ChapterPDF Available

Reliability and Durability of Thermoelectric Materials and Devices

Authors:
Thermoelectric Micro/Nano Generators 2,
coordinated by Hiroyuki AKINAGA, Atsuko KOSUGA,
Takao MORI and Gustavo ARDILA.
© ISTE Ltd 2023.
1
Reliability and Durability of
Thermoelectric Materials and
Devices: Present Status and
Strategies for Improvement
Congcong XU1, *, Hongjing SHANG1, 2, *, Zhongxin LIANG1, *,
Fazhu DING2, and Zhifeng REN1
1 Department of Physics and Texas Center for Superconductivity
(TcSUH), University of Houston, USA
2 Key Laboratory of Applied Superconductivity, Institute of Electrical
Engineering, Chinese Academy of Sciences, Beijing, China
*These authors contributed equally to this chapter.
1.1. Introduction
The performance of a thermoelectric material is usually evaluated by its
operating temperature (T) and three other interrelated parameters: electrical
conductivity (σ), Seebeck coefficient (S) and thermal conductivity (k). These
parameters together determine the thermoelectric dimensionless figure of
merit (zT), 2.
S
zT T
σ
κ
= The thermoelectric conversion efficiency is
COPYRIGHTED MATERIAL
4 Thermoelectric Micro/Nano Generators 2
dependent on the zT. Theoretically, higher zT values can guarantee higher
thermoelectric conversion efficiency, so there has been a great effort to
improve zT over the past several decades, and the zT of some materials has
been reported exceeding 2. Nevertheless, the development of thermoelectric
devices has been somewhat lagging in comparison to the performance
optimization of materials.
Additional scientific and engineering challenges are faced during
thermoelectric device development, including long-term material stability,
material parameter matching, contact layer selection, device structure and
thermoelectric system design, etc. Among these challenges, stability is a
prerequisite for the actual application of materials. Since thermoelectric
devices are operated under large temperature differences, the environment
faced by thermoelectric materials, especially in thermoelectric power
generators, is relatively harsh. Repeated thermal cycles, thermal shocks and
high temperatures will not only reduce the thermomechanical properties of a
material but may also change its composition, which ultimately leads to
deterioration in the thermoelectric performance and even complete device
failure. In addition, the potential thermal stress caused by repeated thermal
expansion and contraction imposes more stringent requirements on the
stability of the device interface and the thermomechanical matching of the
p- and n-type materials. Unfortunately, in comparison to the effort to
improve zT, less attention has been paid to the stability issue for
thermoelectric materials and devices, limiting their practical application.
Here, we will highlight this issue and bridge the gap between
thermoelectric research and actual applications by discussing the durability
and reliability challenges of both thermoelectric materials and devices. We
first address the thermal stability of several thermoelectric materials and
strategies for thermal stability improvement. Secondly, we analyze in detail
thermoelectric device design, including thermal stress and interface issues.
Finally, we summarize recent progress in the design and fabrication of
reliable and durable thermoelectric modules and provide an outlook for
future research needs.
Reliability and Durability of Thermoelectric Materials and Devices 5
1.2. Thermoelectric material stability
For actual applications, a material’s stability is at least as critical as its
thermoelectric properties, if not more important. Different kinds of
instability, including component decomposition, elemental evaporation and
phase transition, can result in the degradation of thermoelectric material
performance, further leading to the decreased energy-conversion efficiency
of thermoelectric modules. In the following section, typical instability issues
of several materials with decent thermoelectric performances are discussed
in detail. Strategies that have been used to improve the thermal stability of
these materials are also summarized and discussed.
1.3. Mg3(Sb, Bi)2
Mg3(Sb, Bi)2 exhibits a CaAl2Si2-type crystal structure consisting of an
octahedrally coordinated Mg2+ cation layer and a tetrahedrally coordinated
anion structure [(Mg2Sb2)2-] (Shuai et al. 2017a; Song et al. 2019; Mao et al.
2018). It had long been considered a p-type semiconductor with
undistinguished performance, but Te-doped n-type Mg3(Sb, Bi)2-based alloys
were recently reported to show high thermoelectric performance,
demonstrating good potential for both power generation and cooling
applications (Tamaki et al. 2016; Zhang et al. 2017). Subsequently, various
strategies, including tuning the carrier scattering mechanism (Mao et al.
2017a; Shuai et al. 2017b), defect engineering (Mao et al. 2017b; Li et al.
2019, 2020a) and doping (Shi et al. 2019b, 2019c; Shu et al. 2019), have been
successfully applied in Mg3Sb2-based alloys, achieving a state-of-the-art
average zT value up to 1.1 in the range of 300~500 K (Imasato et al. 2019a;
Shi et al. 2019a; Shang et al. 2020a; Wood et al. 2019).
Among all defects in Mg3Sb2-based materials, Mg vacancies have the
lowest defect formation energy around the Fermi level, thus resulting in the
intrinsically high concentration of Mg vacancies causing p-type conduction
of these materials (Zhang et al. 2019). The addition of extra Mg into the
matrix has been reported to be successful in reducing the Mg-vacancy
concentration and achieving n-type conduction (Shuai et al. 2018; Kuo et al.
2019). At the operating temperature of 773 K, the corresponding vapor
pressures of elemental Mg, Sb and Bi are 7,390 Pa, 139 Pa and 9.64 Pa,
6 Thermoelectric Micro/Nano Generators 2
respectively. Such a great difference in vapor pressure will result in
significant Mg loss at high temperatures. Jørgensen et al. (2018) studied the
thermal stability of Mg3Sb1.475Bi0.475Te0.05 via high-temperature synchrotron
powder X-ray diffraction and found that a Bi impurity phase forms in
thermal cycles from 300 K to 725 K, suggesting the compound’s
high-temperature structural instability. Shang et al. (2020b) reported that
Mg3Sb2-xBix alloys (Mg3.2Sb1.5Bi0.49Te0.01, Mg3.2SbBi0.99Te0.01 and
Mg3.2Sb0.49Bi1.5Te0.01) showed performance degradation at 673 K and 773 K
via long-term in situ measurements of their thermoelectric properties,
although they demonstrated the stable electrical properties of these
compounds at 573 K, as shown in Figure 1.1. It can be seen that the Seebeck
coefficient of Mg3.2Sb0.49Bi1.5Te0.01 crosses over from negative values
(n-type) to positive values (p-type) at 773 K, which has been attributed to the
Mg loss notable by microstructural analysis and results in the formation of
Mg vacancies that pin the Fermi level, eventually enabling the p-type
conduction. Fortunately, coating the surfaces of the Mg3Sb2-xBix alloys
with boron nitride was found to significantly improve their thermal stability
by effectively limiting Mg loss at high temperatures. The
time-dependent Hall carrier concentration of Mg3Sb2-xBix with cationic
doping by La (Imasato et al. 2018), Yb (Wood et al. 2020) and Y (Shi et al.
2019c) at high temperatures and in a dynamic vacuum was also explored,
and the cation-doped samples showed a delayed decrease in the carrier
concentration compared with their Te-doped counterparts, indicating that
doping with a cation rather than with Te can also improve the thermal
stability of Mg3Sb2-xBix to some extent. Considering the large differences in
the vapor pressures of the elements in Mg3Sb2-xBix, high-temperature Mg
loss that results in structural instability and decreased thermoelectric
performance is common, thus restricting the operating temperature ranges of
these materials. This disadvantage overshadows the high-temperature zT
values and makes practical applications less appealing at temperatures higher
than 673 K. Therefore, determining how to improve the thermal stability of
Mg3Sb2-xBix compounds is more significant than simply focusing on
enhancing their zT values.
Reliability and Durability of Thermoelectric Materials and Devices 7
Figure 1.1. Long-term in situ measurement of thermoelectric properties of
Mg3Sb2-xBix alloys at 573 K, 673 K and 773 K. (a)–(c) σ/σ0, (d)–(f) S/S0 and (g)–(i)
PF/PF0 for Mg3.2Sb1.5Bi0.49Te0.01, Mg3.2SbBi0.99Te0.01 and Mg3.2Sb0.49Bi1.5Te0.01,
respectively. σ0, S0 and PF0 represent the initial values; σ, S and PF represent the
measured values at the specified time. Reprinted from Shang et al. (2020b). For a
color version of this figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
1.4. Zn4Sb3
β-Zn4Sb3 is a well-known p-type thermoelectric material for applications
in the intermediate temperature range (473~673 K) and exhibits high zT
values up to 1.3 at 673 K (Zou et al. 2015). The high thermoelectric
performance of β-Zn4Sb3 and its good potential for thermoelectric applications
mainly result from its low lattice thermal conductivity (~0.3 W m-1 K-1 above
500 K), which is largely caused by its disordered structure that includes
interstitial Zn (Dasgupta et al. 2013).
However, with the increasing temperature, β-Zn4Sb3 starts to decompose
at 473 K due to a loss of Zn, forming ZnSb and Zn as decomposition
8 Thermoelectric Micro/Nano Generators 2
productions and degrading its thermoelectric performance (Hao et al. 2011;
Yang et al. 2017a, 2017b). Thus, it is of great significance to improve the
thermal stability of β-Zn4Sb3 for practical applications. Based on both
experimental results and theoretical calculations, doping with heavier metals
can hamper the Zn migration and subsequently freeze β-Zn4Sb3 in its local
disordered structure, eventually improving its thermal stability. It has also
been found that this doping strategy can optimize the compound’s electrical
properties. For example, Wang et al. reported that Pb-doped β-Zn4Sb3 shows
increased carrier mobility, achieving an optimized zT value of ~1.2 at 660 K
(Wang et al. 2012), while its thermal stability is clearly improved due to
the decreased Zn evaporation at high temperatures. Similar results were
also reported by Deng et al. (2017). Furthermore, the addition of any of
Ag (Song et al. 2018), Cd (Yin et al. 2010), Ge (Moghaddam et al. 2019),
In/Sn (Shai et al. 2015), ZnO (Tang et al. 2015), etc. into β-Zn4Sb3
can also simultaneously improve its thermoelectric performance and
high-temperature stability.
1.5. Skutterudites
In recent years, skutterudite-based materials have attracted significant
attention due to their outstanding performance over a wide temperature
range and their relatively inexpensive constituent elements. Among
the various kinds of skutterudites, CoSb3, a diamagnetic (Co3+, 3d6)
narrow-bandgap semiconductor, demonstrates high carrier mobility and a
relatively large effective electron mass. Generally, CoSb3-based
thermoelectric generators can operate at temperatures up to 773 K. However,
a series of studies have revealed that CoSb3-based skutterudites undergo
oxidation and Sb sublimation when the operating temperature exceeds
673 K, resulting in short service life (Sales et al. 1996; Zhang et al. 2016;
Liu et al. 2020). Therefore, effective and reliable approaches are needed to
address these instability issues.
The most effective method reported thus far is physical isolation, or
inhibiting the sublimation of elements by spraying on a protective coating
such as metal (El-Genk et al. 2006), ceramic (Zhang et al. 2016; Chen et al.
2013), glass (Dong et al. 2012), aerogel (Sakamoto et al. 2008) or a
multilayered composite (Xia et al. 2014; Zhang et al. 2016). Introduction of
the coating under certain conditions can indeed improve the stability of
CoSb3-based skutterudites at high temperatures, but it may result in some
new problems at the same time. For example, the diffusion of a metal
Reliability and Durability of Thermoelectric Materials and Devices 9
coating into the thermoelectric legs may result in decreasing performance
over increasing thermal cycles (Dong et al. 2012). Moreover, the metal
coating would cause an electrical short in actual applications. Although
aerogel coating is highly robust and resistant to high temperatures, it could
include micropores that may serve as escape channels for elemental oxidation
and sublimation. Coating with another material such as glass or a ceramic
may result in a mismatch of coefficient of thermal expansion (CTE) values or
poor adhesion strength. Thus, coating with a multilayered structure consisting
of different kinds of material is proposed. Zhang et al. (2016) reported a more
effective multilayered coating with Mo, SiOx and glass-ceramic layers. As
shown in Figure 1.2, compared with samples without coating and with
single-layered coating, that with the multilayered coating exhibits better
stability in air at 873 K, indicating the effective suppression of both
oxidation and sublimation. Nevertheless, suppressing decomposition in and
improving the stability of skutterudite-based materials remain very
challenging in practical application scenarios.
Figure 1.2. Mass change of coated and uncoated skutterudites at 873 K in air
as a function of time. Reprinted from Zhang et al. (2016). For a color version
of this figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
1.6. Cu2-xX (X = S, Se, Te)
Copper chalcogenides Cu2-xX (X = S, Se, Te) possess complex atomic
arrangements, and the Cu-deficient region in the Cu–Se system (Cu2-xSe)
10 Thermoelectric Micro/Nano Generators 2
exhibits two distinct phases, a low-temperature α phase and an
intermediate-temperature β phase (Liu et al. 2012). The low-temperature α
phase with a low-symmetry crystal structure, where the Cu atoms are
ordered and not superionic, is stable up to ~400 K and eventually transitions
to the intermediate-temperature β phase with a cubic structure at ~400 K. In
the β phase, the superionic Cu ions are kinetically disordered, leading to its
remarkably low thermal conductivity and high thermoelectric performance.
Significant advances have recently been obtained in studies of the p-type
Cu2-xSe materials, and the zT value reportedly reaches 2.4 at 1,000 K for
Cu2-xSe/carbon nanotube (CNT) hybrid materials, showing promise for
applications (Nunn et al. 2017, Lu et al. 2020b, Zhang et al. 2020b).
