Flexible Graphite/PPG Hybrid Composite-Based Resistive Sensor for Sensing Organic Compounds

Our objective in this study was to investigate a sensor for volatile organic compounds based on a graphite (G)/polypropylene glycol (PPG) hybrid composite (HC) for sensing hybrid elements. The G/PPG HC sensor films for organic-matter detection were successfully fabricated on polyethylene terephthalate (PET) film with a simple blade-coating method. The sensing paste based on G/PPG (1:2) HC showed good dispersibility and stability. In addition, G/PPG HC sensor films with organic compounds showed different thickness changes as a function of the G/PPG ratio because of the swelling effect of the polymer. The observed differences in resistance of the G/PPG HC films corresponded to those of common organic compounds, suggesting that the disconnection of graphite caused by the swollen PPG matrix caused explosive resistance change. Moreover, we evaluated the sensitivity of typical hydrocarbon materials, such as benzene and toluene, in the sensor film as well as petroleum materials without moisture-induced malfunctions. This study could provoke knowledge about superior sensing with cost-effective and easily scalable materials using polymer/graphite composite-based sensors to improve the sensitivity, selectivity, and stability of chemical sensor applications.


Introduction
With the recent progress of industries, efforts have been devoted to the resolution of serious medical problems, environmental pollution, and explosion dangers caused by toxic chemical leakage [1,2]. Organic compounds used in a wide array of industrial applications are toxic chemicals with extensive social and environmental significance [3]. They are used either directly or indirectly in many industrial processes. Exposure to them may occur in a range of industrial settings. Because of the deleterious effects of organic compounds used in explosives on human health and the environment, diligent monitoring of organic-compound leakage has become extremely important [4,5]. The development of leak sensors The non-woven cover can improve reactivity and speed by increasing the G/PPG HC film active area when dropping in small amounts of solvent.
Sensors 2020, 20, x FOR PEER REVIEW 3 of 10 can improve reactivity and speed by increasing the G/PPG HC film active area when dropping in small amounts of solvent.

Scheme 1.
Manufacturing process of sensitive flexible organic compounds sensor using G/PPG hybrid composite films.

Sensor Characterization
We carried out Raman spectroscopy (Confotec MR520, SOL Instruments Ltd., Minsk, Republic of Belarus), Fourier-transform infrared (FTIR) (M4000, MIDAC Corporation, California, United States), and X-ray diffraction (XRD-6100, SHIMADZU, Kyoto, Japan) analyses. Since the sensor was a paste type, the paste should have been well dispersed. Therefore, we measured the zeta potential (ELSZ-2000, Otsuka, Tokyo, Japan) to measure the dispersion in the paste. We used a contact angle (SEO phoenix-150, Kromtek Sdn Bhd, Kota Kemuning, Malaysia) to measure the response of organic compounds to G/PPG. When evaluating, we measured the film thickness using a Varnier caliber (ABSOLUTE Digmatic Micrometer, Mitutoyo, Sakado, Japan). For the above measuring methods, the sensor was used without the non-woven fabric cover. Finally, we measured the electrical signals (with the non-woven fabric cover) of resistance changes by multi-meter (789 Process Meter, Fluke, WA, United States).

Sensor Measurement
We need to explain how to measure the resistance of the fabricated G/PPG sensor. The Cu wire was connected to the end of the sensor (with the non-woven fabric cover) to connect to a multi-meter (Fluke 789 Process Meter). Both ends of the sensor were connected to a Cu wire and multi-meter. Then, 1 mL of gasoline was dropped onto the sensor as we started to measure the time. Next, we recorded the resistance of the multi-meter every 10 s for 60 s (Video S1).

Thickness Measurement
We used a sensor without the non-woven fabric cover and measured the thickness using different methods before and after the gasoline reaction. Before the gasoline reaction, the initial thickness was measured by Vernier caliber. After the gasoline reaction, the thickness change was measured by calculating the pixel (px) change rate using contact angle measurement.

