Download PDF

Rare Metals

, Volume 37, Issue 4, pp 259–273 | Cite as

BiCuSeO as state-of-the-art thermoelectric materials for energy conversion: from thin films to bulks

  • Rui Liu
  • Xing Tan
  • Yao-Chun Liu
  • Guang-Kun Ren
  • Jin-Le Lan
  • Zhi-Fang Zhou
  • Ce-Wen Nan
  • Yuan-Hua Lin
Article
First Online:
Received:
Revised:
Accepted:
  • 52 Downloads

Abstract

BiCuSeO-based thermoelectric material has attracted great attention as state-of-the-art thermoelectric materials since it was first reported in 2010. In this review, we update the studies on the BiCuSeO thin films first. Then, we focus on the most recent progress of multiple approaches that enhance the thermoelectric performance including advanced synthesized technologies, notable mechanisms for higher power factor (optimizing carrier concentration, carrier mobility, Seebeck coefficient) and doping effects predicted by calculation. And finally, aiming at further enhancing the performance of these materials and ultimately commercial application, we give a brief discussion on the urgent issues to which should be paid close attention.

Keywords

BiCuSeO State-of-the-art thermoelectric materials Thin films Bulks 
Download article PDF

1 Introduction

The contradiction between the increasing demand for energy and decreasing supplies of fossil fuels needs to be solved in the worldwide range. Therefore, a wide range of alternative energy technologies such as solar and wind energy have been developed. In particular, thermoelectric (TE) materials, a solution for direct conversion from the low-quality heat energy to versatile electricity, have drown vast attention [1, 2]. The performance of TE material is determined by the dimensionless figure of merit (ZT), which is derived from the combinatorial parameters of Seebeck coefficient (S), electrical conductivity (σ), thermal conductivity (κ) and absolute temperature (T): ZT = σS2T/κ. σS2 is defined as the power factor (PF) of the TE material. And σ = nμe, where n is carrier concentration, μ is carrier mobility, and e is carrier charge. To maximize ZT value, we need a large Seebeck coefficient and electrical conductivity with low thermal conductivity. However, these parameters have a complex relationship with each other, which makes it hard to effectively improve thermoelectric performance.

TE devices, composed of solid-state TE materials, can generate power by Seebeck effect or be used as coolers using the Peltier effect. For TE generator, the theoretical efficiency (η) is calculated by:
$$\eta = \left( {\frac{{T_{\text{h}} - T_{\text{c}} }}{{T_{\text{h}} }}} \right)\frac{{\sqrt {1 + {\text{ZT}}_{\text{ave}} } - 1}}{{\sqrt {1 + {\text{ZT}}_{\text{ave}} } + \frac{{T_{\text{c}} }}{{T_{\text{h}} }}}}$$
(1)
where \(\frac{{T_{\text{h}} - T_{\text{c}} }}{{T_{\text{h}} }}\) is the Carnot efficiency between the hot and the cold, ZTave is the average ZT between Tc and Th. To compete with other renewable energy technologies, a ZTave value of more than 3 is needed [3, 4]. Even though the reliability and scalability of thermoelectric modules cannot be neglected [5], the conversion efficiency of TE devices is still limited so far.
To achieve higher efficiency, a wide range of state-of-the-art thermoelectric materials have been developed in the past few years. Many effective strategies were realized successfully in V2VI3 [6, 7, 8] semiconductors, lead chalcogenides [9, 10, 11], skutterudites [12], oxides [13, 14], etc. Nevertheless, most thermoelectric materials contain heavy or rare elements, which may hinder further development of commercial application. Consequently, the TE material of non-toxic, earth-abundant elements is expected to play an important role for energy conversion. BiCuSeO is one of these promising TE materials. Since it was firstly reported in 2010 [15], BiCuSeO has drawn increasing attention (Fig. 1). BiCuSeO belongs to the space group of P4/nmm. Comprised of alternative layers of insulating [Bi2O2]2+ and conducting [Cu2Se2]2− along the c axis, it has an intrinsically low thermal conductivity [15, 16]. And the origin of low phonon conductivity has been explored by further investigation with calculation [17, 18, 19, 20, 21]. Hence, most experimental studies focus on the approaches for higher power factor.
Fig. 1

Number of published papers based on BiCuSeO TE materials since 2010

The reviews by Zhao et al. [22] and Zhang et al. [23] summarized some approaches for enhanced TE performance of bulk BiCuSeO. However, great progress has been achieved in the BiCuSeO thin films so far. Here in this review, not only do we mainly focus on the most recent progress on multiple ways for higher power factor of BiCuSeO bulks including optimizing carrier concentration, carrier mobility, Seebeck coefficient, but also update the development of BiCuSeO TE films. Besides, possible issues and perspectives for the future research have been briefly discussed as well.

2 BiCuSeO TE thin films

It is important to explore TE thin films. For one thing, thin films can be scaled for various capacities. In addition, the single-crystal characteristic of thin film can offer tremendous scope for deep investigations on this material. Moreover, other effects such as interface engineering, quantum confinement, etc., are prominent in thin films, which can act as multiple approaches for ZT enhancement. [24, 25]. Bismuth antimony telluride is believed to be one of the best thermoelectric materials at room temperature. The high ZT values of 2.4 and 1.5 at room temperature of Bi2Te3/Sb2Te3 superlattice and nanocrystalline Bi–Sb–Te attract great attention [26]. Various studies have investigated the synthetic techniques and thermoelectric properties of Bi–Sb–Te films, such as molecular beam epitaxy [27], metal organic chemical vapor deposition (MOCVD) [28], solution synthesis techniques [29], etc. Similarly, BiCuSeO has the layers of insulating [Bi2O2]2+ and conducting [Cu2Se2]2− along the c axis, which can be considered as superlattice structure. However, the studies on BiCuSeO films are not as many as Bi–Sb–Te films. Compared with Bi–Sb–Te films, BiCuSeO shows good thermoelectric performance at higher temperature. Besides, consisting of earth-abundant elements rather than Te, it is beneficial for further commercial application. The intensive studies are thus urgently needed. c axis oriented Bi1−xPb x CuSeO (0 ≤ x ≤ 0.8) TE films were firstly reported by Wu et al. [30]. Grown on (001) SrTiO3 substrates by pulsed laser deposition (PLD), the films exhibit single-crystal-like structure (Fig. 2a, b). The Seebeck coefficient and electrical conductivity can be tuned by Pb doping (Fig. 2c, d). The highest PF of 1.2 mW·m−1·K−2 was achieved at about 673 K for the Bi0.96Pb0.06CuSeO thin film, which is higher than that of any other polycrystalline bulks.
Fig. 2

HRTEM cross-sectional images of a BiCuSeO film on (001) SrTiO3, b film; temperature dependence of ab-plane, c electrical resistivity (ρab), d Seebeck coefficient (Sab).

