Nanostructuring Bi₂Te₃-Based Thermoelectric Thin-Films Grown Using Pulsed Laser Deposition

Abstract

This book chapter reports recent advances in nanostructured Bi2Te3-based thermoelectric (TE) thin-films fabricated by pulsed laser deposition (PLD). By controlling the processing conditions in PLD growths, various fascinating Bi2Te3-based nanostructured films with promising or enhanced TE properties have been successfully fabricated, including super-assembling of Bi2Te3 hierarchical nanostructures, self-assembled Bi2Te3 films with well-aligned 0D to 3D nanoblocks, polycrystalline-nanostructured Bi2Se3 and Bi2Te3 thin-films, etc. In addition, a PLD-growth mechanism for fabricating the super-assembling Bi2Te3 thin-films is presented. This book chapter provides fundamental understanding the relationship amongst processing condition, structure-morphology, and TE property of PLD-growths Bi2Te3-based thin-films. It also presents an overview of TE materials and applications with the challenges and perspectives.

1. Introduction

Thermoelectric materials are solid-state energy converters whose combination of thermal, electrical, and semiconducting properties allows them to be used to convert waste heat into electricity or electrical power directly into cooling and heating [1].

1.1 Thermoelectric effects

When an electric current flows through a pair of p-type and n-type semiconductors connected in series (Figure 1(a), the holes in the p-type material and the electrons in the n-type material carry heat away from the top metal–semiconductor junctions, which leads to a cooling at the junctions called the Peltier effect. When current flows within the module, one side is cooled and the other heated. If the current is reversed, the hot and cold sides reverse also. For each material, the cooling effect is gauged by the Peltier coefficient Π that relates the heat carried by the charges to the electrical current through [1, 2, 4]: Q = Π×I.

Figure 1.

Illustration of TE devices: (a) cooler (Peltier effect), (b) power generator (Seebeck effect). Redrawn after Ref. [2]. (c) Thermoelectric module showing the direction of charge flow on both cooling and power generation [3].

In Figure 1(b), when the two ends of the materials maintain a temperature difference, the higher thermal energy holes and electrons will diffuse from the hot side to the cold side, and consequently a potential difference is created. This is Seebeck effect and it is the principle for thermocouples. The power generation is measured by the Seebeck coefficient α, which relates the voltage generated to the temperature difference through ΔV = −αΔT. The Peltier and the Seebeck coefficients are related through the Kelvin relation [1, 2]: Π = αT.

Thermoelectric devices contain many thermoelectric couples (Figure 1c, bottom), which consist of p-type (containing free holes) and n-type (containing free electrons) thermoelectric elements connected electrically in series and thermally in parallel (Figure 1c, top). A thermoelectric generator uses heat flow across a temperature gradient to power an electric load through the external circuit.

1.2 The thermoelectric figure of merit (ZT)

The performance of the thermoelectric materials is often denoted as figure of merit Z whose unit is K⁻¹, or ZT the dimensionless unit [5, 6].

$$\mathit{ZT}=\frac{{\alpha }^2\sigma }{\kappa }T=\frac{{\alpha }^2\sigma }{{\kappa }_E+{\kappa }_L}T$$E1

where α, σ, T, and κ are the Seebeck coefficient, electrical conductivity, absolute temperature, and thermal conductivity, respectively. The total thermal conductivity can be split into electronic contribution (κ_E) and lattice contribution (κ_L). The thermoelectric power factor (PF) is calculated by the quantity of α²σ. The efficiency of a thermoelectric material is determined by its ZT. Meanwhile, the maximum efficiency (η) of a power generation is expressed by [3, 7]:

$$\eta =\frac{T_h-T_c}{T_h}\cdot \frac{\sqrt{1+Z\overline{T}}-1}{\sqrt{1+Z\overline{T}}+\frac{T_c}{T_h}}$$E2

and the coefficient of performance presents for the efficiency of air-conditioning and refrigeration [7]:

$$\mathit{COP}=\frac{T_c}{T_h-T_c}\cdot \frac{\sqrt{1+Z\overline{T}}-\frac{T_h}{T_c}}{\sqrt{1+Z\overline{T}}+1}$$E3

where T_h and T_c are the hot-end and cold-end temperature of the thermoelectric materials, respectively, and $\overline{T}$ is the average temperature of T_h and T_c. For practical applications, it is important to use high ZT thermoelectric materials.

The best materials so far are alloys of Bi₂Te₃ with Sb₂Te₃ and Bi₂Te₃ with Bi₂Se₃. ZT is of the order of 1 at room temperature. This value gives a COP of about 1 (Figure 2a), which is still far lower than the COP = 2–4 of household refrigerators and air conditioners. Similar situation is true for power generation (Figure 2b) [2, 8]. Thermoelectric cooling and power generation generally still not competitive with the other energy conversion methods.

