P-Type Chemical Doping-Induced High Bipolar Electrical Conductivities in a Thermoelectric Donor–Acceptor Copolymer

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Oct 2021P-Type Chemical Doping-Induced High Bipolar Electrical Conductivities in a Thermoelectric Donor–Acceptor Copolymer Jing Wang, Yizhuo Wang, Qing Li, Zhanchao Li, Kuncai Li and Hong Wang Jing Wang Frontier Institute of Science and Technology, Xi'an Jiaotong University, Xi'an 710054 , Yizhuo Wang Frontier Institute of Science and Technology, Xi'an Jiaotong University, Xi'an 710054 , Qing Li Frontier Institute of Science and Technology, Xi'an Jiaotong University, Xi'an 710054 , Zhanchao Li Frontier Institute of Science and Technology, Xi'an Jiaotong University, Xi'an 710054 , Kuncai Li Frontier Institute of Science and Technology, Xi'an Jiaotong University, Xi'an 710054 and Hong Wang *Corresponding author: E-mail Address: [email protected] Frontier Institute of Science and Technology, Xi'an Jiaotong University, Xi'an 710054 State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an 710054 https://doi.org/10.31635/ccschem.021.202101070 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail Typically, conducting polymers transfer either electrons or holes. It is rare to see high bipolar (p- and n-type) electrical conductivities within a single bulk doped organic polymer without the assistant of gate voltage. Herein, we report that FeCl3-doped solution-processable D–A copolymer poly (2,5-bis(2-octyldodecyl)-3,6-di(thiophen-2-yl)diketopyrrolo[3,4-c]pyrrole-1,4-dione-alt-thieno[3,2-b]thiophen) (DPPTTT) could exhibit a high p-type electrical conductivity of 130.6 S/cm and a good n-type electrical conductivity of 14.2 S/cm by engineering the doping level. Both p- and n-type electrical conductivities were superior to most solution-processable D–A copolymers, including monopolar polymers. The high electrical conductivity resulted in high thermoelectric performance of DPPTTT in both p- and n-type, leading to a high current density of 3 A/cm2 for a fully organic planar p–n junction created with only one material. Structural and spectroscopic tests were performed to provide a fundamental understanding of the polarity-switch mechanism. Our results open up an opportunity of making p- and n-type modules with a single conducting polymer for modern organic electronics in the future and might arouse research interest in exploring novel conducting polymers to enrich the knowledge of charge transport in organic materials. Download figure Download PowerPoint Introduction Conducting polymers have the potential advantages of low manufacturing cost by using traditional printing techniques to produce electronic modules at a large scale,1 which makes them very attractive for innovative electronic applications, especially for flexible electronic devices.2,3 Remarkable progress has been achieved in developing functional conducting polymers in many fields such as organic thermoelectrics,1,2,4 organic solar cells,1 organic field-effect transistors (FETs),5 organic light-emitting diodes,6 and so on, since the conducting polymer was first discovered by Heeger et al. in 1977.7 However, the transport of holes and electrons in conducting polymers is still far from being well understood and controlled. Making p- and n-type modules with a single conducting polymer is still very challenging for organic materials; though, it can be easily realized for inorganics such as doping silicon with phosphorus and boron to make n- and p-type semiconductors, respectively. For modern organic electronics, p- and n-type materials are equally important and desired. Making p- and n-type modules with a single conducting polymer would simplify the fabrication processes of organic devices with high scalability and low cost using traditional printing techniques to pattern a single conducting polymer with p- and n-type dopants. Besides, it may improve the performance of organic devices. For example, Roncali8,9 suggested that single-material solar cells would be the next frontier for organic photovoltaics since they would have a longer lifetime due to the strong stabilization of the morphology of the interface. Toffanin and co-workers10 reported that single-layer light-emitting transistors might ensure good