Comparative Study on the Formation and Oxidation-Level Control of Three-Dimensional Conductive Nanofilms for Gas Sensor Applications.

Investment in wearable monitoring systems is increasing rapidly for realizing their practical applications, for example, in medical treatment, sports, and security systems. However, existing wearable monitoring systems are designed to measure a real-time physical signal and abnormal conditions rather than harmful environmental characteristics. In this study, a flexible chemical sensor electrode based on a three-dimensional conductive nanofilm (3D CNF) is fabricated via facile polymerization with temperature control. The morphology and chemical state of the 3D CNF are modified via electrochemical doping control to increase the carrier mobility and the active surface area of the sensor electrode. The sensor electrode is highly sensitive (up to 1 ppb), selective, and stable for an analyte (NH3) at room temperature owing to the three-dimensional morphology of polypyrrole and the oxidation-level control.

Introduction

Flexible electronics have attracted much considerable attention because of their excellent potential for application in portable devices and displays.[1−4] In partin class="Chemical">cular, investment in wearable monitoring system research has been rapidly increasing for realizing applications, for example, in healthcare, sports, and security systems.[5−7] However, conventional wearable monitoring systems are specialized for physiological measurements to track real-time health status and abnormal conditions rather than hazardous environmental situations. The ability to detect harmful and flammable substances has recently emerged as another technical problem for sensor devices in preventing diseases because of perilous chemicals.[8−11] Therefore, it is necessary to develop small, simple, and accurate sensing materials with low detection limits (down to ppb level) for wearable chemical sensor systems. Considerable research has been performed on nanosized semiconducting materials owing to their favorable characteristics, such as their high surface-to-volume ratio and small size, which lead to amplified signals.[12−16]

Among the semiconducting materials, conducting polymer nanomaterials are used as sensing transducers because of their attractive inherent transport properties (i.e., electrical conductivity and energy migration at room temperature) originating from their conjugated backbone structure.[17−20] Additionally, the oxidation levels of the conducting polymers can be changed via doping processes to facilitate a sensitive and rapid response to specific chemical species.[21−23] Moreover, studies have been performed on the morphology control of conducting polymer nanomaterials in sensor transducers to maximize the interactions with target analytes. For example, Park et al. fabricated a multidimensional carboxyl poly(3,4-ethylenedioxythiophene) nanotube membrane-based biosensor using electrospinning and modified vapor deposition polymerization to detect dopamine molecules.[24] Sakar et al. formulated poly(3,4-ethylenedioxythiophene) polystyrene sulfonate films with quasi-periodic parallel cracks, which had a high charge carrier mobility for use in alcohol and humidity sensors.[25] However, the synthesis steps for morphology control at the nanoscale are complex.

Even though conducting polymer nanomaterials exhibit high performance as sensing transducers, there is a lack of methods for integrating sensor components into flexible electronic cirn class="Chemical">cuits, limiting the mass production of devices.[26,27] In addition, the uniform diffusion of nanosized sensing materials is an obstacle for their application to flexible sensor devices.[28−30] Several approaches have been proposed to achieve integration with a uniform distribution.[31−35] One of the most promising approaches involves the direct generation of a conducting polymer nanostructure on a flexible substrate as a sensing transducer. This method allows the uniform installation of nanostructures on devices at low cost without an integration step.

Herein, we describe the formation of a three-dimensional conductive nanofilm (3D n class="Gene">CNF)-based chemical sensor consisting of vertically aligned polypyrrole nanowires on a graphene substrate, through an electrodeposition process with temperature control. The morphology of the polypyrrole nanostructures is controlled by changing the synthesis temperature because the micelle size and the diffusion area increase with the temperature. Moreover, the doping level of the 3D CNF is changed using an electrochemical oxidation process to enhance the charge-transfer rate during the detection of the target analyte (NH3). The 3D nanostructure and electrochemical doping enhance the active surface area and charge-carrier (hole) transfer of the nanofilm, respectively. Thus, the electrochemically doped nanofilm-based chemical sensor exhibits a high sensitivity to NH3 gas, with low detection limits (down to ppb level) and fast response and recovery times. Additionally, the sensing ability of the chemical sensor electrode is retained after various bending cycles because of the flexibility of the substrates [polyethylene naphthalate (PEN) and graphene] and the polypyrrole nanofilm.

