An Update Review on N-Type Layered Oxyselenide Thermoelectric Materials

Compared with traditional thermoelectric materials, layered oxyselenide thermoelectric materials consist of nontoxic and lower-cost elements and have better chemical and thermal stability. Recently, several studies on n-type layered oxyselenide thermoelectric materials, including BiCuSeO, Bi2O2Se and Bi6Cu2Se4O6, were reported, which stimulates us to comprehensively summarize these researches. In this short review, we begin with various attempts to realize an n-type BiCuSeO system. Then, we summarize several methods to optimize the thermoelectric performance of Bi2O2Se, including carrier engineering, band engineering, microstructure design, et al. Next, we introduce a new type of layered oxyselenide Bi6Cu2Se4O6, and n-type transport properties can be obtained through halogen doping. At last, we propose some possible research directions for n-type layered oxyselenide thermoelectric materials.


Introduction
Thermoelectric (TE) materials can achieve the direct transition between heat and electricity without producing other pollutants, providing an effective solution to the energy crisis and environmental problems [1,2]. The dimensionless figure of merit ZT defines the efficiency of a thermoelectric device, which derives from three related physical quantities: electrical conductivity (σ), Seebeck coefficient (S) and thermal conductivity (κ), ZT = σS 2 T/κ , with absolute temperature T. However, the tightly-coupled relationship among these parameters makes it difficult to improve the overall ZT [2][3][4][5][6].
Intrinsic p-type semiconductor BiCuSeO possesses a layered ZrCuSiAs structure with space group P4/nmm [2,32]. The special layered crystal structure of BiCuSeO is constituted by insulative [Bi 2 O 2 ] 2+ layers and conductive [Cu 2 Se 2 ] 2− layers heaping along the c-axis by turns [33,34]. Due to the weak Van der Waals interaction between layers [9,[35][36][37], the large displacement of the Cu atom [38,39] and the heavy Bi atom [3,40,41], BiCuSeO has intrinsically low thermal conductivity [42], which is an inherent advantage as a thermoelectric material. However, most of the reported BiCuSeO are p-type semiconductor materials at present, and researches on n-type BiCuSeO are relatively few and unsuccessful. The main problems are that no effective electronic dopant was found, and BiCuSeO-based materials with stable n-type transport properties have not been obtained yet.
Different from BiCuSeO, the intrinsic Bi 2 O 2 Se exhibits n-type transport properties due to a large number of Se vacancies in the crystal structure [43,44]. However, the crystal Different from BiCuSeO, the intrinsic Bi2O2Se exhibits n-type transport properties due to a large number of Se vacancies in the crystal structure [43,44]. However, the crystal structure of Bi2O2Se is very similar to BiCuSeO in which insulative [Bi2O2] 2+ layers and conductive [Se] 2-layers stack along the c-axis alternatively [26,45,46]. Similarly, Bi2O2Se also has intrinsically low thermal conductivity due to the weak interlayer interaction [26,[47][48][49], but its intrinsic carrier concentration (~1.5 × 10 15 cm −3 ) is too low, resulting in poor electrical transport performance [50,51]. Therefore, current research is mainly focused on increasing the carrier concentration, thus improving the electrical transport performance of Bi2O2Se.
Bi6Cu2Se4O6 is a new type of layered oxyselenide thermoelectric material [28][29][30]. The crystal structure of Bi6Cu2Se4O6 can be regarded as a 1:2 ratio of BiCuSeO and Bi2O2Se heaping along the c-axis by turns [28], so there are insulative layers [Bi2O2] 2+ and conductive layers [Cu2Se2] 2-and [Se] 2-in the structure. The Bi6Cu2Se4O6 system not only maintains the low thermal conductivity of BiCuSeO but also can utilize the intrinsic electron carrier concentration of Bi2O2Se, considered to be a promising n-type oxyselenide thermoelectric material. At present, the Bi6Cu2Se4O6 system with stable n-type transport properties can be obtained by halogen doping [29], and strategies to improve its thermoelectric performance need to be further explored.
Recently, several studies on n-type layered oxyselenide thermoelectric materials were reported, which motivates us to systematically summarize the recent progress of these researches. The outline of this review is shown in Figure 1. First, several attempts to realize n-type BiCuSeO are summarized. Then, some typical approaches to optimize the thermoelectric performance of Bi2O2Se are presented. Next, a new type layered oxyselenide Bi6Cu2Se4O6 is introduced, and n-type transport properties can be obtained through halogen doping and optimized by introducing transition metal elements. At last, some prospective and outlooks were provided for future research in the end.

