Conductometric ppb-Level CO Sensors Based on In2O3 Nanofibers Co-Modified with Au and Pd Species

Carbon monoxide (CO) is one of the most toxic gases to human life. Therefore, the effective monitoring of it down to ppb level is of great significance. Herein, a series of In2O3 nanofibers modified with Au or Pd species or simultaneous Au and Pd species have been prepared by electrospinning combined with a calcination process. The as-obtained samples are applied for the detection of CO. Gas-sensing investigations indicate that 2 at% Au and 2 at% Pd-co-modified In2O3 nanofibers exhibit the highest response (21.7) to 100 ppm CO at 180 °C, and the response value is ~8.5 times higher than that of pure In2O3 nanofibers. More importantly, the detection limit to CO is about 200 ppb with a response value of 1.23, and is obviously lower than that (6 ppm) of pure In2O3 nanofibers. In addition, the sensor also shows good stability within 19 days. These demonstrate that co-modifying In2O3 nanofibers with suitable amounts of Pd and Au species might be a meaningful strategy for the development of high-performance carbon monoxide gas sensors.


Introduction
The incomplete combustion of fuel in a closed environment, such as home stoves, engines, and fires, will inevitably produce CO, which has extremely higher affinity to human hemoglobin compared with O 2 , and threatens the safety of human life [1]. Currently, the number of poisoning incidents caused by CO is about 15,000 per year [2]. Therefore, it is urgent and meaningful to develop high-performance CO gas sensors with high response and low detection limit.
CO sensors based on metal oxide gas sensors have obvious cross-response to various reducing or combustible gases such as H 2 and CH 4 , especially H 2 . During the use of fuel-cell vehicles using H 2 as fuel, excessive CO will cause poisoning of the fuel cells and affect the operation of the vehicle [3]. In order to ensure the normal movement of fuel-cell vehicles, it is urgent to improve the selectivity of CO sensors to other combustible gases. Metal oxide semiconductors, such as SnO 2 [4], In 2 O 3 [5], and ZnO [6], have been mostly investigated for CO detection. Among them, In 2 O 3 showed excellent selectivity for detecting CO compared with other oxides [7,8]. However, unmodified In 2 O 3 has poor response to CO [9] and requires further modification to improve its sensing performance [5]. Au, Pd, and Pt [8] and their compounds are currently the most commonly used modified noble metals due to their excellent catalytic and electrical properties. Yin  All of the reagents were analytical grade and were used as received without further purification. In 2 O 3 nanofibers with different Au and Pd modifying concentrations were synthesized by an electrospinning method, and then were calcinated in air to obtain the final products [15]. Firstly, solution A was prepared by dissolving 1 mmol In 2 O 3 ·4.5H 2 O and a certain dose of HAuCl 4 and PdCl 2 in a mixture solution containing 4 mL absolute ethanol and 4 mL N, N-Dimethylformamide (DMF), followed by continuous stirring at room temperature on a magnetic mixer for 20 min. The atomic ratios of ln, Au, and Pd in the six solutions A were 1:0:0:, 1:0.02:0, 1:0:0.02, 1:0.01:0.01, 1:0.02:0.02, and 1:0.04:0.04, respectively. After that, 1 g polyvinylpyrrolidone (PVP, Mw = 1,300,000 g/mol) was slowly added to solution A with a water bath for 5 h at 50 • C to obtain solution B. Subsequently, solution B was poured into a 5 mL plastic syringe for electrostatic spinning. The principles and use of typical electrostatic spinning equipment has been described in our previous work [15]. The related parameters were as follows: an electrostatic pressure of 17 kV was applied between the metal needle and collector with a distance of 15 cm, and the feed rate was controlled at 0.02 mL/h. The white as-spun precursor on aluminum foil was collected after 2 h, followed by calcination in a muffle furnace at 450 • C for 2 h with a heating rate of 2 • C/min. The samples were labeled as pure In 2 O 3 , Au2-In 2 O 3 , Pd2-In 2 O 3 , Au1Pd1-In 2 O 3 , Au2Pd2-In 2 O 3 , and Au4Pd4-In 2 O 3 nanofibers.

