High-Performance Cataluminescence Sensor Based on Nanosized V2O5 for 2-Butanone Detection

The development of high-performance sensors is of great significance for the control of the volatile organic compounds (VOCs) pollution and their potential hazard. In this paper, high crystalline V2O5 nanoparticles were successfully synthesized by a simple hydrothermal method. The structure and morphology of the prepared nanoparticles were characterized by TEM and XRD, and the cataluminescence (CTL) sensing performance was also investigated. Experiments found that the as-prepared V2O5 not only shows sensitive CTL response and good selectivity to 2-butanone, but also exhibits rapid response and recovery speed. The limit of detection was found to be 0.2 mg/m3 (0.07 ppm) at a signal to noise ratio of 3. In addition, the linear range exceeds two orders of magnitude, which points to the promising application of the sensor in monitoring of 2-butanone over a wide concentration range. The mechanism of the sensor exhibiting selectivity to different gas molecules were probed by quantum chemistry calculation. Results showed that the highest partial charge distribution, lowest HOMO-LUMO energy gap and largest dipole moment of 2-butanone among the tested gases result in it having the most sensitive response amongst other VOCs.


Introduction
Volatile organic compounds (VOCs) in residential buildings and in the workplace pose serious threats to public health [1][2][3]. As a typical colorless VOC, 2-butanone is produced industrially on a large scale, is widely used as a common solvent, denaturing agent and cleaning agent. It finds wide use in the manufacture of plastics, textiles, paints and other coatings [4]. However, short-term inhalation exposure to 2-butanone causes irritation to the eyes, nose, and throat of humans, and even central nervous system depression [5]. Slight neurological, liver, kidney, and respiratory effects have also been observed in chronic inhalation studies of 2-butanone in animals. Taking measures to protect public health, the Occupational Safety and Health Administration has promulgated an 8-h permissible exposure limit (PEL) of 590 mg/m 3 (200 ppm) for 2-butanone [6]. Therefore, the development of a simple and low-cost method for rapid monitoring of 2-butanone in air environments, especially in workplaces, is of significant importance to avoid its potential hazard.
Gas sensors represent a versatile technology for harmful gas monitoring because of their advantages of low-cost, small size and easy operation [7][8][9][10]. Nowadays, various gas sensors based on different principles, such as electrochemistry [11][12][13][14], surface acoustic wave [15,16], quartz crystal microbalance [17,18], and cataluminescence (CTL) [19][20][21], have been designed for the detection of different analytes. Among them, gas sensors based on CTL have attracted a great deal of scientific interest. CTL is a phenomenon of light emission produced by the catalytic oxidation of analyte on solid nanomaterial surface. During the process of CTL emission, only sample and oxygen in air are consumed; the solid nanomaterial is not consumed [22][23][24]. This characteristic means that CTL-based sensors have the advantage of good stability that is essential to long-term monitoring. With the development of material science, many advanced nanomaterials have been introduced into the design of CTL-based sensor. For instance, F-doped cage-like SiC-based sensors for hydrogen sulfide [25], Y-doped metal-organic framework-5-based sensors for iso-butanol [26], sensors based on hydrotalcite-supported gold nanoparticles for acetaldehyde [27], and sensors based on Y 2 O 3 with multi-shelled hollow structure for methanol [28]. The expanded availability of nanomaterials has greatly advanced the development of CTL-based sensors.
Vanadium oxide (V 2 O 5 ) has gained much research interest because of its excellent structural flexibility, lower bandgap and high energy density [29,30]. These features enable V 2 O 5 to be utilized for numerous applications including gas sensing. Nanosized V 2 O 5 with different structures have been synthesized for the electrochemical sensing of ethanol [31], ammonia [32] and xylene [33]. However, the application of nanosized V 2 O 5 in the design of CTL-based sensors has seldom been reported. In this paper, nanosized V 2 O 5 was synthesized by a simple hydrothermal method. The as-prepared V 2 O 5 was found to be a promising candidate for CTL sensing of 2-butanone. The CTL sensor based on V 2 O 5 exhibited high sensitivity, good selectivity, rapid response speed and recovery speed to 2-butanone. Other VOCs, including acetone, 3-pentanone, n-hexane, methanol, ethanol and so on, only produced a very weak or no response when they flowed through V 2 O 5 surface. It is well known that CTL-based sensors have a good selectivity due to the harsh conditions for CTL emission. However, the gas-sensing selectivity mechanism remains uncertain. Herein, the possible gas-sensing selectivity mechanisms were investigated based on quantum chemistry calculation. Figure 1 shows the XRD patterns of nanosized V 2 O 5 and precursor before calcination. The XRD pattern of precursor shows a poor crystallinity, the diffraction peaks cannot be assigned to any known single compound, meaning the precursor may be an adduct, or a complex mixture of oxides and hydroxides. For the XRD pattern of nanosized V 2 O 5 , all the diffraction peaks match the orthorhombic crystalline phase of V 2 O 5 (JCPDS card no. .

