Rapid thermal evaporation for cadmium selenide thin-film solar cells

Abstract

Cadmium selenide (CdSe) belongs to the binary II-VI group semiconductor with a direct bandgap of ~1.7 eV. The suitable bandgap, high stability, and low manufacturing cost make CdSe an extraordinary candidate as the top cell material in silicon-based tandem solar cells. However, only a few studies have focused on CdSe thin-film solar cells in the past decades. With the advantages of a high deposition rate (~2 µm/min) and high uniformity, rapid thermal evaporation (RTE) was used to maximize the use efficiency of CdSe source material. A stable and pure hexagonal phase CdSe thin film with a large grain size was achieved. The CdSe film demonstrated a 1.72 eV bandgap, narrow photoluminescence peak, and fast photoresponse. With the optimal device structure and film thickness, we finally achieved a preliminary efficiency of 1.88% for CdSe thin-film solar cells, suggesting the applicability of CdSe thin-film solar cells.

Graphical abstract

Keywords

cadmium selenide (CdSe) / rapid thermal evaporation (RTE) / solar cells / thin film

1 Introduction

Recent progress in chiral photonics, i.e., bianisotropic metamaterials [1–7] and bianisotropic metasurfaces [8–15], has significantly advanced our understanding of light transport in complex photonic structures and spurred numerous applications. Indeed, bianisotropic metamaterials have been used to realize novel resonators [1,2,6] and construct photonic topological insulators [4,5,7]. Bianisotropic metasurfaces have been used to manipulate the polarization of light [8,9] and to achieve exotic refraction and transmission of light [10–14]. In metamaterials or metasurfaces, it is common to use optically complicated structures inside each unit cell together with a very large number of inclusions to obtain certain functionalities. Thus the complete multiple scattering of light among all the inclusions is hardly traceable. Therefore, under proper approximation, the chiral metamaterials [15–20] or chiral metasurfaces [21,22] can be described using effective constitutive parameters (permittivity and permeability), which in principle accounts for the bianisotropic response of the complicated unit cell. Notably, the grading effect in a wide range of metasurfaces, e.g., with a rotated orientation of each unit cell structure, can also be included in the spatial dependent bianisotropic constitutive parameters in the effective medium description.

However, the available means of calculating photonic bianisotropic medium is limited within analytical methods, which are usually complicated and only apply to simple structures, i.e., interfaces or slab structures [23–26]. A generic numerical approach that calculates the electromagnetic properties for bianisotropic structure with arbitrary shapes is lacking. In this paper, we intend to fill this gap and propose a finite element method (FEM) to simulate the optical properties of bianisotropic media.

The paper is organized as follows: In Section 2, we derive the balanced formulation of the weak form for the finite element implementation. In Sections 3−5, we illustrate and validate our numerical approach via three examples, i.e., bianisotropic slab structure, bianisotropic photonic crystal, and bianisotropic ring resonator respectively. Finally, Section 6 concludes the paper.

2 Model

The constitute relations of the generic bianisotropic materials are given by $\mathit{D}=\overline{\mathit{ϵ}}\mathit{E}+{\overline{\mathit{\chi }}}_{eh}\mathit{H}$, $\mathit{B}=\overline{\mathit{\mu }}\mathit{H}+{\overline{\mathit{\chi }}}_{he}\mathit{E}$, where $\mathit{D}$ and $\mathit{B}$ are electric displacement vector and the magnetic induction intensity respectively, $\mathit{E}$ and $\mathit{H}$ are electric and magnetic field respectively, $\overline{\mathit{\epsilon }}={\epsilon }_{\mathrm{0}}{\overline{\mathit{\epsilon }}}_r$ and $\overline{\mathit{\mu }}={\mu }_{\mathrm{0}}{\overline{\mathit{\mu }}}_r$ are permittivity and permeability, respectively, ${\overline{\mathit{\chi }}}_{eh}=\sqrt{{\epsilon }_{\mathrm{0}}{\mu }_{\mathrm{0}}}{\overline{\mathit{\chi }}}_{eh}^r$ and ${\overline{\mathit{\chi }}}_{he}=\sqrt{{\epsilon }_{\mathrm{0}}{\mu }_{\mathrm{0}}}{\overline{\mathit{\chi }}}_{he}^r$ are coupling constants. As for the source-free Maxwell’s equations with time-harmonic dependence of ${\mathrm{e}}^{\mathrm{i}\omega t}$, i.e., $\nabla \times \mathit{E}=-\mathrm{i}\omega \mathit{B}$, $\nabla \times \mathit{H}=\mathrm{i}\omega \mathit{D}$, ${\mathrm{i}}^{\mathrm{2}}=-\mathrm{1}$, one reformulated Maxwell’s equations using the normalized field, i.e., $\mathit{e}(\mathit{r})=\mathit{E}(\mathit{r})$ and $\mathit{h}(\mathit{r})=\sqrt{{\displaystyle \frac{{\mu }_{\mathrm{0}}}{{\epsilon }_{\mathrm{0}}}}}\mathit{H}(\mathit{r})$, as follows

