Side-Chain-Assisted Transition of Conjugated Polymers from a Semiconductor to Conductor and Comparison of Their NO2 Sensing Characteristics

To investigate the effect of a side chain on the electrical properties of a conjugated polymer (CP), we designed two different CPs containing alkyl and ethylene glycol (EG) derivatives as side chains on the same conjugated backbone with an electron donor-acceptor (D-A) type chain configuration. PTQ-T with an alkyl side chain showed typical p-type semiconducting properties, whereas PTQ-TEG with an EG-based side chain exhibited electrically conductive behavior. Both CPs generated radical species owing to their strong D-A type conjugated structure; however, the spin density was much greater in PTQ-TEG. X-ray photoelectron spectroscopy analysis revealed that the O atoms of the EG-based side chains in PTQ-TEG were intercalated with the conjugated backbone and increased the carrier density. Upon application to a field-effect transistor sensor for PTQ-T and resistive sensor for PTQ-TEG, PTQ-TEG exhibited a better NO2 detection capability with faster signal recovery characteristics than PTQ-T. Compared with the relatively rigid alkyl side chains of PTQ-T, the flexible EG-based side chains in PTQ-TEG have a higher potential to enlarge the free volume as well as improve NO2-affinity, which promotes the diffusion of NO2 in and out of the PTQ-TEG film, and ultimately resulting in better NO2 detection capabilities.


Introduction
Conjugated polymers (CPs) are promising active materials that can be used in various electronic devices, such as organic photovoltaic cells, field-effect transistors (FETs), and thermoelectric devices [1][2][3][4][5][6]. The electrical properties of CPs strongly depend on their chemical structures, which determine the degree of p-orbital overlap in the conjugated frameworks [7][8][9]. Recently, the insertion of electron-accepting dopants into electrondonating semiconductors has been reported to efficiently induce electron transfer or form a charge transfer (CT) complex, resulting in electrical conductors [10]. However, because infinite p-orbital overlap through conjugated structures is entropically unfavorable, most linear CPs exhibit a kinked chain configuration, resulting in semiconducting properties unless charge carriers are intentionally generated via chemical or electrical doping processes [11][12][13][14]. Few examples of CPs exhibiting essentially conductor-like properties have been reported. As a result, the design strategies for devising conductive CPs have not been fully established.
CPs have received great attention as versatile gas detection platforms because of their ease of structural modification to improve the detection suitability for target gases [15][16][17]. For example, NO 2 is a toxic gas that is abundantly released from industrial sources, and its strong oxidizing properties can critically damage the human respiratory system. Therefore, the accurate and fast detection of NO 2 is necessary to ensure workplace safety. Recently, our research group reported that EG-based side chains have good affinity for polar NO 2 The reagents used in this study were purchased from commercial suppliers (Sigma-Aldrich, Tokyo Chemical Industry, and Alfa Aesar) and used without further purification. Compounds 1a and 1b were synthesized similarly to the previously reported methods [18,38]. The chemical structure of CP was confirmed using Fourier transform infrared (FT-IR) and proton nuclear magnetic resonance ( 1 H-NMR) spectroscopy ( Figures S1 and S2). The molecular weight and polydispersity index (PDI) were determined by gel permeation chromatography with chloroform as an eluent ( Figure S3).

FET and Gas Sensor Fabrication
A heavily n-doped silicon wafer containing a SiO 2 layer (300-nm-thick with a capacitance of 10.8 nF/cm 2 , Fine Science) was cut into 2 cm × 2 cm pieces. After cleaning the pieces by ultrasonication in acetone and isopropyl alcohol for 20 min, they were dried under N 2 flow and then subjected to a 20 min UV-ozone treatment to modify the silicon wafer surface. After forming a self-assembled monolayer of octadecyl trichlorosilane on the substrate, each CP solution dissolved in chloroform (5 mg/mL) was spin-cast onto the self-assembled monolayer at 1500 rpm for 60 s. The formed CP film was dried in a vacuum oven at 25 • C for 4 h to remove residual solvent. To investigate the electrical properties of obtained CP thin films (57.0 ± 5.0 nm), source and drain electrodes (Au, 50 nm) with a channel length of 100 µm and a width of 2000 µm were thermally deposited on the CP layer. The FET active channel was placed in a gas sensor device and wired with Ag wire.