Results and Discussion

Formation Control of 3D CNF

Figure a shows the overall process for the formation of 3D CNFs via the electron class="Chemical">polymerization method. As a template of the conducting polymer nanostructures, monolayer graphene was introduced onto a PEN substrate using the wet transfer method. The graphene-transferred PEN substrate was immersed in a pyrrole-dispersed electrolyte at various temperatures to decorate the surface with polypyrrole nanostructures. In the electrolyte solution, a phosphate buffer maintains the pH of the electrolyte during the polymer synthesis to prevent the influence of the generated hydrogen ions. Additionally, p-toluenesulfonic acid functions as not only a dopant but also a soft template of the pyrrole to form pyrrole-containing micelles.[36] In the polymerization process, the pyrrole-containing micelles are adsorbed on the graphene surface, acting as templates at the beginning of the polymerization (i.e., nucleation sites). The generated nucleation sites inhibit the growth of polypyrrole in directions other than the vertical direction owing to the steric hindrance of adjacent nanostructures.[37] The steric hindrance effect is enhanced with the growth of the polypyrrole nanostructure.

Figure 1

(a) Schematic diagram of the sequential fabrication steps for the 3D CNF. (b) Illustrative formation mechanism of polypyrrole nanostructures with different temperatures on the graphene substrate.

(a) Schematic diagram of the sequential fabrication steps for the 3D n class="Gene">CNF. (b) Illustrative formation mechanism of polypyrrole nanostructures with different temperatures on the graphene substrate.

The morphology of the generated polypyrrole nanostructures changes from a nanowire to a micrograin with temperature variations (Figures and S1). At a low synthesis temperature (5 °C), nucleation sites with a n class="Chemical">diameter of 50 nm are uniformly formed on the surface, and vertically aligned polypyrrole nanowires (ca. 1 μm long, with a diameter of ca. 70 nm) are formed after the growth process. The nucleation sites and nanowires increase in size and become irregular with the increasing synthesis temperature, up to 70 °C. Temperatures higher than 70 °C result in thick-coated polypyrrole films rather than nucleation sites and nanowires on the graphene surface. In particular, at 90 °C, the thickness of the aggregated part of the 3D CNF increases from 55 to 245 nm, and the diameter of the nanostructures enhances to 210 nm (Figure S2). Additionally, at the macroscale, the generated polypyrrole exhibits irregular surfaces at synthesis temperatures higher than 5 °C (Figure ). In a moderate temperature range (45–70 °C), the films display micropores, and the size of the pores increases with the temperature. A synthesis temperature higher than 70 °C yields a bumpy surface rather than pores.

Figure 2

FE-SEM images of the 3D CNF during the electrochemical polymerization process (nucleation and growth step) at different temperatures: (a–c) 5; (d–f) 45; (g–i) 70; (j–l) 90 °C.

Figure 3

Low-magnification FE-SEM images of the 3D CNF at different temperatures after the polymerization process: (a) 5, (b) 45, (c) 70, and (d) 90 °C.

The morphology change of the polypyrrole nanostructures originates from the following mechanism (Figure b). First, the size of the micelles in the electrolyte increases at higher reaction temperatures owing to the enlargement in the hydrodynamic ran class="Chemical">dius.[38−40] Then, the larger containers increase the size of the nucleation sites and enlarge the diffusion area of each nucleation site. The extended diffusion areas of the nucleation sites cause the overlap of the nuclei, resulting in the formation of a mass film on the substrate during the growth process.[41] In addition, the Brunauer–Emmett–Teller (BET) surface areas of the films are decreased with increasing temperatures (33.9 m2 g–1 for 5 °C, 29.9 m2 g–1 for 45 °C, 24.2 m2 g–1 for 70 °C, and 18.2 m2 g–1 for 90 °C), owing to the fact that the number of nanowires per unit area is reduced at higher temperatures (Figure S3).