Characterization
The X-ray powder diffraction (XRD) patterns of six groups of samples were analyzed using a X-ray diffractometer (XRD, Smartlab, Rigaku Corporation, Tokyo, Japan) in the 2θ range of 25 • -80 • . The morphology and size of the six groups of samples were observed by field emission scanning electron microscopy (SEM, Merlin Compact, Carl Zeiss NTS GmbH, Oberkochen, Germany) at an accelerating voltage of 15 kV. Further microscopic surface probing of pure In 2 O 3 and Au2Pd2-In 2 O 3 nanofibers was performed using a transmission electron microscope (TEM, Titan G260-300, FEI, Hillsboro, OR, USA) facility at an accelerating voltage of 200 kV. The sample binding energies and O 1s were analyzed by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Scientifific K-Alpha, Waltham, MA, USA) using a Mg Kα (1358.6 eV) X-ray source.

Fabrication and Measurement of Gas Sensor
The traditional hexapod ceramic tube was used to make the sensor device. The actual image of the hexapod ceramic tube gas sensor and the illustration of the ceramic tube gold electrode are shown in Figure 1. The fabrication and testing process of the device has been introduced in detail in the previous study [16]. The sensor was aged for 48 h at an aging current of 100 mA under a laboratory environment (~25 • C,~35% RH). The sensor was placed in a gas cylinder filled with air and the gas to be measured, and a Fluke (8846A) meter was used to detect the change in the resistance value at both ends of the Au electrode. The humidity test was carried out in a humidity chamber (Shanghai ESPC Environmental Equipment Company, Shanghai, China), and the temperature of the humidity chamber was always controlled at 25 • C. The gas response (S) is defined as the ratio of the stable resistance (R a ) in air to the stable resistance (R g ) in the gas being measured (S = R a /R g ). Furthermore, response and recovery times are defined as the time required for a 90% change in the sensor resistance.

Characterization
The X-ray powder diffraction (XRD) patterns of six groups of samples were analyzed using a X-ray diffractometer (XRD, Smartlab, Rigaku Corporation, Tokyo, Japan) in the 2θ range of 25°-80°. The morphology and size of the six groups of samples were observed by field emission scanning electron microscopy (SEM, Merlin Compact, Carl Zeiss NTS GmbH, Oberkochen, Germany) at an accelerating voltage of 15 kV. Further microscopic surface probing of pure In2O3 and Au2Pd2-In2O3 nanofibers was performed using a transmission electron microscope (TEM, Titan G260-300, FEI, Hillsboro, OR, USA) facility at an accelerating voltage of 200 kV. The sample binding energies and O 1s were analyzed by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Scientifific K-Alpha, Waltham, MA, USA) using a Mg Kα (1358.6 eV) X-ray source.

Fabrication and Measurement of Gas Sensor
The traditional hexapod ceramic tube was used to make the sensor device. The actual image of the hexapod ceramic tube gas sensor and the illustration of the ceramic tube gold electrode are shown in Figure 1. The fabrication and testing process of the device has been introduced in detail in the previous study [16]. The sensor was aged for 48 h at an aging current of 100 mA under a laboratory environment (~25 °C, ~35% RH). The sensor was placed in a gas cylinder filled with air and the gas to be measured, and a Fluke (8846A) meter was used to detect the change in the resistance value at both ends of the Au electrode. The humidity test was carried out in a humidity chamber (Shanghai ESPC Environmental Equipment Company, Shanghai, China), and the temperature of the humidity chamber was always controlled at 25 °C. The gas response (S) is defined as the ratio of the stable resistance (Ra) in air to the stable resistance (Rg) in the gas being measured (S = Ra/Rg). Furthermore, response and recovery times are defined as the time required for a 90% change in the sensor resistance.