(1)$\begin{array}{l}\begin{array}{rr}\hfill \left(\nabla \times +\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{he}^r\right)\mathit{e}(\mathit{r})+\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\mu }}}_r\mathit{h}(\mathit{r})& \hfill =\mathrm{0},\\ \hfill \left(\nabla \times -\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{eh}^r\right)\mathit{h}(\mathit{r})-\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\epsilon }}}_r\mathit{e}(\mathit{r})& \hfill =\mathrm{0.}\end{array}\end{array}$

Eliminating the magnetic field $\mathit{h}(\mathit{r})$, one arrives at the vector wave equation given as follows

(2)$\begin{array}{l}\left(\nabla \times -\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{eh}^r\right)\left[{\displaystyle \frac{\mathrm{1}}{{\overline{\mathit{\mu }}}_r}}\left(\nabla \times +\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{he}^r\right)\right]\mathit{e}(\mathit{r})-k_{\mathrm{0}}^{\mathrm{2}}{\overline{\mathit{\epsilon }}}_r\mathit{e}(\mathit{r})=\mathrm{0.}\end{array}$

Once the electric $\mathit{e}$ field known, the magnetic field $\mathit{h}$ is given by $\mathit{h}(\mathit{r})=-{\displaystyle \frac{\mathrm{1}}{\mathrm{i}{k}_{\mathrm{0}}}}{\overline{\mathit{\mu }}}_r^{-\mathrm{1}}\nabla \times \mathit{e}(\mathit{r})-{\overline{\mathit{\mu }}}_r^{-\mathrm{1}}{\overline{\mathit{\chi }}}_{he}^r\mathit{e}(\mathit{r})$. As for Eq. (2), the vectorial wave equation is essentially determined by operator $\mathit{L}$, i.e., $\mathit{L}=\left(\nabla \times -\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{eh}^r\right)\left[{\displaystyle \frac{\mathrm{1}}{{\overline{\mathit{\mu }}}_r}}\left(\nabla \times +\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{he}^r\right)\right]-k_{\mathrm{0}}^{\mathrm{2}}{\overline{\mathit{\epsilon }}}_r$. Under scaler inner product, we show that the operator $L$ is indeed self-adjoint, which is described as follows

(3)$\begin{array}{l}(\mathit{F},L\mathit{E})=(L\mathit{F},\mathit{E}),\end{array}$

see proof in Supplementary Material A. Importantly, we can further proof that

(4)$\begin{array}{l}{\int }^{\mathrm{}}\mathrm{d}V\left(\left[{\displaystyle \frac{\mathrm{1}}{{\overline{\mathit{\mu }}}_r}}\left(\nabla \times +\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{he}^r\right)\right]\mathit{F}\cdot {\left[\nabla \times -\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{eh}^r\right]}^{\mathrm{T}}\mathit{E}-k_{\mathrm{0}}^{\mathrm{2}}{\overline{\mathit{\epsilon }}}_r\mathit{F}\cdot \mathit{E}\right)=(\mathit{F},L\mathit{E})\\ =(\mathit{L}F,\mathit{E})={\int }^{\mathrm{}}\mathrm{d}V\left(\left[{\displaystyle \frac{\mathrm{1}}{{\overline{\mathit{\mu }}}_r}}\left(\nabla \times +\mathrm{i}{k}_{\mathrm{0}}{\overline{\mathit{\chi }}}_{he}^r\right)\mathit{E}\right]\cdot \left[\nabla \times -\mathrm{i}{k}_{\mathrm{0}}{\left({\overline{\mathit{\chi }}}_{eh}^r\right)}^{\mathrm{T}}\right]\mathit{F}-k_{\mathrm{0}}^{\mathrm{2}}{\overline{\mathit{\epsilon }}}_r\mathit{E}\cdot \mathit{F}\right),\end{array}$