charge transport together with an efficient light-emission in the solid-state. Wang et al.6 reported that a novel organic Schottky barrier diode created in a single planar polymer film exhibited a remarkable current density of 30 A/cm2 that is 2–3 orders in a higher magnitude and superior to that of previously reported organic materials. This is because it excluded interfaces that generally existed between p- and n-type modules, resulting in smooth current flow and subsequent improvement in the performance of the devices. These promising results of single-material devices light the enthusiasm in making p- and n-type modules with a single conducting polymer. However, not many conducting polymers can transport both holes and electrons. In general, the carrier polarity of a conducting polymer strongly depends on its molecular structure. Once a conducting polymer is synthesized, it may prefer to transport either electrons or holes. Most conducting polymers exhibit a unipolar transport property and can transport only holes or electrons. In contrast, ambipolar conducting polymers that exhibit both positive and negative Seebeck coefficients have been reported in limited cases.4,11–13 Especially, it is very challenging to achieve conducting polymers with high bipolar electrical conductivities (possessing high electrical conductivity for both p- and n-type). Currently, a popular strategy employed to make ambipolar conducting polymer is to co-polymerize electron-rich groups (donors, D) and electron-deficient groups (acceptors, A).14–16 However, only a small portion of the D–A copolymers exhibit appreciable bipolar transport properties.17–19 Although D–A copolymers have been reported to have very high motilities of over 20 cm2/(V·s) recently,19,20 the electrical conductivity for the D–A copolymers are often in the range of 10−3–10−5 S/cm.21,22 These ambipolar conducting polymers with low electrical conductivities are mainly used in organic FETs or organic solar cells. Moreover, most of these ambipolar conducting polymers exhibit bipolar electrical conductivities in FET devices, which require the assistance of a gate voltage.23 Very rare examples were reported to have bipolar electrical conductivities for bulk-doped conducting polymers without the assistant of the gate voltage.19,24 To the best of our knowledge, bipolar electrical conductivities (both p- and n-type) of over 10−2 S/cm have never been reported for D–A copolymers to date. Here, we report that a solution-processable D–A copolymer, poly(2,5-bis(2-octyldodecyl)-3,6-di(thiophen-2-yl)diketopyrrolo[3,4-c]pyrrole-1,4-dione-alt-thieno[3,2-b]thiophen) (DPPTTT; Figure 1), could achieve high bipolar electrical conductivities after being doped with FeCl3 ( DPPTTT FeCl 3 ). This fabricated copolymer exhibited the highest p-type electrical conductivity of 130.6 S/cm and a high n-type electrical conductivity of 14.2 S/cm, superior to most solution-processable D–A copolymers ( Supporting Information Tables S1 and S2). Its high electrical conductivity led to the establishment of a high p-type thermoelectric power factor of 23.4 μW/(mK2) and a high n-type thermoelectric power factor of 0.66 μW/(mK2), which are the highest attainable for p- and n-type solution-processable ambipolar D–A copolymers, respectively ( Supporting Information Figure S7). The conversion mechanism was addressed after relevant structural and spectroscopic tests. A p–n junction was created in a planar thin film, exhibiting a high rectification ratio of 2 × 102 at ±5 V for fully printed organic diodes, which further demonstrated the conversion of the p-type D–A copolymer to n-type. The rectification performance met the requirement for high-frequency radio-frequency identification (R-ID) tags.25,26 These results might uncover an opportunity for developing new organic electronic devices with single organic materials. Figure 1 | (a) Molecular structure of the polymers, DPPTTT and P3HT, and commercially available dopants. (b) Electrical conductivities of DPPTTT after being p-type doped by six commercially available dopants at MDR = 1. Download figure Download PowerPoint Experimental Section Materials DPPTTT (Mw = 22.47 kDa) was purchased from Derthon Optoelectronic Materials Science Technology Co. Ltd. (Shenzhen, China). Poly(3-hexylthiophene) (P3HT) (Mw = 85,000 KDa) was purchased from 1-material (Dorval, Quebec, Canada). Ferric chloride was purchased from Macklin Biochemical Co., Ltd. (Shanghai, China). Nitromethane was purchased from Damao Chemical Reagent Factory (Tianjin, China). Iron sulfate hydrate and iron nitrate nonahydrate were purchased from Aladdin Industrial Corp. (Shanghai, China). 