Characterizations of Nanosized V 2 O 5
Molecules 2020, 25, x FOR PEER REVIEW 2 of 14 solid nanomaterial surface. During the process of CTL emission, only sample and oxygen in air are consumed; the solid nanomaterial is not consumed [22][23][24]. This characteristic means that CTL-based sensors have the advantage of good stability that is essential to long-term monitoring. With the development of material science, many advanced nanomaterials have been introduced into the design of CTL-based sensor. For instance, F-doped cage-like SiC-based sensors for hydrogen sulfide [25], Y-doped metal-organic framework-5-based sensors for iso-butanol [26], sensors based on hydrotalcite-supported gold nanoparticles for acetaldehyde [27], and sensors based on Y2O3 with multi-shelled hollow structure for methanol [28]. The expanded availability of nanomaterials has greatly advanced the development of CTL-based sensors. Vanadium oxide (V2O5) has gained much research interest because of its excellent structural flexibility, lower bandgap and high energy density [29,30]. These features enable V2O5 to be utilized for numerous applications including gas sensing. Nanosized V2O5 with different structures have been synthesized for the electrochemical sensing of ethanol [31], ammonia [32] and xylene [33]. However, the application of nanosized V2O5 in the design of CTL-based sensors has seldom been reported. In this paper, nanosized V2O5 was synthesized by a simple hydrothermal method. The as-prepared V2O5 was found to be a promising candidate for CTL sensing of 2-butanone. The CTL sensor based on V2O5 exhibited high sensitivity, good selectivity, rapid response speed and recovery speed to 2-butanone. Other VOCs, including acetone, 3-pentanone, n-hexane, methanol, ethanol and so on, only produced a very weak or no response when they flowed through V2O5 surface. It is well known that CTL-based sensors have a good selectivity due to the harsh conditions for CTL emission. However, the gassensing selectivity mechanism remains uncertain. Herein, the possible gas-sensing selectivity mechanisms were investigated based on quantum chemistry calculation. Figure 1 shows the XRD patterns of nanosized V2O5 and precursor before calcination. The XRD pattern of precursor shows a poor crystallinity, the diffraction peaks cannot be assigned to any known single compound, meaning the precursor may be an adduct, or a complex mixture of oxides and hydroxides. For the XRD pattern of nanosized V2O5, all the diffraction peaks match the orthorhombic crystalline phase of V2O5 (JCPDS card no. 41-1426).  The grain sizes of V 2 O 5 is was calculated by using the Debye-Scherrer equation [34]:

Characterizations of Nanosized V2O5
where κ is the shape factor and its value is 0.94, λ is the wavelength of X-ray radiation (1.5406 Å), β is the full width at half the maximum intensity and θ is the Bragg angle in Radian. The d-spacing (Å) was calculated by using Bragg' equation [35]: where n is the order of diffraction equal to 1. The dislocation density (nm −2 ) was calculated according to the following equation [35]: The microstrain (ε) was calculated by using the following equation [36]: Multiple peak fit was performed to determine the θ and β values from the XRD data. The results are summarized in Table 1. The average grain size was calculated to be 30.46 nm. The average d-spacing, δ and ε were determined to be 2.330 Å, 9.832 × 10 −3 nm −2 and 5.275 × 10 −3 , respectively. The lattice constants were determined according to the following formula [37]: where h, k, l are the miller indices. Crystal planes of (200), (002) and (001) were chosen to determine lattice constants, and a, b and c were calculated to be 1.155 nm, 0.356 nm and 0.4373 nm, respectively.
There calculated values are in very good approximation to standard values. [29,31] Lower magnification the transmission electron microscopy (TEM) image ( Figure 2a) shows the nanosized V 2 O 5 displays a nearly hexagonal structure. Figure where h, k, l are the miller indices. Crystal planes of (200), (002) and (001) were chosen to determine lattice constants, and a, b and c were calculated to be 1.155 nm, 0.356 nm and 0.4373 nm, respectively.
There calculated values are in very good approximation to standard values. [29,31] Lower magnification the transmission electron microscopy (TEM) image ( Figure 2a) shows the nanosized V2O5 displays a nearly hexagonal structure. Figure 2b narrates the high-resolution TEM (HRTEM) image of nanosized V2O5. The d-spacing of 0.44 nm is attributed to the (001) plane of V2O5. The selected area electron diffraction (SAED) shows the ring pattern mainly arising from the (001) and (002) planes of V2O5 structures, indicating the crystalline nature of V2O5 nanoparticles. Energydispersive X-ray spectroscopy (EDAX) with color mapping analysis was carried out to gain insight into the chemical composition and positional distribution of V and O in V2O5, the results are shown in Figure 2d-f. The results of the XRD pattern and TEM images indicate that highly crystalline V2O5 has been successfully synthesized.

Optimization of Sensing Conditions
CTL is produced during the heterogeneous catalytic reaction emitting photoemissions, the CTL spectrum usually is a continuum emission band ranging from the ultraviolet to blue wavelength. Under the conditions of operating temperature at 231 °C and flow rate of 220 mL/min, the influence