Figure 1.3. TG-DTA curves for single-crystal Cu2Se with a heating
rate of 10 K/min. Reprinted from Liu et al. (2019)
Unfortunately, Cu2-xSe materials suffer from thermal instability issues
when they are applied at a high temperature and exposed to a large current,
limiting their application in large-scale thermoelectric power generation (Shi
et al. 2020). Thus, it is urgent to find out the mechanism of the instability
issues of Cu2-xSe materials and improve their performance liability. Bohra
et al. (2016) studied the effects of thermal cycles on the thermoelectric
properties of Cu2-xSe in the temperature range of 320~1,070 K and found
that electrical resistance increased with increasing thermal cycles, resulting
in a reduction in the power factor from 13 μW cm-1 K-2 for the first cycle to
Reliability and Durability of Thermoelectric Materials and Devices 11
8 μW cm-1 K-2 for the third cycle at ~1,070 K. This change was attributed to
Se loss and phase transition with increasing thermal cycles. Xue et al. (2019)
reported that a high synthesis temperature not only enhances the
thermoelectric performance of Cu2-xSe but also effectively improves its
thermal stability; specifically, the Seebeck coefficient, electrical conductivity
and thermal conductivity were found to be nearly constant between the
heating and cooling cycles over the temperature range of 300~723 K for a
sample fabricated at 1,273 K. Furthermore, a Ni/Mo/Cu1.96Se0.8S0.2
thermoelectric unileg was able to exhibit a very steady power output over
400 thermal cycles between 473 K and 873 K through the control of its
chemical composition (Mao et al. 2019b). Other dopants, such as Li (Kang
et al. 2017), CNTs (Nunn et al. 2017), W (Bohra et al. 2020) and Cd
(Lu et al. 2020a), can also improve the thermal stability of Cu2-xSe
materials, allowing them to be applied for power generation. Additionally,
single-crystal α-Cu2-xSe was fabricated using the Bi-flux method and no
weight loss occurred below the melting point (~1,403 K), as shown by the
thermogravimetric-differential thermal analysis (TG-DTA) results displayed
in Figure 1.3. The DTA curve shows the endothermic peak at ~400 K, which
has been attributed to the Cu2Se αβ phase transformation. The power factor
of the single-crystal samples did not obviously decrease after three thermal
cycles in the temperature range of 300~690 K, demonstrating their improved
thermal stability compared with the polycrystalline materials (Liu et al.
2019). However, single crystals are usually more expensive than the
polycrystalline materials, so they are not preferred for large-scale
applications.
1.7. GeTe
GeTe-based thermoelectric materials were first studied in the 1960s, and
they re-emerged as a subject of research interest only recently. Thanks to
band engineering and defect engineering strategies, the zT values exceeding
2 have been achieved in GeTe-based materials, showing their promise for
applications (Hong et al. 2020; Zhang et al. 2020a). Nevertheless, the actual
application of GeTe-based modules have a major obstacle to overcome
concerning how the pristine GeTe undergoes a reversible transition between
a low-temperature rhombohedral structure and a high-temperature cubic
structure around 700 K, as illustrated in Figure 1.4. It should be noted that
such structural instability can seriously reduce the reliability of devices if it
is not handled properly. Additionally, there is a considerable change in
the coefficient of thermal expansion (CTE) between the two phases
12 Thermoelectric Micro/Nano Generators 2
(11.2×10
-6
K
-1
for the rhombohedral structure and 23.4×10
-6
K
-1
for the cubic
structure), and such CTE change is extremely detrimental in device
applications since it can cause significant thermal stress during operation
(Xing et al. 2021). Thus, methods to suppress the phase transition are
required.
Fortunately, Liu et al. (2018) successfully obtained GeTe-based material
that is in the cubic phase over the entire operating temperature range through
simple Mn and Bi co-doping at the Ge site. Soon afterwards, Zheng et al.
(2018) reported that the lower-temperature cubic GeTe can also be achieved
by Mn and Sb co-doping and found that the transition temperature was
reduced from 700 K closer to room temperature. The above two studies
show that Mn plays a significant role in inhibiting the phase transition.
Bi and Sb can also adjust the crystal structure to a certain extent, but they are
more responsible for regulating the carrier concentration. Recently, resonant
bonding properties induced by a symmetrized crystal lattice were found in
the Ti-doped GeTe-based material (Li et al. 2020c). Ti doping not only can
reduce the lattice constant c/a ratio to achieve a cubic structure but can also
enhance the Seebeck coefficient due to the increased band degeneracy
(Shuai et al. 2019).
Figure 1.4. Structure of GeTe: (a) cubic and (b) rhombohedral.
Reprinted from Li et al. (2021). For a color version of this
figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
1.8. Outlook on thermoelectric materials stability
Admittedly, the instability issue is universal in thermoelectric materials
because of their weak chemical bonding and highly disordered content. In
this section, different typical instability issues faced by several kinds of
Reliability and Durability of Thermoelectric Materials and Devices 13
materials were briefly introduced, and the corresponding strategies to
address these issues were also discussed in detail. In addition to the
above-discussed thermoelectric materials, other promising materials with
excellent thermoelectric performance, including PbTe (Wang et al. 2018),
BiCuSeO (Li et al. 2014) and half-Heuslers (Rausch et al. 2016), also suffer
from thermal instability, which markedly impedes their commercial
applications, especially for the generation of thermoelectric power.
Generally, the instability issue can be solved either chemically or physically.
From a chemical point of view, doping or heat treatment can improve the
stability of a thermoelectric material’s constituent elements and structure on
a microscopic level. For some special materials with phase transitions,
reasonable chemical doping can alter the phase-transition point, thereby
avoiding the instability of the material in the operating temperature range.
From a physical point of view, isolating a material from the external
environment, such as with a coating or through vacuum packaging, is also an
effective measure to effectively enhance its stability. These strategies should
be thoroughly evaluated and rationalized in future studies with the help of
advanced characterization tools and theoretical simulations. Without these
methods, it would be difficult to understand the origin of phase-transition
regulation or structure-stability preservation when introducing external
dopants. In addition, for long-term stability, developing a rapid and scalable
method for preparing thermoelectric materials with fewer defects is
advantageous. In short, to achieve the practical uses of thermoelectric
materials, greater efforts should be made with enhancing the thermal
stability of existing materials and searching for new materials with both high
performance and good stability. In particular, reporting the stability data of
high-zT materials together with their transport property measurements is
recommended for at least the papers on device performance.
1.9. Thermoelectric device design analysis
1.9.1. Thermal stress analysis
Despite the tremendous improvements in thermoelectric material
performance over the last several decades, research into building thermally
and mechanically reliable devices has been very slow. Since thermoelectric
modules are subjected to large temperature gradients over long periods, the
thermomechanical properties and thermal stability of various constituent
components must be considered in device design and manufacturing (Karri
et al. 2018). Significant thermal stress, deformation and even failure may
14 Thermoelectric Micro/Nano Generators 2
occur in thermoelectric devices due to mismatches in the thermal,
mechanical and rheological properties of the constituent components caused
by material specificity, cyclic thermal loading and possible mechanical
vibration (Music et al. 2016). As a result, in designing a thermoelectric
device, it is critical to conduct temperature distribution and thermal stress
analyses in order to effectively reduce thermal stress through material
selection or structure optimization. The main aim of such a
thermomechanical stress study is to maintain the thermoelectric device’s
high output performance under a long-term operation.
In this section, thermomechanical stress in thermoelectric modules is
discussed in depth. We first discuss the coefficient of thermal expansion,
which is a significant factor in determining thermal stress, as well as other
design considerations relevant to thermal stress, including module
configuration and geometry. We then present an analysis of the lowest value
of thermal conductivity based on the maximum allowable thermal shear
stress, emphasizing the importance of balancing the high zT generated by
phonon scattering with mechanical reliability. These analyses elucidate the
relationship between the above factors and the performance, reliability and
durability of thermoelectric devices, shedding light on the optimal design of
reliable thermoelectric devices.
1.9.1.1. Coefficient of thermal expansion (CTE)
The coefficient of thermal expansion (CTE,
α
) of a material is its length
change in response to temperature variation. Differing expansion of
constituent parts of a thermoelectric device under a considerable temperature
gradient would result in high thermal stress and hence a greater risk of
interfacial damage during long-term operation (He et al. 2018). For example,
cracking or exfoliation at the interface of an oxide-based module is due to
the widely differing CTEs of oxides and metallic contact materials (Liu et al.
2015).
The temperature-gradient-induced mechanical stress ()
ξ
T of a
thermoelectric material can be approximated by:
()()
() (1 ( ))
αΔ
ξν
=
TET T
TT [1.1]
where
α
, E and
ν
are its CTE, Young’s modulus and Poisson’s ratio,
respectively (Ni et al. 2013).
Reliability and Durability of Thermoelectric Materials and Devices 15
Figure 1.5. (a) Coefficient of thermal expansion (CTE, α) and elastic modulus-CTE
product (Eα) for representative thermoelectric materials. Data obtained from Ni et al.
(2013), Zhao and Tan (2014) and Music et al. (2016). (b) CTE and electrical
conductivity (r) for pure metals. Data obtained from Chowdhury et al. (2009). For a
color version of this figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
Clearly, a large CTE value and a high Young’s modulus are undesirable
for reduced thermal stress. According to calculation results, a CTE mismatch
between different components in a thermoelectric module can result in
excessive thermal stress and in situ fracture (Ravi et al. 2009). In other
words, CTE has a significant impact on the module’s thermomechanical
robustness. Hence, obtaining thermal expansion statistics for commonly used
thermoelectric materials and investigating the underlying relationship
between the structure and the thermal expansion coefficient of each
are critical for achieving thermally match systems in thermoelectric
devices. The measured CTE results for commonly investigated
thermoelectric materials with high zT values are summarized in Figure 1.5a.
Clearly, most high-performance thermoelectrics have CTE values of more
than 15 × 10-6 K-1, which can be attributed to the weak chemical bonding in
these materials that also results in their low lattice thermal conductivity. The
large CTE values emphasize the importance of selecting appropriate metal
electrodes and designing reliable contact layers for the thermoelectric legs in
a module in order to reduce thermal stress (Ni et al. 2013). Additionally, the
CTE values of thermoelectric compounds should be optimized to establish
better matches between p- and n-type materials or between thermoelectric
materials and electrodes by tuning the compositions, filler atoms, grain sizes
(Rogl et al. 2017) and additives (Rogl et al. 2010). However, these solutions
16 Thermoelectric Micro/Nano Generators 2
need to be carefully examined to ensure that high zT values are retained. The
CTE and electrical conductivity values of pure metals are shown in
Figure 1.5b as a reference to choose suitable contact layers for both CTE
matching and low electrical resistivity. Music et al. (2016) used the first-
principles calculation method to investigate the thermomechanical response of
20 thermoelectrics. They found that oxides exhibit larger products of linear
thermal expansion coefficient and elastic modulus, implying that
thermoelectric oxides are susceptible to thermally induced stress, shock and
fatigue. For convenience in prediction and matching, it was also suggested that
equilibrium volume, an easily measurable parameter, can be used to estimate
this product and the thermal response of thermoelectric materials.
1.9.1.2. Module configuration and geometry design
Realizing high conversion efficiency or output power density in a
thermoelectric module relies on a specific configuration and geometry, so its
fabrication always involves the design and optimization of these factors (Wu
et al. 2014). The principle of thermoelectric device configuration design is to
minimize the internal stress caused by temperature difference and thermal
shock while making full use of the heat source.
The π-shaped, O-shaped and Y-shaped structures, schematically
illustrated in Figure 1.6, are the most common thermoelectric-module
configurations (Zhang et al. 2016). In a typical π-shaped module as shown in
Figure 1.6a, many thermoelectric unicouples consisting of an n-type leg and
a p-type leg are connected electrically in series and thermally in parallel. In
order to maximize the use of temperature differences, the direction of heat
flow is usually perpendicular to the substrate. It should be noted that when
this design is applied at higher temperatures, the thermally induced stress,
expansion and diffusion of different materials should be thoroughly
considered because the repeated heat shock may cause severe deformation
and even device failure, especially at the contacts. An O-shaped module
(Figure 1.6b) consists of a coaxial assembly of many annular flat gaskets
consisting of alternating n- and p-type thermoelectric rings. Compared to the
π-shaped configuration, the O-shaped configuration makes it easier to
achieve device integration, efficient heat transfer and low-stress structure
design, but the manufacturing and joining technology needed for the
specially shaped thermoelectric elements and the metal electrodes is very
difficult to master if not impossible. The Y-shaped module configuration
(Figure 1.6c) has been designed mainly to eliminate the impact of thermal
expansion mismatch between different constituent components while
Reliability and Durability of Thermoelectric Materials and Devices 17
maintaining excellent thermoelectric performance. In this configuration,
rectangular n- and p-type thermoelectric materials are alternately connected
to electrode plates in a sandwich structure, which allows the thermoelectric
elements to be maintained in a non-fully constrained state and reduces the
thermal expansion and deformation caused by the constraints of the hot and
cold sides. The horizontal series connection of the n- and p-type materials
also avoids the stress concentration caused by the difference in their thermal
expansion coefficients.
Figure 1.6. Schematic diagrams of (a) π-shaped, (b) O-shaped and (c) Y-shaped
modules. Reprinted from Zhang et al. (2016). For a color version
of this figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
In addition to the shape of the module, the unicouple geometry plays an
important role in maximizing performance and maintaining it under large
thermal stress (Shittu et al. 2020). Recent thermodynamic and thermal stress
studies of thermoelectric generators have revealed that the highest stress
level is found at the edges of a module, particularly those on its hot side
(Al-Merbati et al. 2013). By tuning the pin geometry, the temperature
variation can be improved, and thus the lifetime of the module can be
18 Thermoelectric Micro/Nano Generators 2
prolonged. Fabián-Mijangos et al. (2017) compared the effects of various leg
shapes, including square, hexagon, octagon and circular prisms, on overall
thermal stress and found that cylindrical legs are preferable under the fixed
hot side condition, while square prisms are more reliable under the free hot
side condition. Although modifying the leg dimensions can improve a
module’s thermal reliability, a unilateral focus on thermodynamic properties
may result in low output-power performance. Erturun et al. (2015) found
that there is an inverse relationship between output performance and
thermal integrity by using three-dimensional finite-element analysis and
experiments, indicating the importance of determining the optimal shape for
balancing performance and stability. More specifically, Sarhadi et al. (2015)
recommended that the thermoelectric legs should be long enough to achieve
high efficiency and low thermal stress while short enough to maintain a high
output power density. A segmented module presents a more complicated
situation since more factors are involved in its structural optimization. In
most cases, a compatibility factor match between segmented materials is
sufficient to produce an effective segmented device. However, Jia et al.