Figure 2.

Comparison of thermoelectric technology with other energy conversion methods for (a) cooling and (b) power generation [2, 8].

1.3 Challenges in enhancing ZT

A concept of “phonon-glass electron-crystal” (or PGEC in short) was proposed for designing efficient thermoelectric materials. This is a controversial concept from the aspect of materials science that the materials should have a high electrical conductivity as in a crystal and a low lattice thermal conductivity as in a glass [9]. However, the TE parameters are strongly interdependent, which makes the enhancement efforts of ZT very challenging. A normal approach for the enhanced properties of TE materials is to increase the power factor α²σ by optimizing the carrier concentration n, and/or to reduce the lattice thermal conductivity κ_L by introducing the scattering centers. These parameters are the function of carrier effective mass m^* and carrier mobility μ, scattering factor r, and their interconnectivity limit ZT to approximately 1 in large bulk materials [10].

The kinetic definition of α is the energy difference between the average energy of mobile carriers and the Fermi energy [11]. When carrier concentration (n) is increased, both the Fermi energy and the average energy increase, but the Fermi energy increases more rapidly than the average energy as n is increased. Consequently, α decreases and thus α²n is dragged down rapidly. Therefore, the carrier concentration (n) increases electrical conductivity (σ) but reduces the Seebeck coefficient (α) for most of the homogeneous materials. For this reason, in metals and degenerate semiconductors with energy-independent scattering approximation, the Seebeck coefficient can be expressed as [3, 12]:

$$\alpha =\frac{8{\pi }^2{k}_B^2}{3{\mathit{eh}}^2}{m}^{\ast }T{\left(\frac{\pi }{3n}\right)}^{2/3}$$E4

Where, the parameter m^* is density of states effective mass, and an increase of m^* can raise the Seebeck coefficient according to the Eq. (4). However, most high m^* materials have generally low μ which limits the α by a weighted mobility with a factor proportional to (m^*)^3/2μ. Moreover, there is no such thing as an optimal effective mass. There are high mobility low effective mass semiconductors (SiGe, GaAs) as well as low mobility high effective mass polaron conductors (oxides, chalcogenides) [3].

Noticeably, the defects scatter not only the phonons but also the electrons. When a thermoelectric material is designed for reducing lattice thermal conductivity, its carrier mobility is usually suppressed. Hence, the ratio of μ/κ_L determines the improvement of ZT [5, 10]. The ratio is observed to increase experimentally through a more reduction in κ_L rather than that in μ, but some fundamental issues in this mechanism are not understood well [10].

The electrical resistivity (ρ) and electrical conductivity (σ) are related to n through the carrier mobility μ:

$$1/\rho =\sigma =\mathit{ne\mu }$$E5

The electronic contribution to the thermal conductivity is proportional to the electrical conductivity (σ) of the materials according to Wiedemann–Franz Law [3], and the relationship is expressed as follows:

$$\mathit{\kappa e}=\mathit{L\sigma T}=\mathit{\text{neμLT}}$$E6

where ‘e’ is electron charge, and L is Lorenz factor 2.48 × 10⁻⁸ J²/K²C² for free electrons and this can vary particularly with carrier concentration [3, 13].

Figure 3 shows the compromise of σ, κ and α in thermoelectric materials that must be optimized to maximize the figure of merit ZT. Indeed, the lower carrier concentration will result in the lower σ and a decreasing ZT. Typically, the PF and ZT peaks occur at carrier concentrations of 10¹⁹–10²¹ cm⁻³ (depending on the material system), which falls in between common metals and heavily doped semiconductors [3]. High mobility carriers are most important for high value of electrical conductivity. Again from the Eq. (4), an increase of the carrier effective mass lead to increase the α but reduce the μ and hence the σ according to the Eq. (5). In case of the narrow semiconductor, the thermal excitation of carrier from valence band to conduction band creates holes and electrons. However, the concentration of the major carrier does not vary much. When two types of carriers are present, or bipolar effects takes place, and this is notorious to achieve effective thermoelectrics [4]. For example, the Seebeck coefficient for different carrier types is given by a weighted average of their electrical conductivity values (σ_e and σ_p) [13].

Figure 3.

Maximizing the efficiency (ZT) of a thermoelectric involves a compromise of thermal conductivity (κ; plotted on the y-axis from 0 to a top value of 10 Wm⁻¹ K⁻¹) and Seebeck coefficient (α; 0–500 μVK⁻¹) with electrical conductivity (σ; 0–5000 Ω⁻¹ cm⁻¹) [3].

$$\alpha \approx \frac{{\alpha }_e{\sigma }_e+{\alpha }_p{\sigma }_p}{\left({\sigma }_e+{\sigma }_p\right)}$$E7

In short, any attempt to increase σ, will increase κ_e which contributes to thermal conductivity (κ). In order to counter the increment of κ_e, various approaches are reported to reduce κ_L. However, decreasing κ_L with phonon scattering by adding defects results in decrease in n and σ. These are the major conflicts in the properties of bulk thermoelectric materials which have been addressed in the researches for more than a half century [10].