Optimization of Sensing Conditions
CTL is produced during the heterogeneous catalytic reaction emitting photoemissions, the CTL spectrum usually is a continuum emission band ranging from the ultraviolet to blue wavelength. Under the conditions of operating temperature at 231 • C and flow rate of 220 mL/min, the influence of detection wavelength on sensing of 2-butanone was investigated in the range of 400-535 nm. Figure 3a shows the dynamic response curves of 2-butanone measured by using different interference filters. The maximum apparent CTL intensity (I) is observed at 460 nm of wavelength. The CTL response signal (S), background noise (N) and signal to noise ratio (SNR) at different wavelengths are plotted in Figure 3b. It can be seen that although the maximum S value appears at 460 nm, the background noise emitted by thermal radiation increases with the increase in wavelength, resulting in SNR reaching its maximum at 440 nm. Therefore, 440 nm was selected as the optimal wavelength for the CTL sensing of 2-butanone.
Molecules 2020, 25, x FOR PEER REVIEW 5 of 14 Figure 3a shows the dynamic response curves of 2-butanone measured by using different interference filters. The maximum apparent CTL intensity (I) is observed at 460 nm of wavelength. The CTL response signal (S), background noise (N) and signal to noise ratio (SNR) at different wavelengths are plotted in Figure 3b. It can be seen that although the maximum S value appears at 460 nm, the background noise emitted by thermal radiation increases with the increase in wavelength, resulting in SNR reaching its maximum at 440 nm. Therefore, 440 nm was selected as the optimal wavelength for the CTL sensing of 2-butanone. The operating temperature plays a critical role in gas sensing. Figure 3c shows the dynamic response curves of 2-butanone measured at different operating temperatures. It can be seen that the I value increases with the increase in operating temperature before 237 °C, and then decreases when the operating temperature is above 237 °C. The Figure 3d shows the change trends of S, N and SNR versus operating temperatures. It shows that the S value increases almost exponentially with increasing operating temperature until 237 °C, and then decreases significantly with a further increase in operating temperature. According to the classical theory, the luminous intensity is proportional to luminous efficiency and reaction rate [38]. Therefore, the change trend of S value perhaps can be attributed to the reaction rate, increasing exponentially with increasing operating temperature, but higher operating temperatures accelerate the molecular motion that leads to quenching of CTL emission. In addition, the background noise emitted by thermal radiation also increases with the increase in operating temperature, rendering a decrease in S value and SNR. Because the maximum SNR is observed at 231 °C, a temperature of 231 °C was selected as the optimal operating temperature for the CTL sensing of 2-butanone. The operating temperature plays a critical role in gas sensing. Figure 3c shows the dynamic response curves of 2-butanone measured at different operating temperatures. It can be seen that the I value increases with the increase in operating temperature before 237 • C, and then decreases when the operating temperature is above 237 • C. The Figure 3d shows the change trends of S, N and SNR versus operating temperatures. It shows that the S value increases almost exponentially with increasing operating temperature until 237 • C, and then decreases significantly with a further increase in operating temperature. According to the classical theory, the luminous intensity is proportional to luminous efficiency and reaction rate [38]. Therefore, the change trend of S value perhaps can be attributed to the reaction rate, increasing exponentially with increasing operating temperature, but higher operating temperatures accelerate the molecular motion that leads to quenching of CTL emission. In addition, the background noise emitted by thermal radiation also increases with the increase in operating temperature, rendering a decrease in S value and SNR. Because the maximum SNR is observed at 231 • C, a temperature of 231 • C was selected as the optimal operating temperature for the CTL sensing of 2-butanone.
