Effects of interfacial area and energetic barrier on thermoelectric performance of PEDOT:PSS–MXene composite films

Thermoelectric (TE) devices based on conducting polymers have significant potential for low-temperature energy harvesting. To enhance the TE performance, the incorporation of low-dimensional inorganic fillers into the polymer matrix has been considered as a promising strategy by exploiting the energy filtering effect. Since the energy filtering effect is strongly influenced by the carrier scattering at the interface between polymer and inorganic fillers, the TE properties are likely to be affected by the interfacial properties of two constituents. In this study, we investigated the TE performance in the composite films of poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) and two-dimensional Ti3C2 MXene, in order to reveal the effects of the interfacial area and the energetic barrier on the TE performance by controlling the MXene sizes and the oxidation level of PEDOT:PSS. We found that the composite film with smaller MXene exhibits a higher power factor (PF) than that with larger MXene, originating from the increased interfacial area which facilitates the energy filtering effect. We also showed that an optimal energy barrier (0.14 eV) between PEDOT:PSS and MXene can accelerate the energy filtering effect, which allows to maximize the PF of the composite films up to 69.4 μW m−1 K−2. We believe that our study not only contributes to the development of the composite-based TE devices utilizing the energy filtering effect, but also helps to understand the charge transport in polymer–inorganic composites.


Introduction
Thermoelectric (TE) generation based on conducting polymers is one of the promising technologies that can harvest low-grade waste heat, which accounts for 50% of the total waste heat [1][2][3]. Their unique properties such as light-weight, low-cost, non-toxicity, solution processability, and flexibility provide compelling advantages over the expensive bulk inorganic counterparts at low-temperature heat sources such as the human body [4,5]. In addition, their relatively lower thermal conductivity (κ) is also advantageous for a higher TE figure of merit ZT = α 2 σT/κ, which determines the heat-to-electricity conversion efficiency, where α and σ are the Seebeck coefficient and the electrical conductivity, respectively [6]. Thereby, multilateral efforts in material synthesis [7,8], (de)doping methods [9,10], and device physics [11][12][13] have been made in polymer-based TE devices, leading to rapid advances in their performance [14]. Nevertheless, the performance of polymer TE devices in terms of the power factor (PF) α 2 σ is still lower than that of the inorganic materials, which requires further improvement for practical applications in wearable energy harvesters [15] or Internet of Things sensors [16].
One of the widely used strategies to increase the PF of the conducting polymers is to control the charge carrier concentration (n) by doping. Owing to the development of various dopants and doping methods [17], the electronic structure of conducting polymers can be tuned from non-degenerate to degenerate states [18]. For instance, the σ of pristine poly (3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) is approximately 0.1 S cm −1 but it can be improved to thousands siemens per centimeter through various secondary doping methods [19][20][21], based on increased bipolaronic states and/or enhanced crystallinity of PEDOT chains [22][23][24]. However, an increase in n inevitably causes a decrease in α, and this strong interplay between α and σ limits the PF in a single material [25]. Therefore, embedding inorganic fillers to a polymer matrix has been an important strategy to overcome the limitation by introducing energy filtering effects [25][26][27][28][29] -creating energy barriers that can selectively scatter low-energy charge carriers while allowing high-energy charge carriers to pass through, thereby increasing α which is proportional to the average entropy transported per charge-carrier with little loss of σ [30]. In particular, low-dimensional materials (e.g., nanosheets [31], nanowires [32,33], and nanoparticles [34,35]) are excellent candidates for inorganic fillers due to their larger surface-to-volume ratio than bulk materials [35], which effectively accelerates the effective charge carrier scattering at the interfaces between polymer and inorganic fillers. Since the charge carriers are mainly scattered at the interface between polymer and inorganic fillers [32,35], rationally designing the energy barrier height and increasing the interfacial area between them can enhance the TE performance by facilitating the energy filtering effects.
To date, various low-dimensional materials have been introduced to enhance the performance of PEDOT: PSS-based TE devices, such as Bi 2 Te 3 nanoparticles [35], Te nanowires [25], and two-dimensional (2D) materials (e.g., reduced graphene oxide [36], SnSe [31], MoS 2 [37], MoSe 2 [38], and Ti 3 C 2 MXene [39]). Among them, 2D MXene has significant potential due to its excellent electrical conductivity [40], thermodynamic stability [41], and flexibility [42]. In addition, the hydrophilic nature of MXene [43], owing to surface functional groups such as −OH, −O and −F, makes it compatible with PEDOT:PSS solutions [41,44], thereby facilitating solution processibility [45]. Moreover, it has been reported that 2D MXene can form efficient interfaces with the PEDOT:PSS matrix [39,46], which is beneficial for facilitating energy filtering effects [39]. In 2020, Ouyang et al incorporated 2D MXene into the PEDOT:PSS matrix and reported a dramatic enhancement in TE performance [39]. However, the investigation of how the interfacial area and the energy barrier between PEDOT:PSS and MXene affect the energy filtering effects and the resulting TE performance is highly needed, as it can further help to develop a novel strategy to enhance the TE performance of the PEDOT:PSS and MXene composites.
In this work, we investigate the effects of the interfacial area and the energetic barrier between PEDOT:PSS and MXene on the TE properties of their composite films. The incorporation of different sizes of MXene prepared by ball milling allows to observe the role of the interfacial area on the TE properties of the PEDOT:PSS-MXene composites. We found that a smaller size of MXene results in a higher TE performance based on a higher α. Analysis of the temperature-dependent σ (σ(T)) based on a three-dimensional (3D) variable range hopping (VRH) model reveals that the carriers are scattered more efficiently as the MXene size decreases due to the increased interfacial area between PEDOT:PSS and MXene. Additionally, by systematically controlling the energy barrier between the PEDOT:PSS and MXene using a sequential de-doping method, we could directly observe that the TE performance of the composite films is maximized when the energy level difference between two constituents reaches a specific value. We believe that our study provides valuable insight into the optimization of the interfacial area and the energetic barrier in PEDOT:PSS-MXene composite films, paving the way for the development of high-performance polymer and polymer-inorganic hybrid TE devices.