1.4 Nanostructuring thermoelectric materials

In classical physics, the coefficients α, κ_e and σ are interrelated in such a way that it is impossible to increase one without affecting the others. Therefore, a compromise has to be achieved to find the maximum ZT value. Three different strategies have appeared to improve the ZT [14]:

1. An approach for increasing α while keeping the values of σ and κ_e by looking for new materials with complex band structures, like heavy fermion compounds.

2. Controlling the disorder in materials (such as Skutterudites or Clathrates) to present a rattling effect which causes, (↑) σ and decreases (↓) κ_L (see for instance ref. [15]).

3. Developing nanostructured materials that could lead to (↑) α due to quantum confinement effects, while ↓κ_L due to the scattering of phonons at the interfaces. The latest improvements in the ZT of different materials has been achieved by this approach.

In 1993, Hicks and Dresselhaus pioneered the concept of nanostructuring in design of thermoelectric materials (i.e. Bi₂Te₃). The addition of the dimensionality and size of the system is added as a new parameter that affects the coupling of the electrical conductivity, Seebeck coefficient, and thermal conductivity, leading to substantially enhanced ZT [16, 17, 18]. Two ideas are dominant for the low-dimensional materials approach for improving ZT. Firstly, the presence of nanoscale constituents would introduce quantum confinement effects to enhance Seebeck coefficient and the power factor α²σ. Secondly, the numerous internal nanoinclusions and interfaces found in nanostructures would be designed so that the thermal conductivity would be reduced more than the electrical conductivity, based on differences in their respective scattering lengths [16].

As the dimensionality is decreased from 3D crystalline solids to 2D (quantum wells) to 1D (quantum wires) and finally to 0D (quantum dots), the spatial confinement are introduced that create the possibilities to tune the TE properties α, σ, and κ independently. When the system size decreases and approaches the scale comparable to the feature length of electron behavior (e.q. mean free path and wavelength) in any direction, the electronic density of states (D.O.S.) can split and become narrow as well as increase substantially (Figure 4a), resulting in the enhancement of α. Meanwhile, the thermal conductivity is also reduced because of the extensive phonon scattering at the surface, interfaces, and grain boundaries, as any dimension is less than the mean free path of phonons. Figure 4(b) illustrates examples of different nanostructuring with different dimensionalities [14]. A schematic diagram is shown in Figure 4(e) capturing these various phonon scattering mechanisms, along with the electrical transport within a thermoelectric material. For example, in material embedded nano-inclusions (nanoparticles), atomic defects are effective at scattering short wavelength phonons, but larger embedded nanoparticles are required to scatter mid- and long-wavelength phonons effectively. Grain boundaries can also play an effective role in scattering these longer-wavelength phonons [20].

Figure 4.

(a) Electronic density of states (D.O.S.) for a bulk 3D crystalline semiconductor, a 2D quantum well, a 1D nanowire or nanotube, and a 0D quantum dot [16]. (b) Examples of different nanostructuring with different dimensionalities [14]. (c) A spike in the density of states (solid line) above the bulk value (dashed line) occurs due to resonant states in Tl-doped PbTe [19]. (d) The measured ZT of Tl-PbTe and Na-PbTe samples for 300–800 K indicates an improvement due to the addition of Tl [19]. (e) Schematic diagram illustrating various phonon scattering mechanisms within a thermoelectric material, along with electronic transport of hot and cold electrons [20].

Figure 5 plots major milestones achieved for ZT over the past several decades as a function of both year and temperature [20]. In the 1950s, Bi₂Te₃ was first investigated as a material of great thermoelectric with ZT∼0.6 near room temperature [5, 6]. It was quickly realized that alloying with Sb₂Te₃ and Bi₂Se₃ allowed for the fine tuning of the carrier concentration alongside a reduction in lattice thermal conductivity. These compounds have played a dominant role in the field of thermoelectrics through today. The alloys of Bi₂Te₃ with Sb₂Te₃ (such as Bi_0.5Sb_1.5Te₃; p type) and of Bi₂Te₃ with Bi₂Se₃ (such as Bi₂Te_2.7Se_0.3; n type), with a ZT ∼ 1 at room temperature are traditional cooling materials [6]. In recent year, great enhancements in ZT owning to low dimension and nanostructure materials have been reported [19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32] and achieved the highest ZT value of approximately 2.4.

Figure 5.

