Fabrication and Evaluation of Sb Te /PVDF Hybrid

IN

DEGREE PROJECT ENGINEERING PHYSICS, SECOND CYCLE, 30 CREDITS

,

STOCKHOLM SWEDEN 2020

Fabrication and Evaluation of

Sb Te /PVDF Hybrid

₂

₃

Thermoelectric Films

BO YU

Fabrication and Evaluation of Sb

Te

/PVDF Hybrid

Thermoelectric Films

BO YU

Supervisor: Hazal Batili batili@kth.se

Examiner: Prof. Muhammet S. Toprak

toprak@kth.se

Biomedical and X-Ray Physics KTH/Albanova SE-106 91 Stockholm TRITA-SCI-GRU 2020: 196

Abstract

Thermoelectric (TE) materials and devices have been extensively studied for harvesting waste heat energy. There is more interest to fabricate TE materials and devices at low cost and highly-efficient ways. Besides, TE materials and devices should be flexible and light-weight to address the requirements of emerging applications such as wearable TE power generators, health sensors, and powering small mobile wireless devices. To fabricate flexible and light-weight TE films with low cost and high-efficiency, organic-inorganic hybrid materials have been introduced. Organic-organic-inorganic hybrid materials not only show the advantages of organic materials, such as good flexibility, light-weight, low cost, but also reflects the good TE properties of inorganic TE materials. The aim of this project was to synthesize and investigate the p-type hybrid TE films.

Recently, many different methods and mechanisms have been tested to improve TE performance of materials. One of these mechanisms called defect engineering, which has been used to enhance TE performance by increasing the carrier concentration. Another approach can be the introduction of interfaces and grain boundaries which can improve TE performance by suppressing lattice thermal conductivity. In addition, nanomaterials and low-dimensional materials are also used for improving TE performance, classical size effect suppresses lattice thermal conductivity by limiting the mean free path, while quantum size effect increases the Seebeck coefficient by creating a clear electronic density of state function.

Sammanfattning

Termoelektriska (TE) material och anordningar har studerats för att skörda spillvärmeenergi. Det är mer intresse att tillverka TE-material och enheter till låga kostnader och på ett högeffektivt sätt. TE-material och enheter bör dessutom vara flexibla och lätta att bämöta krav på nya applikationer som bärbara TE-kraftgeneratorer, hälsosensorer och driva små mobila trådlösa enheter. För att tillverka flexibla och lätta TE-filmer med ett billigt och högeffektivt sätt har organiska-oorganiska hybridmaterial införts. Organiska-oorganiska hybridmaterial visar inte bara fördelarna med organiska material, såsom god flexibilitet, lätt vikt, låg kostnad, utan återspeglar också de goda TE egenskaperna hos oorganiska TE material. Syftet med detta projekt var att syntetisera och undersöka p-typ hybrid-TE-filmer.

Nyligen uppståd många metoder och mekanismer för att förbättra TE prestanda. Den första är defektteknik, vilket ökar bärarkoncentrationen för att förbättra TE prestanda. Dessutom kan införandet av gränssnitt och korngräns förbättra TE prestanda genom att undertrycka gitterens värmeledningsförmåga. Dessutom används nanomaterial och material med låg dimension också för att förbättra TE prestanda, klassisk storlekseffekt undertrycker gitterens värmeledningsförmåga genom att begränsa den genomsnittliga fria vägen, medan kvantstorlekseffekt ökar Seebeck-koefficienten genom att skapa en tydlig elektronisk tillståndstäthet (DOS) fungera.

List of Abbreviations

TE Thermoelectric NP Nanoparticles

ZT Thermoelectric figure of merit % Seebeck coefficient

$ Electrical conductivity κ Thermal conductivity PF Power Factor

PEDOT Poly(3,4-ethylene dioxythiophene) PANI Polyaniline

PVDF Poly(vinylidene fluoride) NMP 1-Methyl-2-pyrrolidinone SEM Scanning Electron Microscope

Content

1.2 Aim of the thesis ... 8

1.3 Thermoelectric properties ... 8

1.3.1 Electrical conductivity (!) ... 9

1.3.2 Seebeck coefficient (") ... 9

1.3.3 Thermal conductivity (#) ... 10

1.4 Recent advances in TE materials research ... 10

1.4.1 Defect mechanisms ... 10

1.4.2 Interface and grain boundaries ... 11

1.4.3 Size effect ... 11

1.5 Thermoelectric materials ... 12

1.5.1 Inorganic TE materials ... 12

1.5.2 Conducting polymers and hybrid TE materials ... 13

1.6 Fabrication of TE films ... 17

2. MATERIALS AND METHODS ... 19

2.1 Materials ... 19

2.2 Fabrication of TE films ... 19

2.2.1 Grinding TE pastes ... 19

2.2.2 TE films via doctor blading ... 20

2.3 Characterizations ... 20

2.3.1 Scanning electron microscopy (SEM) ... 20

2.3.2 Two-probe IV-measurements ... 21

3. RESULTS AND DISCUSSION ... 23

3.1 Scanning Electron Microscopic (SEM) analysis ... 23

3.2 Fourier transform infrared (FT-IR) spectroscopy analysis ... 26

3.3 Electrical properties ... 28

4. CONCLUSIONS ... 30

5. FUTURE WORK ... 31

6. ACKNOWLEDGEMENT ... 32

1. Introduction

1.1 Background

The energy crisis is a serious problem which we are facing because of the increase in the world’s population and the shortage of fossil fuel sources. Therefore, scientists spend more time and effort to solve the energy crisis with sustainable and renewable energy sources. Currently, the most common source of energy is fossil fuel, such as coal, petroleum, and natural gas. These energy sources are widely studied and used to provide electricity, but more than half of the energy is wasted as thermal energy. TE materials and devices can be used for collecting the waste thermal energies and convert them into electrical power with a simple and environmental friendly way [1]. It is a good way to reusing and reducing the waste thermal energy, thus reducing the carbon footprint of various processes.

