Regulation of N-type In₂O₃ Content on the Conductivity Type of Co₃O₄ Based Acetone Sensor

Indium nitrate hydrate (In(NO₃)₃·4H₂O), Cobalt nitrate hydrate (Co(NO₃)₂·6H₂O), anhydrous ethanol, 2N-dimethylformamide (DMF) and polyvinylpyrrolidone (PVP, Mw = 1,300,000) were all used as purchased, without further purification.

Typically, the In₂O₃/Co₃O₄ NFs were synthesized via electrospinning and subsequent sintering. Firstly, 900 mg Co(NO₃)₂·6H₂O was dissolved into 6 ml ethanol and 7 ml DMF, then 900 mg PVP was added into the above solution, with magnetic stirring for 10 h to fully dissolve the PVP. Electrospinning precursor solutions of In₂O₃ NFs were obtained by the same method. Then, the electrospinning precursor solution of In₂O₃ NFs and Co₃O₄ NFs was put into two identical 10 ml syringes. The molar ratio of Co(NO₃)2·6H₂O to In(NO₃)₃·4H₂O was 1:1. The ratio of In to Co could be controlled by setting different propulsion speed. The spinning voltage was set to 20 kV and thus the platform moved back and forth at a speed of 20 cm/min. The prepared PVP/In(NO₃)₃·4H₂O/Co(NO₃)₂·6H₂O was torn off from the aluminum plate collector and placed in a crucible, sintering in a muffle furnace with a heating rate of 2°C/min. The calcination process was divided into two stages: the first stage was maintained at 300°C for 2 h to ensure the volatilization of PVP, and the second stage was maintained at 500° C for 2 h to crystallize and form oxide composites. In order to explore the effects of different content ratios of cobalt oxide on the structure, morphology, and gas sensing properties of the composites, four kinds of In₂O₃/Co₃O₄ NFs with different molar ratios – (1) In: Co=1:3, (2) In: Co=1:1, (3) In: Co=3:1, (4) In: Co=6:1 – were prepared. The obtained samples were named as In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3 and In₂O₃/Co₃O₄-4. Intrinsic In₂O₃ and intrinsic Co₃O₄ NFs were prepared by the same process with only one syringe.

The X-ray powder diffraction (XRD) analysis was performed on Bruker D8 Advance over the 2θ range from 10°–80° in steps of 0.03°. The field emission scanning microscopy (FE-SEM) images were obtained on a Sigma500. The elemental distribution of the sample was characterized by X-ray Energy Dispersive Spectrometer (EDS) Analysis. High-resolution transmission electron microscopy (HRTEM) images were acquired on a Fei Tecnai G2 F20.

The calcined sample was mixed with a small amount of deionized water in an agate mortar. The mixture was then ground to a paste in the agate mortar. It should be noted that the grinding needed to be carried out in the same direction to make it uniform. The ground paste sample was evenly coated on the silver palladium electrode, which was placed at room temperature for 10 h to dry and volatilize deionized water, and then the silver palladium electrode coated with the sample was aged in a muffle furnace at 300°C for 10 h three times. Gas-sensing measurements were carried out by the CGS-MT integrated test platform, consisting mainly of static gas distribution and a data acquisition system (SA3101 Configuration), which recorded the change in the resistance value. The target gas volume was calculated using software (EtLiquidVolume) and then injected into the test chamber using a microinjection needle. The conversion formula between the injection volume of the liquid (V) and the concentration of the gas (C) was Q = (V × C × M)/(22.4 × d × ρ) × 10⁻⁹ × (273 + TR)/(273 + TB), where Q is the volume of liquid to be taken (ml), V is the volume of test bottle (ml), M is the molecular weight (g), d is the purity of liquid, C is the concentration of the gas to be prepared (ppm), ρ is the density of the liquid (g·cm⁻³), TR is the test ambient temperature (°C), and TB is the temperature in the test bottle (°C) [25]. The schematic diagram of the whole experimental process using homopolar double-jet electrospinning is shown in Figure 1.

Fig. 1.