Thermoelectric figure-of-merit ZT as a function of temperature and year illustrating important milestones [20]. Although there have been several demonstrations of ZT > 1 in the past decade (2001–2010), no material has yet achieved the target goal of ZT ≥ 3. The material systems that have achieved ZT > 1 have all been based on some form of nanostructuring.

1.5 Overview of thermoelectric applications

The solid-state devices based on TE effect have the inherent advantages of reliability, silent and vibration-free operation (no moving fluids or moving parts), a very high power density, and the ability to maintain their efficiency in small scale applications where only a moderate amount of power is needed [19].

Commercial use has been made mostly from Peltier’s thermoelectric cooling (TEC) effect in applications, as demonstrated in Figure 6 [35]:

• Small refrigerator devices are used for camping and outdoor activities. For example, the cooler/warmer TE device (Engel Thermo 8) has volume 8 L and weighing just over 3 kg. Its features include cooling performance up to 22°C below ambient temperature and warming up to +65°C.

• Gentherm designed and developed Automotive Climate Control Seat [36], which has TE heat pumps in the back and bottom cushions. The TE system makes conditioned flowing air through channels to the occupant for providing on-demand cooling or heating. As shown in the first panel in Figure 6, the seat has the heat pump consisting of a TE module (green box) and a fan (orange).

• Thermal management of tiny laser diodes is used in fiber optic telecom, datacom backhaul networks. TEC can also be used for contact cooling of semiconductor lasers, infrared detectors, CCD- matrix, and miniconditioners for photomultipliers.

• Localized cooling at hot spots of chips was created. For example, the Intel group is the first to demonstrate both concepts of applying the TE material only to a chip’s hottest spots (Figure 6) [33, 37]. On the substrate, the researchers grew a 100-μm-thick layered structure, called a superlattice, containing bismuth, tellurium, antimony, and selenium. The structure can pump 1300W/cm² heat from the back side of the chip to the heat spreader. The superlattice induced an approximately 6°C temperature drop at the hot spot even before the device was powered up, because it conducts heat better than the grease that bonds the rest of the heat spreader to the chip. Yet, when a 3 A- current went through the thermoelectric cooler, the total temperature change was only of 15°C. Managing heat in electronics is a common issue, and TE coolers can improve electronic systems in thermal performance, cost, noise, weight, size or efficiency.

Figure 6.

Overview of potential thermoelectric cooling (TEC) applications [33, 34].

Figure 7 shows an overview of the present and potential applications of thermoelectric generators (TEGs) [34]. They include (1) heating systems and water boilers with TEG units which generate the electricity for the control units and pumping systems, (2) the long term perspective of waste heat recovery for medium-scale industrial facilities, (3) waste heat recovery in automobiles and other combustion-engine-powered vehicles for enhanced efficiency and electric current supply of the electronic system, (4) miniaturized autarkic sensor systems powered by an integrated TEG with a wireless data transmitter, (5) ventilated wood stove powered by a thermoelectric generator with enhanced oxygen supply, improves burning process.

Figure 7.

Overview of potential thermoelectric generator (TEG) applications [33, 34].

2. Nanostructured Bi₂Te₃-based thermoelectric thin films grown using pulsed laser deposition

2.1 PLD growths of nanostructured Bi₂Te₃-based thin films

PLD is one of the most convenient thin film growth techniques that uses a high intensity pulsed laser beam as an external energy source to ablate a target, form a plume, and deposit thin films onto a substrate. In practice, a large number of variables affect the properties of the film, such as substrate temperature (T_s), background gas pressure (P) and laser fluence. Figure 8 shows a PLD system for preparing thermoelectric thin films [38, 39]. The substrate was heated and maintained at desired T_s using a thermocouple and a proportional-integral-derivative temperature controller. The thermocouple was buried inside a substrate holder which was heated by a tungsten lamp or electrical resistance heating. The pressure of ambient gas (He, Ar) could be fine-tuned by the needle valve. Laser source can be KrF excimer laser beam (λ = 248 nm) and Q-switched Nd:YAG laser (λ = 355 nm) with properly selected laser fluence (e.g., 3.8, 6.2, or 8.3 J/cm²) pulsed duration of 5–20 ns, repetition rate of 5–10 Hz [38, 40, 41, 42]. The laser beam was guided by several UV mirrors and focused on a stoichiometric polycrystalline target (e.g., Bi₂Se₃, Bi₂Te₃, Bi_0.5Sb_1.5Te₃, etc.) inside the vacuum chamber by the UV lens. The deposition chamber was evacuated to a base pressure of ∼10⁻⁶ Torr, and high-purity ambient gas (He or Ar) was then introduced until obtaining a target pressure (e.g., usually 10⁻⁵ – 3×10⁻¹ Torr).

Figure 8.

A schematic illustration of a PLD system.