A TE device can convert thermal energy from a temperature gradient into electricity, through the Seebeck effect. The Seebeck effect was first observed in 1821 [2]. It can be simply schematized by Fig. 1(a), when the temperature difference is applied to two ends of a TE device (which made from one n-type TE material and one p-type TE material), the charge carriers (electrons in n-type TE material, holes in p-type TE materials) diffuse from the hot ends to cool ends, and form that current flow in a circuit. While the reverse counterpart of the Seebeck effect was discovered by Peltier in 1834 and called the Peltier effect [1]. The Peltier effect is shown in Fig. 1(b), when applied the voltage to the end of a TE device, a temperature difference will be created between the two ends of the TE device [2].

Figure 1. Schematic illustrations of TE couple for (a) An applied temperature difference causes electricity power. (b) An applied voltage creates a temperature difference between two ends of a

A TE device is a clean and green option to be the power generator because they work without any mechanical movement and yield of harmful byproducts [4]. However, the conversion ratio of TE materials is between 5-20% currently, which is lower than other energy-conversion technologies, such as alkaline metal cell, solar, and nuclear. And it can be further enhanced by doping, introduction of nanostructures, and through organic-inorganic hybrid materials [3].

1.2 Aim of the thesis

In this project, main focus was to fabricate hybrid TE films composed of Sb2Te3 nanoplates and PVDF, which has high dielectric constant and electroactive response, by using doctor blading method. After the fabrication, different techniques had been used for characterization of the films and evaluation of their TE properties.

1.3 Thermoelectric properties

With the increase of TE studies, there are more discoveries that are related to TE material properties. In 1911, Altenkirch was the first person who introduced the term “TE figure of merit (ZT)” which can be considered a generalized figure of merit for a TE device at each temperature [5, 6]. ZT is used to measure the TE performance of materials. ZT value is given by the followed equation:

!" =

Where α is the Seebeck coefficient, σ is electrical conductivity, κ is thermal conductivity and T is the absolute temperature in Kelvin. From the above equation, ZT value is a dimensionless unit. Normally, an ideal TE material should have a high Seebeck coefficient, high electrical conductivity to minimize Joule heating, and low thermal conductivity to maintain a large thermal gradient. Some factors related to %, $ and κ are; carrier mobility, carrier concentration, and bandgap [6].

In the microscopic level, promising TE properties result from phase stability and instability, structural order and disorder, bond covalency and ionicity, band convergence and splitting, itinerant and localized electronic states, and coordination between carrier mobility and effective mass [7]. Therefore, research on TE materials can be defined as the application of driven multidisciplinary basic research involving the charge, spin, orbit, and lattice degrees of freedom of materials (as shown in Fig.2) [7].

For ideal TE materials, ZT should be above 1 to obtain a conversion efficiency of > 10% [8]. For example, bismuth telluride (Bi2Te3), whose ZT value is about 1 near room temperature, is commercially used in Peltier cooling devices [9].

1.3.1 Electrical conductivity (*)

The σ of TE materials depends on carrier concentration and mobility, which is given by:

$ = ,(._! / + ._#1)

where ,, ._! , ._#, /, and 1 are the electrical charge, electron mobility, hole mobility, electron concentration, and hole concentration, respectively [8]. Hall effect measurements are used to determine the carrier concentration, mobility, and the type of carrier. Enhanced $ can be achieved via the improvement of carrier concentration and mobility, resulting in a higher ZT value.

Carrier concentration increases with higher temperatures. It can also be improved by doping, and the dopant concentration and the nature of dopant affect the carrier concentration and carrier type of materials [10]. For TE materials, carrier concentration determines not only $ but also %. With the increase of carrier concentration, the $ increases, while % decreases [8]. Therefore, it is hard to increase $ and % at the same time.

Carrier mobility is determined by lattice and impurity scattering. Both lattice and impurity scattering are increasing as temperature increases, which results in a decrease in carrier mobility [8].

1.3.2 Seebeck coefficient (3)

The α is the ratio of a voltage difference (∆5) between hot and cool ends of TE couple to temperature difference (∆6) between these two ends [4], which is given by:

For the p-type TE materials, % should be positive; while that of n-type TE materials is negative. An ideal TE material should have a high %, meaning high voltage-generating ability.

The % of TE materials have three contributors, which are the electron, phonon, and electron-phonon [11]. Metals have a large number of free electrons, so electrons contribute to the % dominantly. Phonons contribute to the % dominantly at low temperature (less than 200 K) due to the mean free path of phonon is large. When the mean free path of the electron is closed to that of phonon, electron-phonon becomes mainly contributor [12], such as in a conducting polymer with high crystalline structure. 1.3.3 Thermal conductivity (9)

The κ affects ZT of TE materials directly, the lower κ is, the higher ZT value is. κ is calculated by the following equation:

κ = :;<_$

where :, ;, and <_$ are density, thermal diffusivity, and specific heat, respectively [13]. For a TE material, κ depends on two contributions: (1) electrons and holes transporting heat and (2) phonons traveling through lattice [14]. To improve ZT value, κ is reduced by substituting the crystal lattice with an amorphous structure [1].

1.4 Recent advances in TE materials research

This section is a review of several mechanisms used for improving TE performance, which includes introduction of defects, interfaces, and sized effect.

1.4.1 Defect mechanisms

Defects are result from the breaking of translational or rotational symmetry of the crystal lattice, while defects are beneficial in some aspects. Thus, defect engineering arises which aims to deal with two or more adversely interdependent band structure and improves their transport properties [7]. Taking the doping as an example, doping can increase the carrier concentration to enhance the PF, and the resulting point defects scatter heat-carrying phonons to suppress the lattice thermal conductivity [15]. Defects can activate the other hibernating degrees of freedom as well [15].

review the role of intrinsic points in point defects in V2VI3 compounds has shown that intrinsic point defects can be manipulated in terms of composition, mechanical and thermal aspects [16].

