Preparation, morphology and thermoelectric performance of PEDOT/CuI nanocomposites

Abstract

Incorporating inorganic nanomaterials into a polymer matrix is one of the most effective ways to create thermoelectric performance for applications where physical flexibility is essential. In this study, flexible thermoelectric nanocomposite films were synthesized by incorporating inorganic copper iodide (CuI) nanosheets as the filler into poly (3,4-ethylene dioxythiophene): poly (styrene sulfonate) (PEDOT: PSS). The process involved the preparation of bulk CuI from precursors and, subsequently, the nanosheet synthesis by dissolving the bulk CuI in dimethyl sulfoxide (DMSO). The morphology of the nanosheets and the nanocomposite films was thoroughly examined, and the film’s thermoelectric performance was evaluated using a standard thermoelectric measurement system, ZEM-3. The morphological observation revealed a triangular nanosheet geometry for CuI, with an average lateral dimension of ~33 nm. The PEDOT/CuI nanocomposite films were prepared by mixing CuI nanosheets with PEDOT: PSS through ultrasonication and filtration on a PVDF membrane. The film with 6.9 vol% of CuI nanosheets exhibited an electrical conductivity and Seebeck coefficient of 852.07 S·cm^-1 and 14.95 µV·K^-1, respectively. This resulted in an enhanced power factor of 19.04 µW·m^-1·K^-2, much higher than the individual composite components. It demonstrated a trend of increasing power factor with the nanosheets up to 6.9 vol% due to improved electrical conductivity. The increase in electrical conductivity can be attributed to the screening effect induced by DMSO, which leads to a conformational change in the PEDOT chains. Furthermore, an optimal fraction of CuI nanosheets also contributes to this conformational change, further enhancing the electrical conductivity.

Graphical Abstract

Introduction

Since naturally available energy sources as the primary fuel are limited, several renewable technologies have been established as alternatives over the past fifty years. However, all these technologies suffer a healthy fraction of energy loss when converted to electricity, i.e., heat conduction loss. Moreover, wastage is as high as 70% due to inefficiencies in converting energy to electricity from traditional sources such as petroleum, natural gas, and coal [1, 2]. From this perspective, thermoelectric materials have the potential to recover a considerable amount of wasted energy by converting heat to electricity.

Although thermoelectric technology holds such promise, it is limited by low energy efficiency. The synthesis processes are usually complex and involve high temperatures with toxic and rare earth elements. To address these issues, many studies have been conducted to enhance thermoelectric efficiency over the last several decades, focusing on appropriate material selection, rational design, and more importantly, facile synthesis techniques for cost-effective thermoelectric materials [3].

The performance of thermoelectric materials can be realized by a unitless quantity called the Figure of merit, ZT. It is expressed as (S²σ/k) *T, where S, σ, k, and T are the Seebeck coefficient, electrical conductivity, thermal conductivity, and absolute temperature, respectively. Organic materials and their composites typically exhibit very low thermal conductivity, and in many cases, accurate measurement is challenging. As a result, the term S²σ, known as the power factor, is often used to evaluate the thermoelectric performance of polymer nanocomposites [4,5,6]. However, achieving high-performance thermoelectric materials is challenging since the Seebeck coefficient and electrical conductivity are interdependent; increasing one often results in the decrease of the other. Thus, decoupling these parameters poses a significant hurdle in developing highly efficient thermoelectric materials.

The past years have witnessed a remarkable improvement in thermoelectric materials, especially in obtaining inorganic thermoelectric materials with high ZT values. For example, a high ZT value of 2.6 has been reported in tin selenide by Zhao et al. [7]. On the other hand, organic conducting polymers have also shown promising thermoelectric performances. They can resolve the shortcomings (e.g., being rare, expensive, and toxic) of inorganic materials. In addition, they have favorable properties of low density and cost and ease of synthesis and processing into versatile forms.