The enhancement of the PF of Bi₂Te₃-based thin films is challenging due to the coupling among TE material properties [3], and the difficulty in growing stoichiometric films [38]. Indeed, stoichiometry is a key factor for obtaining better TE properties [5, 38, 43, 44, 45]. Yet, both tendency for re-evaporation of volatile elements (i.e., Te, Se) at elevated T_s [45, 46, 47, 48] and the low sticking coefficient Te (< 0.6 for Bi₂Te₃) at T_s beyond 300°C [49, 50] constrain to grow stoichiometric Bi₂Te₃-based films (Figure 9a and b).

Figure 9.

(a) Vapor pressures of Bi, Sb, Te, Se, Bi₂Se₃, and Bi₂Te₃ as a function of temperature [46]. (b) The variation of sticking coefficient K_s (Bi, Te) as a function of substrate temperature T_s at fixed flux ratio F_R = 4.5 [49].

2.2 Super-assembling of Bi₂Te₃ hierarchical nanostructured thin films

C.-H. Chen et al. [41] reported the PLD growths of super-assembling of Bi₂Te₃ hierarchical nanostructured thin films on the SiO₂/Si substrates and their thermoelectric properties. Interesting Bi₂Te₃ super-assemblies were successfully grown using PLD with controlling the substrate temperatures from 350–600°C and at a fixed Ar ambient pressure of approximately 10⁻³ Torr. SEM images in Figure 10 clearly shows the morphological characteristics of the superassembling Bi₂Te₃ nanostructured thin films [41]. At lower deposition temperatures (< 450°C), the films are mainly composed of vertically aligned nanoscaled flakes, but flakes are horizontally stacked for 600°C-film (Figure 10a–d). Moreover, the bottom of each of the deposited super-assemblies has a relatively continuous and dense layer, and this layer thickness increases with increasing substrate temperature from 350–450°C (Figure 10a–d). The top-view SEM images confirm for the high uniformity and presents the unique super-assembling features of the repetitively and regularly assembled nano-flakes (Figure 10e–h). These four films are uniformly composed of spindle-like (Figure 10e), worm-like (Figure 10f and g) and island-like (Figure 10h) hierarchical nanostructures. Magnified top-view SEM images (Figure 10i–k) further show that the nanoflakes are composed of oriented and regular assemblies of numerous rice-like and elongated primitive nanoparticles [41]. At a higher substrate temperature, thin- and large-size nanoflakes are formed from packing of dense rice-like nanoparticles, driving by the relatively sufficient thermal energy for diffusion. In addition, the out-of- plane superassembly structure (600°C) has a limited column width, which is not always consistent along the out-of-plane direction (Figure 10d). Also, the parallel nano-flakes (at 600°C) are evidently formed by flake stacking along c-axis orientation or epitaxial-like growth. The special three-dimensional mesh-like structure of 600°C-film would also be an effective design for scattering phonons, and it’s extremely smooth top surface is certainly beneficial for subsequent analyses and applications [41].

Figure 10.

(a)–(d) The cross-sectional and (e)–(l) the corresponding top-view SEM images of the Bi₂Te₃ superassemblies deposited at 350°C, 400°C, 450°C, and 600°C, respectively [41].

Figure 11(a) shows the crystal structure of Bi₂Te₃, which is usually described by a hexagonal cell that consists of 15 layers of atoms stacking along the c-axis with a sequence [5], namely ···Te⁽¹⁾–Bi–Te⁽²⁾–Bi–Te⁽¹⁾ ··· Te⁽¹⁾–Bi–Te⁽²⁾–Bi–Te⁽¹⁾···Te⁽¹⁾–Bi–Te⁽²⁾–Bi–Te⁽¹⁾ ···. The superscripts refer to two different types of bonding for Te atoms. The 5-atomic-layer thick lamellae of–(Te⁽¹⁾–Bi–Te⁽²⁾–Bi–Te⁽¹⁾)– is called quintuple layers, QLs. The Te⁽¹⁾…Te⁽¹⁾ refers Van der Waals force between Te atoms, whereas the Te⁽¹⁾–Bi and Bi–Te⁽²⁾ are ionic-covalent bonds. This weak binding between the Te⁽¹⁾…Te⁽¹⁾ accounts for the anisotropic thermal and electrical transport properties of Bi₂Te₃. For example, the thermal conductivity along the c-axis direction (∼0.7 Wm⁻¹K ⁻¹) is approximately a haft of the value along the plane perpendicular to the c-axis (∼1.5 Wm⁻¹K ⁻¹) [5, 6, 13]. The weak binding of Te⁽¹⁾…Te⁽¹⁾ also make the ease of cleavage along the plane perpendicular to the c-axis.

Figure 11.