Results and Discussion
Structures of the G/PPG HC with different concentrations of graphite were characterized by ATR-FTIR spectroscopy ( Figure 1a). All composites showed different absorption peaks at 2969, 2867, 1452, 1373, and 1220 (asymmetric, symmetric bending vibrations and deformation bands of C-H), and at 1093 (C-O-C stretching vibrations in the -OCH2CH2unit) cm −1 , indicating the formation of PPG [27,28]. Scheme 1. Manufacturing process of sensitive flexible organic compounds sensor using G/PPG hybrid composite films.

Sensor Characterization
We carried out Raman spectroscopy (Confotec MR520, SOL Instruments Ltd., Minsk, Republic of Belarus), Fourier-transform infrared (FTIR) (M4000, MIDAC Corporation, California, United States), and X-ray diffraction (XRD-6100, SHIMADZU, Kyoto, Japan) analyses. Since the sensor was a paste type, the paste should have been well dispersed. Therefore, we measured the zeta potential (ELSZ-2000, Otsuka, Tokyo, Japan) to measure the dispersion in the paste. We used a contact angle (SEO phoenix-150, Kromtek Sdn Bhd, Kota Kemuning, Malaysia) to measure the response of organic compounds to G/PPG. When evaluating, we measured the film thickness using a Varnier caliber (ABSOLUTE Digmatic Micrometer, Mitutoyo, Sakado, Japan). For the above measuring methods, the sensor was used without the non-woven fabric cover. Finally, we measured the electrical signals (with the non-woven fabric cover) of resistance changes by multi-meter (789 Process Meter, Fluke, WA, United States).

Sensor Measurement
We need to explain how to measure the resistance of the fabricated G/PPG sensor. The Cu wire was connected to the end of the sensor (with the non-woven fabric cover) to connect to a multi-meter (Fluke 789 Process Meter). Both ends of the sensor were connected to a Cu wire and multi-meter. Then, 1 mL of gasoline was dropped onto the sensor as we started to measure the time. Next, we recorded the resistance of the multi-meter every 10 s for 60 s (Video S1).

Thickness Measurement
We used a sensor without the non-woven fabric cover and measured the thickness using different methods before and after the gasoline reaction. Before the gasoline reaction, the initial thickness was measured by Vernier caliber. After the gasoline reaction, the thickness change was measured by calculating the pixel (px) change rate using contact angle measurement.