Capable of recovering low-grade waste heat, conducting polymers and their nanocomposites are often synthesized from abundant elements using low-cost manufacturing techniques. Their inherent low thermal conductivity and flexibility could pave the way for new applications and device designs. Noteworthy is that flexibility has been achieved by converting bulk semiconductor particles into nanosheets, such as antimonene and bismothene [8, 9]. Due to its outstanding electrical conductivity, PEDOT: PSS has been extensively studied not only for supercapacitors [10,11,12] but also for thermoelectric applications [5, 6, 13,14,15,16].

Organic/inorganic nanocomposites can potentially harness the combined benefits of two crucial components: the high electrical conductivity of conducting polymers and the elevated Seebeck value of inorganic thermoelectric materials, resulting in a significant increase in the power factor. Consequently, numerous research studies have been conducted to synthesize and explore the potential of these organic/inorganic nanocomposites for enhancing thermoelectric performances [17,18,19]. Most of the composites consist of inorganic fillers, and the fillers in most cases are toxic and perform well only at high temperatures. To overcome these issues, room-temperature and non-toxic inorganic thermoelectric materials (e.g. Bi₂Te₃ [20, 21], Cu₂Se [22]) and their composites (e.g. Bi_0.5Sb_1.5Te₃/PEDOT: PSS [23]) have been explored in recent years. Nevertheless, extensive research is required to enrich the library of room-temperature thermoelectric materials.

Copper iodide (CuI) is an environment-friendly p-type semiconductor compound that exhibits promising Seebeck coefficient at room temperature [24, 25]. Salah et al. reported that Sn-doped CuI exhibited a Seebeck coefficient of 1260 S·cm^-1 at room temperature [26]. Lately, Xiaowen’s group prepared a CuI/nylon film demonstrating a high Seebeck coefficient of 600 µV·K⁻¹ [27]. On the other hand, PEDOT: PSS is a well-established, highly electrically conductive polymer that has been reported to exhibit an electrical conductivity of 0.8–1418 S·cm^-1 [28, 29]. However, despite the promising properties of both PEDOT: PSS and CuI, there has been no previous research on the development of PEDOT/CuI nanocomposites for thermoelectric applications. Hence, motivated by the impressive room temperature Seebeck of CuI and the electrical conductivity of PEDOT: PSS, this work sets out to synthesize and investigate the thermoelectric performance of PEDOT/CuI nanocomposites.

In this work, bulk CuI is prepared from precursors, and then the CuI nanosheets are synthesized by dissolving bulk CuI in DMSO. The PEDOT/CuI nanocomposites are synthesized by physical mixing with ultrasonication. The compositions and the microstructures of the inorganic filler and its nanocomposite films are investigated by X-ray diffraction, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The effect of CuI fractions on thermoelectric performance is studied. The Seebeck coefficient, electrical conductivity, and power factor are measured by a standard facility called ZEM-3 and the obtained results were used to propose the mechanisms of improved thermoelectrical properties of PEDOT/CuI.

Experimental methods

Materials

PEDOT: PSS solution (1.3 wt%, conductive grade) was purchased from Sigma-Aldrich Pty Ltd. Precursors for copper iodide (CuI), i.e. copper sulphate pentahydrate (CuSO₄·5H₂O) and potassium iodide (KI), were supplied by Southern Cross Science Pty Ltd. PVDF membrane (with a nominal pore size of 0.22 µm) and PTFE filter (with a nominal pore size of 0.45 µm) were bought from Proscitech Pty Ltd. All the materials were used without further purification.

Synthesis of bulk CuI

CuI was prepared from precursors by a facile precipitation method based on the following reaction.

$$2\;{\mathrm{CuSO}}_{4} \cdot 5\;{\mathrm{H}}_{2}\mathrm{O }+ 4\;\mathrm{ KI }\to 2\;\mathrm{ CuI }\downarrow + {\mathrm{I}}_{2} + 2\;{\mathrm{K}}_{2}{\mathrm{SO}}_{4} + 10\;{\mathrm{H}}_{2}\mathrm{O}$$

In specific, 655.51 mg of CuSO₄·5H₂O was dissolved in 150 mL of DI water in a beaker by magnetic stirring. In another beaker, a stochiometric quantity of KI was dissolved in 75 mL of DI water by magnetic stirring. After that, the dissolved KI was added dropwise to the first beaker. The precipitation was collected by centrifugation followed by washing with DI water and ethanol for several times to remove K₂SO₄. Finally, bulk CuI particles were dried in a vacuum oven at 60ºC for 8 h.