which is coined as the balanced weak form. This balanced weak form contains a few implications that deserves further discussion: (1) In practical implementation of FEM, the Galakin procedures is usually adopted. In the Galakin procedures, the test function $\mathit{F}$ is by default selected as the basis functions that are used to interpolate the electric field. Thus, it is essential to keep the test function space and the expansion function spaces undergone the same operations, i.e., transformation in exactly the same manner; (2) Applying standard FEM procedure leads to the unbalanced formulation of weak form in bianisotropic media as used in reference [27], which gives rise to artificial numerical errors due to the spatial differentiation of constitutive parameters, i.e., giving rise to the unusual imaginary part of the propagation constant of bianisotropic waveguides; (3) In this paper, the balanced formulation is adopted, it turns to be extremely important in the finite element modeling in bianisotropic medium and overcome the problem in unbalanced formulation from the comparison between the standard weak form and balanced weak form formulation that is not shown in the paper.

3 Reflection and transmission of light in layered bianisotropic medium

First, we benchmark our finite element model of light reflection and transmission against the analytical calculations. We consider a slab geometry in which light is incident in normal direction, as shown in Fig. 1(a). The slab in dark gray surrounded with air in light gray has ${\epsilon }_r=\mathrm{4}$, ${\mu }_r=\mathrm{1}$ and the magnetoelectric coupling constants

${\overline{\mathit{\chi }}}_{he}^r=-{\left({\overline{\mathit{\chi }}}_{eh}\right)}^{\mathrm{T}}=\left(\begin{array}{lll}\mathrm{0}& \mathrm{0}& \mathrm{0}\\ \mathrm{0}& \mathrm{0}& \mathrm{0}\\ \mathrm{0}& \mathrm{0}& \mathrm{i}{\chi }_{\mathrm{33}}\end{array}\right).$

The slab has a finite thickness $L=\mathrm{1}\mathrm{\hspace{0.17em}}\mathrm{m}$ in horizontal direction, and the incident light contains in-plane ($p$) polarization and out-of-plane ($s$) polarization. The practical implementation is realized by modifying the original weak form in COMSOL Multiphysics [28] into our balanced weak form (the right term in Eq. (4)) for the bianisotropic medium.

Fig.1 (a) Schematic diagram of simulating reflection and transmission spectrums of bianisotropic slab with normal incident light. A bianisotropic slab (dark gray) is surrounded by air (light gray), the boundaries are vertical with y-axis and light propagates along y direction. Horizontal ($p$) and vertical ($s$) polarized light are excited (received) at port 1 (2) and 3 (4) respectively. (b)−(e) Reflection ($r$) and transmission ($t$) spectrum of horizontal ($p$) and vertical ($s$) polarized light. The results from FEM and semi-analysis method are represented by blue circles and red lines respectively

Full size|PPT slide

The numerical calculation of the ${\chi }_{\mathrm{33}}$ dependent reflection/transmission spectrum at vacuum wavelength $\lambda =\mathrm{1}\mathrm{\hspace{0.17em}}\mathrm{m}$ from our reformulated weak form are perfectly coincides with semi-analysis method (see details in Supplementary Material B) displayed in Figs. 1(b)−1(e). As a side remark, there is no difference between $p$ and $s$ polarization for normal incidence at ${\chi }_{\mathrm{33}}=\mathrm{0}$. As for nonzero ${\chi }_{\mathrm{33}}$, the reflection/transmission of $p$ and $s$ polarization differ from each other. This is reasonable since, inside the bianisotropic slab, the $p$ and $s$ polarization are not eigen-polarization and gets mixed due to ${\chi }_{\mathrm{33}}$, and the true eigen-polarization essentially has different effective refractive index. As a result, the incident $p$ and $s$ polarization can be decomposed into the two eigen-polarizations that has different refractive index depending on ${\chi }_{\mathrm{33}}$, which leads to the variation of reflection and transmission against ${\chi }_{\mathrm{33}}$, as confirmed both by our numerical simulation and analytical calculations.