2,3,5,6-Tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4TCNQ), iron p-toluenesulfanate (PTS-Fe), and O-dichlorobenzene were purchased from Energy Chemical (Shanghai, China). All chemicals were used as received. Hexacyano-trimethylene-cyclopropane (CN6-CP) was synthesized in our lab, as described previously.27 Film preparation The polymers and dopants solutions were prepared in advance by dissolving them in O-dichlorobenzene and nitromethane. Then mixed solutions were obtained at different weight ratios by adding the dopants solutions to the polymers. After each mixture was vigorously stirred for ∼15 min, it was dropped onto a glass substrate and dried at room temperature. All the experiments were performed in a glove box. The detailed film preparation process has been described in the Supporting Information. P–N junction preparation The DPPTTT O-dichlorobenzene solution with 21 wt % FeCl3 was dropped onto half of the copolymer-dopant glass substrate. After drying at room temperature, the DPPTTT O-dichlorobenzene solution with 84 wt % FeCl3 was dropped onto the rest half of the glass substrate. The p–n junction was tested after both sides of the films were dry. Computational methods Ionization potential (IP) and electron affinity (EA) potentials were used to evaluate the ambient stability and the charge injection properties of the DPPTTT and P3HT polymers, respectively.28 The adiabatic IP and EA (AIP, AEA) were calculated according to the following formulas: AIP = E + ( G + ) − E 0 ( G 0 ) (1) AEA = E 0 ( G 0 ) − E − ( G − ) (2)where G+, G−, and G0 represent the geometric configurations of optimized cationic, anionic, and neutral states, respectively. E+, E−, and E0 are the energy of cationic, anionic, and neutral states, respectively. The calculation was implemented in the Gaussian 09 Program Package. All of the optimized and the frequency analysis of cationic, anionic, and neutral states were obtained using functional theory methods of B3LYP with a 6-31G(d) basis set. By taking the calculation efficiency and cost into account, only the DPPTTT monomer and P3HT polymers were calculated. Also, the long alkyl side chains of P3HT and DPPTTT were replaced with methyl groups, which would reduce the calculation cost in the determination of the electronic structure of the polymer without significantly decreasing the calculation accuracy. Results and Discussion The molecular structures of DPPTTT, P3HT, and the p-type dopants:F4TCNQ, PTS-Fe, CN6-CP are shown in Figure 1a. A solution-processable method was used to prepare the DPPTTT and P3HT films on glass substrates, as shown in Supporting Information Figure S1. Six commercially available p-type dopants were studied at the same molar doping ratio (MDR = 1, repeat units: dopant ratio of 1:1). Pristine DPPTTT is almost an insulator with a poor electrical conductivity that is not measurable. Its electrical conductivity increased after being doped with the six p-type dopants, with their electrical conductivities arranged in the following order σ FeCl 3 > σ CN 6 − CP > σ PTS − Fe > σ CuCl 2 > σ F 4 TCNQ > σ Fe 2 ( SO 4 ) 3 (Figure 1b). The trend is mainly related to the doping ability of the p-type dopants relative to their molecular orbital occupancy. For example, CN6-CP had the lowest unoccupied molecular orbital (LUMO) of −5.87 eV than F4TCNQ (LUMO = −5.24 eV),27 leading to a higher electrical conductivity of DPPTTT CN 6 − CP than that of DPPTTT F 4 TCNQ . DPPTTT FeCl 3 showed the highest p-type electrical conductivity among the six p-type dopants, which might be due to the small size and high EA of FeCl3, as reported in the previous literature.29 Electrical conductivity and Seebeck coefficient of DPPTTT doped with FeCl3 at different concentrations were measured at room temperature, as shown in Figure 2. The electrical conductivity of DPPTTT FeCl 3 increased with a rise in FeCl3 concentration until