The influence of flow rate on the CTL sensing of 2-butanone was investigated by changing the flow rate ranging from 150 to 225 mL/min. All the 2-butanone samples were introduced into the sensor 4 s after data acquisition. The dynamic response curves of 2-butanone at different flow rates are shown in Figure 4a. It can be seen that the peak width of the dynamic response curve decreases as the flow rate is raised. The response time (t Res ) and recovery time (t Rec ) are two important performances to a gas sensor. Here, the response time is defined as the time taken to reach the maximum response signal after injection of sample, and the recovery time is defined as the time taken for maximum response signal to recover to baseline.
performances to a gas sensor. Here, the response time is defined as the time taken to reach the maximum response signal after injection of sample, and the recovery time is defined as the time taken for maximum response signal to recover to baseline.
In order to further evaluate the relationship between flow rate and sensing performances, the dynamic response curve at each flow rate was measured in three replicates, the change trends of tRes and tRec versus flow rate are shown Figure 4b, and the relationship between S value and flow rate is shown in Figure 4c. Both tRes and tRec decreases as the increasing of flow rate, and then almost remain stable. For S value, it increases almost proportionally with increasing flow rate before 220 mL/min, and then decreases with a further increase in flow rate. These results show that the catalytic oxidation reaction of 2-butanone on nanosized V2O5 surface is controlled by diffusion rate when the flow rate is below 220 mL/min; that is the total reaction rate is controlled by the rate of 2-butanone in the gas phase transfer to the catalyst surface, and thereby the increase in flow rate causes increasing S value. However, the total reaction rate is controlled by the oxidation rate of 2-butanone on the nanosized V2O5 surface when the flow rate exceeds 220 mL/min, resulting in tRec remaining stable. In addition, under high flow rate conditions, 2-butanone molecules flow through the catalyst surface too quickly to be sufficiently oxidized, rending a decrease in S value. In the case of 200 mg/m 3 of 2-butanone, the short response time of 2 s and quick recovery time of 9 s were observed under 220 mL/min. Therefore, 220 mL/min was used for the sensing of 2-butanone as the sensitive response, as well as the satisfactory response and recovery time under this condition.

Evaluation of Selectivity
Selectivity is a very crucial performance for a gas sensor. The selectivity of a gas sensor can be defined as [39]: The selectivity of nanosized V2O5 sensors was investigated by measuring 200 mg/m 3 of 2butanone and 19 other kinds of gases commonly existing in workplace. Results showed that only 2butanone, acetone, 3-pentanone and n-hexane produce CTL responses. The dynamic response curves of 2-butanone, acetone, 3-pentanone and n-hexane on nanosized V2O5 surface are shown in Figure 5a. A much stronger response of 2-butanone was observed. Figure 5b displays the selectivity of the V2O5based sensor calculated by equation 6. The CTL responses of acetone, 3-pentanone, n-hexane are about 8.8%, 1.4% and 0.8% of that of 2-butanone, respectively. Other gases, including n-heptane, methanol, ethanol, ethyl acetate, benzene, toluene, o-xylene, m-xylene, p-xylene, formaldehyde, In order to further evaluate the relationship between flow rate and sensing performances, the dynamic response curve at each flow rate was measured in three replicates, the change trends of t Res and t Rec versus flow rate are shown Figure 4b, and the relationship between S value and flow rate is shown in Figure 4c. Both t Res and t Rec decreases as the increasing of flow rate, and then almost remain stable. For S value, it increases almost proportionally with increasing flow rate before 220 mL/min, and then decreases with a further increase in flow rate. These results show that the catalytic oxidation reaction of 2-butanone on nanosized V 2 O 5 surface is controlled by diffusion rate when the flow rate is below 220 mL/min; that is the total reaction rate is controlled by the rate of 2-butanone in the gas phase transfer to the catalyst surface, and thereby the increase in flow rate causes increasing S value. However, the total reaction rate is controlled by the oxidation rate of 2-butanone on the nanosized V 2 O 5 surface when the flow rate exceeds 220 mL/min, resulting in t Rec remaining stable. In addition, under high flow rate conditions, 2-butanone molecules flow through the catalyst surface too quickly to be sufficiently oxidized, rending a decrease in S value. In the case of 200 mg/m 3 of 2-butanone, the short response time of 2 s and quick recovery time of 9 s were observed under 220 mL/min. Therefore, 220 mL/min was used for the sensing of 2-butanone as the sensitive response, as well as the satisfactory response and recovery time under this condition.