(2014) noted that when considering the strength requirements, a high
conversion efficiency in a segmented device is not always possible. In short,
multi-parameter optimization of thermoelectric generators is required to
achieve both maximum efficiency and structural reliability.
Modulating the hot-side design of a thermoelectric device by taking into
consideration the connection structure and heat flux direction is another
strategy to reduce thermal stress and obtain a durable interface. Sun et al.
(2021) examined a new design that uses a step structure of ceramic plates
with thermoelectric legs inserted into them, which separates the heat and
current conduction paths at the hot-side electrode (Figure 1.7). As a result,
interconnect cracking is only affected by the connection between the
hot-side ceramic plate and the metallic interconnect adjacent to it, which
reduces the thermal stress damage of the interconnect and increases the
number of operation cycles prior to failure. Additionally, the design was
calculated to have a favorable impact on the module’s power output
performance, showing simultaneous improvement in thermomechanical
stability and thermoelectric performance. In the future, a real module should
be prepared and evaluated to substantiate this new design and initiate further
improvements, and also to show the cost difference since the new design will
be more expensive.
Reliability and Durability of Thermoelectric Materials and Devices 19
Figure 1.7. Schematic diagrams of thermoelectric modules featuring (a) the
conventional thermoelectric unicouple geometry and (b–e) the new hot-side design
geometry with different parameters. Reprinted from Sun et al. (2021). For a color
version of this figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
When considering the substrate design specifically, there are two main
strategies for protecting the module from thermal stress. Increasing the
thickness of the ceramic plates can relieve thermal stress while ensuring that
the device’s output performance remains nearly constant. Another effective
way to reduce thermal stress is to replace rigid substrates with compliant
pads for electrical insulation (Kambe et al. 2010). Nevertheless, in this case,
the derived module will be mechanically weak and will need additional
encapsulation. Most thermoelectric modules currently investigated use
alumina substrates because of their high thermal conductivity. A Cu
electrode strip can be deposited onto each substrate prior to assembly. The
thickness of the Cu electrode strips can be adjusted according to the
maximum expected output current of the module.
1.9.1.3. Materials design for reduced thermal stress: engineering
thermal conductivity κeng
Decreasing lattice thermal conductivity by introducing multi-scale
microstructures has been widely applied to improve the performance of
thermoelectric materials. In particular, nanostructured grains are proposed to
20 Thermoelectric Micro/Nano Generators 2
scatter phonons significantly without degrading electrical transport, which is
beneficial for the overall improvement of zT values. Intuitively, the
thermoelectric performance of a material would be significantly improved if
its lattice thermal conductivity could be reduced to the minimal value
predicted by theoretical calculations. However, because of the corresponding
weak chemical bonding and significant anharmonicity, materials with
exceptionally low lattice thermal conductivity are usually brittle, limiting
their practical applications. In other words, even though carrier mobility is
unaffected, the low lattice thermal conductivity attained by creating vast
defects may degrade mechanical strength and stability since thermoelements
will be subjected to large thermal gradients and thermal stress. Thus,
reconsidering the phonon scattering technique is critical to balancing zT
values and mechanical brittleness. Inspiringly, Kim et al. (2016)
hypothesized that when exposing materials to shear thermal stress, there is a
minimal value of thermal conductivity which is safe to employ (Figure 1.8).
To begin, they proposed that there is an appropriate range of leg length (L),
where the lower limit of leg length (Lmin) is dictated by the yield strength and
the upper limit of leg length (Lmax) may be computed using thermoelectric
material parameters and boundary conditions. For materials with higher
thermal conductivity, a larger value of Lmax could be expected for a fixed
cold-side temperature. If Lmax is greater than Lmin, it is safe to reduce the
thermal conductivity so that Lmax is closer to Lmin, therefore enhancing the
efficiency of the device. Engineering thermal conductivity (κeng) is the
thermal conductivity value when Lmin and Lmax are the same. Thus, κeng can
be defined as the lower limit of a material’s thermal conductivity that still
assures the device’s thermal reliability in a steady-state operation. On the
other hand, since thermoelectric devices are subjected to thermal loading and
thermal shock in real applications, it is also important to investigate the
minimum thermal conductivity of thermoelectric materials during transient
operation, which can be estimated from the dynamic boundary conditions
and the thermomechanical properties of the materials. Based on these
findings, Kim et al. (2016) investigated a Bi2Te3-based module as a case
study and found that further reduction in thermal conductivity for zT
improvement should not be attempted because the thermal conductivity of
current-generation p-type Bi2Te3 already approaches the steady-state κeng. In
contrast, more reliable performance improvement may rely on boosting its
electrical properties.
Reliability and Durability of Thermoelectric Materials and Devices 21
Figure 1.8. Flow chart for calculating engineering thermal conductivity (κ
eng
)
in the steady state. Reproduced from Kim et al. (2017)
1.9.2. Interface analysis, design and fabrication
The interfaces between thermoelectric materials and metal electrodes in a
thermoelectric device undergo complex physical and chemical processes
during operation due to repeated thermal cycling and thermal shock. In
addition to the thermal stress mentioned above, more engineering problems,
such as bonding strength, contact resistance and high-temperature diffusion,
are confronted when considering interfaces. These issues have a great impact
on the reliability and durability of the device. In the following section, the
relationship between the interface properties and the device performance, as
well as the failure behavior of the interfaces are first analyzed in detail,
demonstrating the great importance of the interfaces. Subsequently, general
selection and design principles for high-performance and reliable interfaces
are introduced.
1.9.2.1. Interface-induced performance degradation and device failure
Contact thermal and electrical resistance at the interfaces between
different components of the module will result in lower power output
performance because of the wasted electricity and heat. Such a performance
22 Thermoelectric Micro/Nano Generators 2
discrepancy can occur at either a pristine imperfect joining or an interface
that is increasingly degraded over long-term thermal cycles. In order to
evaluate the effect of the interface contacts on device performance, Kim
(2018) developed a numerical method for the performance analysis of
thermoelectric generators with thermal and electrical contact resistance, in
which the effectiveness was defined as the ratio of the thermoelectric
generator’s power output with any resistance to that without resistance.
Effectiveness thus indicates how much of the potential performance is
actually achieved in the presence of any resistance, and the corresponding
mathematical results and numerical modeling for this parameter are shown
in Figure 1.9. It is clear that the effectiveness increases with increasing
thermal contact conductance (i.e. decreasing thermal contact resistance) and
decreases with increasing electrical contact resistivity. According to the
numerical simulation, low contact thermal conductance can cause
performance degradation of up to ~30%, while large electrical contact
resistivity can cause performance degradation of more than 40%. The results
of this study emphasize the importance of interfaces on the actual
performance of thermoelectric generators.
Figure 1.9. Effect of (a) thermal contact conductance and (b) electrical contact
resistivity on thermoelectric device performance. Data obtained from Kim (2018).
For a color version of this figure, see www.iste.co.uk/akinaga/thermoelectric2.
zip
For thermoelectric cooling devices, the theoretical value of the coefficient
of performance (COP) without considering thermal or electrical contact can
be obtained by the formula:
Reliability and Durability of Thermoelectric Materials and Devices 23
h
c
c
hc
OP 1
C1
1
=
+−
++
T
T
T
TT
ZT
ZT
. [1.2]
This formula is based on the theory developed by Ioffe in 1957 and is not
highly accurate in evaluating thermoelectric modules (Ioffe 1957). In 1996,
Rowe and Min proposed a more complete method, in which the thermal and
electrical contacts are considered (Rowe 1995). The output voltage V and
current I can be expressed by:
hc
c
c
()
2
1
κ
κ
=
+
N
ST T
Vl
l
, [1.3]
and
hc
cc
c
()
22
2( )(1 )
ρ
ρρ
κ
κ
=
++
AS T T
Il
ll
[1.4]
where N, S, A, l, lc,
κ
, c
κ
, ρ and ρc are the number of thermocouples in the
module, the Seebeck coefficient, the cross-sectional area of the
thermoelectric material, the length of the thermoelectric leg, the thickness of
contact layer, the thermal conductivity of the thermoelectric leg and of the
contact layer and the electrical resistivity of the thermoelectric leg and of the
contact layer, respectively. ratio c
κκκ
= and ratio c
2
ρρρ
= are standard
thermal and electrical contact parameters, respectively. Generally,
ratio ~0.2
κ
and ratio ~0.1
ρ
mm are appropriate values for commercially
available Peltier modules. The modified conversion efficiency COPrev can be
derived as:
hch
rev
2
c
ratio h ratio ratio chch
()/ ,
(
COP
1
( 2 ) {(2 [ ] (4 / )[(1 ) (1 2 )]}
2)/
l
lZTl
l
TTT
TTT
κκρ
=
+− +++
[1.5]
24 Thermoelectric Micro/Nano Generators 2
where 2
ρλ
=ZS . COPrev could be further derived for operation in the
Peltier mode:
1/2
h
c
c1/2
h
ratio ratio c
rev ratio c
ratio
c
1
1
COP 21
l
ll
ll
l
ZTl T
T
T
TT l
ZTl
ρ
ρ
κ
κ

+−




++



+

=
+
+
, [1.6]
where T is the mean temperature across the thermoelectric module. Based
on this formula, the relationship between COPrev and the length of the
thermoelectric material at different ratio
κ
and ratio
ρ
values is shown in
Figure 1.10 for c0.5=l mm, Z = 2×10-3 K-1 and 25
Δ
=T K with Th at 300 K.
It is clear that COPrev varies under different ratio
κ
and ratio
ρ
values. The
COPrev of a 1-mm leg increases by ~18.6% when ratio
κ
decreases from 0.1 to
0.001 (i.e. the thermal conductivity of the contact layer increases), while it
decreases by ~12.6% when ratio
ρ
increases from 0.1 mm to 0.2 mm (i.e. the
electrical resistivity of the contact layer increases). A dramatic decrease can
be observed in COPrev when the length of the thermoelectric leg is less than
2 mm. The thermal and electrical resistance of the contact layer become
increasingly more important once the length of the thermoelectric leg is
relatively short. Therefore, in thermoelectric cooling devices, for which the
service environment is not as harsh as that for the thermoelectric generators,
a relatively slight increase in contact resistance still leads to notable
degradation in the cooling performance.
In actual applications, device failure frequently occurs at the interfaces
which undergo repeated thermal shocks, which place greater demands on the
interfaces. Figure 1.11 shows the results of long-term thermal stability
testing of a MnSi1.77/Mg2Si1-xSnx-based thermoelectric module (Mejri et al.
2020) and a Bi2Te3-based thermoelectric leg (Barako et al. 2012). For the
MnSi1.77/Mg2Si1-xSnx-based thermoelectric generator, the thermoelectric legs
were soldered to metalized ceramic plates using silver, as shown in
Figure 1.11a. The barrier layer between the legs and the solder consists of a
Reliability and Durability of Thermoelectric Materials and Devices 25
thin layer of Au/Ti (300/100 nm). During testing, thermal cycling between
423 K and 673 K was applied at the hot side of the module while its cold
side was maintained at 323 K. After 400 cycles (~100 hours) of testing,
serious cracks occurred at the hot-side material interface, as shown in
Figure 1.11b. In another example, Bi2Te3-based thermoelectric legs were
tested by quickly switching the temperature between 253 K and 419 K at one
side for ~45,000 cycles, while the other side was maintained at the ambient
temperature, and the results are shown in Figures 1.11c and 11d. It can be
clearly seen that cracks developed at the interface between the
thermoelectric leg and the electrode, which would undoubtedly cause
performance degradation. In summary, whether due to thermal shock or
large interface resistance, imperfections in the contact layer can have a great
impact on device performance, suggesting the importance of interface
research for reliable thermoelectric modules.
Figure 1.10. COPrev as a function of thermoelectric-leg length (l).
Data obtained from Rowe (1995). For a color version of this
figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
26 Thermoelectric Micro/Nano Generators 2
Figure 1.11. MnSi
1.77
/Mg
2
Si
1-x
Sn
x
-based thermoelectric generator (a) before and (b)
after thermal cycling between 423 K and 673 K for 100 h. Reprinted from Mejri et al.
(2020). Comparison of the interface between a Cu electrode and a Bi
2
Te
3
leg (c)
before and (d) after thermal cycling between 253 K and 419 K for ~45,000 cycles.
