The Influence of 2-Methoxyethanol as Capping Agent on WO3-Based Carbon Monoxide Gas Sensor Characteristics

Carbon monoxide (CO) gas detection using a modified WO3-based sensor is being developed. The solvent for solvothermal use, a combination of 2-Propanol and 2-Methoxyethanol, was employed as a capping agent before graphene was used as a component of nanocomposites. Following the creation of the powder, it is combined with ethyl glycol and applied to an alumina substrate using the Doctor Blade process. By X-ray diffraction research, it was discovered that the solvent combinations of 2-Propanol and 2-Methoxyethanol formed monoclinic WO3 in the amounts of 40-0 and 30-10, respectively, while the solvent combinations of 20-20 are thought to have produced W18O49 and 10-30, WO2.9, respectively. According to the SEM examination, the WO3 generated was first present as nanowires and nanorods before being calcined at 500 degrees Celsius, and it then appeared as nanoparticles. The sensor may work at a low temperature of 150°C, and the best sensitivity is found when the sensor is used at a temperature of 250°C, according to the CO gas test findings.


Introduction
Carbon Monoxide is an odorless, tasteless, and colorless toxic gas for human health [1,2].It was estimated that in the early 1980s, over one million people a year were treated in hospitals because of CO poisoning and 3,000 people died annually [3].It might even cause premature death [4].Interestingly, CO was being researched in low concentrations as healing and therapeutic material [5,6].CO is even one of the gases that should be able to be detected in recent Coronavirus electronic nose research [7,8].Along with Stannic oxide, Tungsten oxide is already used for commercial gas sensor devices [9,10].WO3 has been used as material for the gas sensor to detect various gases such as H2S [11][12][13], NH3 [14], alcohol [15], acetone [16-18], NO2 [19][20][21][22][23][24], SO2 [25,26], H2 [27], and Carbon Monoxide [22,[28][29][30].Tungsten oxide has excellent physical and chemical characteristics, including a high oxygen vacancy level and a large effective surface area [13].However, a very low response phenomenon of WO3-based gas sensor to carbon monoxide compared to other gases is observed [31][32][33].Researching a novel WO3 that uses a capping agent instead of a noble metal during synthesis is crucial to providing a superior reaction to carbon monoxide.This study aims to investigate this nanocomposite's potential for catalytic and carbon monoxide sensing applications.Furthermore, to achieve an excellent response to carbon monoxide, a capping agent devoid of a noble metal must be added during the nanocomposite's synthesis.Because of their catalytic qualities, noble metals like gold and platinum are frequently employed as capping agents in the production of nanocomposite materials [34].Nevertheless, using noble metals can be costly and raise environmental issues.It is, therefore, quite interesting to investigate capping agents that are not made of noble metals.There are several benefits to using a capping agent that doesn't contain a noble metal.First off, it can improve the nanocomposite's stability and dispersibility, resulting in a uniform distribution of WO3 nanoparticles on the graphene surface [35].Better catalytic activity and gas-sensing capabilities may result from this.Second, it can lessen the synthesis process's expense and negative environmental effects [36].The total cost of production can be lowered by utilizing an environmentally acceptable and widely accessible capping agent, which increases the nanocomposite's viability for large-scale applications.

Synthesis
The synthesis was carried out with WCl6 dissolved in 2-Propanol (C3H8O) and 2-Methoxyethanol (C3H8O2) solvents.The chemicals used were purchased from Sigma-Aldrich, at analytical grade (99.9% purity).Then, 0.5 mmol of WCl6 and 2-Propanol were dissolved with constant magnetic stirring at room temperature, until the yellow solution turned dark purplish blue.After that, 2-Methoxyethanol with different volumes of 0, 10, 20, and 30 ml (noted as 40-0, 30-10, 20-20, and 10-30), respectively was added to the above solution with the same total volume of 40 ml and then stirred again for 10 minutes.Then, the solution was put into a Teflon 100 ml stainless steel autoclave and solvothermal at 180 o C for 15 hours.After that, the resulting dark blue-grey precipitate was washed with Ethanol three times in a centrifuge and dried in an oven for 16 hours at 60 o C. The synthesized powder was crushed in a mortar until smooth, then calcined in a furnace with an increase rate of 1 o C/minute to a temperature of 500 o C and kept constant for 2 hours.The resulting powder was then deposited onto an Alumina electrode substrate using the doctor blade technique.Pure WO3 can be observed in Figure 1 (a), where monoclinic WO3 can be detected.The associated peaks are 23.10 o , 23.60 o , and 24.33 o , which correspond to the orientations (0 0 2), (0 2 0), and (2 0 0), respectively [37].Similar peak patterns were also obtained in other studies [38], where a solvothermal Intensity (a.u.)