Band convergence is another defect engineering used to improve TE performance. It is mainly through doping to actively arrange the electronic energy bands to achieve a higher degree of energy band degeneracy [17]. When the energy difference between the energy bands is equivalent to KBT (KB is the Boltzmann constant), the energy bands will converge, causing an increase in degeneracy. At the same time, the jumping and movement of carriers between degenerate energy valleys does not need to cross the energy barrier. Therefore, the mobility of carriers is generally not affected by degeneracy. There are light bands and heavy bands in many TE materials with highly symmetric crystal structures. If the carrier concentration reaches a certain concentration through doping. There is degeneracy of band convergence. The heavy bends will also participate in the power transmission. In this way, each energy band maximum or minimum value (peak or valley) will be occupied by carriers. It will contribute to % and $ . Therefore, the simultaneous optimization of % and $ can be effectively achieved [18]. However, the efficacy of band convergence is limited by the temperature-induced band shift, inter-band scattering, and solubility of dopants [17]. 1.4.2 Interface and grain boundaries

The interface and grain boundaries are two-dimensional planar defects actually. Ideally, the interface should provide three preferential scattering: (i) scatter phonons more efficiently than charge carriers, which can reduce κ because phonons are main thermal carriers; (ii) scatter low-energy charge carriers more efficiently than high-energy carriers, which result in the increase of %, and it is called the boundary conducting energy-filtering effect [19]; (iii) more scattering a few charge carriers is more effective than most charge carriers.

In general, coherent or semi-coherent grain boundaries [20] and proper intergranular energy band arrangement [21] tend to maintain carrier mobility, while coarse grain boundaries effectively suppress lattice thermal conductivity [22]. Interface and grain boundaries are used as carrier energy filter, it can filter out low energy carriers to obtain a higher % [7].

1.4.3 Size effect

The size effect is based on nanocomposites and low-dimensional TE materials. It can be divided into classical and quantum size effects, the former concerns the scattering-limited mean free path, while the latter concerns confinement-induced change in the dispersion relation [7].

microstructures, the lattice thermal conductivity will be suppressed. Because when the phonon mean free path gets as small as the interatomic spacing and heat is carried by random-walking Einstein modes [20].

The quantum size effect of charge carriers can be used to create a clear electronic DOS function and is favored by the high % when the EF is placed correctly. When the de Broglie wavelength of the charge carrier is equivalent to the physical size of the material in one or more directions, the charge carrier is strictly restricted in these directions, and therefore its size is reduced [7].

Although it is difficult to distinguish the coexistence of classical and quantum size effects, most of the advances in TE nanocomposites come from interfacial scattering of thermally loaded phonons, rather than from enhanced DOS, which shows that classical size effects dominate quantum size effects [7].

1.5 Thermoelectric materials

1.5.1 Inorganic TE materials

In the past two decades, many inorganic TE materials have been studied and their TE properties have been improved.

Figure 3. Timeline of the maximum ZT values for several representative families of TE materials [7].

Using Bi2Te3 as TE materials was first proposed in 1954 [2]. Using Bi2Te3 as TE materials was first proposed in 1954 [2]. Bi2Te3 has a rhombohedral tetradymite-type crystal structure with the space group R-3m, which can be described by a hexagonal unit cell. Bi2Te3 is a low-temperature TE material which can operate up to 227 °C [2]. It displays unique TE properties such as a high the % (~ 220 .V/K), good $ (~ 400 S/cm), and low κ (1.5 W/m K). And its ZT value can reach 1. However, they pose high κ, and some of them contain heavy atoms (Bi, Sb, and Te) which are toxic and unstable at high temperatures (~ 1000 K) [23].

Antimony telluride (Sb2Te3) is the member of V2VI3 TE materials. It also is a kind of p-type three-dimensional topological insulator [24]. The crystal structure of Sb2Te3 is composed of layers of identical atoms following the sequence of a quintet -Te1-Sb-Te2-Sb-Te1- stack along the c-axis of the Sb2Te3 unit cell. The weak van der Waals bonding between Te1-Te1 of two quintets has been attributed to their anisotropic transport properties [24].

1.5.2 Conducting polymers and hybrid TE materials

In 1977, a major breakthrough in conductive polymers was the redox doping of a “plastic” polyacetylene, whose conductivity is comparable to metals such as aluminum and copper [25]. Numerous studies over the past few decades have shown that conductive polymers show great potential in organic optoelectronics, especially as transparent conductive films for advanced devices [26]. Recently, conductive polymers have made remarkable achievements as active modules of organic TE generators, which has triggered a new wave of research on conducting polymers. The conducting polymer is attractive because of low κ, good flexibility, light weight, and sufficient source in the earth [27].

Figure 4. Chemical structure of PEDOT:PSS [8].

Many additives and post-treatments are used for improving the TE performance of PEDOT:PSS films by removing non-conducting PSS, increasing crystalline of PEDOT:PSS film, and controlling the oxidation state of PEDOT, as shown in Table 1. Table 1. TE performance of hybrid PEDOT:PSS samples at room temperature.