Reproduced with permission of Ref. [30]. Copyright (2017): Royal Society of Chemistry

Similarly, BiCuSeO thin films have been grown on the commercial silicon wafers directly without any pretreatment [31]. The high electrical conductivity and moderate Seebeck coefficient (compared with bulks), which can be seen in Fig. 3a, make it possible for novel TE thin-film devices. Compared with polycrystalline bulks, the electrical resistivity is much lower (Fig. 3b).
Fig. 3

a Cross-sectional HRTEM image of interface between BiCuSeO and Si, b temperature dependence of ab-plane electrical resistivity (ρab) and Seebeck coefficient (Sab).

Reproduced with permission of Ref. [31]. Copyright (2017): Wiley Company

The researches on BiCuSeO TE films just get started in recent years. We can see the amazing physical phenomena through the films. The development on TE films requires more attention and resources.

3 BiCuSeO polycrystalline bulks

3.1 Advances in synthesis and processing techniques for BiCuSeO

Even though BiCuSeO was found to be promising TE materials only a few years ago, it has been proven that BiCuSeO could be prepared by high-pressure and high-temperature (HPHT) synthesis, sol–gel, solid-state reaction (SSR), mechanical alloying (MA) and self-propagating high-temperature synthesis (SHS).

3.1.1 High-pressure and high-temperature process

Compared with other methods, HPHT has the advantages of ability to tune rapidly, shortening the synthesis time, retaining the high-pressure properties, etc. It has been reported that PbTe [32], AgSbTe2 [33] and skutterudites [34] can be synthesized successfully. Study of HPHT for BiCuSeO was reported by Zhu et al. [35]. Bi, Cu, Bi2O3, Se mixed powders were initially pressed into cylindrical shape, after which, the process was carried out in the apparatus [34], as shown in Fig. 4. Three pairs of hydraulic rams actuate six tungsten carbide anvils and thus form a cubic cell, in which BiCuSeO can be prepared at 973 K by the uniform compressing.
Fig. 4

Schematic diagram of a cubic anvil high-pressure apparatus.

Reproduced with permission of Ref. [34]. Copyright (2015): Royal Society of Chemistry

3.1.2 Sol–gel

Sol–gel is an important process to prepare TE materials with various nanostructures. It has been proved to synthesize oxides, skutterudites and metal compositions [36, 37], in which it is possible to control the desired morphology. Bhaskar et al. [38] reported the sol–gel method to prepare BiCuSeO to avoid required high temperature and prolonged time of SSR. Bi(NO3)3·5H2O, Cu(NO3)2·5H2O and Se were used as precursors for the reaction. The chemical reaction details can be found in the mentioned reference. Compared with the SSRed samples, the reduced Fermi energy of the samples prepared by sol–gel shifts to a closer position to the valence band edge (Table 1), which leads to an increase in power factor. Even though the chemical reaction is complicated and has a lot of steps, it offers another new way to synthesize BiCuSeO, and might be beneficial to further enhance ZT of doped BiCuSeO.
Table 1

Some physical and transport properties (room temperature) of BiCuSeO prepared by SSR and sol–gel, where m0 being free electron mass

Parameters

SSR [39]

SSR [38]

SG-1 [38]

SG-2 [38]

Electrical resistivity/(mΩ·cm)

889

1069

134

27

Thermopower/(μV·K−1)

350

337

309

250

Power factor/(μW·cm−1·K−1)

0.14

0.11

0.72

2.28

Mobility/(cm2·V−1·s−1)

22.00

0.34

1.53

3.83

Carrier concentration/cm−3

1.1

2.7

28.0

83.8

Reduced Fermi energy/η

–

− 1.8

− 1.5

− 0.7

Density of states effective mass/(m*/m0)

0.60

0.87

3.42

4.36

3.1.3 Solid-state reaction

Indeed, BiCuSeO bulks were fabricated by SSR in majority of studies [15, 39, 40, 41, 42]. It requires two-step heating. The mixed powders, which were sealed in evacuated ampules, were calcined at 573 K and then 973 K for several hours. SSR is a very simple and useful method to prepare oxides. However, it is hard to control the purity of the product. As SSR is one of the most conventional methods, it is not introduced in detail here. Note that this method is time- and energy-consuming, which motivates people to find other approaches to synthesize BiCuSeO instead of SSR.

3.1.4 Mechanical alloying

In the mechanical alloying process, the initial element powders can form the compound powders by the mechanochemical effect [43]. BiCuSeO compounds were synthesized using MA process by Wu et al. [44]. Bi, Se and CuO powders (not Bi2O3, Cu, Se and Bi) were put in a dry milling form, so that the compound powders could be prepared by the high energy generated by the collision between balls.

According to XRD and DSC results of powders with different mechanical milling time (Fig. 5a–c), the reaction mechanism can be shown in Fig. 5d
$${\text{Bi}} + {\text{Se}}\mathop{\longrightarrow}\limits^{\text{MA}}{\text{BiSe}}$$
(2)
$${\text{BiSe}} + {\text{CuO}}\mathop{\longrightarrow}\limits^{\text{MA}}{\text{BiCuSeO}} .$$
(3)
Fig. 5

XRD results for mixed raw powders of BiCuSeO with different milling time: a 0–30 min, b 1–13 h; c DSC curves of powders with different milling time; d schematic drawing of reaction of BiCuSeO.

Reproduced with permission of Ref. [44]. Copyright (2016): Wiley Company

A metastable phase of BiSe was obtained first and when the energy was high enough to form the equilibrium compound, the reaction went on. Compared with the two-step SSR, MA is very competitive as a facile method to prepare BiCuSeO powders.