Results and Discussion
Structures of the G/PPG HC with different concentrations of graphite were characterized by ATR-FTIR spectroscopy ( Figure 1a). All composites showed different absorption peaks at 2969, 2867, 1452, 1373, and 1220 (asymmetric, symmetric bending vibrations and deformation bands of C-H), and at 1093 (C-O-C stretching vibrations in the -OCH 2 CH 2 -unit) cm −1 , indicating the formation of PPG [27,28]. The G/PPG HC shows absorption bands between 3000~2800 cm −1 and 1500~1250 cm −1 originating from C-H stretching and bending vibrations, respectively (Figure 1b) [27]. The strong band at 1155 cm −1 can be associated with the C-O-C symmetric stretching vibration in C-O modes caused by ethyl cellulose [29]. In addition, the absorption peak at 1652 cm −1 is related to the stretching vibration of the C=C bond in graphite [30]. New absorption peaks were observed at 1155 and 1652 cm −1 for the G/PPG HC, suggesting that the graphite powder was successfully incorporated into the PPG. To investigate whether there was an interfacial interaction between PPG and graphite in the G/PPG HC film, we carried out Raman spectroscopic analysis of the G/PPG HC in the vibration region (800~3000 cm −1 ). Characteristic features of the Raman spectra for pristine carbon can be observed at 1580 cm −1 (G band) and 2700 cm −1 (2D band) [31]. Raman peaks near 875 and 925 cm −1 are caused by extensions to the inside of C-O-C symmetric stretching, and the one near 1455 cm −1 is caused by -COO [32,33]. The Raman spectrum of PPG exhibited strong peaks at 2883 and 2939 cm −1 , corresponding to stretching vibrations of the alkyl chains [34,35]. In addition, peaks associated with the G/PPG HC occurred at 1352, 1567, and 2678 cm −1 , reflecting D (sp3), G (sp2), and 2D bands, respectively, caused by graphite [36]. Two sharp peaks at 2880 and 2934 cm −1 corresponding to the C-H stretching vibration of the G/PPG HC were also observed. This band slightly shifted about ~3 cm −1 from its original position (at 2883 and 2939 cm −1 in pure PPG) because of the formation of hydrogen-bridge bonds between the -OH group of graphite and the hydroxyl group of PPG [37].
The zeta potential is the potential between the liquid layer adjacent to the solid phase and the liquid layer constituting the bulk liquid phase. It is a measure of the magnitude of the electrostatic repulsion or attraction between particles. We measured the zeta potentials of the PPG and graphite particles to investigate the affinity between them in the G/PPG HC paste from the point of view of colloidal chemistry (Table S1). The zeta potential measured for graphite particles in 2-ethoxyethanol was 1166 mV. However, the zeta potential values for graphite particles as a function of the G/PPG ratio (2:1, 1:1, 2:1) steadily increased with the G/PPG ratio. The zeta potential of the G/PPG (1:2) HC paste was +1237.82 mV, indicating that this paste had a more suitable dispersion than other ratios of G/PPG HC paste.
The plausible sensing mechanism of the G/PPG HC film for sensing organic compounds is schematically shown in Figure 2a. A well-aligned G/PPG HC film (25 μm) was fabricated by blade coating, indicating that the graphite was well dispersed in the PPG matrix with sufficient dispersion stability. When the sensor including a G/PPG HC was exposed to organic compounds, a radical chemical reaction occurred. The swelling of the G/PPG HC film under the influence of organic compounds led to a volume phase transition with mechanical stress and separation between the graphite. The volume phase transition and separation of each graphite in the polymer matrix play an important role in detecting the fluctuation of resistance in the G/PPG HC film. In addition, the evaporation of the PPG with organic compounds was observed; it exposed the graphite on the surface The G/PPG HC shows absorption bands between 3000~2800 cm −1 and 1500~1250 cm −1 originating from C-H stretching and bending vibrations, respectively (Figure 1b) [27]. The strong band at 1155 cm −1 can be associated with the C-O-C symmetric stretching vibration in C-O modes caused by ethyl cellulose [29]. In addition, the absorption peak at 1652 cm −1 is related to the stretching vibration of the C=C bond in graphite [30]. New absorption peaks were observed at 1155 and 1652 cm −1 for the G/PPG HC, suggesting that the graphite powder was successfully incorporated into the PPG.