4 Band structure of bianisotropic photonic crystals

Bianisotropic metamaterials and metasurfaces have been used to realize photonic topological insulator or photonic valley Hall effect. Dong and his coworkers [29] explored the photonic valley in honeycomb photonic crystals, in which the inversion symmetry is broken via the material bianisotropy displayed in Fig. 2(a). The unit cell in this honeycomb lattice has two different rods with radii $r$ in purple and blue, respectively, and lattice constant $a$. These two rods have uniaxial permittivity and permeability ${\bar{\epsilon }}_r={\bar{\mu }}_r=\mathrm{diag}({\epsilon }_{//},{\epsilon }_{//},{\epsilon }_{\perp })$, and nonzero bianisotropic tensor

${\overline{\mathit{\chi }}}_{eh}^r={\overline{\mathit{\chi }}}_{he}^r=\left(\begin{array}{lll}\mathrm{0}& \mathrm{i}\kappa {\epsilon }_{//}& \mathrm{0}\\ -\mathrm{i}\kappa {\epsilon }_{//}& \mathrm{0}& \mathrm{0}\\ \mathrm{0}& \mathrm{0}& \mathrm{0}\end{array}\right),$

where the parameter $\kappa$ in rods 1 and 2 have opposite sign. With the modified weak form, we calculated band structure of this photonic lattice in Fig. 2(b), where the dispersion curves for two pseudo-spins (spin-up and spin-down) of light on the high symmetry line of Brillouin zone are shown and have perfect agreement with the data from Ref. [29]. The two spins have almost the same band structure in low frequency, but the spin-up state has a smaller bandgap near $K$ and a larger bandgap around $K\mathrm{\text{'}}$ than that of the spin-down state. As a result, in the range of frequency from 0.375 to 0.435, the bandgap vanishes at $K$ valley for the spin-up state; the same is true for spin-down state at $K\mathrm{\text{'}}$ valley. Thus, the two spin states propagate in different directions determined by the associated valley index, as illustrated with two arrows in Fig. 2(a).

Fig.2 (a) Bianisotropic honeycomb photonic crystal. The blue and purple circles represent two kinds of rods, and a unit cell is extracted as the green hexagon with $r=\mathrm{0.25}a$. ${\epsilon }_{//}=\mathrm{8}$, ${\epsilon }_{\mathrm{prep}}=\mathrm{1}$, $\kappa =-\mathrm{0.5}$ in rod 1 but $\kappa =\mathrm{0.5}$ in rod 2. (b) Band structure on the high symmetry line. Red and blue points represent spin-down and spin-up sampled form literature, and blue circles is the result from COMSOL with modified weak form. The inset shows the Brillouin zone where $G_{\mathrm{1}}$ and $G_{\mathrm{2}}$ are reciprocal lattice vectors. $K$ and $K\mathrm{\text{'}}$ are two kinds of vertex

Full size|PPT slide

5 Bianisotropic ring resonator

We consider a bianisotropic slab waveguide as shown in Fig. 3(a), which extends infinitely along $x$ and $z$ direction and is sandwiched by air. The bianisotropic slab has a relative permittivity ${\epsilon }_r=\mathrm{1}$, relative permeability ${\mu }_r=\mathrm{1}$, and magnetoelectric coupling tensor

${\bar{\chi }}_{eh}^r=-{\left({\bar{\chi }}_{he}^r\right)}^{\mathrm{T}}=\left(\begin{array}{lll}\mathrm{0}& \mathrm{0}& \mathrm{0}\\ \mathrm{0}& \mathrm{0}& \mathrm{0}\\ \mathrm{0}& \mathrm{0}& \mathrm{i}{\chi }_{\mathrm{33}}\end{array}\right).$