PEDOT/CuI nanocomposite film fabrication

The preparation procedure of PEDOT/CuI nanocomposites is schematically illustrated in Fig. 1. Aqueous PEDOT: PSS solution was first filtrated with a PTFE membrane to remove large particles. A specific amount of bulk CuI particles (1.9–10.4 vol%) was then added into 20 µL of DMSO, followed by ultrasonication at 200 W for 20 min; 2 mL of DI water was added dropwise, followed by addition of 200 µL of the filtrated PEDOT: PSS. The above steps were executed to form nanosized CuI, and the mechanisms will be discussed in detail in the results and discussion section. Afterward, the mixture was further sonicated for 20 min to obtain a visually uniform mixture of CuI and PEDOT: PSS. The mixture was then treated by vacuum filtration on a PVDF membrane, resulting in a PEDOT/CuI nanocomposite film. Finally, the nanocomposite film was dried at 80ºC for 4 h in a vacuum oven prior to characterization and measurement. As will be discussed later, approximately 19.50% of PSS molecules were removed during the filtration and drying process.

To convert wt% to vol%, the densities of CuI and PEDOT: PSS were taken as 5.67 g cm-3 and 1.18 g cm-3. In this way, we converted wt% to vol% as shown in Table 1.

Fig. 1

Schematic representation of the PEDOT/CuI nanocomposite film preparation process

Table 1 Conversion of wt% to vol% of CuI in the PEDOT/CuI nanocomposites

Characterization

X-ray diffraction (XRD) analysis was performed using an Empyrean diffractometer (Malvern PANalytical) with CuKα radiation (wavelength = 1.54 Å) operating at 40 kV and 30 mA. The scanning rate was set at 2°/min, covering a range of 4–80°. Scherrer's formula was utilized to estimate the crystallite size of the CuI nanosheets from the XRD data. X-ray photoelectron spectroscopy (XPS) measurements were conducted on a Kratos XPS instrument equipped with an Al monochromatic X-ray source (hν = 1486.6 eV) operating at 10 kV. The survey scan was performed with a pass energy of 160 eV, while high-resolution scans utilized a pass energy of 20 eV. The chamber's base pressure during the analysis was maintained at 5.3 × 10⁻⁶ Pa. Raman spectra of both the pristine PEDOT: PSS and the PEDOT/CuI nanocomposite were recorded using a HORIBA LabRAM HR Evolution Raman spectrometer. The measurements covered a scan range of 1000-1800 cm⁻¹, and a 532 nm laser was used for excitation. Carrier concentration and mobility were measured using an Ecopia Hall Effect Measurement Systems at room temperature.

Morphology and microstructure analyses were conducted using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The SEM observations were performed utilizing a Zeiss Merlin FEG SEM. Meanwhile, TEM characterization was carried out using a JEOL JEM-2100F-HR instrument. For TEM sample preparation, a 0.02 g DMSO-CuI solution was diluted with 10 mL of ethanol and then subjected to 5 minutes of ultrasonication. This process ensured the uniform dispersion of CuI nanosheets in the solvent, facilitating their successful deposition onto the copper grid with holey carbon support for TEM characterization.

Thermoelectric measurement

A four-point probe method was used to measure the electrical conductivity of all the samples. Temperature-dependant electrical conductivity, Seebeck coefficient, and power factor were measured by a commercial measurement system called “Seebeck Coefficient/Electric Resistance Measuring System (ZEM-3M10, ADVANCE RIKO, Yokohama, Japan)” in He atmosphere, with a test error of ±5%. The thickness of the film on the PVDF membrane was measured by a digital micrometre and an optical microscope. The thickness of the neat PVDF membrane is 100 ± 25 µm, while the pristine and nanocomposite films have thicknesses of 2.7 ± 0.5 µm.