Reprinted from Barako et al. (2012). For a color version of this figure, see
www.iste.co.uk/akinaga/thermoelectric2.zip
1.9.2.2. Interface design principles
A thermoelectric device can be reduced to its essential component, a
complete circuit consisting of p- and n-type thermoelectric legs (i.e. a
unicouple), as schematically illustrated in Figure 1.12. The contact interface
of a unicouple is a complicated region containing many layers. The design of
such a multi-layer interface usually requires a more complex processing
technique. Ideally, the interface design should be as simple as possible
since, in practical applications, a simple interface structure facilitates
manufacturing and processing. Therefore, in early studies on thermoelectric
devices, researchers attempted to connect the thermoelectric legs directly to
a conductive-strip (e.g. Cu) substrate. However, numerous experimental
Reliability and Durability of Thermoelectric Materials and Devices 27
results showed that the loss of thermoelectric energy conversion efficiency
could be as high as 40% due to the large contact resistance caused by the
diffusion between the bonding material and the thermoelectric legs (Ren
et al. 2018). Additionally, this diffusion phenomenon also leads to performance
deterioration in the thermoelectric legs, thereby shortening the service life of
the device. Taking the Bi
2
Te
3
-based thermoelectric cooling devices as a
typical case, it was found that although the Bi
2
Te
3
legs and Cu strips can be
joined at 503 K by using a Sn–Sb–Bi solder, the diffusion of Sn into the
thermoelectric legs leads to high electrical resistance. Although the electrical
resistance was reduced in later studies by replacing the Sn–Sb–Bi solder
with a Sb–Ag solder, it was found that Ag diffusion into the thermoelectric
legs could also be detrimental to the device (Rosi et al. 1962). In addition,
Cu is also highly diffusible into Bi
2
Te
3
at the operating temperature of
~523 K (Yang et al. 2013). Thus, the concept of adding a metal contact layer
to serve as both the electrode and the solder diffusion barrier was proposed.
For example, a Ni-metalized layer applied in a Bi
2
Te
3
-based thermoelectric
device can effectively prevent the diffusion of the Sn from the Sn–Sb–Bi
solder into the thermoelectric legs (Lin et al. 2017).
Figure 1.12. Schematic diagram of the structure design of a
thermoelectric unicouple. For a color version of this figure,
see www.iste.co.uk/akinaga/thermoelectric2.zip
28 Thermoelectric Micro/Nano Generators 2
A good metal contact layer and the corresponding connection technology
should have the following characteristics for either a thermoelectric power
generator or a thermoelectric cooler (He et al. 2018; Liu and Bai 2019): (1)
the coefficient of thermal expansion (CTE) of the metal contact layer must
match that of the thermoelectric leg; (2) the metalized layer must have high
thermal conductivity and electrical conductivity to reduce extra energy loss
at the interface; (3) low contact resistivity at the interface between the metal
contact layer and the thermoelectric legs is necessary; (4) the connection
technology between the metal electrodes and the thermoelectric legs should
be simple, and the bonding strength should be high enough to ensure the
stability of the device while operating at high temperatures and (5) avoiding
doping or alloying effects due to high-temperature atomic diffusion at the
interfaces is important since these could significantly alter the compositions
of the thermoelectric legs and the electrode materials, leading to
performance deterioration. The metal electrode should be inert to the
thermoelectric material to avoid serious chemical reactions. Although
diffusion is necessary for establishing a strong bond, it should be terminated
when the joining is complete. In fact, it is very difficult to meet all of the
above criteria at the same time. Therefore, determining how to balance
various influencing factors in interface design is an important subject that
needs to be studied for the development of thermoelectric devices. In
general, the interface design needs to be adjusted according to actual
application requirements.
From the perspective of a practical application, CTE matching is the most
important issue when selecting the metal contact material since the mismatch
of CTEs could cause severe internal stress, resulting in the failure of the
device. Mismatched CTEs could also increase the difficulty in
manufacturing thermoelectric devices. Ravi et al. (2009) carried out detailed
measurements and analysis of the CTEs of some advanced thermoelectric
materials, and their results confirmed that the CTE has a profound effect on
the thermomechanical robustness of the devices in which these materials are
incorporated. Therefore, CTE matching requires careful engineering to meet
performance expectations such as long-term durability. The CTEs of various
thermoelectric materials are quite different from one another, so the metal
electrode materials require targeted optimization such as composition tuning
in the contact layer that is sensitive to the elemental ratio of the
thermoelectric legs (Liu and Bai 2019). Additionally, multi-layer
Reliability and Durability of Thermoelectric Materials and Devices 29
contact materials can be applied to gradually match the CTE of the
thermoelectric materials.
In addition to CTE matching, a strong bond between the metal contact
layer and the thermoelectric legs is essential for device reliability and
durability. Generally, the metal contact layer can be made through either a
chemical route (i.e. electrochemical plating or electrochemical deposition) or
a physical route (i.e. sputtering, spraying, or direct hot pressing) (Zhang
et al. 2016). An interface made by electrochemical plating or
electrochemical deposition is very thin, and although this reduces the total
electrical and thermal resistance at the contact region, the bonding strength is
relatively low (~10 MPa) (Weitzaman 1967; Feng et al. 2013). This bonding
strength could be adequate for refrigeration applications like commercial
Bi2Te3-based devices, but it is far lower than that needed for power
generation under thousands of thermal cycles and thermal shocks. Many
studies have confirmed that increasing the surface roughness through
chemical etching or sandblasting could effectively enhance the bonding
strength because it would result in a larger contact surface area and an
enhanced anchor point effect (Zhang 2018; Liu and Bai 2019). Compared to
the chemical methods for metal contact layer production, some physical
techniques produce a stronger bond between a metal electrode and the
thermoelectric legs; these include thermal spraying (e.g. 18 MPa~25 MPa
bonding strength between a Ni-metalized layer and a Bi2Te3 thermoelectric
leg) (Zhang et al. 2014) and direct hot pressing (e.g. 50 MPa bonding
strength between a Ti-metalized layer and a half-Heusler thermoelectric leg)
(Joshi and Poudel 2016). Unfortunately, greater bonding strength is
accompanied by higher contact resistance since it is usually achieved by
deep interdiffusion or interface reactions in which an intermetallic layer
could be formed between the metal contact layer and the thermoelectric leg.
Although moderate diffusion is beneficial to interface bonding strength,
uncontrolled diffusion and chemical reactions may cause increased thermal
and electrical resistance, as well as significant changes in the composition of
the metal contact layer and the thermoelectric legs, which will eventually
cause performance deterioration in the device, or even device failure.
Therefore, a balance should be found between bonding strength and other
factors such as diffusion rate and contact resistance.
30 Thermoelectric Micro/Nano Generators 2
Using a single-layer metal contact can simplify the production process,
but it makes meeting all of the above requirements at the same time difficult
in some cases. For instance, there may be a large mismatch between the
CTEs of the metal electrode and the thermoelectric legs; the metal contact
layer could react with the thermoelectric legs quickly to form an
intermetallic layer, which may lead to the deterioration of the entire device;
or it may not be possible to produce a strong bond between the metal contact
layer and the thermoelectric leg. Therefore, multi-layer interface technology
has been introduced (Xia et al. 2014; Zhang et al. 2016; Xing et al. 2021), in
which a barrier layer (also known as a transition layer) between the electrode
and the thermoelectric leg serves as a buffer not only to slow down the
diffusion behavior but also to improve the wettability of the material in order
to facilitate brazing. Meanwhile, the barrier layer itself should also have high
thermal and electrical conductivity to reduce the extra energy loss at the
interface and thus improve the reliability of the device.
1.9.2.3. Joining
Joining processes are of great significance for the interface properties of
thermoelectric modules. In particular, a proper joining technology can
reduce the interface stress and improve the bonding strength, thereby
increasing the reliability of the thermoelectric device to a certain extent.
Over the past decades, several methods, such as the chemical and physical
routes mentioned above, have been applied to join metal electrodes and
thermoelectric legs. Among these methods, two major ones stand out as the
most reliable and suitable for industrial manufacturing: diffusion welding
and thermal spraying (see Figure 1.13). Diffusion welding, also known as
one-step sintering, has been successfully applied in the fabrication of
Mg3Sb(Bi)2 Zintls, MgAgSb, skutterudites, half-Heuslers, etc. (Zhao et al.
2012; Joshi and Poudel 2016; Zhang et al. 2016; Yin et al. 2020; Ying et al.
2021). As shown in Figure 1.13a, for the diffusion welding process, the
electrode material powder and the thermoelectric material powder
(sometimes including barrier material powder) are added into a graphite die
in order, and the die is subsequently heated by using a direct hot press or a
spark plasma sintering method. Using this process, the sandwich structure
composed of the electrode material, barrier material and thermoelectric
material powder is thus formed at one time. The main advantages of this
method include its simplicity and the resulting high interface bonding
strength. Thermal spraying, a newer technology, is more efficient and
Reliability and Durability of Thermoelectric Materials and Devices 31
scalable as compared to diffusion welding. As shown in Figure 1.13b, for the
thermal spraying process, the thermoelectric material is cut to the desired
dimensions for the thermoelectric legs and then plugged into a prefabricated
frame made of a polymer or a ceramic. The metal electrode material and the
barrier layer material can then be directly sprayed onto the contact surfaces
of the thermoelectric legs. Thermal spraying is conducive to the
simultaneous realization of interface processing and module integration for
the device. Additionally, joins prepared by this method show higher stability
at high temperatures (Zhang et al. 2016).
Figure 1.13. Schematic diagrams of two major technologies for joining thermoelectric
legs and metal contact layers: (a) diffusion welding. Reprinted from Ying et al. (2021)
and (b) thermal spraying. Reprinted from Zhang et al. (2016). For a color version of
this figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
After the metal electrode layer and the barrier layer are successfully
connected to the thermoelectric legs, this working structure needs to be
connected to the base (i.e. the substrate and the Cu conductive strips, which
are generally joined in advance) by soldering or brazing. The selections of
solder, interface pretreatment and soldering method are critical since the
solder exhibits the highest CTE and the lowest yield strength among the
interface components. The solder, in the form of a powder, paste or foil, is
placed at the interface between the working structure and the base and
should be heated and pressed to promote melting, and it is then cooled to
firmly join the two parts (Kähler et al. 2014; Tewolde et al. 2015). Based on
the above, one basic requirement for the solder is that its melting point
should be lower than that of the thermoelectric legs and the conductive
32 Thermoelectric Micro/Nano Generators 2
strips. Given that the maximum stress level is located on the thermoelectric
legs near the hot-side plate, the use of a low-yield soldering alloy at the hot
side and the promotion of plastic deformation in the solder can result in
considerable thermal stress relief. The melting point of the solder should also
be higher than the operating temperature to prevent interface softening or
failure of the module during service. Additionally, soldering thickness has an
impact on the temperature gradient and thermal stress distribution. An
optimal soldering thickness will achieve both high conversion efficiency and
low thermal stress (Wu et al. 2014). Currently, the most used solder
materials are Sn-based alloys, such as Sn–Ag, Sn–Cu and Au–Sn.
Figure 1.14. Contact angle of a water drop on the surface of a nanostructured
Bi
2
Te
3-y
Se
y
thermoelectric alloy (a) before and (b) after post-CMP (chemical
mechanical polishing) cleaning (reprinted from Feng et al. (2011))
It is worth noting that the quality of the soldering interface is not only
associated with the type of solder and the soldering process but also with the
cleanness of the interface. Careful surface cleaning is beneficial for
removing surface oxides and hence reducing the contact resistance. In
addition, surface cleaning can increase wettability significantly, which is
beneficial for promoting welding. In a typical example, the contact angle of
water dripped onto a Bi
2
Te
3
surface was reduced from 79.7° to 53.6° (i.e.
hydrophilicity increases) after a five-step treatment combining chemical and
physical methods, as shown in Figure 1.14 (Feng et al. 2011). An interface
incorporating solder has an inevitable shortcoming since it is easy to
introduce contamination during the soldering process and form undesired
phases which are harmful to the thermoelectric legs. In addition to applying
Reliability and Durability of Thermoelectric Materials and Devices 33
a metal contact layer and a barrier layer, which can reduce the deterioration
on the thermoelectric legs caused by solder to a certain extent, some new
joining processes, such as non-soldering-based joining, have been developed
and investigated (Malik et al. 2017). The solder composition could also be
adjusted appropriately for different thermoelectric compositions.
1.10. Advanced thermoelectric module case studies
In the previous section, an overview of thermoelectric device design and
the major related issues was provided. Generally, the design principles
described above can be applied to most thermoelectric materials. However,
for particular materials, targeted strategies are required to address issues
such as thermal expansion, thermal stress, bonding strength, contact
resistance and interfacial reaction or diffusion due to the differences in these
materials’ unique characteristics and specific operating temperature ranges.
Hence, the reliability-related issues of some advanced thermoelectric
modules and the corresponding strategies to address these issues are
introduced in the following section.
1.10.1. Bi2Te3
Since they are among the most successful commercial thermoelectric
materials, Bi2Te3-based materials have been studied extensively for decades.
Usually, Ni is applied as a metal contact layer (3~10 μm in thickness) in
Bi2Te3-based devices and the diffusion of Ni at the interface is negligible
when the operating temperature is under 473 K. Additionally, Ni can
effectively inhibit the formation of brittle Sn–Te intermetallic compounds
resulting from the reaction between Sn-containing solder and the
thermoelectric matrix. Therefore, as a metal contact layer, Ni is appropriate
for refrigeration applications. However, when it is applied to thermoelectric
power generators, a number of problems appear. Park et al. (2012) carried
out a thermal cycling test from 303 K to 433 K on a commercial Bi2Te3
module, as shown in Figure 1.15a. During the first 2,000 cycles, the power
output was found to remain consistent with the prediction value, but between
2,000 and 6,000 cycles, there were 11% and 12% reductions in the power
output for load resistance values of 47 Ω and 1,000 Ω, respectively, as
shown in Figure 1.15b.
34 Thermoelectric Micro/Nano Generators 2
Figure 1.15. (a) Thermal cycling test system configuration for a commercial Bi
2
Te
3
module. Reprinted from Park et al. (2012). (b) Power output as a function of load
resistance. Two power output data point (maximum and minimum values) for each
load resistance value indicate the degradation of the thermoelectric module during
thermal cycling. Data obtained from Park et al. (2012). For a color version of this
figure, see www.iste.co.uk/akinaga/thermoelectric2.zip
Liu et al. (2013) studied the interface reaction between Ni and n-type
Bi
2
Te
2.7
Se
0.3
and that between Ni and p-type Bi
0.4
Sb
1.6
Te
3
by analyzing the
interface microstructure of each sample hot-pressed at different
temperatures, as shown in Figure 1.16. It is clear that the thickness of the
interface reaction layer increases with increasing hot-pressing temperature.
