Table 1. The RC modeling.

		Under 500 ppm of Ethanol
	Film	CTOas	CTOann
1St Semicircle	Rb × 106 Ω	6.44	1257
	Cb × 10−10 F	1.812	1.88
	N	0.996	0.988
2nd Semicircle	Rgb × 106 Ω		79.25
	Cgb × 10−7 F		0.391
	N		0.874

, and fit well with the AC experimental results.

The previous results of the AC investigation allow us to affirm that the response enhancement of the CTO_ann sensor is mainly controlled by the oxidation reactions with oxygen adsorbed in grain boundary regions formed by Co₃O₄/Co₂SnO₄ p–n heterojunctions. In the other sensor regions there is probably a coexistence of Co₃O₄/Co₃O₄ p–p homojunctions and Co₂SnO₄/Co₂SnO₄ n–n homojunctions, so their electrical (response) behaviors operate against each other.

The CTO_as sensor, which presents a pure phase of Co₂SnO₄ according to the XRD pattern and TEM analyses (Figure 3 and Figure 4a), was probably formed only by Co₂SnO₄/Co₂SnO₄ n–n homojunctions. The current variation under ethanol is less than that of the CTO_ann sensor formed by Co₃O₄/Co₂SnO₄ p–n heterojunctions. In effect, the mechanism could be related to the contact at the junction interface between Co₃O₄ and Co₂SnO₄ in the heterojunctions and between Co₂SnO₄ and Co₂SnO₄ in the homojunctions, so the sensing property variations for the CTO_as and CTO_ann semiconductors are controlled by the energy barrier changes “qV” at homojunctions or heterojunctions under ethanol. This energy barrier, in turn, depends on work function differences between Co₃O₄ and Co₂SnO₄ given by the following equation:

$$qV=q\left({\varnothing }_{Co2SnO4}-{\varnothing }_{Co3O4}\right)$$

(14)

where ∅_Co3O4 and ∅_Co2SnO4 are the work functions for Co₃O₄ and Co₂SnO₄, respectively.

For the Co₂SnO₄/Co₂SnO₄ n–n homojunctions of a pure Co₂SnO₄ layer, the barrier changes will be negligible, whereas for a Co₃O₄/Co₂SnO₄ p–n heterojunction, in the interfaces of the depletion layer the conductivity “G” is mainly determined by the energy barrier changes, presented by the following equation:

$$\mathrm{G}\propto {\mathrm{Ae}}^{\left(-\frac{\mathrm{qV}}{\mathrm{kT}}\right)}$$

(15)

where A is a constant, T is the absolute temperature, k is the Boltzmann constant and qV is the effective energy barrier at the interface.

In Figure 7 we have presented the schematic diagram of our proposed sensing mechanism and the energy band structure diagram for the sample CTO_ann which contains the highest density of Co₃O₄/Co₂SnO₄ p–n heterojunctions. This is in order to illustrate the effect of oxygen adsorption on the energy band level and its variations after contact in the grain boundary region before ethanol vapor injection and under ethanol vapor.

Figure 7b schematizes the oxygen adsorption reactions given by Equations (6) and (10) when the sample is under a fixed flow of dry air before ethanol injections, and Figure 7c schematizes the possible oxidation reactions given by Equations (7) and (11) from the interactions between ethanol and adsorbed oxygen on the surface of the layer in the grain boundaries.

3.2.2. Response Stability under Ethanol

After mechanism deduction, an investigation of the annealing process effect on the response stability under ethanol was also conducted. Figure 8 presents the response stability of each sample under an ethanol vapor concentration of 500 ppm at the optimum temperature, for a duration of three months. As presented in Figure 7a, we note that for CTO_as the stability deteriorated after the first month, and the layer response drift increased to reach 35% at the third month, whereas CTO_ann response drift did not exceed 5%. The result shows an improvement in stability following the annealing process, as well as sensibility toward ethanol vapor. In addition, in our previous work [37] we demonstrated that humidity has a neglected effect on the response stability of Co₂SnO₄. Thus, we think that the Co₂SnO₄/Co₃O₄ composite which presented better stability will be not affected by humidity; this prediction could be investigated in future research. Moreover, a comparative study of CTO_ann sensing parameters with those of other sensing materials under the concentration range 500–600 ppm of ethanol vapor, which is situated in our concentration range, are presented in

Table 2. Comparative study of Co₂SnO₄/Co₃O₄ composite (CTO_ann) ethanol sensing parameters with other ethanol sensing materials.

Sensor Material	Fabrication Approach	Conc. (ppm)	Response “S”	Sens Temp. (°C)	τres/τrec (s)	Ref.
Co2SnO4 spinel thin film	Spray pyrolysis	500	3.5	150	20/24	[37]
SnO2 (triton)	Spin coating	500	3	RT	15/720	[38]
Zn2SnO4 spinel cube	Hydrothermal	600	5.5	325	18/45	[39]
α-Fe2O3 nanorods	Hydrothermal	500	8	220	−/−	[40]
Porous SiC	Electrochemical etching	Saturated vapors	1.2	RT	85/57	[41]
NiCo2O4 spinel nanoparticles	Hydrothermal	500	0.6	RT	20/26	[42]
TiO2 NWs	Electrospinning	500	7.5	500	23/1	[43]
Cu2O nano-columnar	GLAD technique	500	8.12	200	60/180	[44]
Zn2SnO4cuboctahedra	Chemical method	600	0.01	325	18/45	[45]
Co2SnO4/Co3O4 composite	Spray pyrolysis	500	13.5	200	35/30	This work

. These further reveal the good performance of our Co₂SnO₄/Co₃O₄ composite (CTO_ann).

4. Conclusions

In conclusion, the annealing process of Co₂SnO₄ film synthesized by spray pyrolysis method, denoted as CTO_as, is beneficial for ethanol detection. After annealing, we successfully obtained the Co₂SnO₄/Co₃O₄ composite, denoted as CTO_ann. The new composite presents an improvement in response under ethanol at the optimal temperature of 200 °C, due to the increases in its porosity and grain boundary density, which offers a more specific surface for oxygen adsorption. Further, we observed an improvement in the film’s stability compared with the non-annealed one. We also noted that after the annealing process the Co₂SnO₄ film underwent a switching from n- to p-type conductivity under ethanol, due to the p-type Co₃O₄ phase formation in the Co₂SnO₄ structure. The sensing mechanism was found to be an oxidation reaction which depends on the working temperature, resulting in negative charges, such as a carbonate anion for CTO_as because the oxidation reaction of ethanol occurs with O²⁻ at the optimal temperature of 150 °C, and an electron for CTO_ann because the oxidation reaction of ethanol occurs with O⁻. at the optimal temperature of 200 °C. The resulting charges will be moved to the bulk of the samples, which raises the current intensity due to the increase in the electron density in the CTO_as layer (n-type conductivity) and decreases the current intensity for CTO_ann due to the decrease in hole density after recombination with electrons (n-type conductivity). The improvement in responses was explained by the increase in grain boundary heterojunctions in the CTO_ann film, wherein the conductivity was controlled by the energy barrier changes.

Finally, the stability of CTO_as and CTO_ann over a period of three months was investigated; the results show an improvement in stability after the annealing process.

In future work, we intend to investigate the effects of layer thicknesses and electrode architecture on the sensor responses and selectivity. The primary results of our investigation into the effect of the annealing process on the film’s sensing properties under ethanol vapor presented in the present work show that CTO_ann is promising for application as a sensitive active layer for ethanol sensor devices in the future.