Characterization was necessary to verify the composition and crystallinity of the as-prepared In₂O₃/Co₃O₄ NFs, XRD. As shown in Figure 2, the crystal structure information of In₂O₃ NFs, Co₃O₄ NFs and In₂O₃/Co₃O₄ NFs (In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3 and In₂O₃/Co₃O₄-4) was characterized. The XRD spectrum was analyzed according to the PDF#06-0416 for In₂O₃ (space group la-3), and PDF#43-1003 for Co₃O₄ (space group Fd-3m). The diffraction peaks of intrinsic In₂O₃ were observed at 2θ of 21.498°, 30.580°, 35.466°, 51.037°, and 60.676°, and assigned to (211), (222), (400), (440) and (622) crystal planes of tetragonal rutile In₂O₃, respectively. While the diffraction peaks of Co₃O₄ at 2θ values of 19.000°, 31.271°, 36.845°, 38.546°, 44.808°, 59.353°, and 65.231° corresponded to the (111), (220), (311), (222), (400), (511) and (440) crystal faces, respectively, indicating the absorption peaks matched well with the standard figure of Co₃O₄ phase. Meanwhile, there were still characteristic diffraction peaks of Co₃O₄ in the curves of In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3 and In₂O₃/Co₃O₄-4. With the increase of Co₃O₄ loading, the characteristic diffraction-peak intensity of Co₃O₄ increased, indicating that Co₃O₄ has been successfully loaded on In₂O₃. No diffraction peaks of other phases were detected, indicating high purity in the sample.

Fig. 2.

Figure 3A–L show the SEM images of In₂O₃ NFs, Co₃O₄ NFs, In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3, In₂O₃/Co₃O₄-4 before and after sintering. For the as-spun fibers, all the samples have a uniform diameter of about 150–250 nm. For pure In₂O₃ and Co₃O₄ NFs, as shown in Figures 2G and H, the surface of the Co₃O₄ NFs was rougher than that of In₂O₃ NFs. The diameter of both the Co₃O₄ and In₂O₃ NFs was about 60–70 nm. Figure 3I–L shows the cross-sectional views of In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3 and In₂O₃/Co₃O₄-4. It was found in Figure 3I–K that the majority of In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2 and In₂O₃/Co₃O₄-3 were hollow structures. As In₂O₃ increased, nanofibers with hollow structures became fewer. As shown in Figure 3L, most of the In₂O₃/Co₃O₄-4 were solid structures. In order to further prove the existence of In₂O₃ and Co₃O₄ in the sample, the In₂O₃/Co₃O₄-3 was characterized by EDS, as described in the next section.

Fig. 3.

According to the TEM images of In₂O₃/Co₃O₄-3 NFs in Figure 4A–B, the size and morphology were similar to the SEM characterization results in Figure 3K. For the purpose of verifying whether In₂O₃ and Co₃O₄ were composited on a single nanofiber, the samples were further characterized by HRTEM, as shown in Figure 4C. Analyzing the lattice fingerprint information in HRTEM showed that the lattice spacings of 0.2928 nm and 0.4122 nm corresponded to the (222) and (211) planes of In₂O₃, respectively, and the lattice spacings of 0.2868 nm and 0.2445 nm corresponded to the (220) and (311) planes of Co₃O₄, respectively. The lattice fingerprints of In₂O₃/Co₃O₄-NFs were consistent with the XRD patterns.

Fig. 4.

The elements of Au, In, Co and O appeared on a single nanofiber, as seen from Figure 5A–E. The existence of Au element came from the gold sprayed on the sample surface to enhance the conductivity of material. From above analysis, it could be concluded that In₂O₃/Co₃O₄ NFs had been compounded in the process of double-jet electrospinning. Under the influence of positive high voltage electrodes, homopolar double-jet electrospinning made the two materials contact alternately during collection. The EDS of In₂O₃/Co₃O₄-3 was shown in Figure 5F. The weight ratio and atom ratio of Co:In were 17.5% and 34.0%, respectively.

Fig. 5.

It was worth noting that during the calcination process, the electrospinning precursors gradually formed a hollow nanofiber structure due to the Kirkendall effect [26]. The Kirkendall effect is the process of surface ion oxidation crystallization and internal ion as surface exudation oxidation crystallization. When the temperature in the muffle furnace was high, the PVP on the NFs surface was calcined and volatilized, while the PVP inside was not volatilized in time and thus was in a molten state. At the same time, the surface ions were oxidized and crystallized, while the internal ions were not oxidized and crystallized due to the warping of the surface. Thus, there was a gradient difference between the surface and the interior of the NFs, resulting in the internal ions seeping to the surface with the carrier of molten PVP. The PVP as exudate was calcined and volatilized, and the ions were oxidized and crystallized. The process was repeated to form a hollow nanofiber structure. The schematic diagram of calcination process model is shown in Figure 6.

Fig. 6.