Characterization 3.1. XRD Analysis
synthesis method that was almost identical was used, with the following differences: the precursor used in that study was not derived from WCl6 but from Na2WO4.H2O, and the solvent used was not 2propanol but distilled water.Another difference was that the temperature used in the hydrothermal/solvothermal process was 180 o C, which was 20 o C lower than that used in this study.
In Figure 1 (b), the composition of WO3 with a solvent ratio of 2-propanol to 2-Methoxyethanol of 30:10 by volume shows a similar peak pattern, where the three highest peaks, 23.04 o , 23.50 o , and 24.36 o , match the orientations of WO3 (0 0 2), (0 2 0), and (2 0 0), respectively, which are almost the same as those in Figure 1 (a).This phenomenon indicates that slight changes in the solvent ratio do not significantly affect the crystallization of WO3 during the solvothermal process, although differences in morphology determined by SEM will be observed later.
Another interesting observation is that if 2-Methoxyethanol becomes dominant in the solvent ratio of 10:30 between 2-propanol and 2-methoxyethanol, Figure 1 (d) shows XRD results with peaks at 23.08 o , 23.13 o , and 23.96 o .It is suspected that sub-stoichiometric monoclinic WO2.9 is also formed, which is also sub-stoichiometric.The crystallite size is calculated by Scherrer equation [28,39,40]: where Dhkl is the crystallite size in the direction perpendicular to the lattice planes, hkl are the Miller indices of the planes being analyzed, K is the crystallite-shape factor and here 0.94 is used, λ is the wavelength of the X-rays, Bhkl is the full-width at half-maximum (FWHM) of the X-ray diffraction peak in radians and ߠ is the Bragg diffraction angle.  1 shows that based on the Crystallite Size from the smallest to the largest, the order can be made: 10-30, 20-20, 30-10, and 40-0.The effect of the capping agent is apparent in that the more proportion of capping agent added, the smaller the crystallite size.

SEM Analysis
Without a capping agent, the morphology of WO3 produced with 2-Propanol solvent in Solvothermal at 200°C for 16 hours resulted in flower-like clustered nanoparticles (Figure 2).The effect of calcination at 500°C transformed the shape of nanorods and nanowires (especially in samples synthesized with a 20:20 and 10:30 solvent ratio) into nanoparticles with a size of approximately 100 nm.

EDS Analysis
The EDS analysis results obtained the proportion of Oxygen (O) and Tungsten (W) in various solvothermal solvent compositions.Table 2 shows that the more the proportion of EGME solvent composition, the less the Oxygen (O) content in the sample.The sensor response measurement can be defined in two ways.The first is direct comparison [45][46][47]: where Rg is the sensor resistance when exposed to gas (in this case, CO) and Ra is the sensor resistance in the air environment without the test gas.Thus, Response S is the ratio between resistance in the air and when exposed to CO gas.It means that when the value is more than 1, Rg < Ra or the sensor resistance becomes smaller.At 150°C only one solvothermal solvent combination reacted to the presence of Carbon Monoxide gas, namely the WO3 sensor with the 20-20 combination.For other combinations, 40-0, 30-10, and 10-30, the resistance measured at 150oC is too large, so it cannot be measured with our existing resistance meter with a maximum limit of 100 MΩ.
For CO gas concentrations of 20 ppm and 40 ppm, almost the same response/sensitivity is obtained, which is about 1.37.However, when the CO gas concentration was increased to 60 ppm, the sensitivity skyrocketed to 9.57.This is a surprising result because it occurs at relatively low temperatures for metal oxide semiconductors and is still WO3 without valuable metal groups such as Pt, Pd, or Ag, and also without elements such as graphene.
The sensor response at 200 o C can only be observed at a solvent ratio of 20:20.For other solvent comparisons such as 40-0, 30-10, and 10-30, no resistance change can be observed that can be categorized as a response to the presence of CO gas concentration, either at 20 ppm, 40 ppm, or 60 ppm.
In the solvothermal solvent ratio of 20:20, a CO gas concentration of 20 ppm produces a response of 1.32, then when the CO gas concentration is increased to 40 ppm, the response is 3.22, which follows the theory stating that the number of electrons released to the surface or bulk of the sensor is proportional to the CO concentration present on the sensor surface (Figure 3).Then it turns out that when the CO gas concentration is increased to 60 ppm, the sensor response jumps dramatically to 49.09.This is surprising because no research has produced similar results at 200 o C without functionalization with either precious metals or graphene.Suppose the response of WO3 sensors at 300 o C with various capping agent solvent compositions at various concentrations is compared at various CO gas concentrations.In that case, it can be seen that the response increases with increasing CO gas concentration.We obtained the highest results for the sensor whose material was synthesized with 10 ml 2-Propanol and 30 ml 2-Methoxyethanol (10-30) solvent.At concentrations of 20 ppm, 40 ppm, and 60 ppm of CO gas, the responses were 6.26, 7.08, and 8.35, respectively.However, at 60 ppm CO gas concentration, the highest response was obtained by sensors with a combination of 20-20 with 13.48 and 30-10 with 13.33.Then for sensors with 30-10, the response when given 20 ppm, 40 ppm, and 60 ppm respectively are 3.69, 4.52, and 13.33.The 0 2000 4000 6000 8000 10000 0,0 results for sensors with a solvothermal solvent of only 2-Propanol were 1.16, 1.19, and 1.27 respectively.Thus always the lowest response is obtained from the sensor synthesized without capping agent 2-Methoxyethanol.