Electrical and Gas Sensing Properties
The electrical characteristics of the FET devices were measured using a Keithley 4200-SCS semiconductor parameter analyzer (Keithley, Cleveland, OH, USA) connected to a probe station (MS TECH, Suwon, Republic of Korea). To obtain the transfer curve, the gate bias (V G ) was swept from 40 V to −80 V in −1.0 V increments, while the source-drain voltage (V DS ) was fixed at −80 V. The output curve was collected by sweeping V DS from 0 to −80 V in −1.0 V increments while fixing V G at 0 V, −20 V, −40 V, −60 V, and −80 V, respectively. The gas-sensing properties of CPs were measured using a gas sensor (Precision Sensor System Inc., Daejeon, Republic of Korea). The applied voltage (V G and V DS ) was fixed at −10 V in the FET-type sensor, and the resistor-type sensor applied V DS (−10 V) at zero V G to measure the sensitivity to the analyte gas. The gas detection sensitivity was identified as the average value through repeated exposure to the analyte gas (50 ppm) for 50 s and filling with N 2 three times for 900 s.

Characterization
Polystyrene was used as a standard for gel permeation chromatography (Waters, Worcester County, MA, USA) to determine the molecular weights of the obtained CPs. UV-visible (UV-vis) absorption spectra of CP were characterized using UV-vis spectroscopy (Agilent, Santa Clara, CA, USA) in solution (chloroform) and film states, respectively. Ultraviolet photoelectron spectroscopy (UPS, Riken, Tokyo, Japan) was performed to determine the energy levels of the CPs. The surface morphologies of CP thin films were characterized using atomic force microscopy (AFM, Park Systems, Suwon, Republic of Korea) in a non-contact mode. The degree of CP chain assembly in the film state was analyzed using 2D-GIXRD (Xenocs, Grenoble, France), and the atomic binding energies of the CPs were determined using X-ray photoelectron spectroscopy (XPS, Thermo-Fisher, Seoul, Republic of Korea). All measurements using the CP thin films were performed with the same thickness (57.0 ± 5.0 nm) applied to the gas sensor device.

Results and Discussion
As shown in Scheme S1 and Figure 1a, two CPs containing the same thiadiazoloquinoxaline-based conjugated backbone but different side chains were designed to compare their electrical properties. Both CPs were obtained through a Stille-type cross-coupling reaction between thiadiazolo-quinoxaline and thiophene monomers, and their chemical structures were confirmed using FT-IR and 1 H-NMR spectroscopy. Both CPs exhibited the same skeletal vibrations corresponding to thiophene (1603 cm −1 ) and quinoxaline (1513 cm −1 ) moieties in the FT-IR spectra. However, the stretching vibrations corresponding to C-H (2700-3000 cm −1 ) and C-O-C (1025-1100 cm −1 ) were clearly distinguished ( Figure S1) because the side chains introduced into each CP are different. In addition, judging from the peak position and the integral ratio of aromatic and aliphatic protons in the 1 H-NMR spectra ( Figure S2), it was confirmed that polymerization was carried out successfully. Specifically, the integral ratio of aromatic and aliphatic protons was 1:7.1 for PTQ-T and 1:4.5 for PTQ-TEG, which was consistent with the theoretical values at a similar level (1:5.9 for PTQ-T and 1:4.1 for PTQ-TEG). The electrical properties of CPs have been known to be partially affected by the side chain. For example, the aggregation and chain assembly propensities of CPs depend on the side chains introduced onto the conjugated backbone, which critically affect the charge-carrier mobility of the CPs [7]. To compare the electrical properties of the CPs, alkyl and flexible EG-based side chains were introduced into the same conjugated backbone. In the conjugated framework of the obtained CPs, the thiadiazolo-quinoxaline derivative is a strong electron-accepting moiety, and the connected thiophenes have electron-donating characteristics. Therefore, the obtained CPs have an electron donor-acceptor (D-A) chain configuration. The typical characteristics of D-A type CPs include a bimodal-shaped absorption spectrum and narrow energy band-gap [39,40]. It has been known that, in the bimodal-shaped absorption spectrum of the D-A type CPs, the absorption in the shorter wavelength region is from π-π* transitions, and the red-shifted absorption originates from the intramolecular charge transfer (CT) between electron donating and accepting moieties that are covalently interconnected in the conjugated skeleton [41].

