Preparation and Properties of Flexible CuI/Polyvinylpyrrolidone Nanocomposite Thermoelectric Film

Abstract

A facile preparation method for flexible p-type CuI/polyvinylpyrrolidone (PVP) nanocomposite thermoelectric (TE) film is developed. First, CuI powder was synthesized by a one-pot method; second, PVP was coated in situ with the CuI powder; third, the CuI/PVP nanocomposite film was prepared on a nylon membrane by vacuum filtration and then hot-pressing. Transmission electron microscopy (TEM) observation indicates that the film consists of CuI nanograins with an average size of ~15 nm and PVP distributed at the inner wall of nanopores and the surface of the CuI nanograins. The composite film shows a large Seebeck coefficient of ~605 µVK⁻¹ and a power factor of ~8.05 µWm⁻¹K⁻² at 300 K. The nanocomposite film also exhibits excellent flexibility (~96% of initial electrical conductivity retention after being bent 1000 times along a 4 mm radius rod). A single-leg TE module outputs a voltage of ~3.6 mV when the temperature difference is 6 K. This work provides a fast, simple, and environmentally friendly method by which to prepare flexible CuI/PVP nanocomposite TE film with a large Seebeck coefficient, which could be used as a wearable sensor.

1. Introduction

Owing to the excessive pursuit of industrialization, the global energy crisis is becoming more and more serious, and the use of fossil fuels has also caused serious harm to the environment, such as global warming, acid rain, pollution of the soil environment, and animal extinction. Therefore, it is urgent to explore sustainable new energy to replace traditional fossil fuels [1,2]. Thermoelectric (TE) material is a kind of material with wide application prospects. It can convert heat energy into electrical energy directly by the transport of carriers within solids under a low-temperature difference [3]. The performance of TE materials is evaluated by a dimensionless figure of merit, ZT = S²σTk⁻¹, where S is the Seebeck coefficient, σ is the electrical conductivity, T is the absolute temperature, k is the total thermal conductivity, and S²σ is called the power factor (PF). A good TE material should have a high electrical conductivity, a large Seebeck coefficient, and a low thermal conductivity. However, these three parameters are strongly coupled to each other and hence it is a great challenge to obtain a high ZT value. Various approaches have been chosen to improve TE performance, such as doping, nanostructuring, annealing, alloying, and so on, and much progress has been made in several material systems [4], such as SnSe [5,6], Cu₂Se [7,8], and GeTe [9,10].

In recent years, the demand for wearable electronic devices has explosively increased. But most of the devices are powered by traditional batteries, which need to be replaced or recharged frequently. And this may limit further development of the wearable devices [11,12]. Recently, wearable sensors, which can be installed or worn on the human body to transmit, sense, and process information, have been widely used in consumer electronics, fitness, and health fields. The wearable sensors play an important role, especially in healthcare and treatment. TE materials with good TE properties and flexibility have also attracted attention in the development of wearable sensors [13,14,15,16]. Therefore, more and more researchers pay attention to flexible TE materials [17,18].