3.1.5 Self-propagating high-temperature synthesis

SHS has been proved to be an effective method to fabricate numerous TE materials, such as Bi2Se3, Bi2Te3, Cu2Se, etc. [45]. It is an autowave process, which is similar to the propagation of a combustion wave. After ignition, the heat generated by the combustion process at the ignition point keeps the reaction self-sustaining [4] (Fig. 6a), which is beneficial to commercial application. BiCuSeO was firstly prepared with SHS successfully by Ren et al. [46]. The reaction during SHS process occurs as follow:
$${\text{Cu}} + x{\text{Se}} \to {\text{Cu}}_{x} {\text{Se}}$$
(4)
$${\text{Bi}} + {\text{Bi}}_{2} {\text{O}}_{3} + {\text{Cu}}_{x} {\text{Se}} \to {\text{BiCuSeO}}$$
(5)
Fig. 6

a Schematic diagram self-propagation high-temperature synthesis; b DSC curves of SHS process of BiCuSeO with different heating rates; c difference of lattice thermal conductivity (κL) between SSRed and SHSed BiCuSeO caused by d TEM image of multiple nanophases generated by SHS.

Reproduced with permission of Refs. [4, 46, 48]. Copyright (2017): American Association for Advancement of Science; Copyright (2015): Royal Society of Chemistry; Copyright (2017): Royal Society of Chemistry

The required ignition temperature and heating rate for SHS process were determined by differential scanning calorimetry (DSC) [46] (Fig. 6b). It was found that the ignition temperature should be higher than 540.7 K and heating rate should be 75 K·min−1 for SHS process of BiCuSeO. It should be noted that SHS is the simplest and fasted method for preparation of materials, which needs only several minutes, but it can be only applied to certain materials, for instance, some selenides. Besides, as SHS is a non-equilibrium process, the synthesized samples by SHS may possess rich nanostructures. As an example for comparison, the lattice thermal conductivity (κL) of SSRed [47] and SHSed [48] Bi1-yPb y CuSeO samples at 323 and 623 K is shown in Fig. 6c. For SSRed samples, the experimental results agree well with the calculation data (solid line) indicating there are no additional microstructures that can scatter phonons. But the κL of SHSed samples is obviously lower than that of SSRed samples and calculated ones. As seen in Fig. 6d, Pb-doped BiCuSeO has large number of nanodots embedded in the matrix homogeneously. The abundant nanostructures including nanodots and nanophases can induce strong phonon scattering, which is beneficial to enhance ZT value.

3.2 Tuning carrier concentration

3.2.1 Acceptor doping

BiCuSeO has an impressively low thermal conductivity, but electrical conductivity is far from that of good TE performance. At the initial stage, divalent ions M2+ (Sr [15], Mg [16], Pb [42], Ca [49], Ba [39]) doping at Bi3+ site were adopted to successfully optimize carrier concentration and thus electrical conductivity. And up to now, many studies still focus on this. Note that the carrier concentration boosts from ~ 1 × 1018 of pristine BiCuSeO to ~ 1 × 1020 after M2+ doping.

The monovalent doping at Bi site is also shown as an effective method to optimize carrier concentration. Recently, Lan et al. [50, 51] found that TE performance of BiCuSeO could be enhanced by alkalis elements doping at Bi site with SHS method. Note that A2CO3 (A = Na, K) was used as sources of alkalis elements in the SHS process. As shown in Fig. 7a, the electrical conductivity of Na-doped BiCuSeO is higher than that of K-doped samples, which is caused by the higher carrier concentration of Na-doped samples. This phenomenon is possibly due to large structure crystal distortion caused by the larger difference of ionic radius between K+ (0.151 nm) and Bi3+ (0.096 nm). The behavior of lattice thermal conductivity was similar. The enhanced phonon scattering is induced by point defects of NaBi (KBi). Consequently, the increase electrical conductivity, moderate Seebeck coefficient and decreased lattice thermal conductivity (Fig. 7b, c) caused by point defects collectively achieve higher TE properties. Besides, it offers a fast and efficient way to fabricate alkalis-element-doped BiCuSeO TE materials. Compared with monovalent ions doping, the divalent ions doping shows a larger enhancement of TE performance. The higher solubility limit of divalent ions (more than 10%) may be the possible reason, which can greatly increase concentration and induce more point defects. It should be noted that the solubility limit of monovalent elements is lower than 6%.
Fig. 7

Temperature dependence of a electrical conductivity (σ), b Seebeck coefficient (S), c power factor of Bi1−xNaxCuSeO.

Reproduced with permission of Ref [50]. Copyright (2017): Elsevier

3.2.2 Vacancies at Bi and Cu sites

Apart from acceptor doping, vacancies of Bi or Cu can also give extra holes and optimize carrier concentration for p-type BiCuSeO TE materials. This can be explained in detail by the following defect equation:
$$2{\text{Bi}}_{1 - x} {\text{Cu}}_{1 - y} {\text{SeO}} = \left( {{\text{Bi}}_{2(1 - x)} {\text{O}}_{2} } \right)^{{\left( {2 \, + \, 3x} \right) + }} + \, \left( {{\text{Cu}}_{2(1 - y)} {\text{Se}}_{2} } \right)^{{\left( {2 \, + y} \right) - }} + x{\text{V}}_{\text{Bi}}^{3 - } + y{\text{V}}_{\text{Cu}}^{ - } + \, \left( {6x + \, 2y} \right)h^{ + }$$
(6)
The electrical conductivity of BiCu0.975SeO can achieve 53 S·cm−1 at 923 K [40], which is almost one order higher than that of pristine BiCuSeO. Li et al. [52] first studied the optimization of TE performance through Bi/Cu dual vacancies. As shown in Fig. 8, the introduction of Bi and Cu vacancies significantly enhanced the electrical conductivity, for instance, Bi0.975Cu0.975SeO has an enhanced conductivity of 47 S·cm−1 at 750 K.
Fig. 8

Temperature dependence of electrical conductivity (σ) of BiCuSeO with Bi or Cu vacancies.

Reproduced with permission of Ref. [52]. Copyright (2015): American Chemical Society

3.2.3 Band gap tuning

The reported band gap of BiCuSeO is 0.8 eV [53]. The decreased band gap can promote the minority carriers (electrons) to jump across the band gap with temperature increasing, and thus, carrier concentration will be largely enhanced. Equivalent doping of Te at Se site [54] and Sb at Bi site [55] can decrease band gap. Through 20% Te substitution, the band gap decreases to 0.65 eV (Fig. 9a). At the room temperature, the change in carrier concentration and electrical conductivity can be neglected. The thermal excitation appears at 550 K, which leads to greatly enhanced carrier concentration, and this effect is more obvious with narrower band gap caused by Te substitution. Similar experimental and calculated results of Sb doping at Bi site are also reported (Fig. 9b, c).
Fig. 9

a Schematic figure of band gap tuning of Te-doped BiCuSeO, where CB being conduction band, VB being valence band; calculated band structure of b BiCuSeO and c Sb substitution at Bi site.