Oxidation-Level Control of 3D CNF

The 3D CNF n class="Chemical">polymerized at 5 °C presents different chemical states and morphologies with an electrochemical oxidation-level control. During the oxidation control, transformations of the polymer chains (such as neutral, polaron, bipolaron, and overoxidation) occur with an increasing applied potential (Figure S4). Figure a shows the electrical conductivity of the 3D CNF with different oxidation levels (ranging from −1.4 to +1.4 V). The electrical conductivity increases with the oxidation level because a more positive potential improves the carrier density (0.2 × 1018 to 1.31 × 1019 cm–3) in the polymer chains. There are two improvements of the electrical conductivity—from −1.4 to −1.0 V and from 0 to 0.6 V—because of the changes in the charge-transport property from neutral to polarons and from polarons to bipolarons, respectively. The highest value of the electrical conductivity is 101.8 S cm–1 for an applied potential of +1.0 V. High applied voltages (from +1.3 V) cause overoxidation of the films, with a reduction of the electrical conductivity (to 1.3 × 10–3 S cm–1 at 1.4 V) despite the high carrier densities.[42] Additionally, the carrier mobility (μ), which originated from the electrical conductivity (σ) and carrier density (n0), differs with the increasing applied potential. The carrier mobility rapidly decreases in the overoxidation region (above 1.3 V) owing to the destruction of the polymer chain structures, despite the high carrier density.

Figure 4

(a) Electrical conductivity (black), carrier density (blue), and carrier mobility (red) of the 3D CNF with an applied voltage variation. (b) Raman spectra of the 3D CNF with different oxidation levels (black: −1.2 V, red: −0.2 V, blue: +1.0 V, and pink: +1.4 V).

(a) Electrical conductivity (black), carrier density (blue), and carrier mobility (red) of the 3D CNF with an applied voltage variation. (b) Raman spectra of the 3D n class="Gene">CNF with different oxidation levels (black: −1.2 V, red: −0.2 V, blue: +1.0 V, and pink: +1.4 V).

Raman spectroscopy was performed to further characterize the 3D CNFs. Figure b shows Raman spectra with n class="Chemical">different oxidation levels ranging from −1.2 to +1.4 V. The bands at 1049 and 1332 cm–1 correspond to the C–H in-plane and C–C stretching vibrations, respectively. Moreover, the strong bands at 1415 and 1562 cm–1 are attributed to the C=C symmetric and C=C asymmetric vibrations, respectively. These strong bands are shifted to 1421 and 1591 cm–1, respectively, with an increase in the oxidation level, indicating the transformation of the polymer chains from the benzenoid structure to a quinoid structure. Specifically, positive-doped species (polarons or bipolarons) are located on geometrical defects and form the quinoid structure. Therefore, more positive voltages increase the amount of (bi)polarons and change the polymer chains to quinoid structures.[43]

Furthermore, the chemical composition of the 3D CNF was characterized using X-ray photoelectron spectroscopy (XPS). Figure presents the new bond of the n class="Chemical">polymer chains in the N 1s region with different applied voltages. Peaks at ca. 400 eV [related to the neutral amine N of the pyrrole unit (−NH−)] and ca. 398 eV (related to imine (−N=)) are observed in all cases. For a potential of −0.2 V, there is a peak corresponding to the positively charged N (−NH+−) at 401.4 eV, indicating the formation of polarons. For more positive potentials (+1.0 and +1.4 V), there is a bipolaron N peak (=NH+−) at 403.0 eV. Thus, increasing the oxidation level of the 3D CNF enhances the amount of positively charged N in the structure.

Figure 5

N 1s XPS spectra of the 3D CNF with different oxidation levels: (a) −1.2; (b) −0.2; (c) +1.0; (d) +1.4 V.

The morphology of the 3D CNF is reversibly changed for n class="Chemical">different oxidation levels using the osmotic effect of the counterion (bis(trifluoromethanesulfonyl)amide anion) (Figure ).[44] When electrochemical oxidation occurs, electrons are ejected from the 3D CNF, and a positive charge is developed in the polymer chain. Then, the anions in the electrolyte can move and create a volume expansion in the polymer through the insertion of counterions and the subsequent uptake of the solvent. However, excess electro-oxidation (i.e., overoxidation) causes an irreversible volume expansion of the polymer nanostructure. Consequently, from a macroscale viewpoint, the CNF with overoxidation (at +1.4 V) shows cracks in the structure, and a part of it fell off onto the PEN substrate (Figure S5). The overoxidized nanofilm is very weak to external impacts and displays a degradation of its electroactivity. Moreover, the diameter of the individual nanowires is enhanced from 67.7 nm at −1.2 V to 85.6 nm at +1.4 V, with an increase in the deviation (Figure S6). Additionally, the morphology of the individual nanowires changes as the intermediate holes are extended to nanotubes at high voltages because of counterion insertion. As with the field effect scanning electron microscopy (FE-SEM) images in Figure , larger pores are observed as the oxidation levels increase (7 nm for −1.2 V, 10 nm for −0.2 V, 15 nm for +1.0 V, and 30 nm for +1.4 V) (Figure S7). The BET surface area also increases, that is, from 33.9 m2 g–1 (for −1.2 V) to 41.5 m2 g–1 (for +1.4 V) at more positive potentials on account of the fact that polymer structures are more expanded at higher oxidation levels.