Doping Fe at Cu Sites
Pan et al. obtained n-type transport properties in the BiCuSeO system in the lowtemperature range (300-500 K) through doping Fe at Cu sites [52]. Figure 2a,b display the temperature-dependent electrical conductivity (σ) and Seebeck coefficient (S) for Fedoped samples. The σ of all samples increases monotonically with temperature, showing semiconductor conduction behavior. Meanwhile, at the same temperature, the σ increases with the increasing Fe content, indicating that Fe is an effective dopant for pristine BiCuSeO matrix. The peak values of S for samples BiCu 1-x Fe x SeO (x = 0 to 0.03) are larger than 150 µV K −1 and shift towards higher temperature with the increasing Fe content. The band gap can be estimated by the peak value of S through the Goldsmid-Sharp law as Formula (1) [53,54]:

Doping Fe at Cu Sites
Pan et al. obtained n-type transport properties in the BiCuSeO system in the lowtemperature range (300-500 K) through doping Fe at Cu sites [52]. Figure 2a,b display the temperature-dependent electrical conductivity (σ) and Seebeck coefficient (S) for Fedoped samples. The σ of all samples increases monotonically with temperature, showing semiconductor conduction behavior. Meanwhile, at the same temperature, the σ increases with the increasing Fe content, indicating that Fe is an effective dopant for pristine Bi-CuSeO matrix. The peak values of S for samples BiCu1-xFexSeO (x = 0 to 0.03) are larger than 150 μV K −1 and shift towards higher temperature with the increasing Fe content. The band gap can be estimated by the peak value of S through the Goldsmid-Sharp law as Formula (1) [53,54]: The curves of (a) electrical conductivity, and (b) Seebeck coefficient varying with temperature for BiCu1-xFexSeO (x = 0 to 0.04); (c) Fe content dependences of Hall coefficient for BiCu1-xFexSeO (x = 0 to 0.04) at room temperature. Data presented in (a-c) were adopted from Reference [52]. Copyright 2018, The Royal Society of Chemistry. (d) Calculated band structure of pristine BiCuSeO near the Fermi level. Copyright, the Royal Society of Chemistry. (d) Reproduced with permission from Reference [55]. Copyright 2019, Elsevier Masson SAS.
The peak Seebeck coefficient and corresponding temperature increase with the increase in Fe content, indicating the increase of band gap in those samples. Clearly, n-type transport properties can be found in the x = 0.03 sample at a low-temperature range, indicating that appropriate Fe content can achieve n-type BiCuSeO in a circumscribed temperature range. The curves of (a) electrical conductivity, and (b) Seebeck coefficient varying with temperature for BiCu 1-x Fe x SeO (x = 0 to 0.04); (c) Fe content dependences of Hall coefficient for BiCu 1-x Fe x SeO (x = 0 to 0.04) at room temperature. Data presented in (a-c) were adopted from Reference [52]. The peak Seebeck coefficient and corresponding temperature increase with the increase in Fe content, indicating the increase of band gap in those samples. Clearly, n-type transport properties can be found in the x = 0.03 sample at a low-temperature range, indicating that appropriate Fe content can achieve n-type BiCuSeO in a circumscribed temperature range.
The variation of Hall coefficient (R H ) at room temperature is presented in Figure 2c. Expectedly, BiCuSeO-based materials exhibit p-type electrical transport properties, and R H should be positive. However, the R H of BiCu 0.97 Fe 0.03 SeO exhibits a negative value, confirming the n-type behavior conjectured by the negative S. The transition of electrical transport properties can be explained by the following two formulas [56,57]: where, subscripts h and e express hole and electron carriers, respectively. In the above two formulas, n h and n e are both positive, while S h and S e are positive and negative, respectively. The decline of R H and S after doping Fe in Cu sites can be explained by the competition between the intrinsic holes and the electrons introduced by Fe substitution, which indicates that the substitution of Fe for Cu is a kind of donor doping. When the Fe content increases to 0.03, σ increases due to the increasing total carrier concentration (n t = n h + n e ); while S declines because the negative contribution of electrons partly counterbalances the positive contribution of holes to S, ultimately switching to negative when electrons become majority carriers. Figure 2d exhibits the calculated band structure of pristine BiCuSeO. The valence band maximum (VBM) and conduction band minimum (CBM) are located along the Γ-M and Γ-Z direction, respectively. Larger degeneracy of VBM than CBM leads to a less favorable compromise between large mobility and large effective mass, so the large S values are not easy to obtain n-type transport.