At present, there are three typical types of high-performance flexible TE material. The first one uses organic polymers, usually conductive polymers, such as poly (3,4-ethylenedioxythiophene)/poly(styrenesulfonate) (PEDOT: PSS), polypyrrole (PPy), polyaniline (PANI), and so on [19,20,21,22]. PEDOT: PSS, in which PSS is a styrene sulfonic acid group, is the most widely studied polymer. PEDOT: PSS exhibits p-type semiconductor characteristics. The electrical conductivity of PEDOT: PSS is between ~10⁻⁴ and 10³ Scm⁻¹ [23,24]. By optimizing its composition and removing the PSS group, its PF can approach ~500 μWcm⁻¹K⁻² and the ZT reaches ~0.42 [25]. These organic materials have good flexibility; however, their TE properties are much lower than those of traditional brittle inorganic TE materials. The second one involves exfoliation of the inorganic TE materials by a top-down method. For example, most recently, Lu et al. [26] demonstrated good pliability over 1000 bending cycles and high power factors of 4200 (p-type) and 4600 (n-type) μW m⁻¹ K⁻² in Bi₂Te₃-based films that were exfoliated from corresponding single crystals. This method can completely retain the excellent TE properties of inorganic materials; however, the process is costly and it is difficult to scale up production. The third one involves preparation of organic–inorganic composite TE materials, which have both good flexibility due to the organic component and high TE properties due to the inorganic component. Organic–inorganic composite TE materials can be divided into self-supporting films and flexible substrate-supported films. Generally, the composite films based on a flexible substrate show higher TE properties than the self-supporting films. A variety of methods can be used to prepare such films, such as vacuum filtration, screen printing, vapor deposition, rotary coating [27], electrochemical deposition, and magnetron sputtering [28,29,30,31], and there are many flexible substrates, such as polyimide (PI) membrane [32], nylon filter membrane, cellulose fiber, and even carbon paper. Huang et al. [28] prepared flexible Cu₂Se films by vapor deposition of Cu₂Se on a PI surface. The ratio of Cu and Se can be adjusted by controlling the thermal evaporation time. The maximum PF of Cu_2.09Se film is ~5.3 μWcm⁻¹K⁻² and the ZT is ~0.35 at room temperature. After bending 800 times at a bending radius of 6.5 mm, the electrical conductivity of the film remains 97% of the original. Jin et al. [33] deposited Bi₂Te₃ on cellulose fibers by magnetron sputtering to form a flexible composite film with a nano-micron-sized porous structure. The porous structure can introduce strong phonon scattering; hence, the thermal conductivity of the Bi₂Te₃/cellulose fiber composite was significantly low. At 473 K, the ZT is ~0.38. When the bending radius is 10 mm, the resistance increases to 1.05% after bending 100 times. However, as Se is scarce, and Te is toxic and expensive, it is very desirable to explore low-cost and environmentally friendly inorganic TE materials near room temperature.

CuI is a semiconductor with a wide direct band gap (E_g = 3.1 eV) [34]. The reserves of iodine and copper are very abundant on the earth, and both of these components are non-toxic. It is noted that CuI is a green material [35,36]. In previous years, research on CuI has focused on optoelectronic and electronic applications [37,38,39,40]. CuI has three crystal structures which are closely related to the temperature. When the temperature is below 643 K, CuI shows the structure of sphalerite (γ phase); when the temperature is between 643 and 673 K, CuI shows the structure of wurtzite (β phase); when the temperature is above 673 K, it shows the structure of a plane cube (α phase). The α phase and β phase are both ionic conductors, while the γ phase at room temperature is a p-type semiconductor. The copper vacancy (V_Cu) has the lowest formation energy and ionization energy among all the native defects in either I-rich or Cu-rich growth conditions, and is primarily responsible for the p-type conductivity of CuI. Yadav et al. [34] calculated the TE properties of different phases of CuI based on the Boltzmann transport theory combined with first principles. It is found that the Seebeck coefficient of CuI increases monotonously with increasing temperature at low hole concentrations, which shows that there is no compensation between the hole and the electron. High Seebeck coefficient values have been found for all three phases of CuI, and the typical Seebeck coefficient of γ-CuI at 300 K is ~237 μVK⁻¹ (at a hole concentration of 10²⁰ cm⁻³). Since then, the TE properties of CuI have attracted increasing attention, especially CuI films [41,42,43]. For example, Murmu et al. [43] deposited CuI films using an ion beam sputtering under a high vacuum at an ambient temperature and annealed the resulting CuI films in a quartz tube filled with argon. The electrical conductivity of the deposited CuI film is ~21.9 Scm⁻¹ and the Seebek coefficient is ~264.7 μVK⁻¹. After annealing at 250 °C, the electrical conductivity of the film decreased to 5.1 Scm⁻¹, which is 76% of that before annealing, while the Seebek coefficient increased to 789.5 μVK⁻¹, which is ~3 times as high as that before annealing, and the final PF was 740.9 μWm⁻¹K⁻². Optical transparency of the CuI film in the visible range is found to be 60~85%. However, the film is inflexible. CuI film has been deposited on polyethylene terephthalate (PET) substrate by magnetron sputtering [44], varying the iodine partial pressure in the sputter chamber during the magnetron sputtering to control the hole concentration of the CuI films. The obtained CuI film has a Seebeck coefficient of ~172 μVK⁻¹, a PF ~359 μWm⁻¹K⁻², and ZT ~0.21 at room temperature, which is attributed to a combined effect of the heavy element iodine and strong phonon scattering. With a compressive bending angle up to 90°, the change in the internal resistance is <3%, and, after repeatedly bending up to 400 cycles, the change in the internal resistance is <0.2%. A prototype of a CuI-based transparent TE module shows good mechanical flexibility with a high power density of ~2.4 mWcm⁻² at a temperature difference (∆T) of 50 K, which shows the great potential of the CuI film as a transparent flexible TE material, and the power density of the CuI film is comparable with those of Bi₂Te₃/Sb₂Te₃-based devices. Klochko et al. [45] synthesized CuI film on a nanocellulose (NC) substrate via a successive ionic layer adsorption and reaction (SILAR) method. The CuI/NC films obtained by repeating 25–40 SILAR cycles consist of cubic (111)-oriented γ-CuI crystals with faceted surfaces of 200–300 nm. The CuI/NC films have a PF of ~140 μWm⁻¹K⁻² at 333 K. When the temperature difference is 40 K, the CuI/NC-based single TE leg can generate a voltage of 3.5 mV and a maximum power of 3.8 nW. However, these methods, including pulsed laser deposition [43], sputtering [44,46], and SILAR [45,47], are expensive and time-consuming.