Figure 6

Low- and high-resolution FE-SEM images of the 3D CNF with different oxidation levels: (a,e) −1.2; (b,f) −0.2; (c,g) +1.0; (d,h) +1.4 V.

Sensor Application

The real-time responsive resistance changes were measured for difn class="Chemical">ferent concentrations of gases at room temperature. The 3D CNF-based sensor electrodes consist of three parts, namely, the substrates (PEN and graphene), the source/drain electrodes (silver paste), and the transducer (3D CNF) (Figure S8). To confirm the sensing performance of 3D CNF, a bare chemical vapor deposition (CVD) graphene-based electrode shows a small response at a high concentration of NH3 gas, but it cannot detect at a low concentration (Figure S9). Therefore, the sensing ability of the electrode at low concentrations originated from the 3D CNF layer. First, the effect of the synthesis temperature on the sensing performance was examined by exposing the sensor electrodes to NH3 gas, which is a reducing gas. As shown in Figure , the response decreases with the increasing reaction temperature because of the reduced active surface area, caused by the larger nanowires and thicker aggregated part in the 3D CNF structure. The sensing performance with respect to the oxidation level is investigated for the 3D CNF polymerized at 5 °C.

Figure 7

Normalized resistance changes of the 3D CNF with different polymerization temperatures (a) upon sequential exposure to NH3 gas and (b) as a function of NH3 concentration (black: 5 °C; red: 20 °C; blue: 45 °C; pink: 70 °C).

Normalized resistance changes of the 3D n class="Gene">CNF with different polymerization temperatures (a) upon sequential exposure to NH3 gas and (b) as a function of NH3 concentration (black: 5 °C; red: 20 °C; blue: 45 °C; pink: 70 °C).

Figure a shows the sensing performance of the electrodes with respect to the oxidation level. The sensor electrode oxidized at +1.0 V exhibits the highest sensitivity (down to 1 n class="Chemical">ppb) to NH3 on account of the high carrier density (hole) and the additional inner channel in the nanowire structure. In contrast, despite the improved active sites in the structure, the electrode oxidized at +1.4 V shows no response to the NH3 gas, owing to the destruction of the conjugated polymer chains. A highly sensitive electrode is achieved in the bipolaron state (at 1.0 V), with a large amount of charge carriers and additional active sites for NH3. Figure b presents the recovery times of the electrodes, indicating a decrease from 59 s (at −1.2 V) to 19 s (at +1.0 V). The large number of charge carriers induces the rapid transfer of electrons from the analyte and reduces the response time to <20 s. However, because the detachment of the target molecule is related to the applied external energy rather than to the charge carriers and the number of active sites, the recovery times of the electrodes are similar (66 s for −1.2 V, 61 s for −0.2 V, and 56 s for +1.0 V) (Figure c). Additionally, to confirm the sensing ability for an oxidizing gas (the opposite charge-transfer direction), methanol (MeOH) molecules were detected (Figure d). As observed in the case of NH3, the electrode oxidized at +1.0 V is superior to the other electrodes, for the same reason. Figure e shows the normalized resistance changes of the electrodes with respect to the NH3 and MeOH gas concentrations. The linear behavior range is larger for the electrode oxidized at +1.0 V (1 ppb to 102 ppm for NH3 and 0.1 to 103 ppm for MeOH) than for the other electrodes. Therefore, the sensor electrode oxidized at +1.0 V can be effectively utilized as a signal transducer for detecting NH3 and MeOH gases at various concentrations.