Chemicals Post-treatment * (S/m) 3 (&V/K) PF (&'/)K-2₎ ZT Ref PEDOT:PSS+DMSO EG 88000 73 470 0.42 [30] PEDOT:PSS+DMSO DMSO+EMIMBF4 7000 25 38.46 0.068 [31]

PEDOT:PSS+SiC-NWs Sulfuric acid (98 wt%) 311300 20.3 1287.3 0.17 [32]

PEDOT:PSS+Bi2Te3

powder ____ 42100 15.3 9.9 0.04 [33]

PEDOT:PSS+SWCNTs ____ 24100 38.9 21.1 __ [34]

In Wang et al.’s work, nanowires were added into PEDOT:PSS, because SiC-nanowires are air-stable semiconductor with low κ. Then the films were immersed into sulfuric acid (98 wt%) to increasing $ by removing PSS. And it exhibits high $ (311300 S/m) and ZT reaches 0.17 at room temperature [32]. In Song et al.’s work, Bi2Te3 was mixed with PEDOT:PSS to improve its TE performance. Bi2Te3 can provide carriers the PEDOT:PSS to attach, and in this way a conformational change of the PEDOT chains occurred from coiled to linear, making it much easier for charge carriers to move among the polymer chains. The ZT of this film is 0.04 at 300 K [33]. Song et

al. also mixed single-walled carbon nanotubes (SWCNTs) with PEDOT:PSS. The

layered nanostructure composites showed both improved $ and % as compared to pure PEDOT:PSS, which could be attributed to the improvement of electron transport and phonon transport because of the bonding disruption of SO3H group with the PSS chains and the use of quantum confinement and interface effects in layered nanostructures. The power factor (PF, PF=%%_{$) is up to 21.1 .A/BK}-2_{, about 4 orders} of magnitude higher than the pure PEDOT:PSS [34].

Polyaniline (PANI) is a promising TE material as well due to its low cost and easy processing. The three different structures observed in PANI are leucoemeraldine (y=1), emeraldine (y=0.5), and pernigraniline (y=0), as shown in Fig. 5 [8].

Figure 5. Chemical structure of PANI.

The $ and % of PANI can be further increased via doping. Because the $ of PANI depends on the oxidation level, molecular arrangement, interchain separation, degree of crystallinity [8]. Graphene nanosheets (GN) was mixed with aniline and grew PANI/GN composition via in situ polymerization, its ZT is 0.002 [35]. Ju et al. prepared PANI/tin sulfide (SnS) nanosheets composition, and increased ZT to 0.0029 [36]. And in Toshima et al.’s work, Bi2Te3 was mixed with PANI with a physical method, the ZT can up to 0.025-0.028 [37].

Figure 6. Chemical structure of PVDF.

PVDF shows five crystalline polymorphs with %, C, D, E, and F phases depending on different processing [40]. The chain conformations of %, C, and D phases are shown in Fig. 7. The α phase is the one commonly obtained directly from the melt by crystallization. The all-trans planar zigzag conformation of β phase can induce a significant dipole moment. Furthermore, a large piezoelectric effect of γ phase is weaker than β phase PVDF due to a gauche bond existed in every fourth repeat units [41]. So the β phase PVDF shows best piezoelectric, ferroelectric, and pyroelectric properties among the five polymorphs of PVDF generally [42].

Figure 7. The chain conformation for %, C, and D crystalline phases of PVDF [40].

Dun et al. reported that layered Bi2Se3/PVDF composition is n-type TE materials, and it shows a high power factor (30 µW/mK2_{) due to addition of Bi2Se3 [38]. In Du et}

al.’s work, porous PVDF/multi-walled carbon nanotube (MWCNT) was prepared, and

porous PVDF/MWCNTs exhibits lower κ (0.15 W/mK) than that of pure PVDF (0.19 W/mK) and PVDF/CNTs (0.5 W/mK) [19].

and polymer resulted in severe phonon scattering at the interface so that the κ of polymer/MWNT composites are low [19].

Table 2. TE performance of hybrid PVDF samples at room temperature. Hybrid Compositions * (S/m) 3 (&V/K) PF (&'/)K-2₎ L (W/mK) ZT Ref. PVDF+Bi2Se3 nanoplates 5100 -80 30 0.42 0.02 [38] Porous PVDF +MWCNTs 16 324.45 1.679 0.15 0.0033 [19]

1.6 Fabrication of TE films

In order to fabricate TE films easily, different techniques have been used. Among them, drop-casting method has been used widely, because it is easy to operate and without any special demands. However, it is hard to fabricate films with controlled thickness due to the concentration of liquid droplets.

Direct dilution–filtration is a novel method to fabricate TE films which used PVDF membrane (pore size 0.45 µm) [43]. PVDF membrane which were washed by DI water and alcohol for few times before dried in a vacuum oven 120 o_{C for 15 min and 80} o_C for 30 min. The one-step process is shown in Fig. 8. The as-prepared films on PVDF membrane were immersed in bath for 20 min, which result in that the films would drop down automatically from PVDF surface.

2. Materials and methods

2.1 Materials

Poly (vinylidene fluoride) ((C2H2F2)x, PVDF, MW≈ 532) was purchased from Sigma-Aldrich and was used directly without any pre-treatment. 1-Methyl-2-pyrrolidinone (C5H9NO, NMP) was purchased from Sigma-Aldrich as well. Pure Sb2Te3 nanoplates were synthesized in KTH-Nanochemistry lab, their average lateral sizes is around 1 µm.

2.2 Fabrication of TE films

2.2.1 Grinding TE pastes

Firstly, mortar was washed with isopropanol and DI water orderly. Then Sb2Te3 nanoplates and PVDF were put into the mortar, and NMP was added drop by drop. Sb2Te3 nanoplates, PVDF, and NMP were fully grinded for 15-20 min to form homogeneous pastes. The viscous black pastes were thus obtained.

There are three different percentages of PVDF and Sb2Te3 NPs (85 wt% Sb2Te3 +15 wt% PVDF, 75 wt% Sb2Te3 + 25 wt% PVDF, 50 wt% Sb2Te3 + 50 wt% PVDF), as shown in Table 3. Different volumes of NMP were added into them to obtain homogeneous pastes.

2.2.2 TE films via doctor blading

Glass substrates were cleaned by sonicating with acetone, isopropanol, and de-ionized (DI) water for 5 min, respectively. After drying, glass substrates were blocked by Kapton tape (65 µm thickness), only leaving active area (10x5 mm2_{, 1x3 cm}2_{) in the} center.

The pastes were printed onto the active area of glass substrate via doctor blading method. Let the glass substrate dry for a few minutes, the Kapton tape was taken off by hand, there was only printed film on the glass substrate, the process is shown in Fig. 9.