Selectivity
Selectivity was measured by comparing the sensor response to various.It was found that the sensor is more sensitive to CO when compared to several gases, e.g.: ethanol, methanol, acetone, and NO.The sensitivity to CO is 1.82 which is relatively bigger than other gases: ethanol (1.23), methanol (1.58), Acetone (1.55), and NO (1.47).

Sensing Mechanism
The surface reaction of oxygen on the surface of metal oxide semiconductors occurs according to the surface temperature [41].The following two reactions occur below 147 o C: O2 (gas) → O2 ( a ds ) ( 1) O2 (ads) + e -→ O2 - If the temperature continues to increase (between 147 o C and 397 o C), the following reactions will occur on the surface: While the following reaction is expected to occur at temperatures above 397 o C: Thus the key is that it is very important to make nanocomposites that can guarantee the stability of these oxygen surface reactions at temperatures below 147 o C. Inside the bulk of metal oxide semiconductor-based sensors, there are grains.An energy barrier exists between the metal oxide semiconductor grains, which blocks the flow of electrons.For this, activation energy is required.Figure 4 shows the sensing mechanism at the junction between grains, at the surface of the grain and free air, and at the junction between the grain and the metal used as an electrode [48][49][50].Table 4 shows that this research has produced a breakthrough by using a capping agent in the solvent used for solvothermal.Especially in the solvent combination of 2-Propanol and 2-Methoxyethanol 20:20, the sensor can respond at 150 o C, and this does not include valuable metals such as Pt, Pd, or Ag and composite components such as graphene.In the combination of 2-Propanol and 2-Methoxyethanol 10:30 solvent, a fairly high response of 79.9 at 250 o C was obtained.Some research shows that WO3 has a low response to CO but the humidity factor plays an important role [28,32].

Conclusions
From the 4 solvent variations used (only 40 ml of 2-Propanol solvent, 40 ml of 2-Propanol solvent plus 8 ml of 2-Methoxyethanol, and 40 ml of 2-Propanol solvent plus 10 ml of 2-Methoxyethanol, the following results were observed: -Adding 20 ml to 30 ml of 2-Methoxyethanol in solvothermal solvent gives a clustered morphology of WO3 nanoparticles with a size of about 100 nm but is more "sticky" than if without a capping agent.
-The optimum sensitivity of the sensor with a solvent ratio of 20-20 at 200 o C, 10-30 at 250 o C, and 30-10 at 300 o C and 350 o C was obtained.
According to many studies, these two results must be followed up with graphene, which can further lower the operating temperature and improve response and selectivity.

Figure 3 .
Figure 3. Sensor Responses to CO in various Operating temperature (a) 150 o C (b) 200 o C (c) 250 o C (d) 300 o C and (e) 350 o C

Figure 4 .
Figure 4. Grain Model and Sensing Mechanism

Table 1
shows that the average crystallite sizes are smaller with more 2-Methoxyethanol added in 40ml 2-Propanol solvent.Table

Table 2 .
Atom % Proportions in WO3 Samples with Various Solvent Compositions

. Sensor Characterization 4.1. Gas Response Characterization Carbon
[41,43,44]CO) gas is a strong reducing gas[41,42], and therefore CO gas can be used to validate the Semiconductor Metal Oxide (SMO) sensing model.For each WO3 and Graphene composition, a wide range of CO gas concentrations were given, namely 10 ppm, 60 ppm, 100 ppm, and 300 ppm.The operating temperatures used were 150 o C, 200 o C,250 o C, 300 o C, and 350 o C. The matrix of WO3 and operating temperature and the presence or absence of response can be seen in Table3with exposure to 60 ppm CO gas.It can be seen that the sensor with WO3 does not produce a response at low temperatures except for the sensor with a solution ratio of 20:20, while the sensor that produces the highest response (79.9) is the sensor with a solvent ratio of 10:30 at 250 o C.This phenomenon is understandable because Metal Oxide Semiconductors need thermal energy at temperatures of 200 o C-350 o C to move electrons from the valence band to the conduction band[41,43,44].

Table 4 .
Comparison with previous research