Figure 8

Reversible and reproducible responses are measured at a constant current value (10–6 A) of the 3D CNF with different oxidation levels. (a) Normalized resistance changes upon sequential exposure to various concentrations of NH3. (b) Response and (c) recovery times of the 3D CNF toward 1 ppm of NH3. (d) Normalized resistance changes to different MeOH concentrations. (e) Calibration lines of the 3D CNF as a function of NH3 and MeOH concentrations. Each applied voltage is as follows: black for −1.2 V; red for −0.2 V; blue for +1.0 V; and pink for +1.4 V. (f) Periodic exposure of the +1.0 V applied 3D CNF to 1 ppb of NH3 gas. Normalized resistance changes of the +1.0 V applied 3D CNF under (g) various bending angles and (h) repeated bending cycles. (i) Sensing performance histogram of the +1.0 V applied 3D CNF to different oxidizing and reducing volatile gases. The concentrations of gases are as follow: 1 ppm for NH3 and MeOH and 100 ppm for others.

Reversible and reproducible responses are measured at a constant current value (10–6 A) of the 3D n class="Gene">CNF with different oxidation levels. (a) Normalized resistance changes upon sequential exposure to various concentrations of NH3. (b) Response and (c) recovery times of the 3D CNF toward 1 ppm of NH3. (d) Normalized resistance changes to different MeOH concentrations. (e) Calibration lines of the 3D CNF as a function of NH3 and MeOH concentrations. Each applied voltage is as follows: black for −1.2 V; red for −0.2 V; blue for +1.0 V; and pink for +1.4 V. (f) Periodic exposure of the +1.0 V applied 3D CNF to 1 ppb of NH3 gas. Normalized resistance changes of the +1.0 V applied 3D CNF under (g) various bending angles and (h) repeated bending cycles. (i) Sensing performance histogram of the +1.0 V applied 3D CNF to different oxidizing and reducing volatile gases. The concentrations of gases are as follow: 1 ppm for NH3 and MeOH and 100 ppm for others.

Outstanding flexibility and cycle stability are required for electrode materials used in practical wearable sensor devices. Figure f presents the resistance change of the oxin class="Chemical">dized electrode at +1.0 V under a periodic exposure to 1 ppb NH3 gas. During five repetitive on–off tests, the resistance changes remained relatively constant, without a delay in the response or recovery. Moreover, the flexibility of the substrates (PEN and graphene) and the polypyrrole nanofilm allows the electrodes to perform consistent detection even if they are modified in several ways. The resistance of the sensor electrodes, along with concave and convex deformations at different angles, shows a similar value at 1 ppb NH3 (Figure g). Additionally, repeated bending deformation (up to 500 cycles) has no significant effect on the sensing performance, owing to the uniformity of each component (CNF and Ag paste) on the substrate (Figures h and S10).

The selectivity of sensor electrodes is also important for practical applications. Figure i presents the normalized resistance changes of the electrode oxidized at +1.0 V under exposure to various volatile gases at low concentrations (1 ppm for n class="Chemical">NH3 and MeOH; 100 ppm for others). Even though it has a lower concentration than the other gases, NH3 yields a signal change more than 10 times larger. Therefore, NH3 can be distinguished from other chemicals according to the extent and direction of the resistance changes. Thus, the sensing performance of the 3D CNF-based electrode is higher than that of other conventional conducting polymer-based sensor electrodes (Table ).[22,45−48]

Table 1

Summary of Representative Sensors for NH3 Detection

sensing material	sensing signal	working temperature (°C)	limit of detection	response/recovery time	references
PEDOT:PSS/FeCl3	resistance	25	0.1 ppm	20 s/—	(22)
PPy/rGO	resistance	25	1 ppm	1 min/5 min	(45)
PANI + rGO	resistance	25	20 ppm	18 min/2 min	(46)
ZnO/rGO	resistance	25	10 ppm	84 s/216 s	(47)
TiO2/rGO	resistance	25	5 ppm	114 s/304 s	(48)
3D CNF	resistance	25	1 ppb	18 s/56 s	this work

Conclusions

3D CNFs were formed through a facile n class="Chemical">polymerization method with synthesis temperature control as transducers for a chemical sensor. The chemical states and morphologies of the 3D CNFs are changed using an electrochemical oxidation method. The 3D CNF oxidized at 1.0 V exhibits the highest charge carrier density (1.31 × 1019 cm–3), as well as an inner channel that enhances the interaction with the target analyte and increases the charge-transfer rate in the sensor electrode. The 3D CNF-based sensor shows a high sensitivity (down to 1 ppb) to NH3 gas and fast response times (<20 s) at room temperature. Moreover, the chemical sensor displays flexibility, cycle stability under various deformations, and selectivity to NH3 gas. Thus, the function of conductive nanostructures is optimized through the shape and chemical-state control, and the proposed method is effective for developing electrical devices and sensor systems.