The thick diffusion layer is beneficial for strong bonding at the interface, but
the diffusion of Ni was found to form a p-type region within the n-type
Bi
2
Te
2.7
Se
0.3
leg, resulting in a large contact resistance for the n-type
interface (~210 μΩ cm
2
), while the contact resistance remains low for the
p-type Bi
0.4
Sb
1.6
Te
3
leg (<1 μΩ cm
2
) (Liu et al. 2013). Applying a barrier
layer between the Ni contact layer and these thermoelectric materials is
advised to address severe interface reactions at high temperatures.
Additionally, seeking out new materials for the metal contact layer could be
an alternative for this interface design. Research on devices based on Bi
2
Te
3
remains ongoing and is currently mainly focused on improving output
performance. For instance, Nozariasbmarz et al. (2020) reported that a
record 8% conversion efficiency and 2.1 W/cm
2
power density were obtained
in an n-/p-type Bi
2
Te
3
-based module for waste heat recovery (298 K~523 K).
In addition, a Bi
2
Te
3
/half-Heusler-based segmented unicouple and a
Bi
2
Te
3
/skutterudite segmented module also show remarkable energy
Reliability and Durability of Thermoelectric Materials and Devices 35
conversion efficiency around 12% (Li et al. 2020d). Unfortunately, results
from aging tests and long-term service tests were not provided in these
reports.
Figure 1.16. Microstructures of contact interfaces made by directly hot-pressing Ni
powder together with p-type Bi
0.4
Sb
1.6
Te
3
or n-type Bi
2
Te
2.7
Se
0.3
thermoelectric
powder. Ni/Bi
0.4
Sb
1.6
Te
3
/Ni hot-pressed at (a) 673 K, (c) 723 K and (e) 773 K.
Ni/Bi
2
Te
2.7
Se
0.3
/Ni hot-pressed at (b) 673 K, (d) 723 K and (f) 773 K. In each image,
the dark region is nickel and the bright region is the thermoelectric material.
Reprinted from Liu et al. (2013)
1.10.2. Mg
3
(Sb, Bi)
2
N-type Mg
3
(Sb, Bi)
2
-based materials have attracted tremendous attention
in recent years because of their promising thermoelectric performance, low
36 Thermoelectric Micro/Nano Generators 2
cost and good mechanical properties. It is worth mentioning that Mg3(Sb,
Bi)2-based devices can be used for different applications in several
temperature ranges, for example, room-temperature thermoelectric cooling
(Mao et al. 2019a; Shu et al. 2019a), low-grade heat recovery (Imasato et al.
2019b; Liang et al. 2021) and power generation in the medium-temperature
range (Xu et al. 2021), by tuning the ratio of Sb and Bi to either Sb-
or Bi-rich. As an emerging material system, Mg3(Sb, Bi)2-based
thermoelectrics have received little attention in terms of interface research.
Zhu et al. (2019) were the first to report on the Mg3(Sb, Bi)2 contact
interface in 2019. In their study, a good junction interface with low contact
resistivity and strong bonding strength was formed between
Mg3.1Co0.1Sb1.5Bi0.49Te0.01 and Fe using a direct hot-pressing method, which
was found to lead to a high energy conversion efficiency of 10.6% between
373 K and 773 K. Fe has since been successfully used as a metal contact
electrode layer in a thermoelectric cooling system (Mao et al. 2019a) and a
low-grade waste heat recovery device (Liang et al. 2021). In a later study, Ni
was found to play a role similar to that of Fe as a contact layer (Mao et al.
2019a). Ni can also ensure a low electrical contact resistance for soldering
once it is electroplated on both sides of the Mg3(Sb, Bi)2 thermoelectric leg
(Shang et al. 2020b; Xu et al. 2021). On the other hand, the Fe contact layer
could degrade upon long-term aging, which may be attributed to the
mismatched CTE values and the undesirable interface reaction between the
thermoelectric material and Fe. Thus, 304 stainless steel was introduced as a
potential replacement by Yin et al. (2020). An interface constructed using
304 stainless steel and a Mg3(Sb, Bi)2-based thermoelectric material shows
smaller changes in its electrical and mechanical properties with aging
compared to that constructed using Fe due to its suppressed diffusion and
enhanced expansion coefficient match, as shown in Figure 1.17. A high
conversion efficiency of up to 9% at a temperature difference of 370 K in a
single leg incorporating this Mg3(Sb, Bi)2-based thermoelectric material and
304 stainless steel was reported in the same study. Unfortunately, no studies
of the long-term operation of an integrated module based on Mg3(Sb, Bi)2
have yet been reported. In short, despite the good thermoelectric properties
of Mg3(Sb, Bi)2-based materials, only with abundant contact layer research
and module evaluation can these compounds achieve a large-scale practical
application.
Reliability and Durability of Thermoelectric Materials and Devices 37
Figure 1.17. Aging-time-dependent contact resistivity of Mg3.2Sb1.5Bi0.49Te0.01/304
stainless steel and Mg3.2Sb1.5Bi0.49Te0.01/Fe junctions aged at 523 K. Data obtained
from Yin et al. (2020). For a color version of this figure, see www.iste.co.uk/akinaga/
thermoelectric2.zip
1.10.3. GeTe
Research on GeTe as a thermoelectric material began as early as the
1960s, but it was not studied in depth for long since it has an excessively
high carrier concentration (~1021 cm
-3) and an undesired phase change at
~700 K. In recent years, GeTe has returned as a focus of thermoelectric
research because it was found that its carrier concentration can be reduced to
an ideal range (~1020 cm-3) by doping with Sb or Bi (Liu et al. 2018; Zheng
et al. 2018). Additionally, its undesired phase transition can be suppressed
by element doping, as discussed in section 1.5. GeTe has subsequently
become a promising candidate for use in thermoelectric generators for waste
heat recovery in the mid-temperature range. Although the thermoelectric
performance of GeTe has been greatly improved, device research is currently
lagging significantly, delaying actual applications. At high temperatures, fast
chemical reactions and diffusion can occur between the metal electrodes and
the thermoelectric legs in GeTe-based thermoelectric modules. Thus, a
barrier layer is required. An Al–Si alloy was recently proposed as a diffusion
barrier for GeTe-based thermoelectric legs by Li et al. (2020b). In their
study, a good connection between Al66Si34 and Ge
0.9Sb0.1TeB0.01 was
obtained using the spark plasma sintering method and the resulting
thermoelectric legs underwent aging testing for 16 days at 773 K. After the
38 Thermoelectric Micro/Nano Generators 2
16-day test, their contact resistivity increased slightly from 20.7 to
26.1 μΩ·cm
2
. The thickness of the interface diffusion layer increased quickly
from ~8 to ~18 μm during the first 4 days of aging but slowed down after
8 days of aging. Overall, the thickness of the interface diffusion layer
increased to ~23 μm after 16 days of aging, as shown in Figure 1.18a–c. The
shear strength of the thermoelectric legs also increased from 26.6 MPa to
41.7 MPa after the aging test, which may be attributed to the thicker
diffusion layer. Xing et al. (2021) chose Mo as an effective diffusion barrier
material for GeTe-based thermoelectric legs by testing 12 pure metals (Cr,
Hf, Nb, Ti, Mo, Ni, Ta, Zr, Al, Co, V and Fe). The result of resistance line
scanning across the interface of GeTe/Mo/GeTe indicates that an
extraordinarily low value (less than 1 μΩ·cm
2
) can be obtained.
Furthermore, a Ge
0.92
Sb
0.04
Bi
0.04
Te
0.95
Se
0.05
/Mo/Ni thermoelectric leg was
constructed and underwent aging testing at 800 K, and its interface
microstructure was found to remain stable after 10 days of testing, as shown
in Figure 1.18d–f. Finally, an eight-couple Ge
0.92
Sb
0.04
Bi
0.04
Te
0.95
Se
0.05
/Yb
0.3
Co
4
Sb
12
thermoelectric module was fabricated and achieved an energy
conversion efficiency of 7.8% and a power density of 1.1 W·cm
-2
under a
temperature difference of 500 K.
Figure 1.18. Scanning electron microscopy images of the interface area of
Ge
0.9
Sb
0.1
TeB
0.01
/Al
66
Si
34
thermoelectric legs after (a) 4 days, (b) 8 days and (c) 16
days of aging at 773 K. Reprinted from Li et al. (2020b). Scanning electron
microscopy images of the interface area of Ge
0.92
Sb
0.04
Bi
0.04
Te
0.95
Se
0.05
/Mo/Ni
thermoelectric legs (d) before aging and after aging (e) 5 days and (f) 10 days at
800 K. Reprinted from Xing et al. (2021). For a color version of this figure, see
www.iste.co.uk/akinaga/thermoelectric2.zip
Reliability and Durability of Thermoelectric Materials and Devices 39
1.10.4. Skutterudites
Filled skutterudites show promising thermoelectric and mechanical
properties for mid-temperature power generation. The contact design of
skutterudites has been extensively examined in recent decades. Pure metals
and alloys like Ni, Ti, Zr, Nb, Mo-Cu and Co-Fe-Ni have been employed
and evaluated as metal contact layers (Zhang 2018; Liu and Bai 2019).
Although Ni is a good candidate for most thermoelectric materials in
interfacial design, it is not suitable for skutterudites due to its larger CTE.
Therefore, a buffer layer is needed when Ni is applied as the metal contact
layer for skutterudites. Ti- and Zr-based contact layers often suffer from a
severe interfacial chemical reaction during the aging process, so they do not
provide adequate stability for long-term power generation service. The
Mo–Cu alloy is another potential candidate for the metal contact layer
because of its high electrical and thermal conductivity. Additionally, its
composition can be easily tuned to match the CTE of the filled skutterudite
(Zhang 2018). However, it is very difficult to bond the Mo–Cu alloy onto the
surface of a skutterudite thermoelectric leg. In fact, the most significant
challenge in choosing a metal contact layer for skutterudites is that their
main element, Sb, is highly reactive, which may cause a severe interfacial
reaction and diffusion at high temperatures, and thus significantly degrade
the device performance. To balance effective bonding and a durable
interface, Chu et al. (2020) recently proposed a criterion to predict the
bonding behavior as well as the interfacial reliability of CoSb3-based filled
skutterudite/metal electrode interfaces by combining the interfacial reaction
energy (EIR) value of a metal with its Sb migration activation energy barrier
(EMig) value, as shown in Figure 1.19a. Generally, EIR should be negative to
promote the progress of the interface reaction and produce a mechanical
force at the interface, while EMig should be large enough to limit Sb diffusion
and prevent the interface reaction layer from growing too quickly. Based on
their simulation results, they defined a “sweet spot” area as shown in Figure
1.19a, and metals in this region were determined to be potential candidates
for the metal contact layer. To validate the proposed criterion, Nb was
chosen as the metal contact layer to fabricate a CoSb3-based filled
skutterudite thermoelectric leg. Their experimental results showed robust
bonding with no obvious diffusion under thermal cycles, confirming that the
interface reaction was inhibited and the interfacial resistivity was reduced.
The derived module also showed excellent stability in output power as well
as internal resistance after 846 hours of service testing, as shown in Figure
1.19b. In addition, work–function matching is another effective guiding
principle for interfacial design that was proposed by Jie et al. (2015). They
40 Thermoelectric Micro/Nano Generators 2
reported that the matching of the work function between a metal contact
layer and a filled skutterudite can play a crucial role in determining the
electrical contact resistance based on their study of the interface behavior
between the p-type Ce
0.45
Nd
0.45
Fe
3.5
Co
0.5
Sb
12
and a Cr–Fe–Co alloy or a
Cr–Fe–Ni alloy.
Figure 1.19. (a) Correlation map of interfacial reaction energy (E
IR
) and Sb migration
activation energy barrier (E
Mig
) values of different metals. (b) Output power and
internal resistance of an eight-pair CoSb
3
-based filled skutterudite module fabricated
using Nb as the barrier layer under long-term service conditions with T
h
= 818 K and
T
c
= 308 K. Data obtained from Chu et al. (2020). For a color version of this figure,
see www.iste.co.uk/akinaga/thermoelectric2.zip
1.11. Summary and outlook
In comparison with zT improvement among various thermoelectric
materials over the past decades, the development process from
thermoelectric materials to devices has been much slower. For the materials’
optimization, more effort should be devoted to improving the thermal
stability and mechanical robustness of the materials with relatively high zT
because commercialization will not be possible until such improvement is
achieved. To accelerate commercialization, a multi-criteria evaluation
method needs to be developed based on the performance and reliability
requirements. Since the interfacial layer plays an important role in
performance maximization and structure integrity, it is meaningful to
construct a rational model based on experiment and simulation to guide the
design of high-performance contact layers. For device development,
evaluating the performance of each lab-prepared module based on multiple
industrial standards (e.g. vibration testing, mechanical shock testing,
Reliability and Durability of Thermoelectric Materials and Devices 41
temperature cycle endurance testing, etc.) can help to determine commercial
viability in realizing actual applications. Most importantly, it is time to try
all possible device applications to demonstrate some of the promise of the
past decades of materials research to build a case for economical adaption of
this technology to real applications. With enough practical applications of
thermoelectric devices, more research will be funded and can be carried out,
which in turn will enhance the device applications. Such a healthy,
self-propelling cycle is what the thermoelectric research community needs
going forward.
1.12. References
Al-Merbati, A.S., Yilbas, B.S., Sahin, A.Z. (2013). Thermodynamics and thermal
stress analysis of thermoelectric power generator: Influence of pin geometry on
device performance. Appl. Therm. Eng., 50(1), 683–692.
Barako, M.T., Park, W., Marconnet, A.M., Asheghi, M., Goodson, K.E. (2012).