(a) Crystal structures of Bi₂Te_3. (b) X-ray diffraction (XRD) patterns of of the Bi₂Te₃ super-assemblies deposited at various deposition temperatures from 350–600°C and at an Ar ambient pressure ∼10⁻³ Torr [41].

Figure 11(b) shows XRD patterns of the Bi₂Te₃ super-assemblies deposited at various substrate temperatures from 350–600°C. Clearly, all the films exhibited rhombohedral Bi₂Te₃ (JCPDS no. 89–4302) without traceable impurities or oxides. When substrate temperature increases, the (00 l) preferential orientation gradually becomes stronger, the 600°C- film is highly (00 l)-preferred orientation, which is consistent with the SEM observation (Figure 10d). The gradually enhanced (00 l) peaks from 350–450°C mainly originate from the increased thickness of the bottom layer (Figure 10a–d), which has similar lamellar morphology with (00 l)-preferential orientation of 600°C-film [41].

Figure 12 presents the proposed growth model of the super-assemling nanostructured Bi₂Te₃ films prepared at various T_S. The growth mechanisms are layer-then-fake for T_S = 350–450°C and layer-by-layer for higher T_S of 600°C. We can only find a monotone morphology and single preferential orientation of (00 l) for 600°C-film, which lead to a fully lamellar morphology with the (00 l) preferential orientation. Meanwhile, drastic changes in morphologies from layer to flake and orientations from (00 l) to (015) are observed at lower temperatures (350–450°C). The (00 l)-preferred orientation should be attributed to the thin bottom layer of the films prepared at 350–450°C. The thickness of this layer increases with increasing T_S. Since the bottom layer at 350°C is extremely thin, the required Ts for obtaining layer growth should be just below 350°C. The drastic change in the morphology and orientation at T_S of 350–450°C, namely, the layer- then-flake growth can be induced by a temperature gradient along the growth direction that the temperature at top surface of the as-deposited film should be slightly lower than at the substrate [41].

Figure 12.

Schematic illustration of the layer-then-flake and layer-by-layer growth models and the resulting Bi₂Te₃ in-plane (350–450°C) and out- of-plane (600°C) super-assemblies. Inset is the optical image of the prepared super-assembled films with a size of 1.5 × 1.5 cm² [41].

Table 1 summarizes the detailed properties of the super-assembling nanostructured Bi₂Te₃ thin-films. Due to such the voided structures, the films exhibited low electrical conductivity from 49 S.cm⁻¹ for worm-like superassembly (450°C) to 160 S.cm⁻¹ for 3D-layered super-assembly (600°C). Seebeck coefficient of the films was in range of 113–138 μV/K. As a result, the power factor (PF) is relatively low in range of 0.93 to 3.0 μW/cmK², primarily due to the low electrical conductivity of the films with voided morphologies.

Material	Morphology	Method	Deposition conditions		n (10^19 cm^−3)	μ (cm^2/Vs)	σ (S/cm)	α (μV/K)	PF (μW/cmK^2)	Ref.
Bi2Te3	Spindle-like super-assembly	PLD	350°C	0.13 Pa	4.0	12.4	79	−113	1.01	[41]
Bi2Te3	Worm-like super-assembly	PLD	400°C	0.13 Pa	1.9	25.9	73	−119	1.03	[41]
Bi2Te3	worm-like super-assembly	PLD	450°C	0.13 Pa	1.2	29.4	49	−138	0.93	[41]
Bi2Te3	3D-layered super-assembly	PLD	600°C	0.13 Pa	5.1	20.3	160	−137	3.0	[41]
Bi2Te3	Nanoparticles	PLD	300°C	20 Pa	9.7	14.8	230	−91	1.90	[51]
Bi2Te3	Nanoparticles	PLD	300°C	1.0 Pa	105	8.3	1390	−60	5.0	[52]
Bi2Te3	Layered Structure	Sputtering	350°C	1.0 Pa	95	12.1	1840	−70	8.8	[52]
	Nanorods		250°C	0.9 Pa	9.1	2.0	29	−81	0.19	[51]
Bi2Te3	Columnar Structure	Sputtering	350°C	1.0 Pa	246	7.5	2990	−46	6.4	[52]
Bi2Se3	Layered HPs	PLD	300°C	40 Pa	7.4	81.4	963.8	−75.8	5.5	[38]
Bi2Te3	Compact film	PLD	300°C	80 Pa	5	102	814.3	−172.8	24.3	[40]
Bi2Te3	Layered-smooth film	PLD	250°C	10 Pa	10.1	90.6	1464	−186	50.6	[43]
Bi3Se2Te	Nanocrystalline film	PLD	250°C	40 Pa	35.5	34.4	1747.5	−68.8	8.3	[53]
Bi2Se3	Bulk	Melting and hot-pressing			─	─	251.9	−175	7.7	[54]
Bi2Se0.3Te2.7	Bulk	Ball milling-hot pressing			─	─	892	−190	32.2	[55]
Bi2Se1.5Te1.5	Bulk	Zone melting			1.2	230	441.6	−193	16.5	[56]
Bi2Se1.8Te1.2	Nano-platelet bulk	Polyol method			─	─	199.6	−80.9	1.3	[57]
Bi2Se2Te	Bulk	Ball milling- hot pressing			─	─	1613	−60	5.8	[55]

Table 1.