Results and discussion

Formation of CuI nanosheets

An essential part of this study is to prepare CuI nanosheets. The morphology and the crystallinity of as-prepared CuI nanosheets were studied by TEM and selected area electron diffraction (SAED). As shown in the Fig. 2a, the TEM image of nanosized CuI exhibits a triangular-like nanosheet structure, and the average lateral dimension of the nanosheets is ~33 nm. Further probing of the sample using high-resolution TEM (HRTEM) in Fig. 2b−c shows the presence of crystalline particles with a lattice spacing of 0.347 nm, corresponding to the (1 1 1) plane of CuI (JCPDS no. 06-0246). In addition, Fig. 2d shows the SAED images exhibiting sharp spots, which are due to the crystalline nature of the nanosheets, as supported by a previous work [30].

Fig. 2

a triangular-like nanosheet structure; b HRTEM image, c crystalline lattice spacing, and d SAED pattern of CuI nanosheets

To verify the successful synthesis of CuI nanosheets, XRD measurement was carried out. The XRD pattern of CuI nanosheets in Fig. 3a is consistent with the pure γ-CuI reported in the literature [31, 32]. Moreover, all the major peaks can be indexed as the standard γ-CuI (JCPDS# 06-0246). High crystallinity of the CuI nanosheets can be demonstrated from the sharp and strong diffraction peaks on the XRD pattern of CuI nanosheets. Using the following Scherrer’s formula [23, 24] and all the dominant diffraction peaks from XRD data, we calculated the average crystallite size (d) of as-prepared CuI nanosheets to be ~30 nm, which is close to what we observed under TEM.

Fig. 3

XRD patterns for a as prepared CuI and its corresponding peaks from JCPDS card, and b films of neat PEDOT: PSS and PEDOT/CuI nanocomposite at 6.9 vol%

$$d=\frac{k\lambda }{\beta\;\mathrm{cos}\theta }$$

Where λ represents the X-ray wavelength, β means the FWHM width of a diffraction peak, θ refers to the diffraction angle, and k is constant. Table 2 provides the average crystallite size calculated for the as-prepared CuI nanosheets, by utilizing the corresponding peaks, 2-theta values, FWHM widths, and calculated sizes as reference points.

Table 2 Average crystallite size of as-prepared CuI nanosheets

The formation of CuI nanosheets was explained as below. As discussed in the experiment section, bulk CuI particles were synthesized from precursors. To prepare CuI nanosheets, bulk CuI was dissolved in DMSO by sonication for 20 min, resulting in a CuI-DMSO solution of pale-yellow colour. When DI water was gradually added, CuI nanosheets would be produced through the decomposition of the [Cu(DMSO)₄]⁺ complex formed in this solution — a passivation layer of Cu⁺–DMSO adsorbed on the particle surface was suggested to be responsible for the nanoscale formation. Moreover, DMSO in the solution can still react with the particle surfaces and would act as a capping agent to stabilize each particle by forming a layer of Cu⁺–DMSO, preventing them from aggregating. A similar phenomenon was observed when CdS nanoparticles were synthesized in DMSO where the Cd²⁺–DMSO complex acted as a capping agent for the formation of CdS nanoparticles [33]. Other organic solvents such as dimethyl formamide (DMF) played a similar role in forming CuI nanoparticles [24, 34, 35].

PEDOT/CuI nanocomposites

The dispersion of CuI nanosheets in the PEDOT: PSS matrix was investigated by SEM. Figure 4a–c contains micrographs of the PEDOT/CuI nanocomposite at 1.9 vol%. In Fig. 4a, micron-size clusters and many isolated particles are observed. When magnified in Fig. 4c, the aggregated particles are found to contain a few CuI nanosheets. Figure 4d–f demonstrates similar morphology of the dispersed CuI nanosheets. However, as can be seen in Fig. 4g−i, the morphology of the nanocomposites at 6.9 vol% of CuI shows more aggregated clusters of CuI nanosheets having larger dimensions. Finally, the aggregation in the nanocomposites at 9.2 vol% in Fig. 4j–l seems more severe. Overall, although CuI is well-composited with the PEDOT: PSS matrix, it can clearly be observed that when the fraction of CuI was 6.9 vol% and above, not all particles are properly embedded within the matrix – some sit on the surface of the PEDOT: PSS matrix.