Thermal cycling, mechanical degradation, and the effective figure of merit of a
thermoelectric module. J. Electron. Mater., 42(3), 372–381.
Bohra, A.K., Bhatt, R., Bhattacharya, S., Basu, R., Ahmad, S., Singh, A., Aswal, D.K.,
Gupta, S.K. (2016). Study of thermal stability of Cu2Se thermoelectric material.
AIP Conf. Proc., 1713, 110010.
Bohra, A.K., Bhatt, R., Singh, A., Bhattacharya, S., Basu, R., Bhatt, P.,
Navaneethan, M., Sarkar, S.K., Anwar, S., Muthe, K.P. et al. (2020). Stabilizing
thermoelectric figure-of-merit of superionic conductor Cu2Se through
nanoinclusions. Phys. Status Solidi (RRL), 14, 2000102.
Chen, L.D., Dong, H.L., Li, X.Y., Huang, X.Y., Jiang, W., Tang, Y.S., Xia, X.G.,
Liao, J.C., He, L. (2013). Skutterudite based thermoelectric material, thermal
protection coating for devices and preparation method thereof. Patent,
CN103146301A.
Chowdhury, I., Prasher, R., Lofgreen, K., Chrysler, G., Narasimhan, S., Mahajan, R.,
Koester, D., Alley, R., Venkatasubramanian, R. (2009). On-chip cooling by
superlattice-based thin-film thermoelectrics. Nat. Nanotechnol., 4(4), 235–238.
Chu, J., Huang, J., Liu, R., Liao, J., Xia, X., Zhang, Q., Wang, C., Gu, M., Bai, S.,
Shi, X. et al. (2020). Electrode interface optimization advances conversion
efficiency and stability of thermoelectric devices. Nat. Commun., 11(1), 2723.
42 Thermoelectric Micro/Nano Generators 2
Dasgupta, T., Stiewe, C., Sesselmann, A., Yin, H., Iversen, B.B., Mueller, E. (2013).
Thermoelectric studies in β-Zn4Sb3-the complex interdependence between
thermal stability, thermoelectric transport, and zinc content. J. Appl. Phys., 113,
103708.
Deng, S., Chen, Z., Li, D., Liu, H., Tang, Y., Shen, L., Deng, S. (2017). Effect of Pb
doped on thermal stability and electrical transport properties of single crystalline
β-Zn4Sb3. Physica B, 526, 155–159.
Dong, H., Li, X., Tang, Y., Zou, J., Huang, X., Zhou, Y., Jiang, W., Zhang, G.-J.,
Chen, L. (2012). Fabrication and thermal aging behavior of skutterudites with
silica-based composite protective coatings. J. Alloys Compd., 527, 247–251.
El-Genk, M.S., Saber, H.H., Caillat, T., Sakamoto, J. (2006). Tests results and
performance comparisons of coated and un-coated skutterudite based segmented
unicouples. Energy Convers. Manag., 47(2), 174–200.
Erturun, U., Erermis, K., Mossi, K. (2015). Influence of leg sizing and spacing on
power generation and thermal stresses of thermoelectric devices. Appl. Energy,
159, 19–27.
Fabián-Mijangos, A., Min, G., Alvarez-Quintana, J. (2017). Enhanced performance
thermoelectric module having asymmetrical legs. Energy Convers. Manag., 148,
1372–1381.
Feng, H.-P., Yu, B., Chen, S., Collins, K., He, C., Ren, Z.F., Chen, G. (2011).
Studies on surface preparation and smoothness of nanostructured Bi2Te3-based
alloys by electrochemical and mechanical methods. Electrochim. Acta, 56(8),
3079–3084.
Feng, S.P., Chang, Y.H., Yang, J., Poudel, B., Yu, B., Ren, Z., Chen, G. (2013).
Reliable contact fabrication on nanostructured Bi2Te3-based thermoelectric
materials. Phys. Chem. Chem. Phys., 15(18), 6757–6762.
Hao, Y., Pedersen, B.L., Iversen, B.B. (2011). Thermal stability of high performance
thermoelectric β-Zn4Sb3 in argon. Eur. J. Inorg. Chem., 17, 2733–2737.
He, R., Schierning, G., Nielsch, K. (2018). Thermoelectric devices: A review of
devices, architectures, and contact optimization. Adv. Mater. Technol., 3(4),
1700256.
Hong, M., Zheng, K., Lyv, W., Li, M., Qu, X., Sun, Q., Xu, S., Zou, J., Chen, Z.-G.
(2020). Computer-aided design of high-efficiency GeTe-based thermoelectric
devices. Energy Environ. Sci., 13(6), 1856–1864.
Imasato, K., Wood, M., Kuo, J.J., Snyder, G.J. (2018). Improved stability and high
thermoelectric performance through cation site doping in n-type La-doped
Mg3Sb1.5Bi0.5. J. Mater. Chem. A, 6(41), 19941–19946.
Reliability and Durability of Thermoelectric Materials and Devices 43
Imasato, K., Kang, S.D., Snyder, G.J. (2019). Exceptional thermoelectric
performance in Mg3Sb0.6Bi1.4 for low-grade waste heat recovery. Energy
Environ. Sci., 12, 965–971.
Ioffe, A.F. (1957). Semiconductor Thermoelements and Thermoelectric Cooling.
Infosearch Ltd, London.
Jia, X. and Gao, Y. (2014). Estimation of thermoelectric and mechanical
performances of segmented thermoelectric generators under optimal operating
conditions. Appl. Therm. Eng., 73(1), 335–342.
Jie, Q., Ren, Z., Chen, G. (2015). Fabrication of stable electrode/diffusion barrier
layers for thermoelectric filled skutterudite devices. Patent, US9722164B2.
Jørgensen, L.R., Zhang, J., Zeuthen, C.B., Iversen, B.B. (2018). Thermal stability of
Mg3Sb1.475Bi0.475Te0.05 high performance n-type thermoelectric investigated
through powder X-ray diffraction and pair distribution function analysis.
J. Mater. Chem. A, 6(35), 17171–17176.
Joshi, G. and Poudel, B. (2016). Efficient and robust thermoelectric power
generation device using hot-pressed metal contacts on nanostructured
half-Heusler alloys. J. Electron. Mater., 45(12), 6047–6051.
Kähler, J., Stranz, A., Waag, A., Peiner, E. (2014). Thermoelectric coolers with
sintered silver interconnects. J. Electron. Mater., 43(6), 2397–2404.
Kambe, M., Jinushi, T., Ishijima, Z. (2010). Encapsulated thermoelectric modules
and compliant pads for advanced thermoelectric systems. J. Electron. Mater.,
39(9), 1418–1421.
Kang, S.D., Pohls, J.-H., Aydemir, U., Qiu, P., Stoumpos, C.C., Hanus, R.,
White, M.A., Shi, X., Chen, L., Kanatzidis, M.G. et al. (2017). Enhanced
stability and thermoelectric figure-of-merit in copper selenide by lithium doping.
Mater. Today Phys., 1, 7–13.
Karri, N.K. and Mo, C. (2018). Structural reliability evaluation of thermoelectric
generator modules: Influence of end conditions, leg geometry, metallization, and
processing temperatures. J. Electron. Mater., 47(10), 6101–6120.
Kim, C.N. (2018). Development of a numerical method for the performance analysis
of thermoelectric generators with thermal and electric contact resistance. Appl.
Therm. Eng., 130, 408–417.
Kim, H.S., Wang, T., Liu, W., Ren, Z. (2016). Engineering thermal conductivity for
balancing between reliability and performance of bulk thermoelectric generators.
Adv. Funct. Mater., 26(21), 3678–3686.
Kim, H.S., Liu, W., Ren, Z. (2017). The bridge between the materials and devices of
thermoelectric power generators. Energy Environ. Sci., 10(1), 69–85.
44 Thermoelectric Micro/Nano Generators 2
Kuo, J.J., Yu, Y., Kang, S.D., Cojocaru‐Mirédin, O., Wuttig, M., Snyder, G.J.
(2019). Mg deficiency in grain boundaries of n‐type Mg3Sb2 identified by atom
probe tomography. Adv. Mater. Inter., 6(13), 1900429.
Li, F., Wei, T.-R., Kang, F., Li, J.-F. (2014). Thermal stability and oxidation
resistance of BiCuSeO based thermoelectric ceramics. J. Alloys Compd., 614,
394–400.
Li, J., Jia, F., Zhang, S., Zheng, S., Wang, B., Chen, L., Lu, G., Wu, L. (2019). The
manipulation of substitutional defects for realizing high thermoelectric
performance in Mg3Sb2-based Zintl compounds. J. Mater. Chem. A, 7(33),
19316–19323.
Li, J., Zhang, S., Jia, F., Zheng, S., Shi, X., Jiang, D., Wang, S., Lu, G., Wu, L.,
Chen, Z.-G. (2020a). Point defect engineering and machinability in n-type
Mg3Sb2-based materials. Mater. Today Phys., 15, 100269.
Li, J., Zhao, S., Chen, J., Han, C., Hu, L., Liu, F., Ao, W., Li, Y., Xie, H., Zhang, C.
(2020b). Al–Si alloy as a diffusion barrier for GeTe-based thermoelectric legs
with high interfacial reliability and mechanical strength. ACS Appl. Mater.
Interfaces, 12(16), 18562–18569.
Li, M., Hong, M., Tang, X., Sun, Q., Lyu, W.-Y., Xu, S.-D., Kou, L.-Z., Dargusch, M.,
Zou, J., Chen, Z.-G. (2020c). Crystal symmetry induced structure and bonding
manipulation boosting thermoelectric performance of GeTe. Nano Energy, 73,
104740.
Li, W., Poudel, B., Nozariasbmarz, A., Sriramdas, R., Zhu, H., Kang, H.B., Priya, S.
(2020d). Bismuth telluride/half‐Heusler segmented thermoelectric unicouple
modules provide 12% conversion efficiency. Adv. Energy Mater., 10(38),
2001924.
Li, M., Sun, Q., Xu, S.D., Hong, M., Lyu, W.Y., Liu, J.X., Wang, Y., Dargusch, M.,
Zou, J., Chen, Z.G. (2021). Optimizing electronic quality factor toward
high-performance Ge1-x-yTaxSbyTe thermoelectrics: The role of transition metal
doping. Adv. Mater., 33(40), 2102575.
Liang, Z., Xu, C., Shang, H., Zhu, Q., Ding, F., Mao, J., Ren, Z. (2021). High
thermoelectric energy conversion efficiency of a unicouple of n-type Mg3Bi2 and
p-type Bi2Te3. Mater. Today Phys., 19, 100413.
Lin, W.-C., Li, Y.-S., Wu, A.T. (2017). Study of diffusion barrier for solder/n-type
Bi2Te3 and bonding strength for p- and n-type thermoelectric modules.
J. Electron. Mater., 47(1), 148–154.
Liu, W. and Bai, S. (2019). Thermoelectric interface materials: A perspective to the
challenge of thermoelectric power generation module. J. Mater., 5(3), 321–336.
Reliability and Durability of Thermoelectric Materials and Devices 45
Liu, H., Shi, X., Xu, F., Zhang, L., Zhang, W., Chen, L., Li, Q., Uher, C., Day, T.,
Snyder, G.J. (2012). Copper ion liquid-like thermoelectrics. Nat. Mater., 11(5),
422–425.
Liu, W., Wang, H., Wang, L., Wang, X., Joshi, G., Chen, G., Ren, Z. (2013).
Understanding of the contact of nanostructured thermoelectric n-type
Bi2Te2.7Se0.3 legs for power generation applications. J. Mater. Chem. A, 1,
13093–13100.
Liu, W., Jie, Q., Kim, H.S., Ren, Z. (2015). Current progress and future challenges
in thermoelectric power generation: From materials to devices. Acta Mater., 87,
357–376.
Liu, Z., Sun, J., Mao, J., Zhu, H., Ren, W., Zhou, J., Wang, Z., Singh, D.J., Sui, J.,
Chu, C.W. et al. (2018). Phase-transition temperature suppression to achieve
cubic GeTe and high thermoelectric performance by Bi and Mn codoping. Proc.
Natl. Acad. Sci. USA, 115(21), 5332–5337.
Liu, W., Shen, L., Shai, X., Tang, Y., Chen, Z., Sun, L., Ge, W., Deng, S. (2019).
Effect of hole modulation on the electrical transport properties and thermal
stability of single crystal α-Cu2Se prepared by Bi-flux method. J. Alloy. Compd.,
791, 60–67.
Liu, Z.-Y., Zhu, J.-L., Tong, X., Niu, S., Zhao, W.-Y. (2020). A review of
CoSb3-based skutterudite thermoelectric materials. J. Adv. Ceram., 9(6),
647–673.
Lu, J., Li, D., Liu, W., Shen, L., Chen, J., Ge, W., Deng, S. (2020a). Thermal
stability and thermoelectric properties of Cd-doped nano-layered Cu2Se prepared
using NaCl flux method. Chin. Phys. B, 29, 127403.
Lu, R., Olvera, A., Bailey, T.P., Uher, C., Poudeu, F.P. (2020b). Nanoscale
engineering of polymorphism in Cu2Se-based composites. ACS Appl. Mater.
Interfaces, 12(28), 31601–31611.
Malik, S.A., Hung, L.T., Van Nong, N. (2017). Solder free joining as a highly
effective method for making contact between thermoelectric materials and
metallic electrodes. Mater. Today Energy, 5, 305–311.
Mao, J., Shuai, J., Song, S., Wu, Y., Dallye, R., Zhou, J., Liu, Z., Sun, J., Zhang, Q.,
Cruz, C.D. et al. (2017a). Manipulation of ionized impurity scattering for
achieving high thermoelectric performance in n-type Mg3Sb2-based materials.