Figure 9. (a) Washed glass substrate; (b) Washed glass substrate with Kapton tape (yellow part); (c) TE paste was printed into active area by moving a glass applicator.

After above steps, as-prepared films were dried on hotplate at 50 o_{C for overnight.} Finally, TE films will be obtained after drying the NMP solvent.

2.3 Characterizations

Some experimental techniques were used in this work, including scanning electron microscopy (SEM), two-probe IV-measurements, and Fourier transform infrared spectroscopy (FT-IR). In this section, the principal of operation of these techniques will be explained in detail.

2.3.1 Scanning electron microscopy (SEM)

SEM, in principle, scans a sample by using a focused high-energy electron beam to excite various electrons, such as backscattered electrons, secondary electrons, Auger electrons, characteristic x-rays, and transmitted electrons. Physical and chemical information can be obtained from these different kinds of electrons. By receiving, magnifying, and displaying imaging of this information, an observation of the surface morphology of the test specimen is obtained [44].

spectrometer (EDS or EDX). SEM can be used for microstructure and surface morphology analysis as well. There are two types of contrast in SEM images; first one is the topographic contrast which is a result of variation in signal levels caused by the differences in geometric features of the specimen surface. The secondary electrons (SE) are the primary source of topographic contrast but backscattered electrons (BSE) also contribute to this type of contrast. The second type is the compositional contrast which is a result of compositional variations in the specimen.

BSE refers to a portion of incident electrons that are reflected/scattered back by the atoms of a solid sample (scattering angle is greater than 90°), and their energy is basically unchanged (energy is a few 103_{- 10}4 _{eV). SE are extra nuclear electrons} bombarded by incident electrons. Because the binding energy between the atomic nucleus and the outer valence electron is very small, when the electron outside the atomic core obtains energy from the incident electron that is greater than the corresponding binding energy, it can leave the atom to become a free electron. SE come from a region 5-10 nm from the surface, and the energy is 0-50 eV. It is very sensitive to the surface state of the sample and can effectively display the micro-morphology of the sample surface. Because it is emitted from the surface of the sample, the incident electrons have not been reflected multiple times, so the area where SE are generated is not much different from the area irradiated by the incident electrons, so the resolution of the SE is higher, which can generally reach 5-10 nm. The resolution of a scanning electron microscope is generally the SE resolution [44].

SEM analyses were performed with FEI Nova 200 Dual beam system which is equipped with a 30 kV SEM field emission gun (FEG) column and a 30 kV focused ion beam (FIB) column with Gallium source.

2.3.2 Two-probe IV-measurements

To test the TE properties of prepared TE films, the data of Seebeck coefficient and electrical conductivity should be collected by using two-probe setup.

and is considered as the cold body. Thin TE films are placed between these Cu-blocks to be able to perform temperature dependent power factor measurements.

Figure 10. The front view of Keithley Tektronix two-probe setup (right), the top view of Keithley Tektronix two-probe setup (left) .

2.3.3 Fourier transform Infrared (FT-IR) spectroscopy

3. Results and discussion

Three flexible TE films were made with three different-percentage TE pastes, photographs of them are shown in Fig. 11. Sizes of films are around 10 x 5 mm2_{, and} the thickness is just below 65 .m. As-prepared TE films are p-type because both Sb2Te3 and PVDF are p-type. And all of them exhibit good flexibility due to the presence of PVDF polymer.

Different drying processes have been tested to obtain the TE films. At the beginning, increasing temperature to 100 o_{C under vacuum, caused the films roll up and some of} them even peeled off. The films made with 85 wt% Sb2Te3 + 15 wt% PVDF. This may be because the polymer shrinks when heated, the film begins to roll up as the proportion of polymer increases. Then, drying them on a hotplate at 50 o_{C, it is the best condition} because the films are flat, but 75 wt% Sb2Te3 + 25 wt% PVDF films rolled up. It may be because the adhesion between pastes and glass is destroyed, when the tape used to define the film area is torn off.

Figure 11. The p-type TE films made with (a) 50 wt% Sb2Te3 + 50 wt% PVDF, (b) 75 wt% Sb2Te3 + 25

wt% PVDF, and (c) 85 wt% Sb2Te3 + 15 wt% PVDF after drying at 50 o

C.

3.1 Scanning Electron Microscopic (SEM) analysis

The morphology and microstructure of p-type TE films were characterized by SEM as shown in Fig. 12, 13, and 14. All p-type TE films contain layered Sb2Te3 nanoplates which have lateral dimension of around 1 .m hexagonal structures. The nanostructure is helpful to improve the power factor (PF) of p-type TE films, as it can introduce some grain boundaries and limit the mean free path which reduce the κ and increase the % simultaneously [20, 45].

carriers while high-energy carriers are unaffected [19]. In addition, porous structures can reduce the κ because of the increase of the air volume fraction in the substrates.

Figure 12. SEM micrographs of p-type TE film made with 85 wt% Sb2Te3 + 15 wt% PVDF were dried

at 50 o

C, at different magnifications.

Fig. 13 clearly indicates morphology and nano-structure of TE film made with 75 wt% Sb2Te3 + 25 wt% PVDF. It still is porous, but its porosity is lower than that of p-type TE film made with 85 wt% Sb2Te3 + 15 wt% PVDF, due to the increase in the percentage of PVDF.

Figure 13. SEM micrographs of p-type TE film made with 75 wt% Sb2Te3 + 25 wt% PVDF were dried

at 50 o

C, at different magnifications.

Figure 14. SEM micrographs of p-type TE film made with50 wt% Sb2Te3 + 50 wt% PVDF were dried

at 50 o

3.2 Fourier transform infrared (FT-IR) spectroscopy analysis

FT-IR spectrum of PVDF have been investigated extensively since it can provide valuable information about the material structure. PVDF is a polymer with different crystalline structures (α, β, and γ phases) which can be distinguished by using FT-IR spectroscopy. Depending on the preparation method, a PVDF sample might contain more than one of these crystalline phases and since some of the band structures are similar in these phases, it is hard to differentiate between them. One should look for the characteristic peaks in the spectrum to identify the phases in the sample.