Our group recently developed a method for the synthesis of Ag₂Se nanowires combined with vacuum filtration and hot-pressing to prepare flexible Ag₂Se film on nylon. This Ag₂Se film has PF ~987 μWm⁻¹K⁻² at room temperature, and it has good flexibility [48]. Based on this process, our group has successfully prepared various kinds of flexible Ag₂Se-based composite TE film [49,50,51,52,53,54], such as polyvinylpyrrolidone (PVP)/Ag₂Se composite film, starting from Ag₂Se nanostructures coated with PVP, the PF of which can reach ~1910 μWm⁻¹K⁻² (corresponding ZT ~1.1) at room temperature [49]. The PVP (molecular formula being (C₆H₉NO)_n) acts as a great additive for improving flexibility (a 5.5% decrease in PF after being bent along a rod with a radius of 4 mm 1000 times) and the PVP-coated Ag₂Se nanostructures can disperse more evenly in ethanol leading to a more homogeneous film. This is because PVP is a non-ionic surfactant that has good solubility in water and most organic solvents and can well disperse inorganic nanoparticles due to forming a uniform PVP coating nanolayer on the nanoparticles. PVP has high solubility in both polar and non-polar solvents, biocompatibility, good complexation, good adhesion, low toxicity, and good spinnability [55]. Although PVP is insulating, which will have a negative effect on electrical conductivity, it also has extremely low thermal conductivity; if only a very small amount of PVP is added, the direct effect of adding PVP on the thermoelectric performance of the composite film could be neglected.

In our previous work [56], we reported the preparation and TE properties of CuI film, but the film was porous and hence with ultralow electrical conductivity and low stability. Hence, in this work, PVP is used to improve the flexibility, electrical conductivity, and stability of the CuI film. PVP-coated CuI powder was prepared in situ by a one-pot method at room temperature, and then flexible CuI/PVP film was prepared on a nylon membrane combined with vacuum filtration and hot-pressing. The TE properties and flexibility of the CuI/PVP nanocomposite films were investigated. In addition, a flexible single-leg TE module was fabricated with the nanocomposite film, and its touch tests were investigated.

2. Materials and Methods

All the raw materials, including copper nitrate trihydrate (Cu(NO₃)₂·3H₂O), potassium iodide (KI), polyethylene glycol (PEG), polyvinylpyrrolidone (PVP), hydrazine hydrate (N₂H₄·H₂O), and ethanol, were purchased from sinopharm Chemical Reagent Co., Ltd., Shanghai, China.

In a typical procedure, PVP-coated CuI powder was prepared as follows: first, 1.3288 g of Cu(NO₃)₂·3H₂O and 2g of PEG were dissolved in 50 mL ethanol to form solution A and 0.83 g of KI and 2g of PEG were dissolved in 50 mL ethanol to form solution B. Second, a certain amount of PVP (0.01 g, 0.03 g or 0.05 g) and 100 μL N₂H₄·H₂O were added to solution B with stirring. Third, solution B was slowly added into solution A with continuous stirring for 20 min. Finally, PVP-coated CuI powder was collected by centrifugation and washed with ethanol and deionized water (DI) several times.