The operating temperature strongly influenced the gas-sensing properties of the MOSs gas sensor. Figure 7 shows the response of Co₃O₄, In₂O₃, In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3 and In₂O₃/Co₃O₄-4 to 40 ppm acetone at different working temperatures. The responses ® of Co₃O₄ and In₂O₃/Co₃O₄-1 were defined as R= R_g/R_a. The respon®(R) of In₂O₃, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3, and In₂O₃/Co₃O₄-4 were defined as R= Ra/Rg, with Ra being the stable resistance in air and Rg being the stable resistance in probe gas. For Co₃O₄, In₂O₃/Co₃O₄-1, and In₂O₃/Co₃O₄-2, response values under their respective optimal operating temperatures are very low, lower than 2. Their optimal operating temperatures are 200°C, 175°C and 300°C, respectively. For In₂O₃/Co₃O₄-4, the sensor reaches its highest response of 6 at 200°C. The response values of In₂O₃ are higher than that of In₂O₃/Co₃O₄-3, but the optimal working temperature of the latter is 100°C lower. Considering both the high response value and low operating temperature, we consider In₂O₃/Co₃O4-3 to be the optimal device, with a high response value of 9.5 at 200°C to 40 ppm acetone.

Fig. 7.

Figure 8A presents the dynamic resistance curve of Co₃O₄ and In₂O₃/Co₃O₄-1 to acetone with a concentration range of 1–100 ppm at 200°C. When the sensors are transferred from an air environment to an acetone environment, their response values increase until they reach a maximum and equalize. When the acetone gas supply stops, the response values decrease and return to baseline resistance. This confirms the sensor’s perceptibility and recovery. As observed, the conductivity type of Co₃O₄ and In₂O₃/Co₃O₄-1 were p-type, while that of In₂O₃, In₂O₃/Co₃O₄-3, and In₂O₃/Co₃O₄-4 were n-type, as shown in Figure 8B. In₂O₃/Co₃O₄-2 had almost no response, indicating that the composite is in the critical state of p-type to n-type transition. From the above analysis, it can be seen that with the increase of indium oxide content, the complex gradually changes from p-type to n-type and is in a critical state when the content of both is equal. Therefore, we can conclude that the content of n-In₂O₃ in composite In₂O₃/Co₃O₄ can regulate the final conductivity type of the sensor. The sensing behavior of the In₂O₃/Co₃O₄-1 was p-type, which might be due to the contribution of holes being greater than that of the electrons [27]. The In₂O₃/Co₃O₄ sensor has a wide detection range for acetone, and the response increases with concentration, obeying a linear relationship. When the acetone concentration increased from 1 to 100 ppm, the gas response, based on the In₂O₃ sensor and In₂O₃/Co₃O₄ sensor, increased from 1.3 and 1.86 to 4.04 and 19.2. This indicates that the presence of Co₃O₄ helps to improve the response of the In₂O₃-based sensor to acetone.

Fig. 8.

The responses of Co₃O₄, In₂O₃, In₂O₃/Co₃O₄-1, In₂O₃/Co₃O₄-2, In₂O₃/Co₃O₄-3, and In₂O₃/Co₃O₄-4 sensors to four interference gases including formaldehyde, ethanol, methanol, and ammonia with a concentration of 40 ppm were examined at 200°C, as shown in Figure 9. The response values of In₂O₃/Co₃O₄-3 sensor for 40 ppm acetone, formaldehyde, ethanol, methanol, and ammonia were 9.5, 4.6, 3.9, 2.3, and 2.3, respectively. The response of In₂O₃/Co₃O₄-3 sensor to acetone is higher than the response to the other interference gases. Therefore, it can be concluded that In₂O₃/Co₃O₄-3 gas sensor has a good selectivity to acetone. In addition, we found that although the response value of In₂O₃ to acetone was comparable to that of In₂O₃/Co₃O₄-3, the selectivity of In₂O₃ to acetone was poor in the presence of interfering gases, especially formaldehyde, ethanol, and methanol.

Fig. 9.

As shown in Figure 10A, the lower detection limit of the In₂O₃/Co₃O₄-3 sensor was 0.7 ppm with a response value of 1.14, and the response/recovery time of the In₂O₃/Co₃O₄-3 sensor to 0.7 ppm acetone at 200°C was 21 s/51 s. In order to explore the stability of the sensor, a periodic test of 10 ppm acetone gas was carried out for In₂O₃/Co₃O₄-3 at 200°C. The test interval was 30 days and lasted for 5 months, as shown in Figure 10B. It can be seen from the figure that the sensitivity of In₂O₃/Co₃O₄-3 sensor has a small fluctuation, but the amplitude is small, indicating that the sensor has good stability and can meet the actual demand. A comparison of the sensing performance of various In₂O₃-based sensing materials and preparation processes is given in Table 1. The results show that the In₂O₃/Co₃O₄-3 sensor has lower operating temperature and higher gas response values. The structural features, including hollow interiors, porous surfaces and micropore size, give the In₂O₃/Co₃O₄-3 a higher surface adsorption sites and more efficient gas diffusion of oxygen and target gases, which may facilitate its enhanced response [28]. In addition, the process of preparing In₂O₃/Co₃O₄-3 materials using doublejet electrospinning is simple, convenient and low cost.