Fig. 15 shows the FT-IR spectrum of PVDF. CF2 bending can be showed as FT-IR peaks at 511 cm-1 _{and 532 cm}-1_{, the FT-IR peaks at 614 cm}-1 _{and 763 cm}-1 _represent CH2 rocking, and the FT-IR peaks at 854 cm-1 _{and 975 cm}-1 _{represent C-H out of plane} [46].

Figure 15. FT-IR spectra of PVDF with identification of the ! and " phases characteristics peaks.

Table 4. Identification of % and C phase characteristics peaks of PVDF. Peak number Wavenumber (cm-1₎ Crystal phase Peak 1 489 % phase Peak 2 510 & phase Peak 3 532 % phase Peak 4 614 % phase Peak 5 763 % phase Peak 6 795 % phase Peak 7 841 & phase Peak 8 855 % phase Peak 9 976 ( phase Peak 10 1279 & phase

Figure 16. (a) FT-IR spectra of three different phase PVDF [47], (b) FT-IR of as-used PVDF sample.

Above all, the PVDF we have used contain both % and C phases. FT-IR results can be used to quantify the content of C phase PVDF which exhibits good electroactive properties. The relative fraction of the C phase in a sample containing just % phase and C phase PVDF can be calculated as follows [47]:

O(C) = P&

Q@_&⁄ SP@_{' '} + P_&

where F(C), represents the C phase content; P_' and P_& the intensity of absorption bands at 766 cm−1 _{and 840 cm}−1_{; @}

' and @& are the absorption coefficients at the

respective wavenumber, @' and @& values are 6.1 × 104 and 7.7 × 104 cm2 mol−1,

phase than C phase again. And β phase PVDF shows best piezoelectric, ferroelectric, and pyroelectric properties among the five polymorphs.

Figure 17. FT-IR of PVDF in absorption mode.

3.3 Electrical properties

Room temperature measurements have been performed to determine the resistance of the as-made films; results are shown in Table 5 for consecutive ten measurements performed in samples with different percentage of Sb2Te3/PVDF. The average resistance of nanomaterials for p-type TE films made with 85 wt% Sb2Te3 +15 wt% PVDF is calculated as 156.97 MΩ, while it was 0.126 MΩ for the sample with 50 wt% Sb2Te3 + 50 wt% PVDF composition, respectively.

The resistivity (:) and $ can be calculated from the resistance and dimension of the TE films. For the p-type TE film made with 85 wt% Sb2Te3 +15 wt% PVDF, the : and $ are 5101.53 Ω·m and 0.0002 S/m. And the resistivity and $ of the p-type TE film made of 50 wt% Sb2Te3 +50 wt% PVDF are 4.095 Ω·m and 0.244 S/m.

Table 6. Electrical properties of different p-type TE films at room temperature. Composition of TE films T (Ω·m) * (S/m)

85 wt% Sb2Te3 +15 wt% PVDF 5101.53 0.0002

50 wt% Sb2Te3 +50 wt% PVDF 4.095 0.244

4. Conclusions

5. Future work

The TE performance of films should be measured, such as the %. But due to COVID-19, I can no longer go to the laboratory to perform any measurement.

Porous structures have an obvious influence on thermal conductivity and electrical conductivity. Thus, it will be interesting to explore how to obtain minimum U and maximum $ by manipulating the porous size and porosity.

Besides, different sizes and morphologies of nano- or micro-sized Sb2Te3 should be mixed with PVDF to prepare p-type TE films. Through this process to study and explore the best size and morphology to enhance TE performance. As shown above, β phase PVDF shows the best electric properties among the five polymorphs. Therefore, the percentage of β phase in PVDF sample should be increased.

6. Acknowledgement

I want to say thank you to everyone who gave me support and help during my master's degree.

First of all, I sincerely appreciate my examiner Prof. Muhammet S. Toprak who provided me opportunity to do my degree project in the Nano-Chemistry group, KTH-School of Engineering Sciences, Department of Applied Physics, Stockholm, Sweden. At the beginning of the project, he gave me patience and guidance on how to build a knowledge framework. When I encountered confusion during the experiment, he was always happy to help me.

And I must say thank you to my supervisor Ms. Hazal Batili. For me, she is the best supervisor (not one of the best supervisors) in the world. I am very grateful to her for helping me with my degree thesis. From the determination of the topic of the thesis, the arrangement of the chapter structure, to the academic style of the thesis content, and even the format of the notes, she has devoted a lot of effort. I often meet all kinds of problems on literatures and lab, she always give me guidance patiently. This is a great help for me so that I can correctly grasp the research direction of the paper and successfully complete the paper. More importantly, she gave me full encouragement and trust, allowing me to write the paper with full confidence.

Besides, I want to say thank you to Mr. Hamawandi for his help when I did the lab. He taught me the doctor blading and how to use TE setup.

Bibliography

[1] M. Hamid Elsheikh et al., “A review on thermoelectric renewable energy: Principle parameters that affect their performance,” Renew. Sustain. Energy Rev., vol. 30, pp. 337–355, 2014, doi: 10.1016/j.rser.2013.10.027.

[2] H. J. Goldsmid, “The Improvement of a Specific Material---Bismuth Telluride,” in Introduction to Thermoelectricity, Berlin, Heidelberg: Springer Berlin Heidelberg, 2016, pp. 85–107, doi: 10.1007/978-3-662-49256-7_6.

[3] J. F. Li, W. S. Liu, L. D. Zhao, and M. Zhou, “High-performance nanostructured thermoelectric materials,” NPG Asia Mater., vol. 2, no. 4, pp. 152–158, 2010, doi: 10.1038/asiamat.2010.138.