Results and Discussion
As shown in Scheme S1 and Figure 1a, two CPs containing the same thiadiazoloquinoxaline-based conjugated backbone but different side chains were designed to compare their electrical properties. Both CPs were obtained through a Stille-type cross-coupling reaction between thiadiazolo-quinoxaline and thiophene monomers, and their chemical structures were confirmed using FT-IR and 1 H-NMR spectroscopy. Both CPs exhibited the same skeletal vibrations corresponding to thiophene (1603 cm −1 ) and quinoxaline (1513 cm −1 ) moieties in the FT-IR spectra. However, the stretching vibrations corresponding to C-H (2700-3000 cm −1 ) and C-O-C (1025-1100 cm −1 ) were clearly distinguished ( Figure S1) because the side chains introduced into each CP are different. In addition, judging from the peak position and the integral ratio of aromatic and aliphatic protons in the 1 H-NMR spectra ( Figure S2), it was confirmed that polymerization was carried out successfully. Specifically, the integral ratio of aromatic and aliphatic protons was 1:7.1 for PTQ-T and 1:4.5 for PTQ-TEG, which was consistent with the theoretical values at a similar level (1:5.9 for PTQ-T and 1:4.1 for PTQ-TEG). The electrical properties of CPs have been known to be partially affected by the side chain. For example, the aggregation and chain assembly propensities of CPs depend on the side chains introduced onto the conjugated backbone, which critically affect the charge-carrier mobility of the CPs [7]. To compare the electrical properties of the CPs, alkyl and flexible EG-based side chains were introduced into the same conjugated backbone. In the conjugated framework of the obtained CPs, the thiadiazolo-quinoxaline derivative is a strong electron-accepting moiety, and the connected thiophenes have electron-donating characteristics. Therefore, the obtained CPs have an electron donor-acceptor (D-A) chain configuration. The typical characteristics of D-A type CPs include a bimodal-shaped absorption spectrum and narrow energy bandgap [39,40]. It has been known that, in the bimodal-shaped absorption spectrum of the D-A type CPs, the absorption in the shorter wavelength region is from π-π* transitions, and the red-shifted absorption originates from the intramolecular charge transfer (CT) between electron donating and accepting moieties that are covalently interconnected in the conjugated skeleton [41]. As shown in Figure 1b, when the absorption tendencies of the obtained CPs were characterized using UV-vis absorption spectroscopy, although both CPs have the same conjugated backbone, PTQ-T exhibited more red-shifted absorption than PTQ-TEG in both solution and film states. In general, alkyl side chains have a relatively rigid nature compared to EG-based side chains due to higher rotation barrier energy. Therefore, introducing alkyl side chains into CP (PTQ-T in this study) is likely to enhance the chain As shown in Figure 1b, when the absorption tendencies of the obtained CPs were characterized using UV-vis absorption spectroscopy, although both CPs have the same conjugated backbone, PTQ-T exhibited more red-shifted absorption than PTQ-TEG in both solution and film states. In general, alkyl side chains have a relatively rigid nature compared to EG-based side chains due to higher rotation barrier energy. Therefore, introducing alkyl side chains into CP (PTQ-T in this study) is likely to enhance the chain stiffness, which can promote non-covalent interactions between the CP chains and result in red-shifted absorption compared to analogous CPs with EG-based side chains [34,42]. Interestingly, although both CPs exhibited bimodal-shaped absorption spectra commonly seen in D-A type CPs [43,44], the absorption in the film state was found to be more red-shifted than that in the solution state, only in the longer wavelength region. This can be attributed to facile intramolecular CT interactions between D-A moieties via partial chain planarization or enhanced non-covalent interactions of CP chains in the film state [45]. When the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of each CP were identified using AC2 and the absorption edge, both CPs were confirmed to have similar molecular energy levels with an extremely narrow band-gap of 0.9 eV (Figure 1c). The narrow band-gap in both CPs implies the strong electron-accepting property of the thiadiazolo-quinoxaline unit because the band-gap of quinoxaline-based CPs with similar structures commonly exceeds 1.0 eV [46,47].
Although PTQ-T and PTQ-TEG share the same conjugated framework, they exhibited completely different electrical properties. To compare the electrical properties of the CPs, a FET device with a bottom-gate top-contact structure was fabricated. The PTQ-T-based FET showed a typical p-type charge-transport behavior, and the hole mobility was determined to be 0.002 cm 2 /V·s from the charge-carrier transfer curve (Figures 2a and S4a). However, as shown in Figure 2b, PTQ-TEG exhibited a considerably high current level with no off-current region in the V G sweep from −80 V to 40 V. In the output curve, even when V G was 0 V, PTQ-TEG exhibited a proportional trend in the output current with increasing V DS ( Figure S4b), indicating electrically conducting characteristics. Typically, the electrical conductivity of semiconducting CPs does not exceed 10 −7 S/cm because of the insufficient charge-carrier density unless they are chemically or electrically doped [48]. However, the obtained conductivity of PTQ-TEG was 3.0 × 10 −4 S/cm, which is too high of a value for PTQ-TEG to be labeled as a conventional semiconducting CP. orbital (LUMO) of each CP were identified using AC2 and the absorption edge were confirmed to have similar molecular energy levels with an extremely nar gap of 0.9 eV (Figure 1c). The narrow band-gap in both CPs implies the strong accepting property of the thiadiazolo-quinoxaline unit because the band-gap o line-based CPs with similar structures commonly exceeds 1.0 eV [46,47].