Experimental Section

Materials

Pyrrole, sodium p-toluenesulfonate (pTSA), and bis(trifluoromethane)sulfonimide lithium salt (LiBis) were purchased from Sigma-Aldrich. Sodium phosphate dibasic dodecahydrate (Dibasic) and sodium phosphate monobasic dihydrate (Monobasic) were purchased from Junsei Chemical Co. All chemical reagents were used as received, without further purification.

Formation of 3D CNF

A 3D CNF was prepared on a n class="Chemical">graphene–PEN film via an electrodeposition reaction in a 100 mL aqueous solution containing 0.1 mol of pyrrole, 0.2 mol of sodium phosphates (1:1 molar ratio of Dibasic/Monobasic), and 0.1 mol of pTSA. The graphene–PEN film was prepared by CVD and a wet transfer method, as described in previous reports.[9,14] Electrodeposition was performed at different temperatures under a constant potential of +0.8 V for 15 min in a three-electrode cell consisting of a Pt wire, Ag/AgCl, and the CVD graphene–PEN film as the counter, reference, and working electrodes, respectively. After the electrodeposition, the 3D CNF was rinsed with deionized (DI) water and dried at room temperature.

Oxidation-level control of the 3D CNF was conducted in the potentiostat mode using a three-electrode cell with 50 mL of a 0.1 M n class="Chemical">LiBis aqueous electrolyte. A Pt wire, Ag/AgCl, and the 3D CNF were used as the counter, reference, and working electrodes, respectively. A constant voltage was applied to the 3D CNF electrode for 3 min, and the voltage was changed from −1.6 to +1.4 V stepwise, with 0.2 V intervals. The electrical resistance was measured via a four-probe method using a source meter immediately after each voltage application phase. The oxidation-level-controlled 3D CNF was rinsed with DI water and dried at room temperature.

Electrical Sensing Measurement of Volatile Organic Compound Gases

Gas sensor electrodes were prepared by painting each end of the 3D CNF with Ag paste and attaching a n class="Chemical">Cu wire to the ends at a distance of 1 cm. The sensor electrode was placed inside a vacuum chamber (100 Torr) and connected to a source meter to monitor the resistance change using a computer. In this experiment, a bubbler was used to vaporize different gases and to mix the vapors with N2 gas using a mass flow controller to modify the concentration of the analytes. The sensor electrode was exposed to various concentrations (1 ppb to 100 ppm) for 100 s and purged with N2 gas for the conductance to recover. The measurement was conducted by applying a constant current of 10–6 A to the electrode, and the response was calculated by measuring the normalized electrical resistance change ΔR/R0 = (R – R0)/R0, where R and R0 represent the measured real-time resistance and initial resistance, respectively. The response time was defined as the time taken for the resistance to reach 90% of the saturated value after the gas exposure, and the recovery time was defined as the time taken for the resistance to reach 90% of the initial value after purging with N2 gas.

Characterization

FE-SEM images were obtained using a JSM-6701F (JEOL Ltd., Japan). Raman spectra were obtained using a DXR2xi (Thermo, USA) installed at n class="Chemical">NCIRF at Seoul National University. XPS data were acquired using a Sigma Probe (Thermo, USA). The electrical conductivity was measured using a Keitheley 2400, and the amount of charge was recorded using an electrochemical workstation (WBCS3000, WonATech, Korea). The electrodeposition and oxidation-level control were performed using WBCS3000. The electrical conductivity, charge carrier mobility, and charge carrier density were calculated using the following equations.

The electrical conductivity was calculated aswhere L represents the length, A represents the area of the n class="Gene">CNF, and R represents the resistance measured by the source meter.

The charge carrier mobility was calculated aswhere n0 represents the charge carrier density and |e| represents the electrical charge of the carrier.

The charge carrier density was calculated aswhere q represents the amount of charge and V represents the volume of the nanofilm.