Reproduced with permission of Refs. [54, 51]. Copyright (2013): Royal Society of Chemistry; Copyright (2017): Elsevier

3.3 Improving carrier mobility

Another limitation of further improvement of BiCuSeO TE performance is the low carrier mobility. The mobility of pristine BiCuSeO is about 20 cm2·V−1·s−1. After optimizing carrier concentration, it can be reduced to ~ 1–2 cm2·V−1·s−1 because of strong phonon coupling. It is much lower than that of any other TE materials. Therefore, multiple methods are required to further improve the low carrier mobility.

3.3.1 Increasing bond covalency

According to the measured length of bond, the bonding in (Cu2Se2)2− was found [53] to be primarily ionic. It is caused by the largely different electronegativity between Cu (~ 1.90) and Se (~ 2.55). It renders low carrier mobility because of strong lattice scattering of charge carriers. Ren et al. [48] demonstrated that introducing Te into Cu–Se layers can enhance the chemical band covalency (decrease in ionicity), which leads to improved carrier mobility. The smaller difference of electronegativity between Cu (~ 1.9) and Te (~ 2.1) attributes to increased bond covalency, thus improving the carrier mobility. As shown in Table 2, even with more point defects, the carrier mobility of the Pb-/Te-doped samples is higher than any other Bi site-doped (including Pb-doped) samples at similar carrier concentration, which indicates a weakened carrier scattering. As a result, a high PF can be maintained with Te doping. The lattice thermal conductivity is still suppressed because of the existence of dual point defects and rich nanostructure induced by SHS (mentioned Sect. 3.1).
Table 2

Carrier concentration and mobility of Bi0.96Pb0.04CuSe1−xTe x O

Samples

n/1020 cm−3

μ/(cm2·V−1·s−1)

x = 0

5.6

4.1

x = 0.025

2.1

7.8

x = 0.050

1.8

8.6

x = 0.075

1.6

9.9

x = 0.100

1.5

11.0

3.3.2 Tuning band structure

Liu et al. [56] report that La-doping (Bi1−xLa x CuSeO) can tune the band structure, which can be evidenced by optical absorption spectra and calculation. With more La substitution (Fig. 10a), the band gap becomes larger than that of pure BiCuSeO. As shown in Fig. 10d, there is band offset between the L valence band (heavy band) and secondary Σ band (light band). With larger band gap caused by La substitution, the heavy band will reduce. The energy separation between the heavy and light bands decreases. Then the secondary light band will play a more important role in heavily doped sample. And the contribution of light band leads to a decreased effective mass at Fermi level, which is beneficial to the higher carrier mobility (Fig. 10b). Note that the reduction in carrier concentration in Fig. 10c is caused by the enlargement of band gap.
Fig. 10

a Optical band gaps of Bi1−xLa x CuSeO; b carrier mobility (μ) and c carrier concentration (n) for Bi1−xLa x CuSeO; d schematic figure of band model.

Reproduced with permission of Ref. [56]. Copyright (2015): AIP Publishing LLC

3.3.3 Modulation doping

By heavy doping of Ba, Bi0.875Ba0.125CuSeO shows a good TE performance with optimized carrier concentration [39]. Pei et al. [57] introduce modulation doping in 3D bulks, which can boost carrier mobility without deteriorating carrier concentration. To obtain the sample of Bi0.875Ba0.125CuSeO with modulation doping, the Bi0.75Ba0.25CuSeO and pristine BiCuSeO are prepared separately, then they are mixed at a mole ratio of 1:1 for 10 min by ball milling. The electrons will diffuse from the doped area to the undoped area (Fig. 11a), in which carrier mobility is higher because of less ionized scattering center. Therefore, the mobility can be higher than that of uniformly doping. As shown in Fig. 11b, the carrier mobility follows an inversely proportional relationship with carrier concentration. However, the mobility of the sample with modulation doping deviates from the curve, which is almost twice than that of uniformly doped sample with similar carrier concentration. Finally, the ZT value is further enhanced from 1.1 of uniformly doping Bi0.875Ba0.125CuSeO to 1.4 of modulation doping sample.
Fig. 11

a Schematic figure of band structures, Fermi levels and carrier transport for pristine BiCuSeO, modulation doped and uniformly doped Bi0.875Ba0.125CuSeO; b carrier concentration dependence of carrier mobility at room temperature.

Reproduced with permission of Ref. [57]. Copyright (2014): American Chemical Society

3.3.4 Textured microstructure by hot-forging

Sui et al. [58] demonstrated that the textured Bi0.875Ba0.125CuSeO showed a better thermoelectric performance because of higher carrier mobility along the direction perpendicular to the pressing direction. To obtain the textured bulks, the samples were hot-pressed again in larger dies with extra space. The authors use four steps of hot-pressing to get the finally textured bulks (Fig. 12a). The thermoelectric properties of radial direction and axial direction in cylinder sample are both tested. With the grains preferentially oriented in ab plane (radial direction in cylinder sample), the carrier mobility increases twice than that of the one without texture (Fig. 12c). As shown in Fig. 12d, the Seebeck coefficient is independent of hot-forging steps and the measured direction. The power factor (Fig. 12e) is thus enhanced by the increased electrical conductivity (Fig. 12b) caused by higher carrier mobility and negligible changes of Seebeck coefficient.
Fig. 12

a Scheme of texture process and temperature dependence of b electrical conductivity (σ), c room-temperature carrier mobility (μ), d Seebeck coefficient (S) and e power factor (PF) of textured samples.

Reproduced with permission of Ref. [58]. Copyright (2013): The Royal Society of Chemistry

3.4 Enhancing Seebeck coefficient

So far, due to the largely increased carrier concentration, the Seebeck coefficient of doped BiCuSeO drops heavily. Wen et al. [59] selected Ba and Ni as dopants at Bi site and Cu site of BiCuSeO, respectively. The magnetic ions can induce degeneracy of electronic spin configurations in real space, which can give extra spin entropy, and as a result, Seebeck coefficient can be enhanced.