Although PTQ-T and PTQ-TEG share the same conjugated framework, th ited completely different electrical properties. To compare the electrical proper CPs, a FET device with a bottom-gate top-contact structure was fabricated. Th based FET showed a typical p-type charge-transport behavior, and the hole mo determined to be 0.002 cm 2 /V·s from the charge-carrier transfer curve (Figures 2a However, as shown in Figure 2b, PTQ-TEG exhibited a considerably high cu with no off-current region in the VG sweep from −80 V to 40 V. In the output cu when VG was 0 V, PTQ-TEG exhibited a proportional trend in the output cu increasing VDS ( Figure S4b), indicating electrically conducting characteristics. the electrical conductivity of semiconducting CPs does not exceed 10 −7 S/cm b the insufficient charge-carrier density unless they are chemically or electrically d However, the obtained conductivity of PTQ-TEG was 3.0 × 10 −4 S/cm, which is t a value for PTQ-TEG to be labeled as a conventional semiconducting CP. The distinct electrical properties of PTQ-TEG, compared to PTQ-T, could the contribution of EG-based side chains because both CPs have the same c framework. To ascertain the effect of the EG-based side chain on the electrical of PTQ-TEG, the film morphologies and chain assembly features of both CPs w acterized using AFM and 2D-GIXRD, respectively, because the electrical proper are known to be sensitively affected by the degree of chain aggregation and The distinct electrical properties of PTQ-TEG, compared to PTQ-T, could be due to the contribution of EG-based side chains because both CPs have the same conjugated framework. To ascertain the effect of the EG-based side chain on the electrical properties of PTQ-TEG, the film morphologies and chain assembly features of both CPs were characterized using AFM and 2D-GIXRD, respectively, because the electrical properties of CPs are known to be sensitively affected by the degree of chain aggregation and assembly [49,50]. As shown in Figure 3a,b, both CPs exhibited featureless smooth surfaces with root-mean-square (RMS) roughness values of 0.478 nm and 0.355 nm for PTQ-T and PTQ-TEG, respectively. Compared to PTQ-T, PTQ-TEG showed a slightly reduced RMS roughness. The reduced RMS roughness can be attributed to the enhanced fluidity of the EG-based side chain, which has a lower rotational barrier energy than the alkyl side chain. In addition, the chain assembly tendencies of the CPs in the film state were examined using 2D-GIXRD. As shown in Figure 3c,d, both CPs exhibit diffused diffraction tendencies, indicating amorphous-like weak crystalline characteristics. When comparing the diffraction intensities in the low-q region of both CPs ( Figure S5a,b), the diffraction pattern in the in-plane direction (xy-axis, q xy ) was confirmed to be more prominent than that in the out-of-plane direction (z-axis, q z ), implying that both CPs prefer face-on-type chain assemblies in the film state. AFM and 2D-GIXRD measurements indicated that the CPs had similar morphologies, including a chain assembly tendency, which implies that the difference in the electrical properties between PTQ-T and PTQ-TEG did not originate from the film morphology. TEG, respectively. Compared to PTQ-T, PTQ-TEG showed a slightly redu roughness. The reduced RMS roughness can be attributed to the enhanced flui EG-based side chain, which has a lower rotational barrier energy than the alkyl In addition, the chain assembly tendencies of the CPs in the film state were using 2D-GIXRD. As shown in Figure 3c,d, both CPs exhibit diffused diffractio cies, indicating amorphous-like weak crystalline characteristics. When compari fraction intensities in the low-q region of both CPs ( Figure S5a,b), the diffracti in the in-plane direction (xy-axis, qxy) was confirmed to be more prominent th the out-of-plane direction (z-axis, qz), implying that both CPs prefer face-on-t assemblies in the film state. AFM and 2D-GIXRD measurements indicated th had similar morphologies, including a chain assembly tendency, which impli difference in the electrical properties between PTQ-T and PTQ-TEG did no from the film morphology. To confirm whether the difference in the electrical properties of PTQ-T TEG was related to their inherent chemical structures, the binding energies ments present in the CPs were analyzed using XPS. As shown in Figure 4, when ing energies of the C, N, and O atoms were compared, the binding-energy patt N atoms were almost the same in both CPs; however, the binding energies o To confirm whether the difference in the electrical properties of PTQ-T and PTQ-TEG was related to their inherent chemical structures, the binding energies of the elements present in the CPs were analyzed using XPS. As shown in Figure 4, when the binding energies of the C, N, and O atoms were compared, the binding-energy patterns of the N atoms were almost the same in both CPs; however, the binding energies of C and O were clearly distinguished. As shown in Figure 4a, the binding-energy difference between the C atoms can be inferred to be due to the C atoms in the EG-based side chain introduced in PTQ-TEG. The C atom bonded to the O atom in the EG-based side chain of PTQ-TEG partially loses electrons and shows a relatively high binding energy because the O atom is more electronegative than the C atom. However, it can be clearly noted that the binding energy of O atoms in PTQ-TEG has shifted to the higher binding-energy region compared to PTQ-T. Upon further analysis of the binding-energy distribution of the O atoms in PTQ-TEG, the binding-energy spectrum could be separated into two different types of O atoms. One type showed the same binding energy as PTQ-T, whereas the other type had a higher binding energy. Furthermore, as depicted in Figure 4b, the O atoms with a higher binding energy (C-O*) occupied a larger portion than the neutral O atoms (C-O) in PTQ-TEG. The shift to a higher binding energy for the O atoms implies that they donate electrons. Although the exact interpretation of the binding-energy shift is limited, charge carriers were likely to be generated, because of electron donation by the O atoms in the EG-based side chains, resulting in the electrical conducting properties of PTQ-TEG.