Proc. Natl. Acad. Sci. USA, 114(40), 10548–10553.
Mao, J., Wu, Y., Song, S., Zhu, Q., Shuai, J., Liu, Z., Pei, Y., Ren, Z. (2017b).
Defect engineering for realizing high thermoelectric performance in n-type
Mg3Sb2-based materials. ACS Energy Lett., 2(10), 2245–2250.
46 Thermoelectric Micro/Nano Generators 2
Mao, J., Liu, Z., Zhou, J., Zhu, H., Zhang, Q., Chen, G., Ren, Z. (2018). Advances
in thermoelectrics. Adv. Phys., 67(2), 69–147.
Mao, J., Zhu, H., Ding, Z., Liu, Z., Gamage, G.A., Chen, G., Ren, Z. (2019a). High
thermoelectric cooling performance of n-type Mg3Bi2-based materials. Science,
365(6452), 495–498.
Mao, T., Qiu, P., Du, X., Hu, P., Zhao, K., Xiao, J., Shi, X., Chen, L. (2019b).
Enhanced thermoelectric performance and service stability of Cu2Se via tailoring
chemical compositions at multiple atomic positions. Adv. Funct. Mater., 30(6),
1908315.
Mejri, M., Malard, B., Thimont, Y., Romanjek, K., Ihou Mouko, H., Estournès, C.
(2020). Thermal stability of Mg2Si0.55Sn0.45 for thermoelectric applications.
J. Alloys Compd., 846, 156413.
Moghaddam, A.O., Shokuhfar, A., Zhang, Y., Zhang, T., Cadavid, D., Arbiol, J.,
Cabot, A. (2019). Ge-doped ZnSb/β-Zn4Sb3 nanocomposites with high
thermoelectric performance. Adv. Mater. Interfaces, 6(18), 1900467.
Music, D., Geyer, R.W., Keuter, P. (2016). Thermomechanical response of
thermoelectrics. Appl. Phys. Lett., 109(22), 223903.
Ni, J.E., Case, E.D., Schmidt, R.D., Wu, C.-I., Hogan, T.P., Trejo, R.M.,
Kirkham, M.J., Lara-Curzio, E., Kanatzidis, M.G. (2013). The thermal expansion
coefficient as a key design parameter for thermoelectric materials and its
relationship to processing-dependent bloating. J. Mater. Sci., 48(18), 6233–6244.
Nozariasbmarz, A., Poudel, B., Li, W., Kang, H.B., Zhu, H., Priya, S. (2020).
Bismuth telluride thermoelectrics with 8% module efficiency for waste heat
recovery application. Science, 23(7), 101340.
Nunn, R., Qiu, P., Yin, M., Chen, H., Hanus, R., Song, Q., Zhang, T., Chou, M.-Y.,
Agne, M.T., He, J. et al. (2017). Ultrahigh thermoelectric performance in
Cu2Se-based hybrid materials with highly dispersed molecular CNTs. Energy
Environ. Sci., 10, 1928–1935.
Park, W., Marconnet, A.M., Asheghi, M., Goodson, K.E. (2012). 13th InterSociety
Conference on Thermal and Thermomechanical Phenomena in Electronic
Systems. IEEE, San Diego.
Rausch, E., Balke, B., Ouardi, S., Felser, C. (2016). Long-term stability of
(Ti/Zr/Hf)CoSb1-xSnx thermoelectric p-type half-Heusler compounds upon
thermal cycling. Energy Technol., 3, 1217–1224.
Reliability and Durability of Thermoelectric Materials and Devices 47
Ravi, V., Firdosy, S., Caillat, T., Brandon, E., Van Der Walde, K., Maricic, L.,
Sayir, A. (2009). Thermal expansion studies of selected high-temperature
thermoelectric materials. J. Electron. Mater., 38(7), 1433–1442.
Ren, Z., Lan, Y., Zhang, Q. (2018). Advanced Thermoelectrics: Materials, Contacts,
Devices, and Systems. CRC Press, New York.
Rogl, G. and Rogl, P. (2017). How nanoparticles can change the figure of merit, ZT,
and mechanical properties of skutterudites. Mater. Today Phys., 3, 48–69.
Rogl, G., Zhang, L., Rogl, P., Grytsiv, A., Falmbigl, M., Rajs, D., Kriegisch, M.,
Müller, H., Bauer, E., Koppensteiner, J. et al. (2010). Thermal expansion of
skutterudites. Int. J. Appl. Phys., 107(4), 043507.
Rosi, F.D. and Bernoff, R.A. (1962). Method and materials for obtaining low
resistance bonds to thermoelectric bodies. Patent, US3037064A.
Rowe, D.M. (1995). Thermoelectrics Handbook. CRC Press, London.
Sakamoto, J.S., Snyder, G.J., Calliat, T., Fleurial, J.P., Jones, S.M., Palk, J.A.
(2008). System and method for suppressing sublimation using opacified aerogel.
Patent, US7461512B2.
Sales, B.C., Mandrus, D., Williams, R.K. (1996). Filled skutterudite antimonides: A
new class of thermoelectric materials. Science, 272(5266), 1325–1328.
Sarhadi, A., Bjørk, R., Pryds, N. (2015). Optimization of the mechanical and
electrical performance of a thermoelectric module. J. Electron. Mater., 44(11),
4465–4472.
Shai, X., Deng, S., Shen, L., Meng, D., Li, D., Zhang, Y., Jiang, X. (2015).
Preparation, thermal stability, and electrical transport properties of In/Sn
codoped b-Zn4Sb3 single crystal. Phys. Status Solidi B, 252(4), 1–5.
Shang, H., Liang, Z., Xu, C., Mao, J., Gu, H., Ding, F., Ren, Z. (2020a). N-type
Mg3Sb2-xBix alloys as promising thermoelectric materials. Research, 1219461.
Shang, H., Liang, Z., Xu, C., Song, S., Huang, D., Gu, H., Mao, J., Ren, Z., Ding, F.
(2020b). N-type Mg3Sb2-xBix with improved thermal stability for thermoelectric
power generation. Acta Mater., 201, 572–579.
Shi, X., Sun, C., Bu, Z., Zhang, X., Wu, Y., Lin, S., Li, W., Faghaninia, A., Jain, A.,
Pei, Y. (2019a). Revelation of inherently high mobility enables Mg
3Sb2 as a
sustainable alternative to n-Bi2Te3 thermoelectrics. Adv. Sci., 6(16), 1802286.
Shi, X., Sun, C., Zhang, X., Chen, Z., Lin, S., Li, W., Pei, Y. (2019b). Efficient
Sc-doped Mg3.05–xScxSbBi thermoelectrics near room temperature. Chem. Mater.,
31(21), 8987–8994.
48 Thermoelectric Micro/Nano Generators 2
Shi, X., Zhao, T., Zhang, X., Sun, C., Chen, Z., Lin, S., Li, W., Gu, H., Pei, Y.
(2019c). Extraordinary n-type Mg3SbBi thermoelectrics enabled by yttrium
doping. Adv. Mater., 31(36), 1903387.
Shi, D., Geng, Z., Shi, L., Li, Y., Lam, K.H. (2020). Thermal stability study of
Cu1.97Se superionic thermoelectric material. J. Mater. Chem. C, 8, 10221–10228.
Shittu, S., Li, G., Zhao, X., Ma, X. (2020). Review of thermoelectric geometry and
structure optimization for performance enhancement. Appl. Energy, 268(150),
115075.
Shu, R., Zhou, Y., Wang, Q., Han, Z., Zhu, Y., Liu, Y., Chen, Y., Gu, M., Xu, W.,
Wang, Y. et al. (2019). Mg3+δSbxBi2−x family: A promising substitute for the
state-of-the-srt n-type thermoelectric materials near room temperature. Adv.
Funct. Mater., 29(4), 1807235.
Shuai, J., Mao, J., Song, S., Zhang, Q., Chen, G., Ren, Z. (2017a). Recent progress
and future challenges on thermoelectric Zintl materials. Mater. Today Phys., 1,
74–95.
Shuai, J., Mao, J., Song, S., Zhu, Q., Sun, J., Wang, Y., He, R., Zhou, J., Chen, G.,
Singh, D.J. et al. (2017b). Tuning the carrier scattering mechanism to effectively
improve the thermoelectric properties. Energy Environ. Sci., 10(3), 799–807.
Shuai, J., Ge, B., Mao, J., Song, S., Wang, Y., Ren, Z. (2018). Significant role of
Mg stoichiometry in designing high thermoelectric performance for
Mg3(Sb,Bi)2-based n-type Zintls. J. Am. Chem. Soc., 140(5), 1910–1915.
Shuai, J., Tan, X.J., Guo, Q., Xu, J.T., Gellé, A., Gautier, R., Halet, J.F., Failamani, F.,
Jiang, J., Mori, T. (2019). Enhanced thermoelectric performance through crystal
field engineering in transition metal-doped GeTe. Mater. Today Phys., 9,
100094.
Song, L., Blichfeld, A.B., Zhang, J., Kasai, H., Iversena, B.B. (2018). Enhanced
thermoelectric performance and high-temperature thermal stability of
p-type Ag-doped β-Zn4Sb3. J. Mater. Chem. A, 6, 4079–4087.
Song, S., Mao, J., Bordelon, M., He, R., Wang, Y., Shuai, J., Sun, J., Lei, X., Ren, Z.,
Chen, S. et al. (2019). Joint effect of magnesium and yttrium on enhancing
thermoelectric properties of n-type Zintl Mg3+δY0.02Sb1.5Bi0.5. Mater. Today
Phys., 8, 25–33.
Sun, W., Sui, R., Yuan, G., Zheng, H., Zeng, Z., Xie, P., Yuan, L., Ren, Z., Cai, F.,
Zhang, Q. (2021). Thermoelectric module design to improve lifetime and output
power density. Mater. Today Phys., 18, 100391.
Reliability and Durability of Thermoelectric Materials and Devices 49
Tamaki, H., Sato, H.K., Kanno, T. (2016). Isotropic conduction network and defect
chemistry in Mg3+δSb2-based layered Zintl compounds with high thermoelectric
performance. Adv. Mater., 28(46), 10182–10187.
Tang, D., Zhu, W., Wei, P., Zhou, H., Liu, Z., Yu, J., Zhao, W. (2015). Preparation
and properties of Zn4Sb2.94In0.06/ZnO composite thermoelectric materials.
J. Electron. Mater., 44, 1902–1908.
Tewolde, M., Fu, G., Hwang, D.J., Zuo, L., Sampath, S., Longtin, J.P. (2015).
Thermoelectric device fabrication using thermal spray and laser micromachining.
J. Therm. Spray Technol., 25(3), 431–440.
Wang, S.Y., She, X.Y., Zheng, G., Fu, F., Li, H., Tang, X.F. (2012). Enhanced
thermoelectric performance and thermal stability in b-Zn4Sb3 by slight
Pb-doping. J. Electron. Mater., 41, 1091–1099.
Wang, X., Veremchuk, I., Burkhardt, U., Bobnar, M., Bottner, H., Kuo, C.-Y.,
Chen, C.-T., Chang, C.-F., Zhao, J.-T., Grin, Y. (2018). Thermoelectric stability
of Eu- and Na-substituted PbTe. J. Mater. Chem. C., 6, 9482–9493
Weitzman, L.H. (1967). Etching bismuth telluride. Patent, US3338765A.
Wood, M., Kuo, J.J., Imasato, K., Snyder, G.J. (2019). Improvement of
low-temperature zT in a Mg3Sb2 -Mg3Bi2 solid solution via Mg-vapor annealing.
Adv. Mater., 31(35), 1902337.
Wood, M., Imasato, K., Anand, S., Yang, J., Snyder, G.J. (2020). The importance of
the Mg–Mg interaction in Mg3Sb2-Mg3Bi2 shown through cation site alloying.
J. Mater. Chem. A, 8(4), 2033–2038.
Wu, Y., Ming, T., Li, X., Pan, T., Peng, K., Luo, X. (2014). Numerical simulations
on the temperature gradient and thermal stress of a thermoelectric power
generator. Energy Convers. Manag., 88, 915–927.
Xia, X., Huang, X., Li, X., Gu, M., Qiu, P., Liao, J., Tang, Y., Bai, S., Chen, L.
(2014). Preparation and structural evolution of Mo/SiOx protective coating on
CoSb3-based filled skutterudite thermoelectric material. J. Alloys Compd., 604,
94–99.
Xing, T., Song, Q., Qiu, P., Zhang, Q., Gu, M., Xia, X., Liao, J., Shi, X., Chen, L.
(2021). High efficiency GeTe-based materials and modules for thermoelectric
power generation. Energy Environ. Sci., 14(2), 995–1003.
Xu, C., Liang, Z., Shang, H., Wang, D., Wang, H., Ding, F., Mao, J., Ren, Z. (2021).
Scalable synthesis of n-type Mg3Sb2-xBix for thermoelectric applications. Mater.
Today Phys., 17, 100336.
50 Thermoelectric Micro/Nano Generators 2
Xue, L., Zhang, Z., Shen, W., Ma, H., Zhang, Y., Fang, C., Jia, X. (2019).
Thermoelectric performance of Cu2Se bulk materials by hightemperature and
high-pressure synthesis. J. Materiomics, 5, 103–110.
Yang, C.L., Lai, H.J., Hwang, J.D., Chuang, T.H. (2013). Diffusion soldering of
Bi0.5Sb1.5Te3 thermoelectric material with Cu electrode. J. Mater. Eng., 22(7),
2029–2037.
Yang, J., Liu, G., Shi, Z., Lin, J., Ma, X., Xu, Z., Qiao, G. (2017a). An insight into
β-Zn4Sb3 from its crystal structure, thermoelectric performance, thermal stability
and graded material. Mater. Today Energy, 3, 72–83.