Fig. 4

SEM images of PEDOT/CuI nanocomposites: (a–c) 1.9 vol% of CuI, (d–f) 4.8 vol%, (g–i) 6.9 vol%, and(j–l) 9.2 vol%

The sample was prepared by cutting with a new razor blade and fracturing at room temperature. Since the matrix has a glass transition temperature far higher than room temperature, the matrix should not deform or deform very little during the sample preparation. The CuI particles observed on the surface mean that the interfacial bonding between the particles and the matrix is not sufficiently strong. As the CuI fraction increases, many micron-size CuI particles are observed, mainly due to the aggregation of CuI nanosheets via the drying process.

To further confirm the formation of a PEDOT/CuI nanocomposite, EDS analysis was performed. Figure 5a shows a backscattered electron image of the nanocomposite film. During the EDS measurement, different areas were focused, and the corresponding peak of one such area is shown in Fig. 5b. All the elements from PEDOT: PSS (e.g., C and S) and CuI (Cu and I) can be seen in the synthesized nanocomposite in the EDS spectrums. In addition, the EDS mapping shown in Fig. 5c-g also indicates the presence of S, C, Cu, and I. The I/Cu atom ratio in the nanocomposite film was calculated to be 1.28, which reasonably agrees with the stoichiometric ratio in CuI.

Fig. 5

a Backscattered electron image with b atomic percentage; EDS mapping showing (c) layered image, and the distribution of d sulfur, e carbon, f copper, and g iodine within the nanocomposite

Moreover, XRD analysis was performed to further characterize the nanocomposite as briefly discussed in Fig. 3. Figure 3b illustrates the XRD pattern of the nanocomposite at 6.9 vol%, and it exhibits diffraction peaks that appear to be a combination of the patterns observed for both CuI nanosheets and neat PEDOT: PSS. The overlapping peaks indicate successful dispersion and integration of CuI nanosheets into the PEDOT: PSS matrix. As summarized in Table S1, shifts in peak positions and changes in intensities at specific 2θ angles, such as at 2θ = 18.28 and 26.44 degrees, suggest structural modifications resulting from the interaction between CuI nanosheets and PEDOT: PSS. The preservation of certain characteristic peaks at 2θ = 4.04, 44.62, 65.00, and 78.12 degrees further confirms the successful formation of the nanocomposite, with CuI nanosheets distributed within the PEDOT: PSS matrix. These XRD findings provide strong evidence for the effective synthesis of the PEDOT/CuI nanocomposite, showcasing its distinctive crystalline features and successful incorporation of CuI nanosheets within the PEDOT: PSS matrix.

Thermoelectric performance of PEDOT/CuI nanocomposite film

Figure 6a shows the electrical conductivity of the PEDOT/CuI nanocomposite film at CuI fractions ranging from 0–10.4 vol%. Initially, the DMSO-treated PEDOT: PSS film exhibits a conductivity of approximately 9.0 S·cm^-1 at room temperature. However, the addition of CuI nanosheets leads to a notable increase in electrical conductivity. For example, the conductivity increases from 35 ± 19 S·cm^-1 at 1.9 vol% of CuI to a maximum value of 852 ± 195 S·cm^-1 at 6.9 vol%. A similar increase was reported for PEDOT: PSS compounded with Bi₂Te₃ [36, 37]. This conductivity improvement is explained below.

Fig. 6

a Electrical conductivity, b Seebeck coefficient, c power factor, d carrier concentrations and mobility of PEDOT: PSS and PEDOT/CuI; and (e) the formation process of linear structured PEDOT/CuI

Firstly, DMSO, an organic solvent, weakens the coulombic interaction between PEDOT (+) and PSS (-), resulting in increased electrical conductivity of PEDOT: PSS due as supported by [38]. Moreover, adding CuI nanosheets could create a conformational change to the PEDOT chains and also provide carriers for PEDOT: PSS to attach [37]. As a result, PEDOT macromolecules would change from coiled to linear structure, making the charge carriers move along the polymer chains more efficiently, as illustrated in Fig. 6e.