Materials 2023, 16, x FOR PEER REVIEW 7 of 12 were clearly distinguished. As shown in Figure 4a, the binding-energy difference between the C atoms can be inferred to be due to the C atoms in the EG-based side chain introduced in PTQ-TEG. The C atom bonded to the O atom in the EG-based side chain of PTQ-TEG partially loses electrons and shows a relatively high binding energy because the O atom is more electronegative than the C atom. However, it can be clearly noted that the binding energy of O atoms in PTQ-TEG has shifted to the higher binding-energy region compared to PTQ-T. Upon further analysis of the binding-energy distribution of the O atoms in PTQ-TEG, the binding-energy spectrum could be separated into two different types of O atoms. One type showed the same binding energy as PTQ-T, whereas the other type had a higher binding energy. Furthermore, as depicted in Figure 4b  To compare the carrier densities quantitatively, the polaron spin densities of both CPs were measured using electron spin resonance (ESR). As shown in Figure 5, both CPs clearly show similar ESR signals without any artificial charge generation treatment, such as doping. Radical cations and anions have been reported to be generated by CT interaction when strong electron donating and accepting molecules are electronically intercalated [10]. Radical generation via CT interactions can also occur when D-A type structures form within a conjugated framework. For example, when pure organic materials contain strong electron acceptors in their conjugated structures, radicals have been generated via CT interactions and exhibit organic magnetism [51,52]. Therefore, it can be speculated that the ESR signals in both CPs originate from the strong intramolecular CT interaction between the strong electron accepting thiadiazolo-quinoxaline and electron donating thiophene, because similar ESR signals appear in both CPs with the same conjugated framework. In addition, the asymmetric ESR signals appearing for both CPs indicate that they are the result of multiple radical species rather than a single type of radical. When further analyzed using a Lorentz fitting, the ESR signal was deconvoluted into two different ESR signals with g-factors of 2.0100 and 2.0027. A g-factor of 2.0100 is comparable to that of radical anions, as commonly shown in organic molecules containing strong electron-accepting moieties, and a g-factor of 2.0027 coincides with unstable carbon radicals [35,37]. When the area of each ESR component was compared through double integration and calibration with a 2,2,6,6-tetramethylpiperidine-1-oxyl free radical (TEMPO) solution (Figure S6), PTQ-TEG exhibited a higher spin density than PTQ-T. In particular, the difference in the electrical properties of the CPs can be interpreted by comparing the g-factor 2.0027 component in each ESR signal because the g-factor of the carbon radical is To compare the carrier densities quantitatively, the polaron spin densities of both CPs were measured using electron spin resonance (ESR). As shown in Figure 5, both CPs clearly show similar ESR signals without any artificial charge generation treatment, such as doping. Radical cations and anions have been reported to be generated by CT interaction when strong electron donating and accepting molecules are electronically intercalated [10]. Radical generation via CT interactions can also occur when D-A type structures form within a conjugated framework. For example, when pure organic materials contain strong electron acceptors in their conjugated structures, radicals have been generated via CT interactions and exhibit organic magnetism [51,52]. Therefore, it can be speculated that the ESR signals in both CPs originate from the strong intramolecular CT interaction between the strong electron accepting thiadiazolo-quinoxaline and electron donating thiophene, because similar ESR signals appear in both CPs with the same conjugated framework. In addition, the asymmetric ESR signals appearing for both CPs indicate that they are the result of multiple radical species rather than a single type of radical. When further analyzed using a Lorentz fitting, the ESR signal was deconvoluted into two different ESR signals with g-factors of 2.0100 and 2.0027. A g-factor of 2.0100 is comparable to that of radical anions, as commonly shown in organic molecules containing strong electron-accepting moieties, and a g-factor of 2.0027 coincides with unstable carbon radicals [35,37]. When the area of each ESR component was compared through double integration and calibration with