it reached a maximum of 130.6 S/cm at the FeCl3 concentration of 21 wt %. Then the electrical conductivity decreases while further increasing the concentration of FeCl3. Scanning electron microscopy (SEM) was performed to understand the electrical conductivity dropping from the FeCl3 concentration of 21 wt %. Figures 2d–2f show SEM images of pristine DPPTTT, DPPTTT FeCl 3 − 21 wt % , and DPPTTT FeCl 3 − 84 wt % , respectively. Pristine DPPTTT film is composed of continuous agglomerates with the size of ∼ 50 × 50 nm (Figure 2d). After being doped by FeCl3, the agglomerates aggregate as shown in the SEM images of DPPTTT FeCl 3 − 21 wt % (Figure 2e) and DPPTTT FeCl 3 − 84 wt % (Figure 2f). The gaps between the agglomerates increase because of the aggregation (inserted images in Figures 2d–2f), which might be one of the main reasons for the decrease of electrical conductivity from FeCl3 concentration of 21 wt % in Figure 2a. Atomic force microscopy (AFM) was performed to identify the roughness of the film. As shown in Supporting Information Figure S2, the root mean square (RMS) values obtained for pristine DPPTTT, DPPTTT FeCl 3 − 21 wt % , and DPPTTT FeCl 3 − 84 wt % were 2.95, 4.47, and 36.6 nm, respectively ( Supporting Information Figure S2). It further confirmed that the aggregation of DPPTTT agglomerates at high FeCl3 concentration over 21 wt %. Temperature dependence of the electrical conductivity of DPPTTT FeCl 3 − 21 wt . % and DPPTTT FeCl 3 − 84 wt % at the temperature range of 150–300 K was performed as shown in Supporting Information Figure S3. The electrical conductivities increased with increased temperature, indicating that the charge transport in DPPTTT FeCl 3 fitted the variable range hopping model as reported for most of the conducting polymers. The activation energies calculated from the plots were 43.64 and 81.44 meV for DPPTTT FeCl 3 − 21 wt % and DPPTTT FeCl 3 − 84 wt % , respectively. The larger activation energy should be due to the larger gaps between the agglomerates, leading to the lower electrical conductivity of DPPTTT FeCl 3 − 84 wt % . Figure 2 | Electrical conductivity (a) and Seebeck coefficient (b) of DPPTTT and P3HT as a function of FeCl3 concentration. (c) Comparison of the maximum p- and n-type electrical conductivity of DPPTTT FeCl 3 with solution-processable D–A copolymers. SEM images of pristine DPPTTT (d), DPPTTT FeCl 3 − 21 wt % (e), and DPPTTT FeCl 3 − 84 wt % (f). For each data point in Figures 2a and 2b, three or more examples were tested. However, the error bars are too small to be seen in Figures 2a and 2b. The original data have been added in Supporting Information Table S3 to help the readers to know the variation range of each data point. Download figure Download PowerPoint Figure 2b shows the Seebeck coefficient of DPPTTT FeCl 3 as a function of the FeCl3 concentration. It is interesting to note that the Seebeck coefficient of DPPTTT FeCl 3 switching from p- to n-type. Pristine DPPTTT is a p-type material, according to previous works.29–31 The Seebeck coefficient of DPPTTT FeCl 3 − 10 wt % was +223 μV/K, which indicated that it is still p-type at low FeCl3 concentration. The Seebeck coefficient decreased while more FeCl3 was added due to an increase in the carrier concentration at a higher p-type doping level. However, the Seebeck coefficient of DPPTTT FeCl 3 became negative after the FeCl3 concentration was over 47 wt %. Further increasing the FeCl3 concentration led to more negativity of the Seebeck coefficient. The negative Seebeck coefficient of DPPTTT FeCl 3 revealed that it became an n-type material. The maximum n-type electrical conductivity and the maximum Seebeck coefficient were 14.2 S/cm ( DPPTTT FeCl 3 − 47 wt % ) and −78.9 μV/K ( DPPTTT FeCl 3 − 84 wt % ), respectively. The maximum p- and n-type electrical conductivities of DPPTTT FeCl 3 were compared with those of previously reported solution-processable D–A copolymers as shown in Supporting Information Tables S1 and S2. Very rare examples were reported to have bipolar electrical conductivities for bulk-doped conducting polymers without the assistant of the gate voltage.19,24 High electrical conductivities over 2 S/cm for both p- and n-type in a single conducting polymer have never been reported in previous