Sample preparation
Ball milling was performed to control the size of MXene. In detail, 1 g of MXene powder and 20 g of stainlesssteel balls were added to 10 ml of DMSO solution, and the mixtures were ball milled at 300 rpm for different durations (i.e., 3 h for 830 nm, 24 h for 230 nm, 72 h for 135 nm, respectively). Then, the shattered MXene was dispersed in DMSO and exfoliated by ultrasonication for an hour. The MXene-dispersed DMSO was then mixed into PEDOT:PSS aqueous solutions at a volume ratio of PEDOT:PSS to DMSO of 95:5. The solid weight ratios of MXene to PEDOT:PSS were 0 to 60 wt%. All substrates, including bare glass and indium-tin-oxide (ITO) glass, were cleaned by sonication with acetone, isopropyl alcohol, and deionized water. After the UV-ozone treatment for 15 min, the prepared PEDOT:PSS-MXene solution was spin-coated on the substrate (24 mm × 20 mm) at 2000 rpm for 30 s, followed by the thermal annealing at 150°C for 30 min, resulting in the thickness of 90-100 nm depending on the MXene ratios. To control the oxidation level of the PEDOT:PSS and PEDOT:PSS-MXene composite films, we placed the composite film in the preheated (60°C) home-made vacuum chamber filled with the TDAE vapor. The oxidation level of the composite films was controlled by varying the exposure time of the TDAE vapor. To characterize the TE performance, Au electrodes (50 nm) were deposited on the films by thermal evaporation under high vacuum conditions (<10 −6 torr) at a rate of 1.0 Å s −1 .

Characterization
The TE properties of the composite films were measured using a custom-built setup, consisting of the two Peltier modules controlled by a Keithley 2200 power supply and a Keithley 2601B source meter. The absolute temperatures of the hot and cold sides of the devices were measured using two T-type thermocouples connected to a Keithley 2000 thermometer. The generated thermovoltage was measured with a Keithley 2182 A nanovoltmeter. The α was calculated by performing a linear regression analysis on the measured voltage versus temperature difference data. The σ was obtained by measuring the current-voltage characteristics using the four-point probe methods with a Keithley 2400 source meter. To obtain the temperature-dependent σ, samples were mounted into a cryostat where the temperature varied from 100 K to 300 K using a proportional-integralderivative temperature controller (Model 331 Cryogenic Temperature Controller, Lakeshore). Field emission scanning electron microscopy (FE-SEM) and energy dispersive x-ray spectroscopy (EDS) images were obtained using a JSM-7800F Prime (JEOL). UV-vis-NIR absorption spectra were obtained from a JASCO V-730 spectrophotometer. Film thickness and surface morphology were characterized by atomic force microscopy (AFM, XE-100, Park Systems) in a non-contact mode. Work functions were characterized by Gaussian fitting of the surface potential contrast spectra obtained with the Pt-coated cantilever from scanning Kelvin probe microscopy (SKPM, NX-10, Park systems). The size of MXene was measured by dynamic light scattering (DLS).