Doping Co at Cu Sites
Ren et al. aimed to increase the Seebeck coefficient by introducing magnetic ions into the BiCuSeO matrix because previous studies on other systems showed that extra spin entropy could be introduced by the magnetic ions, thereby contributing to increasing the Seebeck coefficient [55]. They found that the BiCuSeO system exhibits n-type electrical transport properties below room temperature when the Co content reaches 20%. Figure 3a shows the temperature-dependent Seebeck coefficient (S) for Co-doped samples. For x = 0.05 and 0.10 samples, the S is positive within the entire temperature boundaries, indicating the dominant carrier is p-type. However, when the doping content of Co reaches 15%, the S begins to appear negative, indicating that the system exhibits n-type electrical transport properties. The negative S values of x = 0.20 sample within the entire temperature region indicate stable n-type electrical transport properties in the BiCuSeO system below room temperature.
The VBM of BiCuSeO mainly comes from the hybridization between Cu 3d and Se 4p orbitals, while the CBM is derived from the Bi 6p orbital [2,58,59], as shown in Figure 3b. For BiCu 0.875 Co 0.125 SeO, the DOS close to Fermi level mainly derives from the Co 3d and Se 4p orbitals (Figure 3c), and the number of bands crossing the Fermi level is two (Figure 3d), which are both electron cylinder and hole cylinder, indicating the Co substitution introduces n-type carriers [60]. It can be seen that Co substitution can change the band structure of BiCuSeO and make the Fermi level reach to the CBM. When the Co content is large enough, electrons become majority carriers.

Doping Halogen (Br, I) at Se Sites
Zhang et al. designed a series of continuous experimental steps to obtain n-type BiCuSeO [61]. Firstly, considering that the existence of Bi/Cu vacancies is the main reason for the p-type behavior of pristine BiCuSeO [62,63], extra Bi/Cu was introduced into the matrix to fill the vacancies which may produce holes. Finally, the optimal concentrations of extra Bi and Cu are determined as x = 0.04 and 0.05, respectively. To increase electron carrier concentration, halogen elements (Br, I) were selected as donor dopants at Se sites and introduced into the Bi 1.04 Cu 1.05 SeO matrix. The S as a function of temperature for I/Br doped Bi 1.04 Cu 1.05 SeO samples is presented in Figure 4a,b. The introduction of Br/I can successfully transform Bi 1.04 Cu 1.05 SeO from p-type to n-type within a certain temperature range, and the negative S for I-doped Bi 1.04 Cu 1.05 SeO traverses a narrower temperature range than Br-doped one. In the high-temperature range, p-type behavior appears again, indicating that vacancies reproduced as the temperature rises, which may be relevant to the instability of Cu-Br and Cu-I bonds [64,65]. The VBM of BiCuSeO mainly comes from the hybridization between Cu 3d and Se 4p orbitals, while the CBM is derived from the Bi 6p orbital [2,58,59], as shown in Figure 3b. For BiCu0.875Co0.125SeO, the DOS close to Fermi level mainly derives from the Co 3d and Se 4p orbitals (Figure 3c), and the number of bands crossing the Fermi level is two ( Figure  3d), which are both electron cylinder and hole cylinder, indicating the Co substitution introduces n-type carriers [60]. It can be seen that Co substitution can change the band structure of BiCuSeO and make the Fermi level reach to the CBM. When the Co content is large enough, electrons become majority carriers.

Doping Halogen (Br, I) at Se Sites
Zhang et al. designed a series of continuous experimental steps to obtain n-type Bi-CuSeO [61]. Firstly, considering that the existence of Bi/Cu vacancies is the main reason for the p-type behavior of pristine BiCuSeO [62,63], extra Bi/Cu was introduced into the matrix to fill the vacancies which may produce holes. Finally, the optimal concentrations of extra Bi and Cu are determined as x = 0.04 and 0.05, respectively. To increase electron carrier concentration, halogen elements (Br, I) were selected as donor dopants at Se sites and introduced into the Bi1.04Cu1.05SeO matrix. The S as a function of temperature for I/Br doped Bi1.04Cu1.05SeO samples is presented in Figure 4a,b. The introduction of Br/I can successfully transform Bi1.04Cu1.05SeO from p-type to n-type within a certain temperature range, and the negative S for I-doped Bi1.04Cu1.05SeO traverses a narrower temperature range than Br-doped one. In the high-temperature range, p-type behavior appears again, In order to further explore the p-n-p-type behavior in the obtained system, a heatingcooling measurement was carried out for halogen doing BiCuSeO samples [66], as shown in Figure 4c,d. As can be seen, Bi 1.04 Cu 1.05 Se 0.99 X 0.01 O (X = Br, I) changes completely from n-type to p-type transport behavior after eight heating-cooling cycle measurements. The above results indicate that halogens are effective dopants to obtain n-type BiCuSeO but exhibit poor stability. To improve the stability of n-type transport, metallic particles were introduced into the halogen-doped Bi 1.04 Cu 1.05 SeO. The temperature dependence of S for Bi 1.04 Cu 1.05 Se 0.99 Br 0.01 O + x% Ag samples is negative within the entire temperature range, and the maximum |S| is~125 µV/K (Figure 4e). The maximum ZT~0.05 can be reached at 475 K in Bi 1.04 Cu 1.05 Se 0.99 Br 0.01 O + 15% Ag (Figure 4f).
To further understand the instability of halogen doping in BiCuSeO, the energy integrated Crystal Orbital Hamiltonian Population (ICOHP) values were calculated ( Figure 5) [66]. The more negative value of ICOHP indicates the stronger bond strength [67]. As can be seen in Figure 5, after halogen doping, the ICOHP value decreases from~1.09 eV for Cu-Se to~0.36 eV for Cu-I,~0.25 eV for Cu-Br and~0.13 eV for Cu-Cl, respectively, indicating the weaker bond strength due to halogen doping. The weakened bond strength led to the instability of halogen doping in the BiCuSeO system under the heating-cooling cycle. In order to further explore the p-n-p-type behavior in the obtained system, a heatingcooling measurement was carried out for halogen doing BiCuSeO samples [66], as shown in Figure 4c,d. As can be seen, Bi1.04Cu1.05Se0.99X0.01O (X = Br, I) changes completely from n-type to p-type transport behavior after eight heating-cooling cycle measurements. The