Evaluation of Selectivity
Selectivity is a very crucial performance for a gas sensor. The selectivity of a gas sensor can be defined as [39]: The selectivity of nanosized V 2 O 5 sensors was investigated by measuring 200 mg/m 3 of 2-butanone and 19 other kinds of gases commonly existing in workplace. Results showed that only 2-butanone, acetone, 3-pentanone and n-hexane produce CTL responses. The dynamic response curves of 2-butanone, acetone, 3-pentanone and n-hexane on nanosized V 2 O 5 surface are shown in Figure 5a. A much stronger response of 2-butanone was observed. Figure 5b displays the selectivity of the V 2 O 5 -based sensor calculated by equation 6. The CTL responses of acetone, 3-pentanone, n-hexane are about 8.8%, 1.4% and 0.8% of that of 2-butanone, respectively. Other gases, including n-heptane, methanol, ethanol, ethyl acetate, benzene, toluene, o-xylene, m-xylene, p-xylene, formaldehyde, acetaldehyde, trichloromethane, tetrachloromethane, propyl acetate and ammonia, cannot induce a response under this condition. The above results fully demonstrate that the CTL sensor based on nanosized V 2 O 5 has a good selectivity when sensing 2-butanone.
Molecules 2020, 25, x FOR PEER REVIEW 7 of 14 acetaldehyde, trichloromethane, tetrachloromethane, propyl acetate and ammonia, cannot induce a response under this condition. The above results fully demonstrate that the CTL sensor based on nanosized V2O5 has a good selectivity when sensing 2-butanone.   [6]. The LOD of the present sensor is much lower than the permissible exposure limit; moreover, the linear range exceeds two orders of magnitude, indicating that the sensor has a potential for routine monitoring of 2-butanone in the workplace where 2-butanone levels may vary in concentration from low to the occasional high level.   The Occupational Safety and Health Administration permissible exposure limit of 2-butanone during an 8-h workday is 590 mg/m 3 (200 ppm) [6]. The LOD of the present sensor is much lower than the permissible exposure limit; moreover, the linear range exceeds two orders of magnitude, indicating that the sensor has a potential for routine monitoring of 2-butanone in the workplace where 2-butanone levels may vary in concentration from low to the occasional high level.

Analytical Characteristics
Molecules 2020, 25, x FOR PEER REVIEW 7 of 14 acetaldehyde, trichloromethane, tetrachloromethane, propyl acetate and ammonia, cannot induce a response under this condition. The above results fully demonstrate that the CTL sensor based on nanosized V2O5 has a good selectivity when sensing 2-butanone.   [6]. The LOD of the present sensor is much lower than the permissible exposure limit; moreover, the linear range exceeds two orders of magnitude, indicating that the sensor has a potential for routine monitoring of 2-butanone in the workplace where 2-butanone levels may vary in concentration from low to the occasional high level.  The relative standard was 4.2% for five times sensing of 2-butanone at 200 mg/m 3 (Figure 6b), indicating that the sensor has a good reproducibility. The analytical characteristics of some sensors for 2-butanone are summarized in Table 2. Compared with the other sensors for 2-butanone, the present sensor has a wider linear range and a lower LOD. ND stands for not discussion.