a 2,2,6,6-tetramethylpiperidine-1-oxyl free radical (TEMPO) solution ( Figure S6), PTQ-TEG exhibited a higher spin density than PTQ-T. In particular, the difference in the electrical properties of the CPs can be interpreted by comparing the g-factor 2.0027 component in each ESR signal because the g-factor of the carbon radical is comparable to that of a free electron [53]. When comparing the spin density corresponding to free electrons (g-factor of 2.0027), PTQ-TEG exhibited a 4.5 times higher value than PTQ-T. Therefore, PTQ-TEG has sufficient potential to exhibit higher electrical conductivity than PTQ-T, which could be the reason for its electrical conducting properties. comparable to that of a free electron [53]. When comparing the spin density corresponding to free electrons (g-factor of 2.0027), PTQ-TEG exhibited a 4.5 times higher value than PTQ-T. Therefore, PTQ-TEG has sufficient potential to exhibit higher electrical conductivity than PTQ-T, which could be the reason for its electrical conducting properties. Because NO2 has strong electron-accepting properties, its detection generally utilizes the CT interaction with electron-donating CPs, which can generate hole carriers and increase the current level in electrical sensors, such as FET-and resistive-type electrical devices [18,37]. Because PTQ-T and PTQ-TEG exhibit semiconducting and conducting properties, respectively, their NO2 sensing performances were compared using an FETtype sensor for PTQ-T and a resistive-type sensor for PTQ-TEG. When the NO2 detection capability was evaluated by repeatedly injecting NO2 (50 ppm) for 50 s and N2 for 900 s, the source-drain current (IDS) of both CPs markedly increased. As shown in Figure 6a, when repeatedly exposed to NO2 (50 ppm) under VG and VDS of −10 V, PTQ-T in a FETtype senor exhibited a gradual increase in current change upon each NO2 exposure, indicating that the NO2 from the previous exposure did not completely escape during N2filling for 900 s. The IDS(t)/IDS(0) value of PTQ-T did not fully recover to its initial value during N2 charging, indicating that NO2 was likely trapped in the PTQ-T film. In contrast, in the case of PTQ-TEG applied in a resistive-type sensor, the IDS quickly recovered to its initial level when exposed to N2, demonstrating uniform detectivity during repeated NO2 sensing evaluations. Although NO2 traps in CP chains may overestimate the NO2 detectivity of the PTQ-T-based FET sensor, the NO2 detection capabilities of both CPs were quantified using the average values of sensitivity and recovery for three repeated NO2 detection experiments. PTQ-TEG applied to the resistive-type sensor exhibited a detectivity of 6.9%/ppm and recovery of 96%, which exceeded the detectivity and recovery of PTQ-T applied to the FET-type sensor (4.0%/ppm and 31%). The better NO2 detection characteristics of PTQ-TEG, compared with those of PTQ-T, can be attributed to the difference in the chemical structure rather than the type of electrical sensor applied to each CP. As shown in Figure S7, when the surface energies of both CPs were determined by measuring contact angles using three different solvents (water, diiodomethane, and glycerol), PTQ-TEG (35.16 mN/m) was confirmed to have a higher surface energy than PTQ-T (21.41 mN/m). This result indicates that PTQ-TEG has a higher affinity to polar NO2. In addition, compared to the alkyl side chain in PTQ-T, the EG-based side chain can increase the free volume of the PTQ-TEG film because of its more flexible nature. Indeed, PTQ-TEG exhibited stronger amorphous hollow diffraction than PTQ-T in the high-q region of 2D-GIXRD (Figures 3d and S5c). Therefore, the high affinity for NO2 and the enlarged free volume of PTQ-TEG facilitated NO2 diffusion into the resistive-type sensor, resulting in better NO2 detection ability, including faster recovery, compared to PTQ-T. Because NO 2 has strong electron-accepting properties, its detection generally utilizes the CT interaction with electron-donating CPs, which can generate hole carriers and increase the current level in electrical sensors, such as FET-and resistive-type electrical devices [18,37]. Because PTQ-T and PTQ-TEG exhibit semiconducting and conducting properties, respectively, their NO 2 sensing performances were compared using an FETtype sensor for PTQ-T and a resistive-type sensor for PTQ-TEG. When the NO 2 detection capability was evaluated by repeatedly injecting NO 2 (50 ppm) for 50 s and N 2 for 900 s, the source-drain current (I DS ) of both CPs markedly increased. As shown in Figure 6a, when