Introduce Bi Deficiencies
Due to a large number of Se vacancies in the crystal structure, the pristine Bi2O2Se exhibits n-type semiconductor behavior [43,44]. To improve the conductivity of Bi2O2Se, thereby optimizing its ZT, the general approach is to do donor doping at the Bi/Se sites [47,51,68]. However, Zhan et al. made an innovative attempt that introduces Bi deficiencies into Bi2O2Se by components deviating from the stoichiometric ratio [46]. In fact, the introduction of Bi deficiencies is equivalent to acceptor doping to the matrix, which runs counter to the general method. However, the increment of Seebeck coefficient and the decrease in thermal conductivity caused by the introduction of Bi deficiencies have optimized the thermoelectric properties of Bi2O2Se.
The introduction of Bi deficiencies has little effect on the electric conductivity (σ) of the Bi2O2Se system, but it can significantly change the Seebeck coefficient (S). The absolute values of S increased first and then decreased with temperature but always kept a large peak value (−445.6, −556.6, −490.0 and −568.8 μV/K at 773 K, respectively), as shown in Figure 6a. When the carrier concentration in the semiconductor is very low, the S can be evinced as the following Formula (4) [4]:

Introduce Bi Deficiencies
Due to a large number of Se vacancies in the crystal structure, the pristine Bi 2 O 2 Se exhibits n-type semiconductor behavior [43,44]. To improve the conductivity of Bi 2 O 2 Se, thereby optimizing its ZT, the general approach is to do donor doping at the Bi/Se sites [47,51,68]. However, Zhan et al. made an innovative attempt that introduces Bi deficiencies into Bi 2 O 2 Se by components deviating from the stoichiometric ratio [46]. In fact, the introduction of Bi deficiencies is equivalent to acceptor doping to the matrix, which runs counter to the general method. However, the increment of Seebeck coefficient and the decrease in thermal conductivity caused by the introduction of Bi deficiencies have optimized the thermoelectric properties of Bi 2 O 2 Se.
The introduction of Bi deficiencies has little effect on the electric conductivity (σ) of the Bi 2 O 2 Se system, but it can significantly change the Seebeck coefficient (S). The absolute values of S increased first and then decreased with temperature but always kept a large peak value (−445.6, −556.6, −490.0 and −568.8 µV/K at 773 K, respectively), as shown in Figure 6a. When the carrier concentration in the semiconductor is very low, the S can be where k B , e, h, m*, T and n mean Boltzmann constant, electron charge, Plank constant, the effective mass of carrier, the absolute temperature and the carrier concentration, respectively.
mized the thermoelectric properties of Bi2O2Se. The introduction of Bi deficiencies has little effect on the electric conductivity (σ) of the Bi2O2Se system, but it can significantly change the Seebeck coefficient (S). The absolute values of S increased first and then decreased with temperature but always kept a large peak value (−445.6, −556.6, −490.0 and −568.8 μV/K at 773 K, respectively), as shown in Figure 6a. When the carrier concentration in the semiconductor is very low, the S can be evinced as the following Formula (4) [4]: where kB, e, h, m*, T and n mean Boltzmann constant, electron charge, Plank constant, the effective mass of carrier, the absolute temperature and the carrier concentration, respectively. The S at low temperature increases proportionally with temperature for a given carrier concentration and effective mass. The intrinsic excitation at high temperature is enhanced, and the effective carrier concentration increases, so the S decreases. Furthermore, the S of Bi deficiencies samples are basically larger than pristine Bi2O2Se within the entire temperature range, which can be attributed to the influence of m*. Thanks to the significant improvement of S, the PF peak value reaches ~0.93 μW cm −1 K −2 at 773 K (Figure 6b), which is twice that of the intrinsic sample (~0.45 μW cm −1 K −2 at 773 K). The Bi deficiencies strengthen the point defect scattering and lead to a decrease of κtot (Figure 6c). Integrating the enhanced electrical properties and suppressed thermal properties, the peak ZT value reaches ~0.12 in Bi1.9O2Se at 773 K (Figure 6d). The S at low temperature increases proportionally with temperature for a given carrier concentration and effective mass. The intrinsic excitation at high temperature is enhanced, and the effective carrier concentration increases, so the S decreases. Furthermore, the S of Bi deficiencies samples are basically larger than pristine Bi 2 O 2 Se within the entire temperature range, which can be attributed to the influence of m*. Thanks to the significant improvement of S, the PF peak value reaches~0.93 µW cm −1 K −2 at 773 K (Figure 6b), which is twice that of the intrinsic sample (~0.45 µW cm −1 K −2 at 773 K). The Bi deficiencies strengthen the point defect scattering and lead to a decrease of κ tot (Figure 6c). Integrating the enhanced electrical properties and suppressed thermal properties, the peak ZT value reaches~0.12 in Bi 1.9 O 2 Se at 773 K (Figure 6d).