[4] M. Bharti, A. Singh, S. Samanta, and D. K. Aswal, “Conductive polymers for thermoelectric power generation,” Prog. Mater. Sci., vol. 93, pp. 270–310, 2018, doi: 10.1016/j.pmatsci.2017.09.004.

[5] L. Cigarini, A. Ruini, A. Catellani, and A. Calzolari, “Thermoelectric figure of merit of polymeric systems for low-power generators,” J. Phys. D. Appl. Phys., vol. 50, no. 39, 2017, doi: 10.1088/1361-6463/aa84f7.

[6] A. Bulusu and D. G. Walker, “Review of electronic transport models for thermoelectric materials,” Superlattices Microstruct., vol. 44, no. 1, pp. 1–36, 2008, doi: 10.1016/j.spmi.2008.02.008.

[7] J. He and T. M. Tritt, “Advances in thermoelectric materials research: Looking back and moving forward,” Science (80-. )., vol. 357, no. 6358, 2017, doi: 10.1126/science.aak9997.

[8] C. Gayner and K. K. Kar, “Recent advances in thermoelectric materials,” Prog.

Mater. Sci., vol. 83, pp. 330–382, 2016, doi: 10.1016/j.pmatsci.2016.07.002.

[9] W. H. Clingman, “Entropy production and optimum device design,” Adv. Energy

Convers., vol. 1, no. C, pp. 61–79, 1961, doi: 10.1016/0365-1789(61)90008-X.

[10] Y. Pei et al., “Stabilizing the optimal carrier concentration for high thermoelectric efficiency,” Adv. Mater., vol. 23, no. 47, pp. 5674–5678, 2011, doi: 10.1002/adma.201103153.

[11] O. Bubnova and X. Crispin, “Towards polymer-based organic thermoelectric generators,” Energy Environ. Sci., vol. 5, no. 11, pp. 9345–9362, 2012, doi: 10.1039/c2ee22777k.

[13] N. Toshima, “Recent progress of organic and hybrid thermoelectric materials,”

Synth. Met., vol. 225, pp. 3–21, 2017, doi: 10.1016/j.synthmet.2016.12.017.

[14] A. Pereira Gonalves, E. Branco Lopes, O. Rouleau, and C. Godart, “Conducting glasses as new potential thermoelectric materials: The Cu-Ge-Te case,” J. Mater.

Chem., vol. 20, no. 8, pp. 1516–1521, 2010, doi: 10.1039/b908579c.

[15] Z. Li, C. Xiao, H. Zhu, and Y. Xie, “Defect Chemistry for Thermoelectric Materials,” J. Am. Chem. Soc., vol. 138, no. 45, pp. 14810–14819, 2016, doi: 10.1021/jacs.6b08748.

[16] T. J. Zhu, L. P. Hu, X. B. Zhao, and J. He, “New insights into intrinsic point defects in v2vi3 thermoelectric materials,” Adv. Sci., vol. 3, no. 7, 2016, doi: 10.1002/advs.201600004.

[17] Y. Pei, X. Shi, A. Lalonde, H. Wang, L. Chen, and G. J. Snyder, “Convergence of electronic bands for high performance bulk thermoelectrics,” Nature, vol. 473, no. 7345, pp. 66–69, 2011, doi: 10.1038/nature09996.

[18] X. Chong, L. Zhou, and X. Yi, “Synergistic Optimization of Electrical and Thermal Transport Properties in Chalcogenides Thermoelectric Materials,” J.

Chem. Inf. Model., vol. 53, no. 9, pp. 1689–1699, 2013, doi:

10.1017/CBO9781107415324.004.

[19] F. P. Du et al., “Fabrication of Porous Polyvinylidene Fluoride/Multi-Walled Carbon Nanotube Nanocomposites and Their Enhanced Thermoelectric Performance,” Polymers (Basel)., vol. 10, no. 7, pp. 1–12, 2018, doi: 10.3390/polym10070797.

[20] K. Biswas et al., “High-performance bulk thermoelectrics with all-scale hierarchical architectures,” Nature, vol. 489, no. 7416, pp. 414–418, 2012, doi: 10.1038/nature11439.

[21] K. Biswas et al., “Strained endotaxial nanostructures with high thermoelectric figure of merit,” Nat. Chem., vol. 3, no. 2, pp. 160–166, 2011, doi: 10.1038/nchem.955.

[22] L. D. Zhao et al., “High thermoelectric performance via hierarchical compositionally alloyed nanostructures,” J. Am. Chem. Soc., vol. 135, no. 19, pp. 7364–7370, 2013, doi: 10.1021/ja403134b.

[23] Y. F. Wang, K. H. Lee, H. Ohta, and K. Koumoto, “Fabrication and thermoelectric properties of heavily rare-earth metal-doped SrO(SrTiO3)n (n = 1, 2) ceramics,” Ceram. Int., vol. 34, no. 4, pp. 849–852, 2008, doi: 10.1016/j.ceramint.2007.09.034.

[24] E. Y. Liu, F. H. Lin, Z. R. Yang, and C. J. Liu, “A facile energy-saving route of fabricating thermoelectric Sb2Te3-Te nanocomposites and nanosized Te,” R.

[25] C. K. Chiang et al., “Electrical Conductivity in doped polyactylene,” Phys. Rev.

Lett., vol. 40, no. 22, p. 1472, 1978, doi: 10.1103/PhysRevLett.39.1098.

[26] G. Li, R. Zhu, and Y. Yang, “Polymer solar cells,” Nat. Photonics, vol. 6, no. 3, pp. 153–161, 2012, doi: 10.1038/nphoton.2012.11.

[27] Y. Lu, J. Y. Wang, and J. Pei, “Strategies to Enhance the Conductivity of n-Type Polymer Thermoelectric Materials,” Chem. Mater., vol. 31, no. 17, pp. 6412– 6423, 2019, doi: 10.1021/acs.chemmater.9b01422.