The PVP-coated CuI powder was ultrasonically dispersed in ethanol for 5 min, and then deposited on a porous nylon membrane via vacuum-assisted filtration. The as-prepared film was dried in a vacuum oven for 24 h at 65 °C then hot-pressed at 230 °C and 1 MPa for 30 min to obtain CuI/PVP composite film. Figure 1 shows a schematic diagram of the preparation of the CuI/PVP composite film Schematic illustration of CuI grain growth with PVP when hot-pressing is illustrated in Figure S8. To study the effect of the content of PVP on CuI, PVP content was adjusted from 0.01 and 0.03 to 0.05 g; for simplicity, the corresponding powders were named P1-powder, P2-powder, and P3-powder, respectively, and the corresponding films were named P1-film, P2-film, and P3-film, respectively. Pure CuI powder and film were named P0-powder and P0-film, respectively.

A flexible single-leg TE module was fabricated by first sticking a strip (20 mm × 5 mm) of the CuI/PVP composite film onto polyimide (PI) substrate with a double-faced adhesive tape; then, Ag paste (SPI# 04998-AB) was painted onto two ends of the strip to connect with conducting wires. A finger touch test on a single-leg TE module was tested by a self-made experimental instrument (see Figure S8). As Figure S8 shows, in the single-leg TE module, heat transfers from the hot side to the cold side along the length direction of the module. We used an automatic temperature controlling system to heat a copper plate and one end of the module was put on the plate as the hot side (T + ΔT). The other end of the module was put in the air as the cold side (T). Then, we collected the output voltage at a particular temperature difference.

The electrical conductivity at room temperature was measured using a four-probe technical machine, with the help of a Hall effect measurement system (Ecopia HMS-3000). The Seebeck coefficient at room temperature was measured by the slope of the linear relationship between the thermal electromotive force and temperature difference (ΔT) between two ends on one side of each film. The temperature-dependent TE properties were measured by a Cryoall CTA-3 instrument in He atmosphere, with an instrument test error of ±5% for both electrical conductivity and Seebeck coefficient. The flexibility of composite film was tested by measuring the electrical conductivity after bending the film along a rod with a radius of 4 mm at different times.

The crystallinity and phase composition of the CuI powder and films were examined by X-ray diffraction (XRD) (Bruker D8 Advance). All XRD measurements were recorded in the 2θ range of 10–90° at a scanning speed of 5°/min. The thickness of the film was around 10.87 μm, which was determined by field-emission scanning electron microscopy (FESEM, FEI Nova NanoSEM 450) observation of the cross-section of the film. The FESEM observation was also used to examine the surface morphology of CuI powder and CuI/PVP films. High-resolution transmission electron microscopy (HRTEM) (JEM-2100F) was used to observe the CuI powder. A transmission electron microscope (TEM, Titan Themis G2 60-300, Thermo Fisher Scientific, Shanghai, China) was used to investigate the internal details of the film.

3. Results

X-ray diffraction (XRD) patterns of the P1-film and P1-powder are in shown Figure 2a. As seen, all peaks are indexed to the γ-CuI with no impurity peaks [57]. The CuI thin film exhibits strong peaks at (111), (220), and (311) planes. Note that the (220) and (311) peaks of the P1-film are stronger than those of P1-powder, showing that hot-pressing improves the crystallinity of the CuI. No XRD peaks of PVP are detected, which may be because of its small content and amorphous characteristic [49]. Figure 2b and Figure S1a show scanning electron microscopy (SEM) images of P1 and P0-powder. From the SEM images, P0-powder consists of numerous nanoparticles and thin slices, and P1-powder is composed of many slices with an average size of ~1 μm. It was noticed that adding PVP changes the morphology of the CuI powder. Figure 2c shows a TEM image of the P1-powder. There is an amorphous layer with an average thickness of ~20 nm uniformly coated on the surface of the CuI powder. Considering that a small content of PVP is added in situ when synthesizing CuI and that no impurity peaks (Figure 2a) are detected, we think that the amorphous layer is PVP. Figure 2d and Figure S1c show typical surface SEM images of the P1-film and the CuI film reported in our previous work [56]. By comparing Figure 2d with Figure S1c, the P1-film is denser. This is probably because the PVP-coated CuI powder can disperse more evenly in ethanol to avoid agglomeration, forming a more uniform film after vacuum filtration. Figure 2e is a cross-sectional SEM image of the P1-film, which also shows that the film is dense with a small amount of nanopores.