To investigate whether there was an interfacial interaction between PPG and graphite in the G/PPG HC film, we carried out Raman spectroscopic analysis of the G/PPG HC in the vibration region (800~3000 cm −1 ). Characteristic features of the Raman spectra for pristine carbon can be observed at 1580 cm −1 (G band) and 2700 cm −1 (2D band) [31]. Raman peaks near 875 and 925 cm −1 are caused by extensions to the inside of C-O-C symmetric stretching, and the one near 1455 cm −1 is caused by -COO [32,33]. The Raman spectrum of PPG exhibited strong peaks at 2883 and 2939 cm −1 , corresponding to stretching vibrations of the alkyl chains [34,35]. In addition, peaks associated with the G/PPG HC occurred at 1352, 1567, and 2678 cm −1 , reflecting D (sp 3 ), G (sp 2 ), and 2D bands, respectively, caused by graphite [36]. Two sharp peaks at 2880 and 2934 cm −1 corresponding to the C-H stretching vibration of the G/PPG HC were also observed. This band slightly shifted about~3 cm −1 from its original position (at 2883 and 2939 cm −1 in pure PPG) because of the formation of hydrogen-bridge bonds between the -OH group of graphite and the hydroxyl group of PPG [37].
The zeta potential is the potential between the liquid layer adjacent to the solid phase and the liquid layer constituting the bulk liquid phase. It is a measure of the magnitude of the electrostatic repulsion or attraction between particles. We measured the zeta potentials of the PPG and graphite particles to investigate the affinity between them in the G/PPG HC paste from the point of view of colloidal chemistry (Table S1). The zeta potential measured for graphite particles in 2-ethoxyethanol was 1166 mV. However, the zeta potential values for graphite particles as a function of the G/PPG ratio (2:1, 1:1, 2:1) steadily increased with the G/PPG ratio. The zeta potential of the G/PPG (1:2) HC paste was +1237.82 mV, indicating that this paste had a more suitable dispersion than other ratios of G/PPG HC paste.
The plausible sensing mechanism of the G/PPG HC film for sensing organic compounds is schematically shown in Figure 2a. A well-aligned G/PPG HC film (25 µm) was fabricated by blade coating, indicating that the graphite was well dispersed in the PPG matrix with sufficient dispersion stability. When the sensor including a G/PPG HC was exposed to organic compounds, a radical chemical reaction occurred. The swelling of the G/PPG HC film under the influence of organic compounds led to a volume phase transition with mechanical stress and separation between the graphite. The volume phase transition and separation of each graphite in the polymer matrix play an important role in detecting the fluctuation of resistance in the G/PPG HC film. In addition, the evaporation of the PPG with organic compounds was observed; it exposed the graphite on the surface of film during the swelling. Figure 2b shows time-dependent thickness changes of the hybrid composite films based on various G/PPG ratios upon exposure to gasoline for 60 s. Because the resistance change response might be associated with volume changes, swelling ratios of the G/PPG sensor film in various formulations were calculated.
Sensors 2020, 20, x FOR PEER REVIEW 5 of 10 of film during the swelling. Figure 2b shows time-dependent thickness changes of the hybrid composite films based on various G/PPG ratios upon exposure to gasoline for 60 s. Because the resistance change response might be associated with volume changes, swelling ratios of the G/PPG sensor film in various formulations were calculated. The as-fabricated dispersed graphite without PPG did not show any obvious thickness change, and the G/PPG (2:1) and G/PPG (1:1) HCs showed a slight thickness change when they were exposed to gasoline. However, this slight thickness change is not sufficient for the resistance change ratio. In contrast, Figure 2b clearly shows a 15% increase in the thickness change for a G/PPG (1:2) HC according to Equation (1). It is apparent that the observed thickness changes result from the swelling during 10 s. In addition, after 10 s, thickness decreased owing to the decomposition of the polymer matrix upon exposure to the organic compound ( Figure S2).  Figure S3 compare the sensor response of intrinsic responsivity between graphite and G/PPG (2:1, 1:1, 1:2) HC sensors toward 1 mL of water, gasoline, diesel, and kerosene liquids at room temperature during the 90 s. Baseline resistance of the film (30 cm) at room temperature varied from 70 to 150 kilo ohms, depending on the proportion of thickness present and how effectively the electrode gap was bridged. The as-fabricated dispersed graphite without PPG did not show any obvious thickness change, and the G/PPG (2:1) and G/PPG (1:1) HCs showed a slight thickness change when they were exposed to gasoline. However, this slight thickness change is not sufficient for the resistance change ratio. In contrast, Figure 2b clearly shows a 15% increase in the thickness change for a G/PPG (1:2) HC according to Equation (1). It is apparent that the observed thickness changes result from the swelling during 10 s. In addition, after 10 s, thickness decreased owing to the decomposition of the polymer matrix upon exposure to the organic compound ( Figure S2).  