Doping Cl at Se Sites
In the special crystal structure of Bi 2 O 2 Se, the [Se] 2− layer is considered to be an electron-conducting pathway [43,44,46]. Therefore, effective electron donor dopants can be used to modify the conductive [Se] 2− layer to increase the carrier concentration of the Bi 2 O 2 Se system, thereby enhancing its thermoelectric performance. Tan et al. doped Cl at Se sites and achieved an extraordinary enhancement in the electrical conductivity of the Bi 2 O 2 Se system [51].
At room temperature, the σ hikes from~0.019 S cm −1 for Bi 2 O 2 Se to~101.6 S cm −1 for Bi 2 O 2 Se 0.985 Cl 0.015 , and then declines obviously as the Cl content increases (Figure 7a). One Cl − doped into the Se 2− sites can provide an extra electron, and the measured carrier concentration increased from 1.5 × 10 15 cm −3 to 1.38 × 10 20 cm −3 (x = 0.015). However, when the Cl content exceeds the solubility limit, the formation of the low-conductivity second phase Bi 12 O 15 Cl 6 [69] will reduce the effective doping amount of Cl, thereby deteriorating the σ. The small polaron hopping conduction theory was chosen to study the impact of Cl dopant on the σ. This theory could be expressed as the following Formula (5) [70]: where n, e, µ, C, k B , E a and T express the carrier concentration, carrier charge, carrier mobility, the pre-exponential terms, Boltzmann constant, activation energy and the absolute temperature, respectively. Figure 7b shows the linear relationship between ln(σT) and 1000/T, and the E a can be obtained by calculating the slope of the straight line. As shown in the inset of Figure 7b, the E a of Cl-doped samples is obviously lower than Bi 2 O 2 Se, indicating that the introduction of Cl is conducive to carrier excitation. In summary, the high σ achieved in Cl-doped Bi 2 O 2 Se is estimated to be the result of higher n coupled with lower E a .
In the special crystal structure of Bi2O2Se, the [Se] 2− layer is considered to be an electron-conducting pathway [43,44,46]. Therefore, effective electron donor dopants can be used to modify the conductive [Se] 2− layer to increase the carrier concentration of the Bi2O2Se system, thereby enhancing its thermoelectric performance. Tan et al. doped Cl at Se sites and achieved an extraordinary enhancement in the electrical conductivity of the Bi2O2Se system [51].
At room temperature, the σ hikes from ~0.019 S cm −1 for Bi2O2Se to ~101.6 S cm −1 for Bi2O2Se0.985Cl0.015, and then declines obviously as the Cl content increases (Figure 7a). One Cldoped into the Se 2-sites can provide an extra electron, and the measured carrier concentration increased from 1.5 × 10 15 cm −3 to 1.38 × 10 20 cm −3 (x = 0.015). However, when the Cl content exceeds the solubility limit, the formation of the low-conductivity second phase Bi12O15Cl6 [69] will reduce the effective doping amount of Cl, thereby deteriorating the σ. The small polaron hopping conduction theory was chosen to study the impact of Cl dopant on the σ. This theory could be expressed as the following Formula (5) [70]: where n, e, μ, C, kB, Ea and T express the carrier concentration, carrier charge, carrier mobility, the pre-exponential terms, Boltzmann constant, activation energy and the absolute temperature, respectively. Figure 7b shows the linear relationship between ln(σT) and 1000/T, and the Ea can be obtained by calculating the slope of the straight line. As shown in the inset of Figure 7b, the Ea of Cl-doped samples is obviously lower than Bi2O2Se, indicating that the introduction of Cl is conducive to carrier excitation. In summary, the high σ achieved in Cl-doped Bi2O2Se is estimated to be the result of higher n coupled with lower Ea. Benefitting from both the enhancement of the σ and the depression of the κ tot , the peak ZT value~0.23 at 823 K is achieved in Bi 2 O 2 Se 0.985 Cl 0.015 (Figure 7d), which demonstrates that Cl is an effective dopant to optimize the thermoelectric performance of Bi 2 O 2 Se.