Structural and Morphological Characteristics
XRD patterns of all samples are presented in Figure 2a. According to XRD patterns (Figure 2a), all diffraction peaks of the as-obtained samples are consistent with the standard card of In2O3 (PDF#06-0416). The diffraction peaks belonging to Au and Pd species can be barely observed due to the low modifying contents of Au and Pd species. Figure  2b displays the patterns of the high resolution (440) peak of all samples. As shown in Figure 2b, the half widths of the diffraction peaks of the other samples gradually increase and shift to small angles with an increase in the modifying amount of Pd species [17]. This is because the radius of a Au atom (0.134 nm) is too large (larger than that of In 3+ (0.08 nm)) to incorporate into the In2O3 lattice, while the radius of Pd 2+ (0.085 nm) is only slightly larger than that of In 3+ , and Pd 2+ can replace the position of In ions, which makes the interplanar spacing larger. The grain sizes are calculated according to Scherrer's formula (D = 0.89 λ/β cosθ) ( Table 1). It can be seen that Au modification can only slightly decrease

Structural and Morphological Characteristics
XRD patterns of all samples are presented in Figure 2a. According to XRD patterns (Figure 2a), all diffraction peaks of the as-obtained samples are consistent with the standard card of In 2 O 3 (PDF#06-0416). The diffraction peaks belonging to Au and Pd species can be barely observed due to the low modifying contents of Au and Pd species. Figure 2b displays the patterns of the high resolution (440) peak of all samples. As shown in Figure 2b, the half widths of the diffraction peaks of the other samples gradually increase and shift to small angles with an increase in the modifying amount of Pd species [17]. This is because the radius of a Au atom (0.134 nm) is too large (larger than that of In 3+ (0.08 nm)) to incorporate into the In 2 O 3 lattice, while the radius of Pd 2+ (0.085 nm) is only slightly larger than that of In 3+ , and Pd 2+ can replace the position of In ions, which makes the interplanar spacing larger. The grain sizes are calculated according to Scherrer's formula (D = 0.89 λ/β cosθ) ( Table 1). It can be seen that Au modification can only slightly decrease the crystal size, while Pd 2+ doping can significantly decrease it, indicating that Pd 2+ might enter the In 2 O 3 crystal and effectively inhibit the growth of In 2 O 3 grain size. The reduced grain size is beneficial to the gas-sensing performance [16].   (Figure 3a,b). The surface will become smoother with an increase in comodifying amounts of Au and PdO (Figure 3i-l), which is due to a decrease in the grain size with the increase in the modifying amount. However, the nanofiber structure can be well-preserved after noble metal modification, and the target gas can easily transport in and out of the sensing body, which can ensure the high utility ratio of the gas sensors.   (Figure 3a,b). The surface will become smoother with an increase in co-modifying amounts of Au and PdO (Figure 3i-l), which is due to a decrease in the grain size with the increase in the modifying amount. However, the nanofiber structure can be well-preserved after noble metal modification, and the target gas can easily transport in and out of the sensing body, which can ensure the high utility ratio of the gas sensors.   (Figure 3a,b). The surface will become smoother with an increase in comodifying amounts of Au and PdO (Figure 3i-l), which is due to a decrease in the grain size with the increase in the modifying amount. However, the nanofiber structure can be well-preserved after noble metal modification, and the target gas can easily transport in and out of the sensing body, which can ensure the high utility ratio of the gas sensors.   