As shown in Fig. 13a, based on a multiple valence band model, carried-concentration dependence of Seebeck coefficient at 300 and 923 K was calculated. The experiment data of Mg-doped BiCuSeO data [16] at 300 K agree well with Pisarenko line, indicating that the calculation is reliable. When Ni substitutes for Cu, the hole carrier concentration decreased because of electron doping. The difference between experiment data and calculation data (ΔS) at two temperatures is shown in Fig. 13b. The Pisarenko line was sometimes used to hint the change in band structure. However, there is no space for further band convergence [16, 56], and the following was found to confirm the spin entropy contribution to the Seebeck coefficient: (1) the linear dependence on the density of magnetic ions (Ni) in Fig. 13b; (2) the change of ΔS between 300 and 923 K caused by released spin entropy by thermal excitation; (3) a typical signature of spin entropy: the decrease in Seebeck coefficient caused by partial removal of spin degeneracy under the induced magnetic field (Fig. 13c) [60].
Fig. 13

a Hole concentration (p) dependence of Seebeck coefficient (S) of BiCuSeO at 300 and 923 K, where lines being calculation result and dots being experimental results; b difference between experimental results and calculation results (ΔS) at 300 and 923 K; c temperature dependence of Seebeck coefficient (S) without and with magnetic field for Bi0.875Ba0.125Cu0.85Ni0.15SeO, where inset showing S(4 T)–S(0 T) as a function of temperature.

Reproduced with permission of Ref. [59]. Copyright (2017): Royal Society of Chemistry

3.5 Investigation on doping effects by calculation

3.5.1 Si as an unexpected n-type dopant

The influence of IV elements (Si, Ge, Sn, Pb) on the electronic structures of BiCuSeO material was studied from the first principles by Shen and Chen [61]. The calculated formation energy of defects showed that all the group of IV elements tend to occupy Bi site. Note that the difference of atomic radius between Pb and Bi is the smallest among these elements, Pb has a lower formation energy. As seen in Fig. 14a, BiCuSeO doped with Ge, Sn and Pb has a similar density of states (DOS). The Fermi level slightly moves into the valence band, indicating that these elements serve as p-type dopants. However, when Bi is substituted by Si, Fermi level moves into conduction band, as shown in Fig. 14b. The group of IV elements has a similar electronegativities; thus, the distinctly different doping effect of Si is unexpected. Besides, the unfolded effective band structures of Pb- and Si-doped BiCuSeO (Fig. 14c, d) indicate the large effective mass of the top of the valence band (compared with that of the bottom of the conduction band). Therefore, the Si-doped BiCuSeO has a higher carrier mobility and electrical conductivity. And it requires further exploration in experiments.
Fig. 14

Density of states (DOS) of BiCuSeO with a Ge, Sn, Pb and b Si doping; effective band structure of BiCuSeO with c Pb doping and d Si doping.

Reproduced with permission of Ref. [61]. Copyright (2017): American Chemical Society

3.5.2 Inducing resonant states by In and Tl

Band engineering has been proved as an effective way to increase Seebeck coefficient. Method of inducing resonant levels through impurity doping has been used in many TE materials system such as PbTe [62, 63, 64]. The resonant levels can enhance the Seebeck coefficient by the added density of states (DOS) or a strong energy filtering effect [65]. According to Shen et al. [66], In and Tl doping can induce resonant states at conduction band minimum (CBM) and valence band maximum (VBM), which is based on a band unfolding and density functional theory.

For In-doped BiCuSeO, at the bottom of the conduction band, a non-dispersive impurity band is induced, which leads to a sharp peak in the density in the DOS. The doping effects can be better understood by the partial density of states (PDOS) in Fig. 15a. The localized doping level comes from In 5s, O 2p and Se 4p orbitals, while the In 5s contributes the most. For Tl doped system, a large peak appears near the top of the valence band and the Fermi energy moves into the valence bands (Fig. 15b). The resonant states near the Fermi level can increase the Seebeck coefficient as just mentioned [62]. These studies can offer a guideline for tuning the electrical transport performance in BiCuSeO TE systems.
Fig. 15

Partial density of states of BiCuSeO with a In doping and b Tl doping.

Reproduced with permission of Ref. [66]. Copyright (2017): Royal Society of Chemistry

4 Summary and perspectives

In this review, we introduce the development of BiCuSeO TE thin films and the most recent progress in BiCuSeO bulks including notable mechanisms of optimizing carrier concentration, carrier mobility, Seebeck coefficient to enhance power factor and the calculation results. Apart from the mentioned work, the research on the BiCuSeO systems still has space for further exploration, which at least includes the following.

The studies on the BiCuSeO thin films need more attention and resources. A lot of physical phenomena can be shown in the films. And the thin films can offer other possibilities of investigations including heterostructure, modified layers, etc., which are almost impossible to achieve in the bulks. Besides, technologies of measuring temperature-depended in-plane thermal conductivity of thin films need to be further developed.

For BiCuSeO TE bulks, the urgent issue is to increase Seebeck coefficient and overall power factor. The effective method of band engineering can be extended to BiCuSeO systems. In addition, even though the p-type BiCuSeO TE performance can have a ZT value more than 1, the corresponding n-type BiCuSeO-based thermoelectric materials with high TE properties remains to be developed. The Cl-doped BiCuSeO is still p-type material [67]; however, trials are needed to explore n-type BiCuSeO thermoelectric materials.

Apart from pushing the TE performance of BiCuSeO to a higher level, assembling TE devices is also important. Advanced technologies of assembling different TE materials are needed to weaken the effect of elemental diffusion and contact resistance, which will pave the way for the future commercial application of energy conversion.

Notes

Acknowledgements

This work was financially supported by the National Key Research Programme of China (No. 2016YFA0201003), the National Basic Research Program of China (No. 2013CB632506), the National Natural Science Foundation of China (No. 51772016), the National Natural Science Foundation of China (Nos. 51672155, 51532003) and China Postdoctoral Science Foundation (No. 2016M601020).