Film properties of PEDOT:PSS-MXene composites
In order to vary the interfacial contact areas between Ti 3 C 2 MXene and PEDOT:PSS, different sizes of MXene were prepared by ball milling [47] for different durations (3, 24, and 72 h), resulting in average sizes of 135, 230, and 835 nm, respectively, as confirmed by the DLS data as shown in figure 1(a). A schematic illustration of the fabrication process of the PEDOT:PSS-MXene composite solution is shown in figure 1(b). The nano-sized MXene particles were dispersed in DMSO, and then mixed into the PEDOT:PSS solutions at different weight ratios (0 to 60 wt%), while maintaining a volume ratio of 5% for PEDOT:PSS and DMSO. By spin-coating the prepared solutions onto the glass substrate, we fabricated the PEDOT:PSS-MXene composite films containing different sizes of MXene (see the Materials and Methods section for details). The incorporation of MXene hardly change the film morphologies and surface roughness of the PEDOT:PSS films without MXene as observed in the AFM images shown in figure S1. The absorption spectra of the films also show no noticeable change in their spectral intensity at around 1200 nm, 900 nm, and 600 nm, which correspond to the bipolaron, polaron, and neutral state, respectively, regardless of the size of MXene, as shown in figure S2. This indicates that the electronic state of PEDOT:PSS is not changed by MXene. To confirm the uniform distribution of MXene in the PEDOT:PSS matrix, the surface of the PEDOT:PSS-MXene (10 wt%) composite films was investigated by FE-SEM, as shown in figures 1(c)-(e). It can be observed that the MXene nanoparticles are well mixed; the smaller MXene particles are more widely distributed within the films at the same concentration. Figures 1(f)-(h) shows the EDS mapping profile of Ti atoms, further showing that the Ti 3 C 2 MXene particles are well distributed in the PEDOT:PSS matrix. Thus, it is expected that the smaller MXene will form more interfacial area with PEDOT: PSS. The effect of the MXene size on the TE properties will be discussed below.

TE performance of PEDOT:PSS-MXene composites
To investigate the TE properties of the PEDOT:PSS-MXene composite films, we measured the α and σ using a customized setup as shown in figure 2(a). The temperature gradient was produced on the films using two Peltier devices, and the absolute temperature T and the generated voltage were measured simultaneously using two T-type thermocouples and the Au probes as described in the Materials and Methods section [5,10]. Figures 2(b) -(d) display α, σ, and PF of the PEDOT:PSS-MXene composite films as a function of the MXene concentration. As shown in figure 2(b), α of the films continuously increases with the MXene concentration and then plateaus at a certain concentration, i.e., 20 wt% for 135 nm, 20 wt% for 230 nm, and 40 wt% for 830 nm. The films composed of smaller MXene show higher α at the same concentration due to the increased surface-to-volume ratios; the highest α of the films containing 135 nm MXene is 32.4 μV K −1 , which is approximately 53% higher than that of the film without MXene (21.2 μV K −1 ). Meanwhile, σ presents an opposite trend to α along with the size and concentration of MXene, as shown in figure 2(c)-σ consistently decreases as a function of the MXene concentration and the composite film with smaller MXene exhibits lower σ at the same concentration. As a result, the PEDOT:PSS-MXene films with 135 nm-sized MXene exhibit a much higher PF (54.9 μW m −1 K −2 ) than the films with larger MXene. These results demonstrate that the size of MXene has a significant effect on the TE performance of the PEDOT:PSS-MXene composite films. This difference in the TE performance may be due to the effect of the interfacial area on the charge transport, especially at the interface between PEDOT:PSS and MXene. No noticeable change in the UV-vis absorption spectra by incorporating MXene implies that the difference is not attributed to the charge of electronic states (figure S2). In particular, the higher α and lower σ observed in the films with 135-nm-sized MXene suggest that the energy filtering occurred more efficiently at the interfaces of PEDOT:PSS-MXene than the films with larger MXene. This curiosity prompted us to further investigate how the MXene size affects the charge transport properties of the composite films, which was studied using a temperature-dependent σ measurement as shown below. Figure 3(a) plots the temperature-dependent σ of the PEDOT:PSS film without MXene and the PEDOT:PSS-MXene films with different sizes of MXene (10 wt%) over the temperature range of 100 to 300 K. In all samples, a thermal activation behavior of σ(T), i.e., a constant increase of σ with increasing temperature [48], is measured. Here, the σ(T) data of the films fit well with the 3D VRH model expressed as σ(T) = σ 0 exp[−(T 0 /T) 1/4 ], where σ 0 and T 0 are the electrical conductivity at the infinite temperature and the characteristic temperature, respectively [49]. Thus, the higher T 0 value, the higher the energy barrier between localized states. To investigate the effect of the MXene size on the charge transport, we extracted the T 0 values of the samples, as shown in figure 3(b). The T 0 value increases with the addition of MXene, and interestingly shows an inversely proportional relationship to the MXene size. The increased T 0 indicates that the energy barrier that charge carriers must overcome to be transported becomes larger [39,50], which well explains the decrease in σ and increase in α depending on the MXene size. The similar trend of α and T 0 with respect to the MXene size supports our suggestion that the energy filtering is more efficient in the composite films with smaller MXene compared to those with larger MXene. Based on the results, we can propose how the MXene size affects the energy filtering effect and the TE performance. Figures 3(c) and (d) schematically show the charge transport processes in PEDOT:PSS-MXene composite films with smaller and larger MXene, respectively. As the MXene size becomes smaller, the effective interfacial area between PEDOT:PSS and MXene becomes larger at the same MXene concentration, as confirmed by the EDS data, which promotes the energy filtering effect. Conclusively, this leads to higher α and PF values in the PEDOT:PSS-MXene composite film with smaller MXene.