XPS is carried out to analyze the surface chemical compositions as shown in Figure  5. The In 3d5/2 and In 3d3/2 peaks located at 444.13 eV and 451.68 eV for the pure In2O3 nanofiber shift to higher binding energies of 444.34 eV and 451.89 eV after noble metal modification as shown in Figure 5b, while the Au 4f7/2 and Au 4f5/2 peaks of the Au2Pd2-In2O3 sample located at 83.18 eV and 86.88 eV become lower compared with those of metallic Au located at 84.0 eV and 87.6 eV caused by the electron transfer from In2O3 to Au [18,19]. Figure 5d,e are the high-resolution Pd 3d spectra of the Pd2-In2O3 and Au2Pd2-In2O3 samples, indicating the successful Pd modification in In2O3 crystal [20]. Pd 3d3/2 can be divided into 341.10 eV (Pd 0 ) and 342.40 eV (Pd 2+ ), and Pd 3d5/2 can be decomposed into 335.08 eV (Pd 0 ) and 336.90 eV (Pd 2+ ). It is obvious that Pd species mainly existed in the form of Pd 2+ in the Pd2-In2O3 sample (96.3%) and the Au2Pd2-In2O3 sample (85.6%). Many XPS is carried out to analyze the surface chemical compositions as shown in Figure 5. The In 3d 5/2 and In 3d 3/2 peaks located at 444.13 eV and 451.68 eV for the pure In 2 O 3 nanofiber shift to higher binding energies of 444.34 eV and 451.89 eV after noble metal modification as shown in Figure 5b, while the Au 4f 7/2 and Au 4f 5/2 peaks of the Au2Pd2-In 2 O 3 sample located at 83.18 eV and 86.88 eV become lower compared with those of metallic Au located at 84.0 eV and 87.6 eV caused by the electron transfer from In 2 O 3 to Au [18,19]. Figure 5d,e are the high-resolution Pd 3d spectra of the Pd2-In 2 O 3 and Au2Pd2-In 2 O 3 samples, indicating the successful Pd modification in In 2 O 3 crystal [20]. Pd 3d 3/2 can be divided into 341.10 eV (Pd 0 ) and 342.40 eV (Pd 2+ ), and Pd 3d 5/2 can be decomposed into 335.08 eV (Pd 0 ) and 336.90 eV (Pd 2+ ). It is obvious that Pd species mainly existed in the form of Pd 2+ in the Pd2-In 2 O 3 sample (96.3%) and the Au2Pd2-In 2 O 3 sample (85.6%). Many works have reported the co-existence of Pd species [21,22], which could lead to a significant increase in gas response. Figure 5f-i are high-resolution O 1s spectra resolved into three peaks, including lattice oxygen (O L ) at 529.50 ± 0.6 eV, defective oxygen (O D ) at 530.55 ± 0.4 eV, and adsorbed oxygen (O C ) at 531.65 ± 0.3 eV [23]. It is clear that the relative percentages of O D and O C are significantly increased after noble metal modification, and the Au2Pd2-In 2 O 3 sample possesses the highest relative percentages of O D and O C . Therefore, the highest response of the Au2Pd2-In 2 O 3 sample can be expected [24]. The percentages of the three different O 1s compositions in the four samples are listed in Table 2. works have reported the co-existence of Pd species [21,22], which could lead to a significant increase in gas response. Figure 5f-i are high-resolution O 1s spectra resolved into three peaks, including lattice oxygen (OL) at 529.50 ± 0.6 eV, defective oxygen (OD) at 530.55 ± 0.4 eV, and adsorbed oxygen (OC) at 531.65 ± 0.3 eV [23]. It is clear that the relative percentages of OD and OC are significantly increased after noble metal modification, and the Au2Pd2-In2O3 sample possesses the highest relative percentages of OD and OC. Therefore, the highest response of the Au2Pd2-In2O3 sample can be expected [24]. The percentages of the three different O 1s compositions in the four samples are listed in Table 2.