Gas-Sensing Selectivity Mechanism
The gas sensing process usually relates to the surface adsorption and surface reaction of gas molecules on the metal oxide nanoparticle. When metal oxide nanoparticles are heated in air, O 2 molecules from the air can be adsorbed on the surface of metal oxide nanoparticle, and then chemisorbed oxygen species (O 2− , O − , and O 2− ) are formed by trapping electrons from the conduction band [4,42]. The working temperature has an important impact on the type of chemisorbed oxygen species. An increase in temperature, the chemisorbed oxygen adsorbed on the surface of metal oxide nanoparticle undergoes the following state: Previous studies demonstrated that O − is the most active and stable chemisorbed oxygen species in the range of 100-300 • C [38]. Upon exposure to VOCs, VOCs can be oxidized into CO 2 by the chemisorbed oxygen species on the surface of metal oxide nanoparticle [43]. If an excited intermediate was formed during the above process, and the excited intermediate returned to the ground state via radiative transition, CTL emission could be observed. From Figure 2, it can be seen that nanosized V 2 O 5 exhibits a sensitive CTL response to 2-butanone, and a weak response to acetone, 3-pentanone and n-hexane. Although it is very difficult to identify the luminous species due to the very short-lived excited states, we can conclude that the catalytic reaction of the 2-butanone is more favorable to produce CTL emission than other VOCs (e.g., acetone, 3-pentanone and n-hexane) on nanosized V 2 O 5 surface.
The sensing response is affected by the reactivity and absorptivity of gas molecules on sensing materials. In order to probe the mechanism behind the sensor shows different response to different VOCs, quantum calculations were implemented, and 2-butanone, acetone, and 3-pentanone and n-hexane were selected as representatives. All the optimization of structures and analysis of vibrational frequencies (energy calculation) in this study were performed using Gaussian 09 with B3LYP/6-311G (d, p) basis set.
The 3D molecular structures of the four molecules and their partial charge distribution calculated by Gaussian are shown in Figure 7. It can be seen that C1 site in 2-butanon, C1/C3 sites in acetone, O1 site in 3-pentanone, and C1/C6 sites in n-hexane are the more active sites to occur reactions due to these sites has higher charge dense than other sites in corresponding molecular structures. Because C1 site in 2-butanon has highest charge dense (−0.3249 e), meaning that 2-butanone is easier to react with the chemisorbed oxygen species than other gas molecules. structures. Because C1 site in 2-butanon has highest charge dense (−0.3249 e), meaning that 2butanone is easier to react with the chemisorbed oxygen species than other gas molecules. It has been revealed that the highest occupied and lowest unoccupied molecular orbitals (HOMO and LUMO) play a prominent role in governing chemical reaction. A large energy gap (Eg = EHOMO-ELUMO) between HOMO and LUMO implies high stability, which indicates low chemical reactivity. In turn, a small energy gap implies low stability and thereby high chemical reactivity [44]. Figure 8 illustrates the energy gaps in graph forms respective to the HOMO and LUMO levels of four molecules. The HOMO-LUMO energy gaps for 2-butanone, acetone, and 3-pentanone and n-hexane are 6.11, 6.18, 6.39 and 9.71 eV. It means that in order of reactivity from highest to lowest is 2-butanone, acetone, and 3-pentanone and n-hexane. A high level of consistency exists between the energy gap and the CTL response signal. Moreover, it was reported that the absorptivity of gas molecules on sensing materials is associated with the dipole moment of molecule in its gaseous state. The larger the dipole moment, the higher the attractive force between the gas molecule and sensing material [45,46]. The dipole moments in the molecular structure of 2-butanone, acetone, and 3-pentanone and n-hexane were calculated to be 2.7497, 2.6475, 2.6246 and 0 D, respectively. The dipole moment of 2butanone is larger than other gas molecule, suggesting the interaction between 2-butanone molecules and nanosized V2O5 is strongest among these gases. Therefore, more 2-butanone molecules can react with the chemisorbed oxygen species on the nanosized V2O5 surface, displaying the observed selective response toward 2-butanone. It has been revealed that the highest occupied and lowest unoccupied molecular orbitals (HOMO and LUMO) play a prominent role in governing chemical reaction. A large energy gap (E g = E HOMO -E LUMO ) between HOMO and LUMO implies high stability, which indicates low chemical reactivity. In turn, a small energy gap implies low stability and thereby high chemical reactivity [44]. Figure 8 illustrates the energy gaps in graph forms respective to the HOMO and LUMO levels of four molecules. The HOMO-LUMO energy gaps for 2-butanone, acetone, and 3-pentanone and n-hexane are 6.11, 6.18, 6.39 and 9.71 eV. It means that in order of reactivity from highest to lowest is 2-butanone, acetone, and 3-pentanone and n-hexane. A high level of consistency exists between the energy gap and the CTL response signal. Moreover, it was reported that the absorptivity of gas molecules on sensing materials is associated with the dipole moment of molecule in its gaseous state. The larger the dipole moment, the higher the attractive force between the gas molecule and sensing material [45,46]. The dipole moments in the molecular structure of 2-butanone, acetone, and 3-pentanone and n-hexane were calculated to be 2.7497, 2.6475, 2.6246 and 0 D, respectively. The dipole moment of 2-butanone is larger than other gas molecule, suggesting the interaction between 2-butanone molecules and nanosized V 2 O 5 is strongest among these gases. Therefore, more 2-butanone molecules can react with the chemisorbed oxygen species on the nanosized V 2 O 5 surface, displaying the observed selective response toward 2-butanone. According to the above discussion, we can conclude that the catalytic oxidations of 2-butanone, acetone, and 3-pentanone and n-hexane on nanosized V2O5 surfaces are able to form excited intermediates and favor their radiative transitions. The highest partial charge distribution, lowest HOMO-LUMO energy gap and largest dipole moment of 2-butanone among the tested gases are According to the above discussion, we can conclude that the catalytic oxidations of 2-butanone, acetone, and 3-pentanone and n-hexane on nanosized V 2 O 5 surfaces are able to form excited intermediates and favor their radiative transitions. The highest partial charge distribution, lowest HOMO-LUMO energy gap and largest dipole moment of 2-butanone among the tested gases are responsible for its strongest sensing response. However, more research is still needed to better explain the detailed mechanism relating to excited state chemistry.