repeatedly exposed to NO 2 (50 ppm) under V G and V DS of −10 V, PTQ-T in a FET-type senor exhibited a gradual increase in current change upon each NO 2 exposure, indicating that the NO 2 from the previous exposure did not completely escape during N 2 -filling for 900 s. The I DS (t)/I DS (0) value of PTQ-T did not fully recover to its initial value during N 2 charging, indicating that NO 2 was likely trapped in the PTQ-T film. In contrast, in the case of PTQ-TEG applied in a resistive-type sensor, the I DS quickly recovered to its initial level when exposed to N 2 , demonstrating uniform detectivity during repeated NO 2 sensing evaluations. Although NO 2 traps in CP chains may overestimate the NO 2 detectivity of the PTQ-T-based FET sensor, the NO 2 detection capabilities of both CPs were quantified using the average values of sensitivity and recovery for three repeated NO 2 detection experiments. PTQ-TEG applied to the resistive-type sensor exhibited a detectivity of 6.9%/ppm and recovery of 96%, which exceeded the detectivity and recovery of PTQ-T applied to the FETtype sensor (4.0%/ppm and 31%). The better NO 2 detection characteristics of PTQ-TEG, compared with those of PTQ-T, can be attributed to the difference in the chemical structure rather than the type of electrical sensor applied to each CP. As shown in Figure S7, when the surface energies of both CPs were determined by measuring contact angles using three different solvents (water, diiodomethane, and glycerol), PTQ-TEG (35.16 mN/m) was confirmed to have a higher surface energy than PTQ-T (21.41 mN/m). This result indicates that PTQ-TEG has a higher affinity to polar NO 2 . In addition, compared to the alkyl side chain in PTQ-T, the EG-based side chain can increase the free volume of the PTQ-TEG film because of its more flexible nature. Indeed, PTQ-TEG exhibited stronger amorphous hollow diffraction than PTQ-T in the high-q region of 2D-GIXRD (Figures 3d and S5c). Therefore, the high affinity for NO 2 and the enlarged free volume of PTQ-TEG facilitated NO 2 diffusion into the resistive-type sensor, resulting in better NO 2 detection ability, including faster recovery, compared to PTQ-T. The selectivity and limit of detection (LOD) are important parameters of gas sensors. To demonstrate that the EG-modified polar side chain of PTQ-TEG did not impair the NO2 detection selectivity for other polar gases, the gas-sensing characteristics of the resistive sensor adopting PTQ-TEG were further evaluated using SO2, NH3, and CO2 with electron-withdrawing, electron-donating, and non-polar neutral properties, respectively. As shown in Figures 7a and S8, under the same conditions as NO2 detection, the detection sensitivities were in the following order: NO2 (6.9%/ppm), CO2 (0.19%/ppm), NH3 (0.05%/ppm), and SO2 (0.03%/ppm). This result clearly indicates that the introduced EGbased side chains in PTQ-TEG only marginally impaired the selective detection of NO2. Interestingly, while most CP-based electrical sensors respond sensitively to NH3, the devised resistive sensor adopting PTQ-TEG exhibited a negligible response to NH3. Additionally, when the theoretical LOD was determined by exposure to NO2 at concentrations of 100, 300, 600, and 900 ppb, the PTQ-TEG resistive sensor responded linearly to the NO2 concentration ( Figure S9). From the signal calibration with the signal-to-noise ratio and RMS noise, the slope extracted from the linear curve fitting using the obtained IDS(t)/IDS(0) value of each NO2 concentration marked a theoretical LOD of 1.59 ppb (Figure 7b). The extracted LOD was comparable to that of the most sensitive electrical NO2 sensor that adopted crystalline CPs with good electrical properties [31].  The selectivity and limit of detection (LOD) are important parameters of gas sensors. To demonstrate that the EG-modified polar side chain of PTQ-TEG did not impair the NO 2 detection selectivity for other polar gases, the gas-sensing characteristics of the resistive sensor adopting PTQ-TEG were further evaluated using SO 2 , NH 3 , and CO 2 with electron-withdrawing, electron-donating, and non-polar neutral properties, respectively. As shown in Figure 7a and Figure S8, under the same conditions as NO 2 detection, the detection sensitivities were in the following order: NO 2 (6.9%/ppm), CO 2 (0.19%/ppm), NH 3 (0.05%/ppm), and SO 2 (0.03%/ppm). This result clearly indicates that the introduced EG-based side chains in PTQ-TEG only marginally impaired the selective detection of NO 2 . Interestingly, while most CP-based electrical sensors respond