Sensing Performance
The responses to 100 ppm CO of the gas sensor based on as-obtained samples are shown in Figure 6a. It is clear that both Au modification (Au2-In2O3) and PdO modification (Pd2-In2O3) can increase the gas response to 100 ppm CO compared with pure In2O3, and the sensitization effect of Au is more prominent. However, Au modification does not change the OWT (300 °C), while PdO does lower the OWT to 150 °C [10]. Co-modification

Sensing Performance
The responses to 100 ppm CO of the gas sensor based on as-obtained samples are shown in Figure 6a. It is clear that both Au modification (Au2-In 2 O 3 ) and PdO modification (Pd2-In 2 O 3 ) can increase the gas response to 100 ppm CO compared with pure In 2 O 3 , and the sensitization effect of Au is more prominent. However, Au modification does not change the OWT (300 • C), while PdO does lower the OWT to 150 • C [10]. Co-modification of Au and PdO into In 2 O 3 nanofibers can further increase the gas response, and Au2Pd2-In 2 O 3 possesses the highest response (21.7) among these sensors. Higher (Au4Pd4-In 2 O 3 , 11.2) and lower (Au1Pd1-In 2 O 3 , 11.6) co-modifying contents will decrease the gas response. The OWT of the gas sensors after co-modification of Au and PdO is 180 • C, located between 150 • C (Pd2-In 2 O 3 ) and 300 • C (Au2-In 2 O 3 ), indicating the synergistic effect of Au and PdO. The baseline resistance of the gas sensors based on the as-obtained samples will decrease with an increase in the working temperature (Figure 6b), which is a typical property of semiconductors. of Au and PdO into In2O3 nanofibers can further increase the gas response, and Au2Pd2-In2O3 possesses the highest response (21.7) among these sensors. Higher (Au4Pd4-In2O3, 11.2) and lower (Au1Pd1-In2O3, 11.6) co-modifying contents will decrease the gas response. The OWT of the gas sensors after co-modification of Au and PdO is 180 °C, located between 150 °C (Pd2-In2O3) and 300 °C (Au2-In2O3), indicating the synergistic effect of Au and PdO. The baseline resistance of the gas sensors based on the as-obtained samples will decrease with an increase in the working temperature (Figure 6b), which is a typical property of semiconductors. Figure 7a-c show the dynamic response curves of pure In2O3 and Au2Pd2-In2O3 sensors at their optimal operating temperatures. The responses will increase with an increase in CO gas concentration for both sensors, and the resistances can recover their original value after exposure to different CO concentrations. The response values of the gas sensor based on pure In2O3 nanofibers to 6, 10, 20, 50, 100, 300, and 500 ppm CO gas at 300 °C are 1.19, 1.37, 1.56, 2.05, 2.30, 3.86, and 4.55, respectively. The response values based on the Au2Pd2-In2O3 sample to 0.2, 0.5, 1, 3, 6, 10, 20, 50, 100, 300, and 500 ppm CO gas at 180 °C are 1.23, 1.37, 1.40, 1.83, 2.42, 3.30, 5.44, 11.7, 21.7, 54.0, and 65.2, respectively. It is clear that the Au2Pd2-In2O3 sensor exhibits a higher gas response to the different concentrations of CO than the pure In2O3 sensor. In addition, the Au2Pd2-In2O3 sensor exhibits an extremely lower detection limit (0.2 ppm) than the pure In2O3 sensor (6 ppm), indicating that noble metal modification can not only enhance the gas response, but can also decrease the detection limit. Figures 7b and 6d demonstrate that the gas responses share an almost linear relationship with CO concentration, which is beneficial for accurate gas concentration detection.  It is clear that the Au2Pd2-In 2 O 3 sensor exhibits a higher gas response to the different concentrations of CO than the pure In 2 O 3 sensor. In addition, the Au2Pd2-In 2 O 3 sensor exhibits an extremely lower detection limit (0.2 ppm) than the pure In 2 O 3 sensor (6 ppm), indicating that noble metal modification can not only enhance the gas response, but can also decrease the detection limit. Figures 7b and 6d demonstrate that the gas responses share an almost linear relationship with CO concentration, which is beneficial for accurate gas concentration detection.