Material, type, method, processing conditions, carrier concentration (n), mobility (μ), electrical conductivity (σ), Seebeck coefficient (α), power factor (PF = α²σ) of the Bi₂Te₃ films deposited by PLD and RF sputtering, as compared to properties of Bi₂Se₃, Bi₂Se₃, Bi₂Se_xTe_1-x bulk and films reported in the literature. All the selected values were recorded at room temperature.

In PLD, tightly controlling substrate temperatures (Ts) and ambient pressures (P) enables the morphologies and compositions of films to be manipulated extensively, which offers a new method for enhancing the TE properties of films [38, 43, 51, 58, 59]. For example, self-assembled Bi₂Te₃ films featuring well-aligned zero- to three-dimensional nanoblocks have been fabricated (Figure 13a–d), but the room-temperature PFs of these films remain low (≤ 1.9 μWcm⁻¹ K⁻²) [51]. By contrast, A. Li Bassi et al. [43] obtained several microstructured Bi₂Te₃ films (Figure 13e–h) with high PFs for morphologies: layered-smooth (50.6 μWcm⁻¹ K⁻², Figure 13e,e1), and compact-smooth (21.2 μWcm⁻¹ K⁻², Figure 13f,f1) at room-temperature; whereas the PFs remained low values of 8.8 μWcm⁻¹ K⁻² for 3D crystallite shapes (Figure 13g,g1) and 0.08 μWcm⁻¹ K⁻² for 3D-voided platelets (Figure 13h,h1).

Figure 13.

The morphology and power factor (unit μWcm⁻¹ K⁻²) of nano/micro-structured Bi₂Te₃ thin-films grown by PLD at various substrate temperatures and ambient pressures, reported by (i) Chang and Chen [51] and (ii) Li Bassi et al. [43].

Table 1 summaries the morphology and properties of Bi₂Te₃-based thin-films deposited by PLD, sputtering, in comparison with the properties of TE bulks. Usually, TE nanomaterials possess low σ values due to the separating or voided structure-morphology, but bulk and thin films have superior σ. For example, the compact-polycrystalline Bi₃Se₂Te achieved σ = 1747.5 S/cm [53] or even higher for some other films [52]. Unfortunately, the κ of films are missed in many published works to calculate ZT of the films. Thermal conductivity (κ) of nanocrystalline and nanostructured thermoelectrics is expected to achieve low values thanks to the extensive phonons scattering at interfaces, surfaces and grain boundaries. Indeed, reduced κ values have been noted for the monocrystalline Bi₂Se_0.3Te_2.7 films (κ = 0.8 W/m K for an average grain size of 60 nm) [60], and for Bi-Sb-Te films [61, 62].

2.3 Thermal conductivity κ of Bi₂Se₃ and Bi₂Te₃ and Bi-Te-Se compounds

A transient 3ω technique is usually employed in measuring thermal conductivity of thermoelectric films. The detail of this technique can be found in refs. [62, 63, 64]. Table 2 summarizes thermal transport properties (at room–temperature) of nanocrystalline–nanostructured Bi₂Te₃-based thin films and bulk materials in the literature. Generally, the thermal conductivity κ value for polycrystalline films is expected to be smaller than that of bulk alloys because of the extensive phonons scattering at interfaces, surfaces and grain boundaries [5, 60, 66]. Moreover, the κ of nanocrystalline Bi₂Te₃-based films will further decrease when the grain size of decreases (κ ≤ 0.81 W/mK, Figure 14A) [62, 65]. For Bi₂Te₃/Sb₂Te₃ superlattice films, the coherent backscattering of phonon waves at the superlattice interfaces is outlined for the reduction of lattice thermal conductivity, resulting in the low κ ≤ 0.4 W/mK [67, 69].