References

  1. [1]
    Bell LE. Cooling, heating, generating power, and recovering waste heat with thermoelectric systems. Science. 2008;321(5895):1457.CrossRefGoogle Scholar
  2. [2]
    Snyder GJ, Toberer ES. Complex thermoelectric materials. Nat Mater. 2008;7(2):105.CrossRefGoogle Scholar
  3. [3]
    Shakouri A. Recent developments in semiconductor thermoelectric physics and materials. Annu Rev Mater Res. 2011;41(1):399.CrossRefGoogle Scholar
  4. [4]
    He J, Tritt TM. Advances in thermoelectric materials research: looking back and moving forward. Science. 2017;357(6358):eaak9997.CrossRefGoogle Scholar
  5. [5]
    Su X, Wei P, Li H, Liu W, Yan Y, Li P, Su C, Xie C, Zhao W, Zhai P, Zhang Q, Tang X, Uher C. Multi-scale microstructural thermoelectric materials: transport behavior, non-equilibrium preparation, and applications. Adv Mater. 2017;29(20):1602013.CrossRefGoogle Scholar
  6. [6]
    Hu L, Zhu T, Liu X, Zhao X. Point defect engineering of high-performance bismuth-telluride-based thermoelectric materials. Adv Funct Mater. 2014;24(33):5211.CrossRefGoogle Scholar
  7. [7]
    Hu L, Wu H, Zhu T, Fu C, He J, Ying P, Zhao X. Tuning multiscale microstructures to enhance thermoelectric performance of n-type bismuth-telluride-based solid solutions. Adv Energy Mater. 2015;5(17):1500411.CrossRefGoogle Scholar
  8. [8]
    Zhu T, Hu L, Zhao X, He J. New insights into intrinsic point defects in V2VI3 thermoelectric materials. Adv Sci. 2016;3(7):1600004.CrossRefGoogle Scholar
  9. [9]
    Pei Y, Lensch-Falk J, Toberer ES, Medlin DL, Snyder GJ. High thermoelectric performance in PbTe due to large nanoscale Ag2Te precipitates and La doping. Adv Funct Mater. 2011;21(2):241.CrossRefGoogle Scholar
  10. [10]
    Wu CF, Wang H, Yan Q, Wei TR, Li JF. Doping of thermoelectric PbSe with chemically inert secondary phase nanoparticles. J Mater Chem C. 2017;5(41):10881.CrossRefGoogle Scholar
  11. [11]
    Zhang K, Zhang Q, Wang L, Jiang W, Chen L. Enhanced thermoelectric performance of Se-doped PbTe bulk materials via nanostructuring and multi-scale hierarchical architecture. J Alloys Compd. 2017;725(S):563.Google Scholar
  12. [12]
    Shi X, Yang J, Salvador JR, Chi M, Cho JY, Wang H, Bai S, Yang J, Zhang W, Chen L. Multiple-filled skutterudites: high thermoelectric figure of merit through separately optimizing electrical and thermal transports. J Am Chem Soc. 2011;133(20):7837.CrossRefGoogle Scholar
  13. [13]
    Koumoto K, Wang Y, Zhang R, Kosuga A, Funahashi R. Oxide thermoelectric materials: a nanostructuring approach. Annu Rev Mater Res. 2010;40(1):363.CrossRefGoogle Scholar
  14. [14]
    He J, Liu Y, Funahashi R. Oxide thermoelectrics: the challenges, progress, and outlook. J Mater Res. 2011;26(15):1762.CrossRefGoogle Scholar
  15. [15]
    Zhao LD, Berardan D, Pei YL, Byl C, Pinsard-Gaudart L, Dragoe N. Bi1−xSrxCuSeO oxyselenides as promising thermoelectric materials. Appl Phys Lett. 2010;97(9):092118.CrossRefGoogle Scholar
  16. [16]
    Lan JL, Zhan B, Liu YC, Zheng B, Liu Y, Lin YH, Nan CW. Doping for higher thermoelectric properties in p-type BiCuSeO oxyselenide. Appl Phys Lett. 2013;102(12):123905.CrossRefGoogle Scholar
  17. [17]
    Jingxuan D, Ben X, Yuanhua L, Cewen N, Wei L. Lattice vibration modes of the layered material BiCuSeO and first principles study of its thermoelectric properties. New J Phys. 2015;17(8):083012.CrossRefGoogle Scholar
  18. [18]
    Saha SK. Exploring the origin of ultralow thermal conductivity in layered BiOCuSe. Phys Rev B. 2015;92(4):041202.CrossRefGoogle Scholar
  19. [19]
    Vaqueiro P, Al Orabi RAR, Luu SDN, Guelou G, Powell AV, Smith RI, Song JP, Wee D, Fornari M. The role of copper in the thermal conductivity of thermoelectric oxychalcogenides: do lone pairs matter? Phys Chem Chem Phys. 2015;17(47):31735.CrossRefGoogle Scholar
  20. [20]
    Ji HS, Togo A, Kaviany M, Tanaka I, Shim JH. Low phonon conductivity of layered BiCuOS, BiCuOSe, and BiCuOTe from first principles. Phys Rev B. 2016;94(11):115203.CrossRefGoogle Scholar
  21. [21]
    Kumar S, Schwingenschlogl U. Lattice thermal conductivity in layered BiCuSeO. Phys Chem Chem Phys. 2016;18(28):19158.CrossRefGoogle Scholar
  22. [22]
    Zhao LD, He J, Berardan D, Lin Y, Li JF, Nan CW, Dragoe N. BiCuSeO oxyselenides: new promising thermoelectric materials. Energy Environ Sci. 2014;7(9):2900.CrossRefGoogle Scholar
  23. [23]
    Zhang X, Chang C, Zhou Y, Zhao LD. BiCuSeO thermoelectrics: an update on recent progress and perspective. Materials (Basel). 2017;10(2):198.CrossRefGoogle Scholar
  24. [24]
    Venkatasubramanian R, Siivola E, Colpitts T, O’Quinn B. Thin-film thermoelectric devices with high room-temperature figures of merit. Nature. 2001;413:597.CrossRefGoogle Scholar
  25. [25]
    Ohta H, Kim S, Mune Y, Mizoguchi T, Nomura K, Ohta S, Nomura T, Nakanishi Y, Ikuhara Y, Hirano M, Hosono H, Koumoto K. Giant thermoelectric Seebeck coefficient of a two-dimensional electron gas in SrTiO3. Nat Mater. 2007;6(2):129.CrossRefGoogle Scholar