Doping Te at Se Sites
The ZT value of pristine Bi 2 O 2 Se is primarily restricted by the low electric conductivity (~2.0 S cm −1 ) mainly caused by the low carrier concentration (~10 15 cm −3 ) [50,51]. Fundamentally, this shortcoming can be attributed to the excessively wide band gap (~1.28 eV) [71]. Bi 2 O 2 Te, an isostructure of Bi 2 O 2 Se, possesses a narrow band gap (~0.23 eV) and a moderately high room-temperature carrier concentration (~1.06 × 10 18 cm −3 ) [45]. Additionally, p-type BiCuTeO has a narrower band gap (~0.4 eV) compared with BiCuSeO (~0.8 eV) [72,73], and relevant studies have proved that Te substitution can effectively enhance the electrical conductivity of p-type BiCuSeO by narrowing the band gap [74]. Hence, isovalent Te doping at Se sites could be utilized to the n-type Bi 2 O 2 Se.
The significantly narrowed band gap is conducive for electrons to jump across the band gap and enter the valence band ( Figure 8a) so that more electrons can be excited and participate in the electrical transport. The measured optical absorption spectrum of Bi 2 O 2 Se 1-x Te x (x = 0.02, 0.03, 0.04, 0.06) samples indicates that the band gap is monotonically reduced from~1.77 eV for Bi 2 O 2 Se to~0.78 eV for Bi 2 O 2 Se 0.94 Te 0.06 with the increasing Te content (Figure 8b) [44]. Considering the apparent difference of the band gap between Bi 2 O 2 Se (~1.77 eV) and Bi 2 O 2 Te (~0.23 eV), the band gap engineering can be effectively tuned by a small amount of Te substitution.
Ultimately, the low carrier concentration of pristine Bi 2 O 2 Se (~10 15 cm −3 ) was boosted to~10 18 cm −3 , which was increased by three orders of magnitude [44]. The greatly increased carrier concentration makes the σ of all the Te-doped samples significantly larger than pristine Bi 2 O 2 Se throughout the entire test temperature range (Figure 8c).
To obtain insight into the electrical transport behavior, the small polaron hopping conduction theory mentioned in the previous work [70] was selected to analyze the electric conductivity. The strong linear correlation between ln(σT) and 1000/T is exhibited in Figure 8d, and the curve of calculated activation energy E a varying with the Te content is depicted in Figure 8e. The E a of electronic conduction declines with the increasing Te content, indicating that the Te substitution is beneficial for the intrinsic excitation of electrons, thereby contributing to the optimized σ. Figure 8f plots the temperature dependence of ZT. Due to the increment in σ caused by the narrowing band gap, the thermoelectric performance of Bi 2 O 2 Se was enhanced. Ultimately, the highest ZT reaches 0.28 at 823 K for Bi 2 O 2 Se 0.96 Te 0.04 .