TE performance optimization by energy level tuning
As shown above, the TE performance can be improved by introducing an energy filtering effect. In addition, it is reported that the TE performance, particularly α that is proportional to the average entropy transported per charge-carrier [29], can be further enhanced by selectively accelerating the low-energy carrier scattering [28,32]. The main idea is to control the energetic barrier between two materials, i.e., PEDOT:PSS and MXene. For this, we tuned the oxidation level of the composite films using a TDAE vapor treatment method by varying the treatment time [51]. Figure 4(a) shows the TE properties of the PEDOT:PSS-MXene composite films (135 nm, 10 wt%) as a function of the TDAE treatment time. The change in α of the composite film as a function of the TDAE treatment time is distinguishable from that of the PEDOT:PSS film, indicating that the changes in α and TE performance are not solely due to the de-doping effect of PEDOT:PSS by TDAE (see figure S3). As the treatment time increases from 0 to 2 min, α increases from 31.0 μV K −1 to 38.1 μV K −1 ; however, α turns to a decrease when the treatment time exceeds 2 min, as low as ∼ 29 μV K −1 . Meanwhile, the absorption spectra plotted in figure S4 show the decreasing intensity at around 1200 nm and the increasing intensity at around 600 and 900 nm, indicating that the bipolaron states in the PEDOT chains change to neutral and polaron states. In other words, the TDAE treatment continuously oxidizes the PEDOT:PSS-MXene composite film, so that σ decreases continuously up to the TDAE treatment time of 3 min and is almost maintained with increasing treatment time. Considering the result, the rapid change of α at 2 min is thought to be more related to the   energetic barrier between PEDOT:PSS and MXene rather than to the carrier concentration. This can be verified by measuring the change in the work function (Φ) of the composite film depending on the TDAE treatment time using the SKPM, as shown in figure 4(b). The Φ values were obtained by fitting the surface potential contrast data to the Gaussian function, as shown in figure S5 [10]. As the treatment time increases, Φ of the composite film consistently decreases from 5.08 to 4.76 eV, approaching Φ of the neat MXene film (4.74 eV). Interestingly, we could observe that the difference in the work function (ΔΦ) between PEDOT:PSS and MXene almost disappears (<0.04 eV) after the TDAE treatment time beyond 2 min as shown in figure 4(c). Because a small ΔΦ close to 0 can hardly contribute to the energy filtering effect, resulting in the rapid decrease of α after 2 min. We also should note that the highest PF (69.4 μW m −1 K −2 ) was obtained when ΔΦ is 0.14 eV (at 2 min), which is similar to the barrier height that caused an efficient energy filtering effect previously [52,53]. When ΔΦ is larger (< 2 min), high energy carriers can be scattered in part, resulting in reduction of α and PF due to the decreased average entropy transported per charge-carrier [54]. Based on our systematic study, we could verify that the energy filtering effect can be further optimized by tuning the energy level of composite films.

Conclusion
In this work, TE composites with enhanced performance were demonstrated by incorporating nano-sized 2D MXene into the PEDOT:PSS matrix by exploiting the energy filtering effect. In the PEDOT:PSS-MXene composite film, the effects of the interfacial area and the energy barrier between PEDOT:PSS and MXene were systematically investigated by thoroughly manipulating the size of MXene particles and the oxidation level of the composite films. The films with smaller MXene exhibited higher TE performance as a result of increased interfacial area between PEDOT:PSS and MXene, which in turn enhances low-energy charge carrier scattering. An additional improvement in TE performance was accomplished by tuning the energy barrier difference between PEDOT:PSS and MXene. Conclusively, we could verify that both the interfacial area and the energy barrier between two constituents of a composite play a crucial role in the energy filtering effect-induced enhancement of the composite-based TE properties. We believe that these findings will contribute to the design and development of polymer-based high-performance TE devices for the self-powered devices and sensors.

Acknowledgments
This work is supported by the National Research Foundation of Korea (NRF) grants (NRF-2022R1C1C1010152 and NRF-2022R1A6A3A13069150) funded by the Ministry of Science and ICT and Ministry of Education, respectively.

Data availability statement
The data cannot be made publicly available upon publication because they are not available in a format that is sufficiently accessible or reusable by other researchers. The data that support the findings of this study are available upon reasonable request from the authors.

Conflict of interest
There are no conflicts to declare.