However, the conductivity decreases at higher CuI fractions (above 6.9 vol%). The reduction can be attributed to the less effective dispersion of CuI nanosheets due to the excess amount of CuI in the matrix, which would result in fewer “accelerator” loci, making the charge carrier transport more challenging [39]; that is, excessive CuI nanosheets act as a barrier between conductive PEDOT chains, decreasing the conductivity.

To investigate whether DMSO and CuI nanosheets pose an effect on the inner composition of the PEDOT: PSS film, X-ray photoelectron spectroscopy (XPS) measurements were performed. The S(2p) XPS spectra of neat PEDOT: PSS and its nanocomposite films are compared in Fig. 7a. Neat PEDOT: PSS exhibits two signal bands at 158−162 (doublet peaks) and 162−166 eV, which are assigned to the sulphur atoms of PEDOT and PSS, respectively. The use of DMSO and CuI has shifted the peaks a little bit to the lower energy and strengthened the signals of both PEDOT and PSS chains, indicating that the PEDOT and PSS chains are more separated. The possible compositional change in the nanocomposite films is probed by the PEDOT-to-PSS ratio calculated according to the integrated peak areas of the S(2p) [40], as summarised in Table 3.

Fig. 7

a S(2p) XPS spectra, and b Raman spectra of PEDOT: PSS and PEDOT/CuI

Table 3 Ratios of PSS to PEDOT from the area under XPS graphs

Normalized XPS data was used to evaluate the PSS to PEDOT ratio for PEDOT: PSS and its composites. Although UV-vis spectroscopy was also found useful in this regards as reported by Mardi et al [41], XPS is more often used due to higher accuracy [42, 43]. In Table 3, the ratio of PSS to PEDOT reduces as the CuI fraction increases. For the pristine polymers, the ratio was measured as 2.2275. When DMSO was used with CuI nanosheets added, the ratio started to reduce down to 1.7934 for the nanocomposite at 6.9 vol% of CuI. These results suggest that the nanocomposite composition has evolved during the preparation, likely corresponding to removal of a considerable amount (around 19.50%) of PSS component from PEDOT: PSS. This eventually resulted in an improvement in the electrical conductivity of the nanocomposite film. Nevertheless, the PSS to PEDOT ratios of the nanocomposite films tend to stop decreasing at over 6.9 vol% of CuI. This result is consistent with the trend of electrical conductivity in Fig. 6a.

Raman analysis has been employed to investigate the conformational change in the structure of PEDOT chains. In Fig. 7b, the peaks for the PEDOT/CuI nanocomposite are slightly shifted towards the red region in comparison with those for PEDOT: PSS. The shift indicates a conformational change in the structure of the PEDOT chains from a coiled to an extended coiled or linear structure [44, 45]. This change is likely due to the presence of both the CuI nanosheets and DMSO. The linear structure of the PEDOT chains is known to have a positive effect on the electrical conductivity of their nanocomposite [46]. Additionally, the XRD patterns in Fig. 3b indicate noticeable changes in the crystallization or arrangement of PEDOT and PSS chains in the nanocomposite film. Although the XRD patterns for the nanocomposite film are observed at similar positions to pristine PEDOT: PSS, the pattern intensities are significantly altered. A summary of the intensity and peak position changes can be found in Table S1. These changes in the XRD and Raman data suggest that the PEDOT macromolecules have undergone a conformational change from a coiled to a linear structure in the nanocomposite.

Hall measurements were performed to investigate the charge carrier concentrations and mobility of the nanocomposites. In Fig. 6d the carrier concentration increases with the CuI fractions up to 6.9 vol%, indicating the beneficial effect of CuI in enhancing the number of charge carriers available for conduction. This increase in carrier concentration contributed significantly to the observed enhancement in electrical conductivity.