literature. Figure 2c compares the maximum p- and n-type electrical conductivities of DPPTTT FeCl 3 with those of solution-processable D–A copolymers. As far as we know, the maximum p- and n-type electrical conductivities achieved were the highest for p- and n-type solution doped ambipolar D–A polymers. A few post-doped solution-processable D–A polymers were reported recently to have higher electrical conductivities. Therefore, compared to all the solution-processable D–A polymers, the maximum p-type electrical conductivity of 130.6 S/cm achieved by the DPPTTT FeCl 3 film was among the top seven values of solution-processable state-of-the-art D–A copolymers, revealed by previous literature reports (including unipolar and ambipolar D–A polymers) ( Supporting Information Table S1). Likewise, the maximum n-type electrical conductivity of 14.2 S/cm for DPPTTT FeCl 3 was one of the best among n-type electrical conductivities, as only one example is available in previous reports, with a higher n-type electrical conductivity of over 14.2 S/cm ( Supporting Information Table S2). Compared with previously reported ambipolar D–A copolymers, the maximum bipolar electrical conductivities were ∼ 4–5 orders of magnitude higher than previously reported values. The high bipolar electrical conductivities indicated a potential to make p- and n-type modules for modern organic electronics with a single conducting polymer, which might promote a revolution in the field of semiconducting polymers. The electrical conductivity and Seebeck coefficient of FeCl3-doped P3HT were measured at room temperature for comparison, as shown in Figure 2. The electrical conductivity increased with an increase in FeCl3 concentration, reaching its maximum of 63 S/cm at the FeCl3 concentration of 47 wt % (Figure 2a), after which it started to drop when the FeCl3 concentration was increased further. The Seebeck coefficient of P3HT at the FeCl3 concentration of 10 wt % was +157 μV/K, which decreased with the rising FeCl3 concentration. Unlike FeCl3-doped DPPTTT, no switching from p- to n-type was observed (Figure 2b). All the Seebeck coefficients for P 3 HT FeCl 3 samples were positive. Ultraviolet–visible–near-infrared (UV–vis–NIR) spectrophotometry was performed to monitor the doping process of DPPTTT and P3HT. Figure 3a shows the UV–vis–NIR spectra of DPPTTT FeCl 3 films at different FeCl3 concentrations. A major peak appeared at 810 nm for pristine DPPTTT film, assigned to the π–π* transition of the diketopyrrolo-pyrrole (DPP) unit.32–34 This prominent peak decreased together with an increase in a new broad peak at the NIR region when the FeCl3 concentration was increased, ascribed to the reduction in the neutral state of the DPP units since they were converted into the polaron state or bipolaron state,34 resulting in the emergence and increase of a new peak (>1000 nm) in the NIR region. Figure 3d shows a clear conversion process of P3HT between the neutral, the polaron, and the bipolaron states. The neutral state of P3HT could be converted to the polaron state, which might be converted further to the bipolaron state.35,36 At low FeCl3 concentration, the neutral state of P3HT was converted into the polaron and bipolaron states, leading to a decrease in the neutral state peak at 532 nm, an increase of the polaron state peak at 832 nm, and an increase in the bipolaron state peak in the NIR region (>1100 nm). At high FeCl3 concentration, the decrease in the polaron state peak at 832 nm indicated that the rate of the polaron state P3HT conversion to the bipolaron state was higher than the rate of the generation of the polaron state P3HT. The band gaps obtained from the UV–vis–NIR spectra of pristine DPPTTT and P3HT were 1.28 and 1.9 eV, respectively ( Supporting Information Figure S4), consistent with previous reports.37–42 Figure 3 | UV–vis–NIR spectra of DPPTTT (a) and P3HT (d) as a function of FeCl3 concentration. The CV of pristine DPPTTT (b) and pristine P3HT (e). The UPS results for pristine DPPTTT, DPPTTT FeCl 3 − 21 wt % , DPPTTT FeCl 3 − 84 wt % (c) and pristine P3HT, P 3 HT FeCl 3 − 21 wt % , P 3 HT FeCl 3 − 84 wt % (f). The electronic band diagrams of DPPTTT (g) and P3HT (h) during the FeCl3 doping process. Download figure Download PowerPoint The highest