Doping Ta at Bi Sites
Choosing a suitable dopant to enhance its low carrier concentration has always been an important means to optimize the thermoelectric performance of the Bi 2 O 2 Se system. A pentavalent Ta 5+ cation doping at Bi sites will provide two extra electrons for the matrix. Moreover, Ta is less electronegative than Bi, thereby easily extracting electrons. Therefore, Tan et al. chose Ta as the dopant to increase the carrier concentration, thereby enhancing the electrical transport properties of Bi 2 O 2 Se [68].
Ta doping can significantly increase the σ of Bi 2 O 2 Se, from~0.02 S cm −1 of pristine Bi 2 O 2 Se to~149.3 S cm −1 of Bi 1.90 Ta 0.10 O 2 Se at room temperature (Figure 9a). Meanwhile, the temperature-dependent σ transform from semiconductor behavior to mixed-conducting behavior, and finally, Bi 1.90 Ta 0.10 O 2 Se and Bi 1.88 Ta 0.12 O 2 Se even exhibit degenerate semiconductor behavior. The carrier concentration (n H ) and mobility (µ H ) were measured and exhibited in Figure 9b. Ta doping increases the n H by four orders of magnitude, from 10 15 cm −3 to~10 19 cm −3 . In addition to the reason that Ta replaces Bi to provide extra electrons, authors believe that the formation of the Ta 2 O 5 phase will introduce oxygen vacancies into the matrix, and each oxygen vacancy is compensated by two electrons as the following formula: Ultimately, the low carrier concentration of pristine Bi2O2Se (~10 15 cm −3 ) was boosted to ~10 18 cm −3 , which was increased by three orders of magnitude [44]. The greatly increased carrier concentration makes the σ of all the Te-doped samples significantly larger  Therefore, both Ta doping and oxygen vacancies lead to the increment of nH. Simultaneously, the stable deterioration of μH implies gradually reinforced carrier scattering, but relatively high μH (>40 cm 2 V −1 s −1 ) can be maintained, which is because performing Ta at Bi sites would not introduce lattice defects into the conductive [Se] 2− layers.
Compared with pristine Bi2O2Se, the absolute value of S decreases when the Ta content increases (Figure 9c), which is coincident with the increase in nH. The calculated weighted mobility (μm *3/2 ) is greatly increased from ~5.41 m0 3/2 cm 2 V −1 s −1 for Bi2O2Se to ~15.44 m0 3/2 cm 2 V −1 s −1 for Bi1.90Ta0.10O2Se, revealing that Ta doping in Bi2O2Se can effectively optimize the electrical transport properties. Therefore, both Ta doping and oxygen vacancies lead to the increment of n H . Simultaneously, the stable deterioration of µ H implies gradually reinforced carrier scattering, but relatively high µ H (>40 cm 2 V −1 s −1 ) can be maintained, which is because performing Ta at Bi sites would not introduce lattice defects into the conductive [Se] 2− layers.
Compared with pristine Bi 2 O 2 Se, the absolute value of S decreases when the Ta content increases (Figure 9c), which is coincident with the increase in n H . The calculated weighted mobility (µm *3/2 ) is greatly increased from~5. 41 (Figure 9d). The phonon mean-free-path (l ph ) is calculated by the following Formula (6) [75,76] and plotted in Figure 9e as a function of Ta content.
where, C v and v a represent the specific heat capacity per unit volume and average sound speed, respectively. A highly intense phonon scattering process and decrease of κ lat in Ta-doped Bi 2 O 2 Se can be seen from monotonically reduced l ph from~11.9 Å for Bi 2 O 2 Se to~9.9 Å for Bi 1.90 Ta 0.10 O 2 Se, which mainly results from that Ta substitution introduces multi-scale lattice defects, including the enormous defects, grain boundaries, and phase interfaces [68,77]. The carrier engineering and hierarchical microstructure by Ta doping remarkably enhance the ZT values in Bi 1.90 Ta 0.10 O 2 Se, reaching~0.30 at 773 K, which is an increase of 350% compared to pristine Bi 2 O 2 Se (Figure 9f).

Attempts to Realize a New Kind of n-Type Oxyselenide: Bi 6 Cu 2 Se 4 O 6
Because of the strong phonon scattering caused by layered structure [9], the lone pair electrons of Bi 3+ [78,79], and the local vibration of Cu + [38], BiCuSeO exhibits inherent low thermal conductivity. Another well-known thermoelectric oxyselenide, Bi 2 O 2 Se reveals stable n-type transport properties due to Se vacancies [43,44]. To fully utilize the features of BiCuSeO and Bi 2 O 2 Se, a new type layered oxyselenide Bi 6 Cu 2 Se 4 O 6 was synthesized through solid state reaction (SSR) with the 1:2 ratio of BiCuSeO and Bi 2 O 2 Se (Figure 10a) [28][29][30], and stable n-type conductive transports were observed in this system through halogen element doping [29].

Halogen Element Doping at Se Sites
The σ of Cl-doped Bi 6 Cu 2 Se 4 O 6 is higher than Br-doped one at high temperature for the doping content x = 0.2. When the doping content x is increased to 0.8, the σ was significantly improved to~70 S cm −1 , and the Br-doped sample was better σ than the Cl-doped one at high temperature (Figure 10b). The Bi 6 Cu 2 Se 3.8 Br 0.2 O 6 exhibits p-type semiconductor characteristics below 673 K and transfer to n-type with the temperature increasing; while Bi 6 Cu 2 Se 3.8 Cl 0.2 O 6 has a negative S value within the entire temperature boundaries indicating that a small amount of Cl doping (x = 0.2) can achieve stable n-type semiconductor behavior (Figure 10c). When the doing content raises up to 0.8, the S of Cl/Br-doped samples has little difference, varying from~−60 to −160 µV K −1 . The maximum ZT value~0.15 at 823 K is achieved in Bi 6 Cu 2 Se 3.2 Br 0.8 O 6 ( Figure 10c).