The selectivity of the gas sensors is shown in Figure 8. The ratio of the response values for CO and H 2 is taken to evaluate the selectivity. It is clear that Au modification is more effective for increasing the CO gas response than PdO modification, but the selectivity of the Au-modified sensor (~1.3) is smaller than the PdO-modified sensor (~1.6). In addition, the response to benzene and toluene has also been significantly increased. Co-modification using Au and PdO for the Au1Pd1-In 2 O 3 sensor can further increase the gas response to CO and H 2 with higher selectivity (2.4), and suppress the response to benzene and toluene at the same time. The response for CO and H 2 is further enhanced with an increase in Au and PdO modifying content (Au2Pd2-In 2 O 3 ), and the selectivity is changed slightly (~2.2). At the same time, the response to benzene and toluene has been further suppressed. The response becomes worse and the selectivity becomes lower (1.5) when the co-modifying content of Au and PdO is too high (Au4Pd4-In 2 O 3 ). Therefore, the Au2Pd2-In 2 O 3 sensor is more suitable for CO detection in this work. The selectivity of the gas sensors is shown in Figure 8. The ratio of the response values for CO and H2 is taken to evaluate the selectivity. It is clear that Au modification is more effective for increasing the CO gas response than PdO modification, but the selectivity of the Au-modified sensor (~1.3) is smaller than the PdO-modified sensor (~1.6). In addition, the response to benzene and toluene has also been significantly increased. Comodification using Au and PdO for the Au1Pd1-In2O3 sensor can further increase the gas response to CO and H2 with higher selectivity (2.4), and suppress the response to benzene and toluene at the same time. The response for CO and H2 is further enhanced with an increase in Au and PdO modifying content (Au2Pd2-In2O3), and the selectivity is changed slightly (~2.2). At the same time, the response to benzene and toluene has been further suppressed. The response becomes worse and the selectivity becomes lower (1.5) when the co-modifying content of Au and PdO is too high (Au4Pd4-In2O3). Therefore, the Au2Pd2-In2O3 sensor is more suitable for CO detection in this work. The response/recovery times for pure In2O3 (38 s/220 s) and Au2Pd2-In2O3 (5 s/106 s) at 180 °C can be observed in Figure 9a,b, indicating that noble metal modification can accelerate the response/recovery speed. Such an improvement in the response/recovery speed can be attributed to the catalytic effect of noble metal modification, which has been verified in many works [25,26]. The six reversible cycling dynamic response curves of Au2Pd2-In2O3 100 ppm CO demonstrate that the Au2Pd2-In2O3 sensor has good reproducibility and repeatability (Figure 9c). The stability test of Au2Pd2-In2O3 for 19 days indicates that the baseline resistance in air and response to 100 ppm CO are almost un- The response/recovery times for pure In 2 O 3 (38 s/220 s) and Au2Pd2-In 2 O 3 (5 s/106 s) at 180 • C can be observed in Figure 9a,b, indicating that noble metal modification can accelerate the response/recovery speed. Such an improvement in the response/recovery speed can be attributed to the catalytic effect of noble metal modification, which has been verified in many works [25,26]. The six reversible cycling dynamic response curves of Au2Pd2-In 2 O 3 100 ppm CO demonstrate that the Au2Pd2-In 2 O 3 sensor has good reproducibility and repeatability (Figure 9c). The stability test of Au2Pd2-In 2 O 3 for 19 days indicates that the baseline resistance in air and response to 100 ppm CO are almost unchanged (Figure 9d), which is of great significance in practical applications. Overall, the theAu2Pd2-In 2 O 3 sensor has favorable sensing characteristics in its high gas response and low detection limit for CO detection compared with the existing reports as shown in Table 3.  Humidity is also an important factor affecting the sensing performance of oxide sem iconductor sensors [30]. The dynamic response curves of the Au2Pd2-In2O3 sensor at 30% 50%, 70%, and 90% RH were measured under 25 °C (Figure 10a), as well as the baselin resistance and response of the sensor versus humidity change (Figure 10b). When humid ity was increased from 35% to 98% RH, baseline resistance of the sensor decreased by 42%  Humidity is also an important factor affecting the sensing performance of oxide semiconductor sensors [30]. The dynamic response curves of the Au2Pd2-In 2 O 3 sensor at 30%, 50%, 70%, and 90% RH were measured under 25 • C (Figure 10a), as well as the baseline resistance and response of the sensor versus humidity change (Figure 10b). When humidity was increased from 35% to 98% RH, baseline resistance of the sensor decreased by 42% and the response value decreased by 28%. This reduction in baseline resistance is due to the reaction of water molecules with oxygen on the surface of the material. Secondly, water molecules can compete with gas molecules and oxygen for adsorption sites on the surface of the sensing material, leading to a decrease in the response value [31]. Nevertheless, response of the Au2Pd2-In 2 O 3 sensor to 100 ppm CO is still up to 15.2 under 98% RH, indicating excellent moisture resistance.