Fig. 10.

The gas-sensing mechanism of MOSs is generally explained by the electron depletion layer theory. In this study, when the In₂O₃/Co₃O₄ gas sensor was exposed to air, the oxygen molecules were adsorbed on the surface of MOSs material. These adsorbed oxygen molecules captured free electrons to be converted to the ionized oxygen species Oads (O2−,O−,O2−)$$\left( {{\rm{O}}_2^ - ,{{\rm{O}}^ - },{{\rm{O}}^{2 - }}} \right)$$, as shown in Figure 11A, where “ads” is the abbreviation for adsorbed. The electron depletion layers and hole accumulation layers were formed around the surface of MOSs grains.

Fig. 11.

Upon exposure to acetone gas, CO2, and H2O are formed as follows [33]: 1CH3COCH3 (gas) +8O2− (ads) →3CO2 (gas) +3H2O (gas) +16e−$$\matrix{ {{\rm{C}}{{\rm{H}}_3}{\rm{COC}}{{\rm{H}}_3}{\rm{\;(gas)\;}} + 8{{\rm{O}}^{2 - }}{\rm{\;(ads)\;}}} \hfill \cr { \to 3{\rm{C}}{{\rm{O}}_2}{\rm{\;(gas)\;}} + 3{{\rm{H}}_2}{\rm{O\;(gas)\;}} + 16{{\rm{e}}^ - }} \hfill \cr} $$

For a single In₂O₃ sensor, oxygen molecules in the air at room temperature are converted into adsorbed oxygen ions on the surface of the indium oxide material, forming an electron depletion layer. There are a lot of In₂O₃/In₂O₃ homogeneous junctions between indium oxide grains in this material. According to the carrier concentration formula for the n-type indium oxide sensor, most of the carriers in the material are negatively charged electrons. The electron concentration formula is shown as follows: 2n=niexp(EF−EikT)$$n = {n_i}\exp \left( {{{{E_F} - {E_i}} \over {kT}}} \right)$$

Where n is the electron concentration, n_i is the intrinsic carrier concentration, E_F is the Fermi level, E_i is the intrinsic Fermi level as the reference level, k is the Boltzmann constant, and T is the thermodynamic temperature. When In₂O₃ particles are in the air, ionized oxygen is adsorbed on the particle surface and electrons are lost, leading to a decrease in the conduction band electron concentration. According to Formula 2, E_i increases and bends upward, leading to an increase in barrier height. When exposed to a reducing gas such as acetone, a large number of electrons will be released into the material due to the reaction in Formula 1. When electrons are released, the electron concentration increases, so E_i decreases and the barrier height in the air decreases, as shown in Figure 11C. Similarly, for Co₃O₄ sensors, according to formula 3, when Co₃O₄ particles are in the air, E_i bends downward, and the barrier height in the air is higher than that of VOCs.

3p=niexp(Ei−EFkT)$$p = {n_i}\exp \left( {{{{E_i} - {E_F}} \over {kT}}} \right)$$

In In₂O₃/Co₃O₄, due to the contact between p-type Co₃O₄ grains and n-type In₂O₃ grains, there are a large number of p-Co₃O₄/n-In₂O₃ heterostructures besides homogeneous p-Co₃O₄/p-Co₃O₄ and n-In₂O₃/n-In₂O₃. The Fermi energy levels of n-In₂O₃ and p-Co₃O₄ are different. The electrons flow from n-In₂O₃ (W_f=5.0 eV, E_g=2.8 eV) [34] to p-Co₃O₄ (W_f=6.1 eV, E_g=1.6 eV) [35], forming an electron transfer until the Fermi energy level is balanced by the Fermi energy level balance effect [36, 37]. Therefore, the formation of p-n heterostructures in the composites additionally increases the width of In₂O₃/Co₃O₄ heterostructure depletion layer. In reducing gas atmosphere, as shown in Figure 11B, due to the release of electrons, the net decrease of the width of the depletion layer becomes larger, which improves gas-sensing performance.