Figure 3 shows TEM and high-resolution TEM (HRTEM) images of the P1-film. The P1-film consists of nanograins with an average size of ~15 nm. Figure 3b is an enlarged image of the red box in (a), and Figure 3c is an enlarged image of the green box in (a). Figure 3b,c shows that CuI nanograins are connected by an amorphous PVP phase and there exists a small number of nanopores (~5–10 nm). The insets of Figure 3c show inverse fast Fourier transform (IFFT) images corresponding to grains A and C. Notably, the crystallinity of the CuI grains is good. Figure 3d displays an HRTEM image of another area, which shows several connected grains coated with an amorphous PVP layer. Figure 3e shows a semi-coherent GB. The insets of Figure 3e show the corresponding IFFT and fast Fourier transform (FFT) images, indicating that the lattice spacings of regions A and B are both 0.349 nm, corresponding to the (111) plane of CuI. Figure 3f displays a twin boundary, which can significantly reduce the thermal conductivity and has little effect on the electrical conductivity [58]. Figure 3g shows an HRTEM image containing several grains. Figure 3h,i shows enlarged images of the orange and blue boxes marked in (g). Figure 3h shows a typical triangular GB, indicating the three grains (grains A, B, and C) are well-sintered together. Figure 3i displays grains A, B, and another grain (grain D). It is deduced from the different contrasts of the grains (Figure S6) that grain D lies beneath grains A and B. Namely, in the P1-film, most CuI grains are sintered, and a small number of CuI grains are bonded by PVP.

The TE performances of P0-film, P1-film, P2-film, and P3-film at room temperature are shown in

Table 1. TE performances of the P0–P3 films at room temperature.

Sample	Mass Ratio (PVP:CuI)	S (μVK−1)	σ (Scm−1)	PF (μWm−1K−2)
P0-film	0:1	522	0.18	4.94
P1-film	0.0095:1	605	0.22	8.05
P2-film	0.0286:1	374	0.17	2.38
P3-film	0.0477:1	332	0.14	1.54

. It can be seen from Table 1 that, with increasing PVP content, the Seebeck coefficient increases from 522 to 605 μVK⁻¹ and then gradually decreases, which indicates that the films are p-type conductors. The variation in electrical conductivity is consistent with the variation of the Seebeck coefficient, which may be related to the content of PVP. Compared with the P0-film, the electrical conductivity and Seebeck coefficient of the P1-film are larger. There are two opposing effects of adding PVP. One is that PVP is an insulating polymer, which leads to the degradation of the electrical conductivity of the composite film. The other is that adding PVP makes the composite film more uniform and denser, which is beneficial for increasing the electrical conductivity. For the P1-film, the amount of PVP added is very small and the latter effect of increasing electrical conductivity is more significant, but for the P2-film and P3-film, the amount of PVP added is more, and hence the former effect of decreasing electrical conductivity is more significant. However, the Seebeck coefficient is not sensitive to the density. Dun et al. [59] studied Cu_0.1Bi₂Se₃ nanoplatelet/polyvinylidene fluoride (PVDF) composite films. They think that there could be an energy filtration effect at the Cu_0.1Bi₂Se₃ /PVDF interface, leading to an increase in the Seebeck coefficient. However, because PVDF is an insulator, the energy barrier between PVDF and Cu_0.1Bi₂Se₃ could be so high that no carriers can surmount it. In the present case, PVP is also an insulator; hence, the reason for the P1-film having an increased Seebeck coefficient value is not clear yet. Consequently, P1-film exhibits an enhanced PF of ~8.05 μWm⁻¹K⁻² at room temperature. Therefore, PVP should not be added too much in the synthesis process. In this work, the CuI/PVP mass ratio of ~0.0095:1 was the best, so P1-film was chosen for further study.