However, the carrier concentration at 9.2 vol% of the nanosheets decreases compared to the 6.9 vol%. It is attributed to the formation of large clusters or agglomerates of CuI nanosheets, hindering the efficient interaction between PEDOT and CuI, thus impeding the charge transport.

On the other hand, the carrier mobility exhibits varying behavior with CuI fractions. The presence of numerous heterointerfaces and different phases in the nanocomposites would influence the mobility of charge carriers, leading to variations in transport behavior. Some nanocomposites show a decrease in mobility, while others display an increase.

Considering these findings, we affirm that the addition of CuI nanosheets to the PEDOT: PSS matrix significantly enhanced the electrical conductivity through both the increase in carrier concentration and the formation of a conductive network within the matrix. The use of DMSO treatment also plays a crucial role in the conductivity improvement.

Regarding the Seebeck coefficient, the nanocomposite films exhibited a slight fluctuation in the range of 14.1 to 18.4 µV·K^-1 with varying fractions of CuI nanosheets. In Fig. 6b, the DMSO-treated PEDOT: PSS film exhibits a Seebeck coefficient value of 17.7 µV·K^-1, whereas the nanocomposite film with 6.9 vol% of CuI nanosheets shows 14.95 µV·K^-1. The maximum value of 18.4 µV·K^-1 is exhibited by the nanocomposite with 10.4 vol% of CuI nanosheets. These values are lower in comparison with the Seebeck coefficient of our neat CuI film which was measured as 194 µV·K^-1 at room temperature. A similar phenomenon was reported by [36, 37, 47] when Bi₂Te₃ was composited with PEDOT: PSS. The reason was attributed to that the electrically connecting junctions between the Bi₂Te₃ particles were weakened or destroyed, resulting in no obvious enhancement of the Seebeck coefficient for the PEDOT: PSS nanocomposite films containing Bi₂Te₃. Our group is currently investigating novel strategies to further enhance the Seebeck coefficient of the nanocomposite film.

Figure 6c contains the power factors of PEDOT: PSS and its nanocomposite films. At room temperature, the DMSO-treated PEDOT: PSS film without CuI shows a power factor of 0.27 µW·m^-1·K^-2. As the CuI nanosheets are added, the power factor gradually increases, reaching a peak at 6.9 vol% of CuI nanosheets. However, beyond this concentration, the power factor starts to decrease, reaching 11 ± 1.15 µW·m^-1·K^-2 at 7.8 vol%. Since the power factor is calculated by PF=S²σ and Seebeck coefficient remains relatively stable with the CuI fraction, the electrical conductivity of the nanocomposite films plays a crucial role in determining the power factor. Therefore, the nanocomposite film containing 6.9 vol% of CuI exhibits the highest power factor with a mean value of 19.04 µW·m^-1·K^-2, corresponding to the highest electrical conductivity of 852.07 S·cm^-1. This performance is compared with similar nanocomposites that utilize non-toxic elements and avoid post-treatment. The results show comparable performance, as presented in Table S2 in the supplementary document.

Conclusion

In conclusion, flexible thermoelectric PEDOT/CuI nanocomposite films were prepared by vacuum filtration on a PVDF membrane. The CuI nanosheets, synthesized from non-toxic precursors, were incorporated into the PEDOT: PSS matrix using ultrasonication. The broad range of CuI concentrations (1.9–10.4 vol%) in nanocomposite film was explored, where the nanocomposite film with 6.9 vol% CuI nanosheets exhibited optimal performance. Despite its relatively low Seebeck coefficient of 14.95 µV·K^-1, the nanocomposite exhibited an enhanced power factor of 19.04 µW·m^-1·K^-2 at room temperature, surpassing the performance of the individual constituents of the nanocomposite. This enhancement in the power factor can be attributed to the improved electrical conductivity of 852.07 S·cm^-1. These findings on PEDOT/CuI nanocomposite films as new thermoelectric nanocomposites are showing promising results that would advance research in the field of polymer/inorganic nanocomposites for thermoelectric applications.