Yang, J., Zhang, X., Ge, B., Yan, J., Liu, G., Shi, Z., Qiao, G. (2017b). Effect of Zn
migration on the thermoelectric properties of Zn4Sb3 material. Ceram. Int., 43,
15275–15280.
Yin, H., Christensen, M., Pedersen, B.L., Nishibori, E., Aoyagi, S., Iversen, B.B.
(2010). Thermal stability of thermoelectric Zn4Sb3. J. Electron. Mater., 39,
1957–1959.
Yin, L., Chen, C., Zhang, F., Li, X., Bai, F., Zhang, Z., Wang, X., Mao, J., Cao, F.,
Chen, X. et al. (2020). Reliable N-type Mg3.2Sb1.5Bi0.49Te0.01/304 stainless steel
junction for thermoelectric applications. Acta Mater., 198, 25–34.
Ying, P., He, R., Mao, J., Zhang, Q., Reith, H., Sui, J., Ren, Z., Nielsch, K.,
Schierning, G. (2021). Towards tellurium-free thermoelectric modules for power
generation from low-grade heat. Nat. Commun., 12(1), 1121.
Zhang, J.F., Yang, K., Deng, C.M., Deng, C.G., Liu, M., Dai, M.J., Zhou, K.S.
(2014). Method for preparing thick nickel coating on surface of semiconductor
material. Patent, CN104357784A.
Zhang, Q.H., Huang, X.Y., Bai, S.Q., Shi, X., Uher, C., Chen, L.D. (2016).
Thermoelectric devices for power generation: Recent progress and future
challenges. Adv. Eng. Mater., 18(2), 194–213.
Zhang, J., Song, L., Pedersen, S.H., Yin, H., Hung, L.T., Iversen, B.B. (2017).
Discovery of high-performance low-cost n-type Mg3Sb2-based thermoelectric
materials with multi-valley conduction bands. Nat. Commun., 8, 13901.
Zhang, J., Song, L., Iversen, B.B. (2019). Insights into the design of thermoelectric
Mg3Sb2 and its analogs by combining theory and experiment. Npj Comput.
Mater., 5(1), 76.
Zhang, X., Bu, Z., Lin, S., Chen, Z., Li, W., Pei, Y. (2020a). GeTe thermoelectrics.
Joule, 4(5), 986–1003.
Reliability and Durability of Thermoelectric Materials and Devices 51
Zhang, Z., Zhao, K., Wei, T.-R., Qiu, P., Chen, L., Shi, X. (2020b). Cu2Se-based
liquid-like thermoelectric materials: Looking back and stepping forward. Energy
Environ. Sci., 13, 3307–3329.
Zhao, D. and Tan, G. (2014). A review of thermoelectric cooling: Materials,
modeling and applications. Appl. Therm. Eng., 66(1–2), 15–24.
Zhao, D., Geng, H., Chen, L. (2012). Microstructure contact studies for skutterudite
thermoelectric devices. Int. J. Appl. Ceram. Technol., 9(4), 733–741.
Zheng, Z., Su, X., Deng, R., Stoumpos, C., Xie, H., Liu, W., Yan, Y., Hao, S.,
Uher, C., Wolverton, C. et al. (2018). Rhombohedral to cubic conversion of
GeTe via MnTe alloying leads to ultralow thermal conductivity, electronic band
convergence, and high thermoelectric performance. J. Am. Chem. Soc., 140(7),
2673–2686.
Zhu, Q., Song, S., Zhu, H., Ren, Z. (2019). Realizing high conversion efficiency of
Mg3Sb2-based thermoelectric materials. J. Power Sources, 414, 393–400.
Zou, T., Xie, W., Feng, J., Qin, X., Weidenkaff, A. (2015). Recent developments in
β-Zn4Sb3 based thermoelectric compounds. J. Nanomater., 642909.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Owing to high intrinsic figure-of-merit implemented by multi-band valleytronics, GeTe-based thermoelectric materials are promising for medium-temperature applications. Transition metals are widely used as dopants for developing high-performance GeTe thermoelectric materials. Herein, relevant work is critically reviewed to establish a correlation among transition metal doping, electronic quality factor, and figure-of-merit of GeTe. From first-principle calculations, it is found that Ta, as an undiscovered dopant in GeTe, can effectively converge energy offset between light and heavy conduction band extrema to enhance effective mass at high temperature. Such manipulation is verified by the increased Seebeck coefficient of synthesized Ge1−x−yTaxSbyTe samples from 160 to 180 µV K⁻¹ at 775 K upon doping Ta, then to 220 µV K⁻¹ with further alloying Sb. Characterization using electron microscopy also reveals the unique herringbone structure associated with multi-scale lattice defects induced by Ta doping, which greatly hinder phonon propagation to decrease thermal conductivity. As a result, a figure-of-merit of ≈2.0 is attained in the Ge0.88Ta0.02Sb0.10Te sample, reflecting a maximum heat-to-electricity efficiency up to 17.7% under a temperature gradient of 400 K. The rationalized beneficial effects stemming from Ta doping is an important observation that will stimulate new exploration toward high-performance GeTe-based thermoelectric materials.
Article
Full-text available
Low-grade heat harvesting by thermoelectric technology can contribute to efficient energy utilization. However, the large-scale application of thermoelectric devices is partially hindered by their unsatisfactory performance and relatively high cost. Here we prepared a cost-effective n-type Mg3Bi2-based compound with a peak zT of 1.24 at 573 K. A unicouple comprised of n-type Mg3.2Bi1.29Sb0.7Te0.01 and p-type Bi0.2Sb1.8Te3 was constructed and its energy conversion efficiency was evaluated. The unicouple exhibited a record-high efficiency of (9.0±0.5)% or (8.3±0.4)% [considering the radiation heat loss] under a temperature difference of 265 K at the hot-side temperature of 573 K. Our results demonstrate the great potential of Mg3Bi2-based materials for low-grade heat recovery.
Article
Full-text available
Thermoelectric technology converts heat into electricity directly and is a promising source of clean electricity. Commercial thermoelectric modules have relied on Bi2Te3-based compounds because of their unparalleled thermoelectric properties at temperatures associated with low-grade heat (<550 K). However, the scarcity of elemental Te greatly limits the applicability of such modules. Here we report the performance of thermoelectric modules assembled from Bi2Te3-substitute compounds, including p-type MgAgSb and n-type Mg3(Sb,Bi)2, by using a simple, versatile, and thus scalable processing routine. For a temperature difference of ~250 K, whereas a single-stage module displayed a conversion efficiency of ~6.5%, a module using segmented n-type legs displayed a record efficiency of ~7.0% that is comparable to the state-of-the-art Bi2Te3-based thermoelectric modules. Our work demonstrates the feasibility and scalability of high-performance thermoelectric modules based on sustainable elements for recovering low-grade heat.
Article
Full-text available
N-type Mg3Sb2-x Bi x alloys have been extensively studied in recent years due to their significantly enhanced thermoelectric figure of merit (zT), thus promoting them as potential candidates for waste heat recovery and cooling applications. In this review, the effects resulting from alloying Mg3Bi2 with Mg3Sb2, including narrowed bandgap, decreased effective mass, and increased carrier mobility, are summarized. Subsequently, defect-controlled electrical properties in n-type Mg3Sb2-x Bi x are revealed. On one hand, manipulation of intrinsic and extrinsic defects can achieve optimal carrier concentration. On the other hand, Mg vacancies dominate carrier-scattering mechanisms (ionized impurity scattering and grain boundary scattering). Both aspects are discussed for Mg3Sb2-x Bi x thermoelectric materials. Finally, we review the present status of, and future outlook for, these materials in power generation and cooling applications.
Article
Full-text available
Cu 2 Se is a promising “phonon liquid–electron crystal” thermoelectric material with excellent thermoelectric performance. In this work, Cd-doped Cu 2– x SeCd x ( x = 0, 0.0075, 0.01, and 0.02) samples were prepared using NaCl flux method. The solubility of Cd in Cu 2 Se at room temperature was less than 6%, and a second phase of CdSe was found in the samples with large initial Cd content ( x = 0.01 and 0.02). Field-emission scanning electron microscopic image showed that the arranged lamellae formed a large-scale layered structure with an average thickness of approximately 100 nm. Transmission electron microscopy demonstrated that doping of Cd atoms did not destroy the crystal integrity of Cu 2 Se. A small amount of Cd in Cu 2 Se could reduce the electrical and thermal conductivities of the material, thus significantly enhancing its thermoelectric performance. With the increase in Cd content in the sample, the carrier concentration decreased and the mobility increased gradually. Thermogravimetric differential thermal analysis showed that no weight loss occurred below the melting point. Excessive Cd doping led to the emergence of the second phase of CdSe in the sample, thus significantly increasing the thermal conductivity of the material. A maximum ZT value of 1.67 at 700 K was obtained in the Cu 1.9925 SeCd 0.0075 sample.
Article
Full-text available
The recently discovered n-type Mg3Sb2-xBix alloys with high thermoelectric performance hold great potential for applications in waste heat recovery due to their high average zT in the temperature range between 300 and 773 K. However, systematic studies of their thermal stability remain lacking and are significant for these thermoelectric materials since the possible degradation of their thermoelectric performance could greatly limit their practical applications. Here we studied the thermal stability of the Mg3Sb2-xBix alloys via in situ measurements of their thermoelectric properties at different temperatures, along with microstructural and composition characterizations. Our results show that Mg3Sb2-xBix alloys are unstable when the temperature is above 673 K due to significant Mg loss and changed microstructures. By coating Mg3Sb2-xBix alloys with boron nitride, the Mg loss can be effectively suppressed, thus greatly improving their thermal stability. Additionally, energy conversion efficiency measurements validated the high thermoelectric performance of the Mg3Sb2-xBix alloys and further confirmed the improved thermal stability of the boron-nitride-coated sample. Therefore, our study provides an effective strategy for improving the thermal stability of Mg3Sb2-xBix, thus promoting it as a promising candidate for thermoelectric power generation.
Article
In a traditional thermoelectric π-type module, there is severe thermal stress at the hot side due to the thermal expansion of the hot-side ceramic plate and the mismatch between the coefficient of thermal expansion values among the different components, resulting in the likely development of permeant strain in the module, which leads to short lifetime. In this study, a new module design is proposed to reduce the thermal stress and enhance the lifespan of the module by 68%. Simulation using COMSOL 5.4 Multiphysics software confirmed the advantage. For a given module size, the new design has also a positive effect on the output performance of the thermoelectric module. Such an improvement may be implemented in future module manufacturing for improved performance.
Article
GeTe-based materials have a great potential to be used in thermoelectric generators for waste heat recovery due to their excellent thermoelectric performance, but their module research is greatly lagging behind the material research. In this work, we successfully fabricate a GeTe-based thermoelectric module and report a high energy conversion efficiency of 7.8% under a temperature gradient of 500 K. The Sb-, Bi-, and Se-included GeTe-based material, with the chemical composition of Ge0.92Sb0.04Bi0.04Te0.95Se0.05 and a peak ZT of 2.0 at 700 K, is used to make the p-type legs in the module. By using a high-throughput strategy, Mo is screened from 12 pure metals as an effective diffusion barrier material between GeTe-based material and electrode. Based on the optimal geometry predicted by the three-dimensional numerical analysis, one eight-couple Ge0.92Sb0.04Bi0.04Te0.95Se0.05/Yb0.3Co4Sb12 TE module is fabricated and evaluated, which shows a comparable energy conversion efficiency with those of skutterudites- and half-Heusler-based modules in the similar temperature range. This study opens the door for the development of GeTe-based modules.
Article
The highly efficient n-type Mg3Sb2-xBix thermoelectric materials hold great promise for application in power generation as well as refrigeration. Currently, n-type Mg3Sb2-xBix compounds with high zTs can be easily reproduced on a laboratory scale with ∼10 g per batch. However, scaling up the synthesis of Mg3Sb2-xBix with uniform high thermoelectric performance, which is critical for promoting this compound for practical applications, has yet to be achieved. Here we report a scalable preparation method based on Simoloyer ball-milling, which allows us to obtain over 1 kg Mg3.1Sb1.5Bi0.49Te0.01 powder in a single batch. Subsequently, samples with different diameters (ranging from a half inch to two inches) were successfully prepared and their thermoelectric performance was found to be comparable. In addition, a two-inch sample was sectioned into several parts, and the thermoelectric properties of the separate parts are also similar, indicating the high uniformity of the prepared large-scale sample. Importantly, the single-leg Mg3.1Sb1.5Bi0.49Te0.01 attains a high energy conversion efficiency of ∼12.9% under a temperature difference of ∼480 K at the hot-side temperature of 773 K. This study represents a step toward the practical application of Mg3Sb2-xBix for thermoelectric power generation.
Article
Approaching practical thermoelectric devices require high-performance and machinable thermoelectric materials. However, the currently available materials are usually brittle. In this work, Nd-doped Mg3Sb2-based compounds exhibit not only excellent thermoelectric performance but also superior machinability. Mg3.2Nd0.03Sb1.5Bi0.5 exhibits a high power factor of 20.6 μW cm-1 K-2 at 725 K and a peak zT of 1.8, which mainly originates from the increased n of ~8 × 1019 cm-3 by Nd/Mg substitutional defects. Defect calculations predict that other rare earth elements (Sm, Gd, Tb, Dy and Ho) have the same effect as Nd on Mg3Sb2 and the predicted highest achievable electron concentrations at 700 K are ~1020 cm-3 . The measured hardness, Young's modulus and fracture toughness of Mg3.2Nd0.03Sb1.5Bi0.5 are 1.1 GPa, 49.8 GPa and 1.4 MPa m1/2 , respectively. In addition, the sample can be easily machined into the dog-bone shape with external thread at both ends, indicating the excellent machinability of Mg3Sb2-based materials. This work suggests a bright future of Mg3Sb2-based thermoelectric materials for practical applications and device fabrication.