sensitively to NH 3 , the devised resistive sensor adopting PTQ-TEG exhibited a negligible response to NH 3 . Additionally, when the theoretical LOD was determined by exposure to NO 2 at concentrations of 100, 300, 600, and 900 ppb, the PTQ-TEG resistive sensor responded linearly to the NO 2 concentration ( Figure S9). From the signal calibration with the signal-to-noise ratio and RMS noise, the slope extracted from the linear curve fitting using the obtained I DS (t)/I DS (0) value of each NO 2 concentration marked a theoretical LOD of 1.59 ppb (Figure 7b). The extracted LOD was comparable to that of the most sensitive electrical NO 2 sensor that adopted crystalline CPs with good electrical properties [31]. The selectivity and limit of detection (LOD) are important parameters of gas sensors. To demonstrate that the EG-modified polar side chain of PTQ-TEG did not impair the NO2 detection selectivity for other polar gases, the gas-sensing characteristics of the resistive sensor adopting PTQ-TEG were further evaluated using SO2, NH3, and CO2 with electron-withdrawing, electron-donating, and non-polar neutral properties, respectively. As shown in Figures 7a and S8, under the same conditions as NO2 detection, the detection sensitivities were in the following order: NO2 (6.9%/ppm), CO2 (0.19%/ppm), NH3 (0.05%/ppm), and SO2 (0.03%/ppm). This result clearly indicates that the introduced EGbased side chains in PTQ-TEG only marginally impaired the selective detection of NO2. Interestingly, while most CP-based electrical sensors respond sensitively to NH3, the devised resistive sensor adopting PTQ-TEG exhibited a negligible response to NH3. Additionally, when the theoretical LOD was determined by exposure to NO2 at concentrations of 100, 300, 600, and 900 ppb, the PTQ-TEG resistive sensor responded linearly to the NO2 concentration ( Figure S9). From the signal calibration with the signal-to-noise ratio and RMS noise, the slope extracted from the linear curve fitting using the obtained IDS(t)/IDS(0) value of each NO2 concentration marked a theoretical LOD of 1.59 ppb (Figure 7b). The extracted LOD was comparable to that of the most sensitive electrical NO2 sensor that adopted crystalline CPs with good electrical properties [31].

Conclusions
Two CPs with the same conjugated backbone containing different side chains were designed to investigate the effect of the side chain on their electrical properties. Although both CPs exhibited similar smooth morphologies and weak crystalline chain assemblies in the film state, they exhibited completely different electrical properties. Specifically, PTQ-T with alkyl side chains showed typical p-type semiconducting characteristics, whereas PTQ-TEG with EG-based side chains exhibited electrical conducting behaviors. It was confirmed that, although both CPs have radical species owing to their strong D-A type conjugated structure, the O atoms of the EG-based side chains can additionally intercalate with the conjugated backbone, increase the carrier density, and ultimately generate the conductor-like properties of PTQ-TEG. When PTQ-T and PTQ-TEG were applied to FETand resistive-type sensors, respectively, PTQ-TEG exhibited higher NO 2 sensitivity with a faster recovery tendency than PTQ-T. The flexible EG-based side chains increased the free volume of the CP chains as well as the affinity with polar NO 2 molecules, which facilitated NO 2 diffusion in and out of the PTQ-TEG film, resulting in better sensitivity to NO 2 than PTQ-T. In addition, it was confirmed that the EG-based side chains in PTQ-TEG barely impaired the detection selectivity for other common gases, such as SO 2 , NH 3 , and CO 2 .
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ma16072877/s1, Scheme S1: Detailed chemical structures and polymerization scheme of CPs; Figure S1: 1 H-NMR spectra of each CP; Figure S2: FT-IR spectra of each CP; Figure S3: Molecular weight of obtained CPs, which were determined using gel permeation chromatography with elution of chloroform; Figure S4: Output curve characteristics of both CPs according to V G in FET devices; Figure S5: 1D XRD diffractograms; Figure S6: ESR intensity according to TEMPO concentration; Figure S7: contact angle information used to determine the surface energy of each CP; Figure S8: Response pattern of PTQ-TEG resistive sensor to common gases; Figure S9: Response pattern of PTQ-TEG resistive sensor according to NO 2 concentration. Data Availability Statement: Data sharing not applicable, all data obtained from this study are already given in the article.