Figure 4a shows the TE performance of the P1-film from 20 °C to 105 °C. Over the entire temperature range, the Seebeck coefficient value remains about ~600 μVK⁻¹ as the temperature increases. The electrical conductivity is ~0.22 Scm⁻¹ at room temperature, and it tends to increase slightly before 90 °C and then drops gradually. Finally, the PF of the P1-film increases gradually with the temperature. At room temperature, P1-film exhibits a PF of ~8.05 μWm⁻¹K⁻². The PF of the P1-film is relatively low; however, it is higher than that of a CuI/amorphous carbon composite pellet, which exhibits a PF of ~7.5 μWm⁻¹K⁻² (corresponding electrical conductivity of ~0.5 Scm⁻¹ and Seebeck coefficients of ~390 μVK⁻¹); the PF of the composite pellet markedly decreases to a range of 0.119 to 0.637 μVK⁻¹ after a two- to twelve-cycle measurement from room temperature (RT) to 200 °C [60], which is somewhat lower than that of CuI/bacterial cellulose nanofiber (BCNF) composite film with a PF of ~24.6 μWm⁻¹K⁻² at RT [61] and Cu_1.75Te/polyvinylidene fluoride (PVDF) composite film with a PF of 23 μWm⁻¹K⁻² at RT [62]. Nevertheless, the component Te of the latter is toxic and expensive.

To study the thermal stability of P1-film, we tested temperature-dependent TE properties (the electrical conductivity and Seebeck coefficient) of the same film twice in succession, and the results are shown in Figure 4b and Figure S4. Note that both parameters are almost the same during the two tests, suggesting that the prepared films have much better thermal stability below 100 °C.

Figure 5a is a photo of the P1-film flexibility test. Figure 5b shows the flexibility test result. It indicates that, after being bent around a rod with a radius of 4 mm 1000 times, the electrical conductivity of P1-film decreases by only 4%, which is better than most of the reported flexible TE materials.

Compared with the CuI film reported in our previous work [56], the P1-film has better flexibility, higher electrical conductivity, and higher thermal stability, mainly because CuI nanograins in the film are sintered together and the adhesive effect of PVP.

Figure 5c,d shows the results of one-cycle and five-cycle touch tests on a single-leg TE module, respectively. In Figure 5c, note that when this module touched the heating plate, it immediately generated a voltage of ~1.5 mV in 0.5 s, reached a maximum voltage of about ~3.6 mV in 10 s and then remained the same voltage value. When moving this module away from the heating plate, the voltage quickly decreased to ~0 mV in ~60 s. Before moving the module from the heating plate, there was a slight increase in voltage, which may have been because the contact between the heating plate and the module enhanced somewhat before the module was moved away. It is indicated that the temperature difference (∆T) between the heating plate and room temperature (29 °C) is 6 K. The Seebeck coefficient can be evaluated at about ~600 μVK⁻¹ using the formula S = V/∆T, which indicates the voltage touch test results are reliable. Figure 5d shows the result for multiple cycles. The peak voltage was somewhat different for each touch, which should be because the contact between the heating plate and the module was slightly different each time. The touch test shows that flexible PVP/CuI composite films have great potential prospects in wearable sensors.

4. Conclusions

In conclusion, we have prepared flexible CuI/PVP nanocomposite films by a simple method, and this method is environmentally friendly and highly efficient. In addition, PVP added in situ enhances the density and hence improves the flexibility, electrical conductivity and stability of the CuI film. TEM observation indicates that the film is composed of CuI nanograins (average size of ~15 nm) and PVP distributed at the surface of the CuI nanograins and the inner wall of nanopores. The optimized film shows a large Seebeck coefficient ~605 μVK⁻¹ and PF ~8.05 μWm⁻¹K⁻² at room temperature, and the film exhibits excellent flexibility (~96% of initial electrical conductivity retention after being bent 1000 times along a 4 mm radius rod). A single-leg TE module outputs a voltage of ~3.6 mV when ∆T is 6 K. This work provides a feasible and simple method for preparing dense and low-cost flexible TE nanocomposite films.