Synthesis and Characterizations of Nanosized V 2 O 5
Nanosized V 2 O 5 was synthesized by a simple hydrothermal method. briefly, ammonium metavanadate solution (30 mL, 30 mg/mL) was mixed with dilute hydrochloric acid solution (20 mL, 1.0 mol/L) under vigorous stirring at room temperature till appearing bright yellow clarified liquid. The resultant solution was transferred into a 100 mL Teflon container. The Teflon container was sealed tightly by a stainless-steel autoclave, and then the autoclave was subjected to hydrothermal reaction at 180 • C for 12 h in a vacuum oven. The resultant precipitates were centrifuged and washed with deionized water and ethanol five times. Afterwards, the collected precipitates were dried at 100 • C for 10 h in a vacuum oven. Finally, the dried powders were calcined at 500 • C in the air for 2 h to obtain nanosized V 2 O 5 .
The morphology of the as-prepared nanosized V 2 O 5 was characterized by TEM (FEI Tecnai G2 F20) with an accelerating voltage of 200 kV. Elemental analyses were performed by EDAX (Bruker Nano GmbH). The phase structure was investigated by Powder X-ray diffraction (XRD, Rigaku, Ultima IV) with Cu Kα radiation (λ = 1.5406 Å).

Fabrication of the Sensor System
The schematic diagram of the sensor for 2-butanone is shown in Figure 9. Nanosized V 2 O 5 was sintered onto a ceramic heater to form a catalyst layer. The ceramic heater was inserted into a home-made quartz tube (length = 10 cm, diameter = 0.6 cm) with gas-inlet and gas-outlet. The quartz tube was placed into a sample chamber with a nono-paque widow at the bottom. The temperature of catalyst layer was controlled by adjusting the output voltage of the temperature controller connecting to the ceramic heater, and the detailed temperature could be measured by thermocouple. A photomultiplier (PMT) was assembled into a cylindrical tube under the nono-paque widow for measurement of luminous signal. Interference filters were used to reduce the background noise, and the detection wavelength can be selected by inserting different interference filters between the quartz tube and PMT. The air and sample were delivered by an air pump. The sample reacted with the oxygen in air on the surface of nanosized V 2 O 5 to produce light emission. A commercial BPCL ultra-weak luminescence analyzer was used record and process the signal intensity.
The peak intensity on dynamic response curves directly recorded by the instrument is defined as apparent CTL intensity. The CTL response signal (S) is equal to the apparent CTL intensity minus the background noise, and can be expressed as: where I is the apparent CTL intensity on dynamic response curves, N stands for the background noise. The signal to noise ratio (SNR) is described as: Molecules 2020, 25, x FOR PEER REVIEW 11 of 14 oxygen in air on the surface of nanosized V2O5 to produce light emission. A commercial BPCL ultraweak luminescence analyzer was used record and process the signal intensity. The peak intensity on dynamic response curves directly recorded by the instrument is defined as apparent CTL intensity. The CTL response signal (S) is equal to the apparent CTL intensity minus the background noise, and can be expressed as: where I is the apparent CTL intensity on dynamic response curves, N stands for the background noise. The signal to noise ratio (SNR) is described as:

Methods for Quantum Calculations
Quantum calculations were implemented to reveal the mechanism of the difference of gassensing selectivity mechanism. Three compounds, 2-butanone, acetone, and 3-pentanone and nhexane were selected as representatives in this theoretical study. All the optimization of structures and analysis of vibrational frequencies (energy calculation) in this study were performed by using Gaussian 09 with B3LYP/6-311G (d, p) basis set. After the structural optimization, Mulliken charges were obtained. Based on these values, the dipole moment (μ) along X, Y, and Z axis (μx, μy, and μz) can be calculated. The dipole moment is calculated on the basis of the following equation: (μ) = u + u + u The energy of the HOMO) and LUMO were calculated based on the population analysis using Self-Consistent Field density.

Conclusions
In conclusion, V2O5 nanomaterials were successfully synthesized by a simple and facile hydrothermal method. CTL-based gas sensor fabricated by nanosized V2O5 shows high sensitivity, good selectivity, rapid response and recovery speed to 2-butanone. The detection conditions, including detecting wavelength, operating temperature and flow rate on the sensing of 2-butanone, were investigated in detail. Quantum chemistry calculation was performed to probe the gas-sensing selectivity mechanism. The CTL sensor based on V2O5 shows high sensitivity and selectivity to 2butanone which can be attributed to 2-butanone having a highest partial charge distribution, the lowest HOMO-LUMO energy gap and the largest dipole moment. All these factors favor 2-butanone

Methods for Quantum Calculations
Quantum calculations were implemented to reveal the mechanism of the difference of gas-sensing selectivity mechanism. Three compounds, 2-butanone, acetone, and 3-pentanone and n-hexane were selected as representatives in this theoretical study. All the optimization of structures and analysis of vibrational frequencies (energy calculation) in this study were performed by using Gaussian 09 with B3LYP/6-311G (d, p) basis set. After the structural optimization, Mulliken charges were obtained. Based on these values, the dipole moment (µ) along X, Y, and Z axis (µ x , µ y , and µ z ) can be calculated. The dipole moment is calculated on the basis of the following equation: (µ) = u 2 x + u 2 y + u 2 z (9) The energy of the HOMO) and LUMO were calculated based on the population analysis using Self-Consistent Field density.

Conclusions
In conclusion, V 2 O 5 nanomaterials were successfully synthesized by a simple and facile hydrothermal method. CTL-based gas sensor fabricated by nanosized V 2 O 5 shows high sensitivity, good selectivity, rapid response and recovery speed to 2-butanone. The detection conditions, including detecting wavelength, operating temperature and flow rate on the sensing of 2-butanone, were investigated in detail. Quantum chemistry calculation was performed to probe the gas-sensing selectivity mechanism. The CTL sensor based on V 2 O 5 shows high sensitivity and selectivity to 2-butanone which can be attributed to 2-butanone having a highest partial charge distribution, the lowest HOMO-LUMO energy gap and the largest dipole moment. All these factors favor 2-butanone molecules reacting more easily with the chemisorbed oxygen species on the nanosized V 2 O 5 surface. This work provides a promising method for the monitoring of 2-butanone in the indoor air environment, including in the workplace, in industry, and in houses.