occupied molecular orbitals (HOMOs) of DPPTTT and P3HT were estimated using the cyclic voltammetry (CV) method with a three-electrode electrochemical system. Ferrocene was used as the internal reference (see detailed description in the Supporting Information). Figures 3b and 3e show that the oxidation peak edges (φox) were 0.97 and 0.94 V for DPPTTT and P3HT, respectively. Therefore, the HOMOs ( E HOMO ) were −5.19 and −5.16 eV for DPPTTT and P3HT, respectively, according to the following equation 37: E HOMO = − ( 4.8 − E 1 / 2 Ferrocene + φ ox ) (3)where E 1 / 2 Ferrocene is the difference between the average oxidation peak potential and the reduction peak of the internal reference, ferrocene (shown in Supporting Information Figure S5). The HOMO values obtained were similar to those previously reported, −5.2 eV for P3HT,37,38 and −5.2 eV for DPPTTT.40–42 Based on the CV results and the optical band gap values, the LUMO values for DPPTTT and P3HT were calculated to be −3.91 and −3.26 eV, respectively. Ultraviolet photoelectron spectroscopy (UPS) was performed to provide the work function (WF) changes of DPPTTT and P3HT during the FeCl3 doping process. Figure 3c shows UPS results of pristine DPPTTT, DPPTTT FeCl 3 − 21 wt % and DPPTTT FeCl 3 − 84 wt % along with that of gold as a reference. Their WFs can be obtained from the equation = + | E | − | E | and are high and low energy respectively, and is the energy of the used The achieved for pristine DPPTTT was After being doped with FeCl3 at a concentration of 21 wt the became larger indicating p-type doping of It further increased to eV when the FeCl3 concentration was increased to 84 wt %. Also, the of P3HT increased after FeCl3 as shown in Figure as The WFs for pristine P3HT, P 3 HT FeCl 3 − 21 wt % , and P 3 HT FeCl 3 − 84 wt % were in the order of pristine − P 3 HT P 3 HT − FeCl 3 − 21 wt % P 3 HT − FeCl 3 − 84 wt % As a comparison, the of gold was a of eV ( Supporting Information Figure which is in the range of previously reported values, The electronic band diagrams of DPPTTT and P3HT during the FeCl3 doping process are shown in Figures and respectively. The was obtained from the work = is the which since the the energy barrier for an electron to at the to the The of pristine DPPTTT, from the UPS was eV, to the of HOMO and LUMO of pristine DPPTTT The between these values could be ascribed to doping or in the After being doped by FeCl3, the to = eV for DPPTTT FeCl 3 − 21 wt % , to the HOMO of DPPTTT indicating p-type doping of DPPTTT (Figure the for P 3 HT FeCl 3 − 21 wt % was eV, which was still far from its HOMO of −5.16 These results indicated that it was to DPPTTT than P3HT at the same FeCl3 concentration. This trend was consistent with the calculated IP for DPPTTT and P3HT The lower IP suggested a leading to a lower at the same FeCl3 concentration. the FeCl3 was 84 wt the for DPPTTT FeCl 3 − 84 wt % to eV, which was the HOMO of DPPTTT, while the for P 3 HT FeCl 3 − 84 wt % was eV, which was still the HOMO of P3HT (Figure Based on these we that the polarity switching of DPPTTT films from p- to n-type was due to the from the HOMO to The hopping model reported by was used to the charge carrier transport in conducting For a p-type material, the Seebeck coefficient is = ( E F − E V ) and are the the and the temperature. For a neutral conducting polymer, the in the of LUMO and As shown in Figures and the of pristine DPPTTT and P3HT to the of the HOMO and they were doped by the p-type FeCl3, the For example, the for DPPTTT FeCl 3 − 21 wt % ∼ eV HOMO and the for P 3 HT FeCl 3 − 21 wt % ∼ eV The in led to a between the and which resulted in a decrease of the Seebeck according to (Figure 2b). The Seebeck coefficients for DPPTTT FeCl 3 − 21 wt % and P 3 HT FeCl 3 − 21 wt % were positive because was positive and > For DPPTTT FeCl 3 − 84 wt % , the the HOMO which resulted in a negative Seebeck coefficient of This polarity switching is rare for conducting it has been observed in polymer semiconductors, such as D–A copolymer and shown to from n- to p-type when doped by switching from n- to p-type when it is by an organic In contrast, no polarity switching was observed for P 3 HT FeCl 3 the same To the best of our knowledge, is the first report copolymer from p- to

Topics & Concepts