  26. [26]
    Poudel B, Hao Q, Ma Y, Lan Y, Minnich A, Yu B, Yan X, Wang D, Muto A, Vashaee D, Chen X, Liu J, Dresselhaus MS, Chen G, Ren Z. High-thermoelectric performance of nanostructured bismuth antimony telluride bulk alloys. Science. 2008;320(5876):634.CrossRefGoogle Scholar
  27. [27]
    Boyer A, Cissé E. Properties of thin film thermoelectric materials: application to sensors using the Seebeck effect. Mater Sci Eng B. 1992;13(2):103.CrossRefGoogle Scholar
  28. [28]
    Mzerd A, Aboulfarah B, Giani A, Boyer A. Elaboration and characterization of MOCVD (Bi1−xSbx)2Te3 thin films. J Mater Sci. 2006;41(5):1659.CrossRefGoogle Scholar
  29. [29]
    Scheele M, Oeschler N, Veremchuk I, Reinsberg K-G, Kreuziger A-M, Kornowski A, Kornowski A, Broekaert J, Klinke C, Weller H. ZT enhancement in solution-grown Sb(2−x)BixTe3 nanoplatelets. ACS Nano. 2010;4(7):4283.CrossRefGoogle Scholar
  30. [30]
    Wu X, Wang JL, Zhang H, Wang S, Zhai S, Li Y, Elhadj D, Fu G. Epitaxial growth and thermoelectric properties of c-axis oriented Bi1−xPbxCuSeO single crystalline thin films. CrystEngComm. 2015;17(45):8697.CrossRefGoogle Scholar
  31. [31]
    Wu X, Gao L, Roussel P, Dogheche E, Wang J, Fu G, Wang S. Growth of c-axis-oriented BiCuSeO thin films directly on Si wafers. J Am Ceram Soc. 2016;99(10):3367.CrossRefGoogle Scholar
  32. [32]
    Fan H, Su T, Li H, Li S, Hu M, Liu B, Du B, Ma H, Jia X. Enhanced low temperature thermoelectric performance and weakly temperature-dependent figure-of-merit values of PbTe–PbSe solid solutions. J Alloys Compd. 2016;658(S):885.CrossRefGoogle Scholar
  33. [33]
    Su T, Jia X, Ma H, Yu F, Tian Y, Zuo G, Zheng Y, Jiang Y, Dong D, Deng L, Qin B, Zheng S. Enhanced thermoelectric performance of AgSbTe2 synthesized by high pressure and high temperature. J Appl Phys. 2009;105(7):073713.CrossRefGoogle Scholar
  34. [34]
    Sun H, Jia X, Deng L, Lv P, Guo X, Zhang Y, Sun B, Liu B, Ma H. Effect of HPHT processing on the structure, and thermoelectric properties of Co4Sb12 co-doped with Te and Sn. J Mater Chem A. 2015;3(8):4637.CrossRefGoogle Scholar
  35. [35]
    Zhu H, Su T, Li H, Pu C, Zhou D, Zhu P, Wang X. High pressure synthesis, structure and thermoelectric properties of BiCuChO (Ch = S, Se, Te). J Eur Ceram Soc. 2017;37(4):1541.CrossRefGoogle Scholar
  36. [36]
    Fan FJ, Yu B, Wang YX, Zhu YL, Liu XJ, Yu SH, Ren Z. Colloidal synthesis of Cu2CdSnSe4 nanocrystals and hot-pressing to enhance the thermoelectric figure-of-merit. J Am Chem Soc. 2011;133(40):15910.CrossRefGoogle Scholar
  37. [37]
    Ciriminna R, Fidalgo A, Pandarus V, Béland F, Ilharco LM, Pagliaro M. The sol–gel route to advanced silica-based materials and recent applications. Chem Rev. 2013;113(8):6592.CrossRefGoogle Scholar
  38. [38]
    Bhaskar A, Lai RT, Chang KC, Liu CJ. High thermoelectric performance of BiCuSeO prepared by solid state reaction and sol–gel process. Scr Mater. 2017;134:100.CrossRefGoogle Scholar
  39. [39]
    Li J, Sui J, Pei Y, Barreteau C, Berardan D, Dragoe N, Cai W, He J, Zhao LD. A high thermoelectric figure of merit ZT > 1 in Ba heavily doped BiCuSeO oxyselenides. Energy Environ Sci. 2012;5(9):8543.CrossRefGoogle Scholar
  40. [40]
    Liu Y, Zhao LD, Liu Y, Lan J, Xu W, Li F, Zhang BP, Berardan D, Dragoe N, Lin YH, Nan CW, Li JF, Zhu H. Remarkable enhancement in thermoelectric performance of BiCuSeO by Cu deficiencies. J Am Chem Soc. 2011;133(50):20112.CrossRefGoogle Scholar
  41. [41]
    Li F, Li JF, Zhao LD, Xiang K, Liu Y, Zhang BP, Lin YH, Nan CW, Zhu H. Polycrystalline BiCuSeO oxide as a potential thermoelectric material. Energy Environ Sci. 2012;5(5):7188.CrossRefGoogle Scholar
  42. [42]
    Lan JL, Liu YC, Zhan B, Lin YH, Zhang B, Yuan X, Zhang W, Xu W, Nan CW. Enhanced thermoelectric properties of Pb-doped BiCuSeO ceramics. Adv Mater. 2013;25(36):5086.CrossRefGoogle Scholar
  43. [43]
    Suryanarayana C, Ivanov E, Boldyrev VV. The science and technology of mechanical alloying. Mater Sci Eng A. 2001;304–306(S):151.CrossRefGoogle Scholar
  44. [44]
    Wu J, Li F, Wei TR, Ge Z, Kang F, He J, Li JF. Mechanical alloying and spark plasma sintering of BiCuSeO oxyselenide: synthesis process and thermoelectric properties. J Am Ceram Soc. 2016;99(2):507.CrossRefGoogle Scholar
  45. [45]
    Su X, Fu F, Yan Y, Zheng G, Liang T, Zhang Q, Cheng X, Yang D, Chi H, Tang X, Zhang Q, Uher C. Self-propagating high-temperature synthesis for compound thermoelectrics and new criterion for combustion processing. Nat Commun. 2014;5:4908.CrossRefGoogle Scholar
  46. [46]
    Ren GK, Lan JL, Butt S, Ventura KJ, Lin YH, Nan CW. Enhanced thermoelectric properties in Pb-doped BiCuSeO oxyselenides prepared by ultrafast synthesis. RSC Adv. 2015;5(85):69878.CrossRefGoogle Scholar
  47. [47]