[28] Y. Du, S. Z. Shen, K. Cai, and P. S. Casey, “Research progress on polymer-inorganic thermoelectric nanocomposite materials,” Prog. Polym. Sci., vol. 37, no. 6, pp. 820–841, 2012, doi: 10.1016/j.progpolymsci.2011.11.003.

[29] J. Y. Kim, W. Lee, Y. H. Kang, S. Y. Cho, and K. S. Jang, “Wet-spinning and post-treatment of CNT/PEDOT:PSS composites for use in organic fiber-based thermoelectric generators,” Carbon N. Y., vol. 133, pp. 293–299, 2018, doi: 10.1016/j.carbon.2018.03.041.

[30] G. H. Kim, L. Shao, K. Zhang, and K. P. Pipe, “Engineered doping of organic semiconductors for enhanced thermoelectric efficiency,” Nat. Mater., vol. 12, no. 8, pp. 719–723, 2013, doi: 10.1038/nmat3635.

[31] J. Luo et al., “Enhancement of the thermoelectric properties of PEDOT:PSS thin films by post-treatment,” J. Mater. Chem. A, vol. 1, no. 26, pp. 7576–7583, 2013, doi: 10.1039/c3ta11209h.

[32] X. Wang et al., “High performance of PEDOT:PSS/SiC-NWs hybrid thermoelectric thin film for energy harvesting,” J. Alloys Compd., vol. 734, pp. 121–129, 2018, doi: 10.1016/j.jallcom.2017.11.013.

[33] H. Song et al., “Improved thermoelectric performance of free-standing PEDOT:PSS/Bi 2Te3 films with low thermal conductivity,” J. Electron. Mater., vol. 42, no. 6, pp. 1268–1274, 2013, doi: 10.1007/s11664-013-2587-y.

[34] H. Song, C. Liu, J. Xu, Q. Jiang, and H. Shi, “Fabrication of a layered nanostructure PEDOT:PSS/SWCNTs composite and its thermoelectric performance,” RSC Adv., vol. 3, no. 44, pp. 22065–22071, 2013, doi: 10.1039/c3ra42414f.

[35] Y. Lu, Y. Song, and F. Wang, “Thermoelectric properties of graphene nanosheets-modified polyaniline hybrid nanocomposites by an in situ chemical polymerization,” Mater. Chem. Phys., vol. 138, no. 1, pp. 238–244, 2013, doi: 10.1016/j.matchemphys.2012.11.052.

[36] H. Ju, D. Park, and J. Kim, “Conductive polymer based high-performance hybrid thermoelectrics: Polyaniline/tin(II) sulfide nanosheet composites,” Polymer

(Guildf)., vol. 160, no. August 2018, pp. 24–29, 2019, doi:

[37] N. Toshima and S. Ichikawa, “Conducting Polymers and Their Hybrids as Organic Thermoelectric Materials,” J. Electron. Mater., vol. 44, no. 1, pp. 384– 390, 2015, doi: 10.1007/s11664-014-3312-1.

[38] D. Chaochao et al., “Layered Bi2Se3 nanoplate/polyvinylidene fluoride composite based n-type thermoelectric fabrics,” ACS Appl. Mater. Interfaces, vol. 7, no. 13, pp. 7054–7059, 2015, doi: 10.1021/acsami.5b00514.

[39] A. California et al., “Tailoring porous structure of ferroelectric poly(vinylidene fluoride-trifluoroethylene) by controlling solvent/polymer ratio and solvent evaporation rate,” Eur. Polym. J., vol. 47, no. 12, pp. 2442–2450, Dec. 2011, doi: 10.1016/j.eurpolymj.2011.10.005.

[40] L. Ruan, X. Yao, Y. Chang, L. Zhou, G. Qin, and X. Zhang, “Properties and Applications of the β Phase Poly(vinylidene fluoride),” pp. 1–27, 2018, doi: 10.3390/polym10030228.

[41] R. Song, D. Yang, and L. He, “Effect of surface modification of nanosilica on crystallization, thermal and mechanical properties of poly(vinylidene fluoride),”

J. Mater. Sci., vol. 42, no. 20, pp. 8408–8417, 2007, doi:

10.1007/s10853-007-1787-3.

[42] B. Dutta, E. Kar, N. Bose, and S. Mukherjee, “Significant enhancement of the electroactive β-phase of PVDF by incorporating hydrothermally synthesized copper oxide nanoparticles,” RSC Adv., vol. 5, no. 127, pp. 105422–405434, 2015, doi: 10.1039/c5ra21903e.

[43] J. Xiong, F. Jiang, W. Zhou, C. Liu, and J. Xu, “Highly electrical and thermoelectric properties of a PEDOT:PSS thin-film via direct dilution-filtration,” RSC Adv., vol. 5, no. 75, pp. 60708–60712, 2015, doi: 10.1039/c5ra07820b.

[44] Y. Leng, Marerials Characterization: Intriduction to Microscopic and

Spectroscopic Methods. 2013.

[45] N. Toshima, M. Imai, and S. Ichikawa, “Organic-inorganic nanohybrids as novel thermoelectric materials: Hybrids of polyaniline and bismuth(III) telluride nanoparticles,” J. Electron. Mater., vol. 40, no. 5, pp. 898–902, 2011, doi: 10.1007/s11664-010-1403-1.

[46] B. Wang and H. X. Huang, “Incorporation of halloysite nanotubes into PVDF matrix: Nucleation of electroactive phase accompany with significant reinforcement and dimensional stability improvement,” Compos. Part A Appl.

Sci. Manuf., vol. 66, pp. 16–24, 2014, doi: 10.1016/j.compositesa.2014.07.001.

[47] P. Martins, A. C. Lopes, and S. Lanceros-Mendez, “Electroactive phases of poly(vinylidene fluoride): Determination, processing and applications,” Prog.

TRITA -SCI-GRU 2020:196