Sensing Mechanism
The gas-sensing mechanism can be explained by the space-charge layer, oxygen ad sorption, and surface reaction between the reducing gases and adsorbed oxygen [32,33 The sensing enhancement of the gas response can be mainly attributed to the electroni sensitization and chemical sensitization of the noble metals [24,34,35], as well as the syn ergistic effect of Au and PdO.
The electronic sensitization effect can be seen from the baseline resistance increas after noble metal modification, as shown in Figure 6b. The work functions of In2O3, Au and PdO are about 4.3 eV [36], 5.1 eV [37], and 7.9 eV [38], respectively. The Schottk contact between Au and In2O3 and p-n heterojunction between PdO and In2O3 will forme after co-modification using Au and PdO [39][40][41]. The electrons will flow from In2O3 to A and PdO and lead to the formation of the thicker space-charge layer due to the differen Fermi levels of these materials as shown in Figure 11. Therefore, the baseline resistanc will increase. In addition, Au modification seems to have better effect than PdO modif cation on the increase in the baseline resistance according to Figure 6b. This may be du to the fact that Au atoms mainly exist on the surface of In2O3 nanofibers and many Au In2O3 contacts can be formed, while the majority of Pd ions enter into the In2O3 crystal an the formations of the p-n heterojunctions are fewer. However, the increase in baselin resistance after modification with Au and PdO is beneficial in improving the gas respons [42].

Sensing Mechanism
The gas-sensing mechanism can be explained by the space-charge layer, oxygen adsorption, and surface reaction between the reducing gases and adsorbed oxygen [32,33]. The sensing enhancement of the gas response can be mainly attributed to the electronic sensitization and chemical sensitization of the noble metals [24,34,35], as well as the synergistic effect of Au and PdO.
The electronic sensitization effect can be seen from the baseline resistance increase after noble metal modification, as shown in Figure 6b. The work functions of In 2 O 3 , Au, and PdO are about 4.3 eV [36], 5.1 eV [37], and 7.9 eV [38], respectively. The Schottky contact between Au and In 2 O 3 and p-n heterojunction between PdO and In 2 O 3 will formed after co-modification using Au and PdO [39][40][41]. The electrons will flow from In 2 O 3 to Au and PdO and lead to the formation of the thicker space-charge layer due to the different Fermi levels of these materials as shown in Figure 11. Therefore, the baseline resistance will increase. In addition, Au modification seems to have better effect than PdO modification on the increase in the baseline resistance according to Figure 6b. This may be due to the fact that Au atoms mainly exist on the surface of In 2 O 3 nanofibers and many Au-In 2 O 3 contacts can be formed, while the majority of Pd ions enter into the In 2 O 3 crystal and the formations of the p-n heterojunctions are fewer. However, the increase in baseline resistance after modification with Au and PdO is beneficial in improving the gas response [42].
The chemical sensitization can be verified by the XPS results as shown in Table 2. It is clear that relative chemisorbed oxygen (O C ) percentages will significantly increase after noble metal modification, and Au2Pd2-In 2 O 3 has the highest O C content. Therefore, it is reasonable that the Au2Pd2-In 2 O 3 sensor has the highest gas response. The OWT changes from 300 • C to 180 • C after co-modification using Au and PdO. The lower OWT also indicates the chemical sensitization effect, which leads to a higher gas response. The chemical sensitization can be verified by the XPS results as shown in Table 2. It is clear that relative chemisorbed oxygen (OC) percentages will significantly increase after noble metal modification, and Au2Pd2-In2O3 has the highest OC content. Therefore, it is reasonable that the Au2Pd2-In2O3 sensor has the highest gas response. The OWT changes from 300 °C to 180 °C after co-modification using Au and PdO. The lower OWT also indicates the chemical sensitization effect, which leads to a higher gas response.