    Luu SDN, Vaqueiro P. Synthesis, structural characterisation and thermoelectric properties of Bi1−xPbxOCuSe. J Mater Chem A. 2013;1(39):12270.CrossRefGoogle Scholar
  48. [48]
    Ren GK, Wang SY, Zhu YC, Ventura KJ, Tan X, Xu W, Lin YH, Yang J, Nan CW. Enhancing thermoelectric performance in hierarchically structured BiCuSeO by increasing bond covalency and weakening carrier–phonon coupling. Energy Environ Sci. 2017;10(7):1590.CrossRefGoogle Scholar
  49. [49]
    Li F, Wei TR, Kang F, Li JF. Enhanced thermoelectric performance of Ca-doped BiCuSeO in a wide temperature range. J Mater Chem A. 2013;1(38):11942.CrossRefGoogle Scholar
  50. [50]
    Lan JL, Deng C, Ma W, Ren GK, Lin YH, Yang X. Ultra-fast synthesis and high thermoelectric properties of heavy sodium doped BiCuSeO. J Alloy Compd. 2017;708:955.CrossRefGoogle Scholar
  51. [51]
    Lan JL, Ma W, Deng C, Ren GK, Lin YH, Yang X. High thermoelectric performance of Bi1−xKxCuSeO prepared by combustion synthesis. J Mater Sci. 2017;52(19):11569.CrossRefGoogle Scholar
  52. [52]
    Li Z, Xiao C, Fan S, Deng Y, Zhang W, Ye B, Xie Y. Dual vacancies: an effective strategy realizing synergistic optimization of thermoelectric property in BiCuSeO. J Am Chem Soc. 2015;137(20):6587.CrossRefGoogle Scholar
  53. [53]
    Hiramatsu H, Yanagi H, Kamiya T, Ueda K, Hirano M, Hosono H. Crystal structures, optoelectronic properties, and electronic structures of layered oxychalcogenides MCuOCh (M = Bi, La; Ch = S, Se, Te): effects of electronic configurations of M3+ ions. Chem Mater. 2008;20(1):326.CrossRefGoogle Scholar
  54. [54]
    Liu Y, Lan J, Xu W, Liu Y, Pei YL, Cheng B, Liu DB, Lin YH, Zhao LD. Enhanced thermoelectric performance of a BiCuSeO system via band gap tuning. Chem Commun (Camb). 2013;49(73):8075.CrossRefGoogle Scholar
  55. [55]
    Feng B, Li G, Hou Y, Zhang C, Jiang C, Hu J, Xiang Q, Li Y, He Z, Fan X. Enhanced thermoelectric properties of Sb-doped BiCuSeO due to decreased band gap. J Alloy Compd. 2017;712:386.CrossRefGoogle Scholar
  56. [56]
    Liu Y, Ding J, Xu B, Lan J, Zheng Y, Zhan B, Zhang B, Lin Y, Nan C. Enhanced thermoelectric performance of La-doped BiCuSeO by tuning band structure. Appl Phys Lett. 2015;106(23):233903.CrossRefGoogle Scholar
  57. [57]
    Pei YL, Wu H, Wu D, Zheng F, He J. High thermoelectric performance realized in a BiCuSeO system by improving carrier mobility through 3D modulation doping. J Am Chem Soc. 2014;136(39):13902.CrossRefGoogle Scholar
  58. [58]
    Sui J, Li J, He J, Pei YL, Berardan D, Wu H, Dragoe N, Cai W, Zhao LD. Texturation boosts the thermoelectric performance of BiCuSeO oxyselenides. Energy Environ Sci. 2013;6(10):2916.CrossRefGoogle Scholar
  59. [59]
    Wen Q, Chang C, Pan L, Li X, Yang T, Guo H, Wang Z, Zhang J, Xu F, Zhang Z, Tang G. Enhanced thermoelectric performance of BiCuSeO by increasing Seebeck coefficient through magnetic ion incorporation. J Mater Chem A. 2017;5(26):13392.CrossRefGoogle Scholar
  60. [60]
    Wang Y, Rogado NS, Cava RJ, Ong NP. Spin entropy as the likely source of enhanced thermopower in NaxCo2O4. Nature. 2003;423(6938):425.CrossRefGoogle Scholar
  61. [61]
    Shen J, Chen Y. Silicon as an unexpected n-type dopant in BiCuSeO thermoelectrics. ACS Appl Mater Interfaces. 2017;9(33):27372.CrossRefGoogle Scholar
  62. [62]
    Heremans JP, Jovovic V, Toberer ES, Saramat A, Kurosaki K, Charoenphakdee A, Yamanaka S, Snyder GJ. Enhancement of thermoelectric efficiency in PbTe by distortion of the electronic density of states. Science. 2008;321(5888):554.CrossRefGoogle Scholar
  63. [63]
    Yu B, Zhang Q, Wang H, Wang X, Wang H, Wang D, Wang H, Snyder GJ, Chen G, Ren ZF. Thermoelectric property studies on thallium-doped lead telluride prepared by ball milling and hot pressing. J Appl Phys. 2010;108(1):016104.CrossRefGoogle Scholar
  64. [64]
    Zhang Q, Wang H, Liu W, Wang H, Yu B, Zhang Q, Tian Z, Ni G, Lee S, Esfarjani K, Chen G, Ren Z. Enhancement of thermoelectric figure-of-merit by resonant states of aluminium doping in lead selenide. Energy Environ Sci. 2012;5(1):5246.CrossRefGoogle Scholar
  65. [65]
    Heremans JP, Wiendlocha B, Chamoire AM. Resonant levels in bulk thermoelectric semiconductors. Energy Environ Sci. 2012;5(2):5510.CrossRefGoogle Scholar
  66. [66]
    Shen J, Yu H, Pei Y, Chen Y. Resonant doping in BiCuSeO thermoelectrics from first principles. J Mater Chem A. 2017;5(3):931.CrossRefGoogle Scholar
  67. [67]
    Zhou Z, Tan X, Ren G, Lin Y, Nan C. Thermoelectric properties of Cl-doped BiCuSeO oxyselenides. J Electron Mater. 2017;46(5):2593.CrossRefGoogle Scholar

Copyright information

© The Nonferrous Metals Society of China and Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.State Key Laboratory of New Ceramics and Fine Processing, School of Materials Science and EngineeringTsinghua UniversityBeijingChina
  2. 2.State Key Laboratory of Organic-Inorganic CompositesBeijing University of Chemical TechnologyBeijingChina

Personalised recommendations