Figure S3 compare the sensor response of intrinsic responsivity between graphite and G/PPG (2:1, 1:1, 1:2) HC sensors toward 1 mL of water, gasoline, diesel, and kerosene liquids at room temperature during the 90 s. Baseline resistance of the film (30 cm) at room temperature varied from 70 to 150 kilo ohms, depending on the proportion of thickness present and how effectively the electrode gap was bridged. Data were plotted according to ΔR Equation (2), where Rbefore was the resistance value just prior to the organic solvent exposure, and Rafter was the resistance value observed during the organic material exposure. As shown in Figure 3a, there was no change in the relative electrical resistance change (ΔR) of the graphite sensors upon exposure to organic materials (diesel and kerosene) for 90 s at room temperature. Only gasoline had a slight ΔR variation. This suggests that the graphite short circuit is not enough to be generated solely by the interfacial reaction of the organic solvent and graphite. As shown in Figure 3b, the ΔR of the G/PPG (1:2) HC film rapidly increased with exposure to organic solvents (gasoline, diesel, and kerosene). It was observed that the increased ΔR of the composite with organic solvents reached a value of about 35,000%~45,000% and was saturated within 10 s. The sensor based on G/PPG (1:2) had outstanding sensitivity compared to that of graphite, G/PPG (1:1) and (2:1). The resistance increase was attributed to the free volume change of the PPG polymer matrix as well as the gap between the electrical junction. The conducting graphite increases and widens because of the ratio of the swelling PPG matrix. The saturation phenomenon arises from the excess range of the fluke 789 process (~45 MΩ). Moreover, the ΔR of the G/PPG (1:2) HC film did not respond to water. These results indicate that G/PPG (1:2) HC film could be applied without the risk of moisture-induced malfunction in an external environment. Figure 3c shows the summarized ΔR variation of the graphite and G/PPG (2:1, 1:1, 1:2) HC sensors. As the PPG content increased, the ΔR variation and reactivity of the G/PPG HC sensors increased. The G/PPG (1:2) HC sensor showed high ΔR variations and saturation to all organic solvents, including gasoline, diesel, and kerosene. However, the reactivity of the G/PPG (1:3) HC increased and the HC film separated easily on the PET film because of physical resistance ( Figure S4). In addition, the sensor based on G/PPG (2:1) and (1:1) HCs had slightly lower ΔR variation and reactivity in the diesel. Data were plotted according to ∆R Equation (2), where R before was the resistance value just prior to the organic solvent exposure, and R after was the resistance value observed during the organic material exposure. As shown in Figure 3a, there was no change in the relative electrical resistance change (∆R) of the graphite sensors upon exposure to organic materials (diesel and kerosene) for 90 s at room temperature. Only gasoline had a slight ∆R variation. This suggests that the graphite short circuit is not enough to be generated solely by the interfacial reaction of the organic solvent and graphite. As shown in Figure 3b, the ∆R of the G/PPG (1:2) HC film rapidly increased with exposure to organic solvents (gasoline, diesel, and kerosene). It was observed that the increased ∆R of the composite with organic solvents reached a value of about 35,000%~45,000% and was saturated within 10 s. The sensor based on G/PPG (1:2) had outstanding sensitivity compared to that of graphite, G/PPG (1:1) and (2:1). The resistance increase was attributed to the free volume change of the PPG polymer matrix as well as the gap between the electrical junction. The conducting graphite increases and widens because of the ratio of the swelling PPG matrix. The saturation phenomenon arises from the excess range of the fluke 789 process (~45 MΩ). Moreover, the ∆R of the G/PPG (1:2) HC film did not respond to water. These results indicate that G/PPG (1:2) HC film could be applied without the risk of moisture-induced malfunction in an external environment. Figure 3c shows the summarized ∆R variation of the graphite and G/PPG (2:1, 1:1, 1:2) HC sensors. As the PPG content increased, the ∆R variation and reactivity of the G/PPG HC sensors increased. The G/PPG (1:2) HC sensor showed high ∆R variations and saturation to all organic solvents, including gasoline, diesel, and kerosene. However, the reactivity of the G/PPG (1:3) HC increased and the HC film separated easily on the PET film because of physical resistance ( Figure S4). In addition, the sensor based on G/PPG (2:1) and (1:1) HCs had slightly lower ∆R variation and reactivity in the diesel.
We investigated the sensor response by measuring fractional changes in resistance upon exposure to a set of 14 chemically diverse solvents: ethanol, methanol, acetone, 2-ethoxyethanol, cyclohexane, styrene monomer, N-methyl-2-pyrrolidone (NMP), di-methyl-formamide (DMF), isopropyl alcohol (IPA), toluene, ethyl acetate (EA), benzene, and methyl ethyl ketone (MEK). Response curves representing the change in resistance upon exposure to each organic compound are shown at Figure  S5. We confirmed that it could be used as a sensor for most hydrocarbons and alcohols.