The synergistic effect of Au and PdO can be seen from the OWT after co-modification using Au and PdO. Solo Au modification will not change the OWT obviously (300 °C for Au2-In2O3), while solo PdO modification (Pd2-In2O3) will lower the OWT from 300 °C to 150 °C (Figure 6a). This indicates that the catalytic effect of PdO is much better than that of Au. The OWT after co-modification using Au and PdO is 180 °C . The intermediate OWT between 300 °C and 150 °C indicates the synergistic effect of Au and PdO. Furthermore, The synergistic effect of Au and PdO can be seen from the OWT after co-modification using Au and PdO. Solo Au modification will not change the OWT obviously (300 • C for Au2-In 2 O 3 ), while solo PdO modification (Pd2-In 2 O 3 ) will lower the OWT from 300 • C to 150 • C (Figure 6a). This indicates that the catalytic effect of PdO is much better than that of Au. The OWT after co-modification using Au and PdO is 180 • C. The intermediate OWT between 300 • C and 150 • C indicates the synergistic effect of Au and PdO. Furthermore, we speculate that Au might have higher adsorption ability than PdO to CO gas [43], which might be verified by the higher Oc content (35.2%) of Au2-In 2 O 3 than that (33.6%) of Pd2-In 2 O 3 as shown in Table 2, indicating that Au loading might adsorb more CO gas than Pd modification. However, a high OWT (300 • C) is necessary for solo Au modification to ensure the full reaction between CO gas and adsorbed oxygen due to the low catalytic effect of Au. After PdO modification, the gas response can be significantly increased though fully utilizing the adsorption ability of Au and catalytic effect of PdO with relatively lower OWT. Therefore, both the selectivity and gas response to CO gas can be greatly imporved. In addition, the synergistic effect can be also illustrated by the peak shift of Au 4f 7/2 and Au 4f 5/2 from 83.43 eV and 87.08 eV (Au2-In 2 O 3 , not shown here) to 83.18 eV and 86.88 eV (Au2Pd2-In 2 O 3 ) as shown in Figure 5d, indicating greater electron transportation to the Au cluster after PdO doping. It is obvious that co-modification using suitable amounts of Au and PdO (Au2Pd2-In 2 O 3 ) optimizes the electronic sensitization and chemical sensitization, reduces the activation energy of the reaction, greatly improves the detection limit of CO, and enhances the sensitivity of the material to CO gas.

Conclusions
In this work, the co-modifying effects of Au and PdO on In 2 O 3 nanofibers are systematically investigated for the enhancement of CO detection compared with pure In 2 O 3 and individual Au and PdO doping. The experimental results indicate that a suitable modifying amount of Au and PdO (Au2Pd2-In 2 O 3 ) can not only increase the gas response, but can also lower the detection limit down to 200 ppb with a response value of 1.23 and OWT of 180 • C. The enhancement of the gas response to CO can be attributed to the electronic sensitization due to the formation of p-n heterojunctions between PdO and In 2 O 3 , Schottky contact between Au and In 2 O 3 , the chemical sensitization of PdO verified by lower OWT and higher Oc contents, and the synergistic effect of Au and PdO caused by the high adsorption ability of Au and excellent catalytic effect of PdO. The response to 100 ppm CO for the Au2Pd2-In 2 O 3 sensor is about 8.5 times higher than that for the pure In 2 O 3 sensor. The synergistic effect of Au and PdO also improves the selectivity to CO in different gases, especially the selectivity of CO against H 2 . Co-modification using Au and PdO might be a promising strategy for the enhancement of gas-sensing performance for CO detection.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.