Sample, fabrication method	Avg. grain size	κ (W/m K)	σ (S/cm)	α (μV/K)	PF = σα^2 (μW/cmK^2)	ZT (300 K)	Ref.
Bi2Te2.7Se0.3 nanocrystalline thin film, flash evaporation	60 nm	0.8 (cross-plane)	540	−186.1 (in-plane)	18.7 (in-plane)	0.7	[60]
Sintered bulk Bi2Te3-xSex material, hot-pressing	30 μm	1.6	930	−177.5	29.3	0.6
Nanocrystalline bismuth-telluride-based (Bi2Te3-xSex) thin film	10 nm	0.61	550	−84.0	3.9	0.19	[65]
	27 nm	0.68	540	−138.1	10.3	0.46
	60 nm	0.80	540	−186.1	18.7	0.70
Nanocrystalline Bi-Sb-Te thin film, sputtering	26 nm	0.46	3.3	—	—	—	[62]
	45 nm	0.65	6.7	—	—	—
	84 nm	0.81	33.3	—	—	—
Nanocrystalline BiSbTe (8:30:62) thin film, flash evaporation	150 nm	0.6	—	—	—	—	[66]
Single crystal BiSbTe bulk alloys	—	0.75	—	—	—	—	[5]
Bi2Te3/Sb2Te3 superlattices (period∼5 nm)	—	0.4	—	—	—	—	[67]
Bi2Te3+0.63 bulk	—	2.2	1000	−240	58	0.87	[5]
Bi2(Te0.95Se0.05)3 bulk	—	1.59	901	−223	45	0.85	[5]
Bi2Te3/Bi2(Te0.88Se0.12)3 superlattice film, MBE	80 nm	1.25	639	−204	27	0.60	[68]
Bi2Te3 film, PLD	—	1.1	—	—	—	—	[69]
Bi2Te3/Sb2Te3 superlattices film (layered thickness ∼ 6 nm), PLD.	—	0.11	—	—	—	—
Bi2Te3 films, laser ablation	—	0.2–0.3	—	—	—	—	[70]
BixSb2-xTe3 nanolayer film, PLD	190 nm	1.16	2700	95	25	0.65	[71]
BixSb2-xTe3 nanodisc film, PLD	100 nm	1.00	1100	132	20	0.60
BixSb2-xTe3 nanocolumn film, PLD	70 nm	0.93	280	207	12	0.39

Table 2.

Room–temperature thermal transport properties of nanocrystalline–nanostructured Bi₂Te₃-based thin films and bulk materials in the literature, included: sample and fabrication method, average grain size, thermal conductivity κ, electrical conductivity σ, Seebeck coefficient α, power factor PF (= α²σ), and ZT (at 300 K).

Figure 14.

The morphology and thermal conductivity of Bi₂Te₃-based films with different grain sizes: (A) nanocrystalline Bi₂Te_3-xSe_x films [65], (B) the Bi_xSb_2-xTe₃ films [71].

For PLD Bi₂Te₃-based films, Yamasaki et al. [69] measured thermal conductivity with an ac calorimetric method in the direction across the film, obtaining k∼1.1 W/m K for a Bi₂Te₃ film deposited by PLD in vacuum (Table 2). In addition, Walachova et al. [70] used direct ZT measurement with the Harman method to estimate the κ value, and it is about 0.2–0.3 W/mK for the Bi₂Te₃ films. Recently, Chang et al. [71] reported the κ values of 0.93–1.16 W/mK for Bi_xSb_2-xTe₃ films with the granular-layered morphologies (Figure 14B).

3. Conclusions

In this book chapter, we present an overview of thermoelectric materials and applications, challenging of enhancing TE properties, and the nanostructuring approach in development TE materials. Various interesting nanostructured Bi₂Te₃- based thin films have been grown successfully by PLD with properly controlled substrate temperatures ambient gas pressures. For example, super-assembling of Bi₂Te₃ hierarchical nanostructures were grown at T_S from 350 to 600°C, and the films possessed relative high Seebeck coefficient of 113–138 μV/K, but exhibited low electrical conductivities of 49–160 S.cm⁻¹, and thus they had relatively low PF in range of 0.93 to 3.0 μW/cmK². Our intensive literature review on Bi₂Te₃-based TE materials can make general conclusion that TE nanomaterials possess low σ values when their structure-morphology are separating or voided, meanwhile, bulk and compact-smooth thin films can achieve high σ values. The PF values of Bi₂Te₃-based thermoelectrics varied in a wide range, i.e. below 5.0 μW/cmK² for voided structure-morphology, and reaching intermediate-high PF values of 5.0–50.6 μW/cmK² for compact-smooth or compact-layered structures. An advantage of nanocrystalline and nanostructuring thermoelectrics is the reduced thermal conductivity (possibly below 1 W.m⁻¹ K⁻¹). This book chapter provides fundamental understanding the relationship amongst processing condition in PLD growths, structure-morphology, and TE properties of Bi₂Te₃-based thin films.

Acknowledgments

This research is funded by Vietnam National Foundation for Science and Technology Development (NAFOSTED) under grant number 103.02-2019.374, and the Ministry of Science and Technology, Taiwan under Contract Nos. MOST 109-2221-E-214-016 and MOST 110-2221-E-214-013.