To account for the apparent relationship of the swelling effect to volume change and sensor activity, we investigated the volume changes of G/PPG (1:2) HC film during gasoline adsorption with a contact-angle camera (Figure 4a-c, Figure S6, and Video S2). We found that the thickness of G/PPG (1:2) HC decreased after the adsorption of gasoline. These results indicated the dilation of the polymer network, confirming the swelling effect of G/PPG (1:2) HC. In addition, the exposed graphite on the top of the G/PPG (1:2) HC film suggested that PPG is dissolved and washed away by gasoline. The recycle test of the G/PPG (1:2) HC film using swelling as a sensing medium is another key reason to consider it for practical chemical sensing applications. However, the ∆R value decreased significantly after only one or two reuses (Figure 4d). The variation in resistance change gradually decreased as the test was repeated. This result indicates that the reaction with gasoline can reduce the swelling effect because of the absence of the removed PPG. In addition, the variation in resistance by water was constant (Figure 4e). Furthermore, the Raman spectra showed that the intensity of C-H bonds at PPG decreased, and air peaks appeared because of the reaction between gasoline and PPG, in good agreement with the exposed G and recycle test shown in Figure 4f. We investigated the sensor response by measuring fractional changes in resistance upon exposure to a set of 14 chemically diverse solvents: ethanol, methanol, acetone, 2-ethoxyethanol, cyclohexane, styrene monomer, N-methyl-2-pyrrolidone (NMP), di-methyl-formamide (DMF), isopropyl alcohol (IPA), toluene, ethyl acetate (EA), benzene, and methyl ethyl ketone (MEK). Response curves representing the change in resistance upon exposure to each organic compound are shown at Figure S5. We confirmed that it could be used as a sensor for most hydrocarbons and alcohols.
To account for the apparent relationship of the swelling effect to volume change and sensor activity, we investigated the volume changes of G/PPG (1:2) HC film during gasoline adsorption with a contact-angle camera (Figure 4a-c, Figure S6, and Video S2). We found that the thickness of G/PPG (1:2) HC decreased after the adsorption of gasoline. These results indicated the dilation of the polymer network, confirming the swelling effect of G/PPG (1:2) HC. In addition, the exposed graphite on the top of the G/PPG (1:2) HC film suggested that PPG is dissolved and washed away by gasoline. The recycle test of the G/PPG (1:2) HC film using swelling as a sensing medium is another key reason to consider it for practical chemical sensing applications. However, the ΔR value decreased significantly after only one or two reuses (Figure 4d). The variation in resistance change gradually decreased as the test was repeated. This result indicates that the reaction with gasoline can reduce the swelling effect because of the absence of the removed PPG. In addition, the variation in resistance by water was constant (Figure 4e). Furthermore, the Raman spectra showed that the intensity of C-H bonds at PPG decreased, and air peaks appeared because of the reaction between gasoline and PPG, in good agreement with the exposed G and recycle test shown in Figure 4f.

Conclusions
In summary, G/PPG HC films were successfully prepared and applied to organic compound sensors using a simple blade-coating method. We confirmed that the graphite in PPG solution was well dispersed and that graphite and PPG formed a hybrid composite structure. As a function of graphite and PPG ratio, the optimized G/PPG (1:2) HC had no chance of malfunctioning in water and had a very high sensitivity to several organic compounds. Based on our characterization outcomes,

Conclusions
In summary, G/PPG HC films were successfully prepared and applied to organic compound sensors using a simple blade-coating method. We confirmed that the graphite in PPG solution was well dispersed and that graphite and PPG formed a hybrid composite structure. As a function of graphite and PPG ratio, the optimized G/PPG (1:2) HC had no chance of malfunctioning in water and had a very high sensitivity to several organic compounds. Based on our characterization outcomes, we attributed resistance change with various organic compounds to the deformation and swelling of the polymer in the G/PPG HC films caused by each disconnected G plate. To the best of our knowledge, cost-effective and easily scalable sensors based on hybrid materials, including polymer/graphite composites, can be conducted with superior chemical sensing applications by improving their sensitivity, selectivity, and stability.