Transition Metal Element Doping at Bi Sites
Zheng et al. chose Bi 6 Cu 2 Se 3.6 Cl 0.4 O 6 as the matrix and doped transition metal elements (Zr, Ti and Ce) at Bi sites to enhance its thermoelectric performance [30]. The introduction of transition metal elements can effectively increase the carrier concentration (n H ) and maintain the carrier mobility (µ H ; Figure 11a), thereby optimizing the electric conductivity (σ) of the matrix. The S of all doped samples remains negative throughout the entire temperature range, indicating the stable n-type semiconductor properties (Figure 11b). Thanks to the optimized σ and maintained S, the power factor (PF) is effectively enhanced (Figure 11c). Finally, due to the enhanced electrical transport properties and reduced thermal conductivity [30], the peak ZT value reached~0.16 at 873 K in Bi 5.9 Zr 0.1 Cu 2 Se 3.6 Cl 0.4 O 6 (Figure 11d), which is 60% higher than that in Bi 6 Cu 2 Se 3.6 Cl 0.4 O 6 (~0.10 at 873 K).
As a new type of layered oxyselenide thermoelectric material, Bi 6 Cu 2 Se 4 O 6 maintains the advantages of BiCuSeO and Bi 2 O 2 Se, such as lower-cost and nontoxic elements, better thermal and chemical stability. Moreover, Bi 6 Cu 2 Se 4 O 6 can exhibit stable n-type semiconductor behavior by simple halogen doping and has intrinsic low thermal conductivity due to complex crystal structure. Thereby, Bi 6 Cu 2 Se 4 O 6 is a new kind of n-type layered oxyselenide thermoelectric material with broad development prospects.

Halogen Element Doping at Se Sites
The σ of Cl-doped Bi6Cu2Se4O6 is higher than Br-doped one at high temperature for the doping content x = 0.2. When the doping content x is increased to 0.8, the σ was significantly improved to ~70 S cm −1 , and the Br-doped sample was better σ than the Cl-doped one at high temperature (Figure 10b). The Bi6Cu2Se3.8Br0.2O6 exhibits p-type semiconductor characteristics below 673 K and transfer to n-type with the temperature increasing; while Bi6Cu2Se3.8Cl0.2O6 has a negative S value within the entire temperature boundaries indicating that a small amount of Cl doping (x = 0.2) can achieve stable n-type semiconductor behavior (Figure 10c). When the doing content raises up to 0.8, the S of Cl/Br-doped samples has little difference, varying from ~−60 to −160 μV K −1 . The maximum ZT value ~0.15 at 823 K is achieved in Bi6Cu2Se3.2Br0.8O6 (Figure 10c).

Transition Metal Element Doping at Bi Sites
Zheng et al. chose Bi6Cu2Se3.6Cl0.4O6 as the matrix and doped transition metal elements (Zr, Ti and Ce) at Bi sites to enhance its thermoelectric performance [30]. The introduction of transition metal elements can effectively increase the carrier concentration (nH) and maintain the carrier mobility (μH; Figure 11a), thereby optimizing the electric conductivity (σ) of the matrix. The S of all doped samples remains negative throughout the entire temperature range, indicating the stable n-type semiconductor properties (Figure 11b). Thanks to the optimized σ and maintained S, the power factor (PF) is effectively enhanced ( Figure 11c). Finally, due to the enhanced electrical transport properties and reduced thermal conductivity [30], the peak ZT value reached ~0.16 at 873 K in Bi5.9Zr0.1Cu2Se3.6Cl0.4O6 (Figure 11d), which is 60% higher than that in Bi6Cu2Se3.6Cl0.4O6 (~0.10 at 873 K). As a new type of layered oxyselenide thermoelectric material, Bi6Cu2Se4O6 maintains the advantages of BiCuSeO and Bi2O2Se, such as lower-cost and nontoxic elements, better thermal and chemical stability. Moreover, Bi6Cu2Se4O6 can exhibit stable n-type semiconductor behavior by simple halogen doping and has intrinsic low thermal conductivity due to complex crystal structure. Thereby, Bi6Cu2Se4O6 is a new kind of n-type layered oxyselenide thermoelectric material with broad development prospects.

Summary and Perspective
In this short review, we introduced the latest accomplishments in n-type layered oxyselenide thermoelectric materials, including BiCuSeO, Bi2O2Se and Bi6Cu2Se4O6. For Bi-

Summary and Perspective
In this short review, we introduced the latest accomplishments in n-type layered oxyselenide thermoelectric materials, including BiCuSeO, Bi 2 O 2 Se and Bi 6 Cu 2 Se 4 O 6 . For BiCuSeO, many strategies have been used to enhance the thermoelectric performance of p-type systems, but there are few studies on n-type BiCuSeO, and it is difficult to obtain stable n-type semiconductor behavior. For Bi 2 O 2 Se, carrier engineering, band engineering, microstructure design, etc., achieved performance enhancements of Bi 2 O 2 Se, but the ZT value is still limited to 0.4 [68]. Moreover, a new kind of promising n-type transport properties can be obtained in layered oxyselenide Bi 6 Cu 2 Se 4 O 6 through halogen element doping. Apart from the advancements mentioned above, there is still room left for further research, such as make full utilization of the anisotropic transport properties of those compounds through texturing microstructure and crystals growth.
Author Contributions: All the authors co-edited and wrote the paper. All authors have read and agreed to the published version of the manuscript. Data Availability Statement: The data of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.