Light propagates along $x$, the effective refractive index of the fundamental mode increases for larger ${\chi }_{\mathrm{33}}$, as can be seen from the simplified wave equation in the presence of ${\overline{\mathit{\chi }}}_{eh}^r$ in Supplementary Material C. The fundamental mode is linearly polarized for vanished ${\chi }_{\mathrm{33}}$, and is elliptically polarized for nonzero ${\chi }_{\mathrm{33}}$. By time reversal symmetry, the polarization of light will be converted into its time-reversal partner, i.e., the helicity flipping sign, for opposite propagating direction. As the slab waveguide is rolled to form a ring resonator shown in Fig. 3(b), the $x$, $y$ and $z$ components of an electric field in an ordinary slab waveguide corresponds to the azimuthal ($E_{\theta }$), radial ($E_r$) and $z$ components. Therefore, light propagates along the azimuthal direction in the ring resonator is the same as it transports along $x$ in the slab waveguide. Thus, light propagates clockwise or counter-clockwise, the polarization of light will have opposite helicity. Consequently, the ring resonator could function as a polarization depended insulator as shown in Figs. 3(c)−3(f). The clockwise and counter-clockwise modes in the ring resonator have opposite helicity of polarization in electric field basis ($E_z\mp \mathrm{1.72}\mathrm{i}{E}_{\theta }$) at vacuum wavelength $\mathrm{1.0615}\mathrm{\hspace{0.17em}}\mathrm{\mu m}$. A slab waveguide with the same permittivity and permeability supports transverse electric and transverse magnetic modes, which span the degenerate basis for the polarization. In Figs. 3(c) and 3(d), the polarized optical beam $E_z+\mathrm{1.72}\mathrm{i}{E}_y$ is excited at the left and the right ports of the slab waveguide respectively, it couples with counter-clockwise mode $E_z+\mathrm{1.72}\mathrm{i}{E}_{\theta }$ in the ring resonator with positive direction propagation (from left to right) while it propagates through the slab without any coupling with the ring from right to left. In contrast, its time-reversal partner, i.e., $E_z-\mathrm{1.72}\mathrm{i}{E}_y$, propagates without coupling in a positive direction but couples with clockwise mode $E_z-\mathrm{1.72}\mathrm{i}{E}_{\theta }$ in negative propagation as shown in Figs. 4(e) and 4(f). Therefore, the polarized beams $E_z\pm \mathrm{1.72}\mathrm{i}{E}_y$ are insulated in positive and negative directions, respectively. Furthermore, the polarization ellipticity of the modes in the ring can be manipulated by changing ${\chi }_{\mathrm{33}}$, this structure provides a platform to isolate or pick out any ellipse polarized beam in a simple way. Evidently, our numerical results are fully consistent with the earlier polarization-dependent optical insulation based on the bianisotropic ring resonator.

Fig.3 Bianisotropic ring resonator behaves as polarization depended insulator. Bianisotropic slab waveguide (a) and bianisotropic ring resonator formed by rolling slab waveguide (b). Optical beam with polarization $E_z+\mathrm{1.72}\mathrm{i}{E}_y$ propagates in the slab couples with ring from the left to the right (c) but propagates without coupling from the right to the left (d). Beam with polarization $E_z-\mathrm{1.72}\mathrm{i}{E}_y$ goes through the slab in positive direction (e) but is insulated in reversal direction (f). The width of slab and ring is $\mathrm{0.15}\mathrm{\hspace{0.17em}}\mathrm{\mu m}$, the radius of ring is $\mathrm{2.3272}\mathrm{\hspace{0.17em}}\mathrm{\mu m}$. ${\epsilon }_r={\mu }_r=\mathrm{1}$ and ${\chi }_{\mathrm{11}}=\mathrm{2.7442}$ in bianisotropic ring, ${\epsilon }_r={\mu }_r=\mathrm{2.366}$ in slab. Vacuum wavelength is $\mathrm{1.0615}\mathrm{\hspace{0.17em}}\mathrm{\mu m}$ and background medium is air

Full size|PPT slide

6 Conclusion

We propose a numerical method based on FEM approach to study the optical properties of complex bianisotropic structures. We first benchmark our method in reflection and transmission of a bianisotropic slab, and further validate our numerical calculation of the band structure in a bianisotropic photonic crystal that plays a relevant role in topological photonics. Our results have perfectly coincided either with the analytical results or the literature data. Finally, we illustrate that our method can be used to design a bianisotropic ring resonator, which functions as a polarization-dependent optical insulator.

References

[1] Hermle M, Feldmann F, Bivour M, Goldschmidt J C, Glunz S W. Passivating contacts and tandem concepts: Approaches for the highest silicon-based solar cell efficiencies. Applied Physics Reviews, 2020, 7(2): 021305–021312

[2] Meillaud F, Shah A, Droz C, Vallat-Sauvain E, Miazza C. Efficiency limits for single-junction and tandem solar cells. Solar Energy Materials and Solar Cells, 2006, 90(18–19): 2952–2959

[3] Rickus D B E. The CdSe thin-film solar cell. Conference Record of the IEEE Photovoltaic Specialists Conference, 1980, 1: 629–632

[4] Lu S, Chen C, Tang J. Possible top cells for next-generation Si-based tandem solar cells. Frontiers of Optoelectronics, 2020, 13(3): 246–255

[5] Yamaguchi M, Lee K H, Araki K, Kojima N. A review of recent progress in heterogeneous silicon tandem solar cells. Journal of Physics D, Applied Physics, 2018, 51(13): 133002–133015

[6] Todorov T, Gunawan O, Guha S. A road towards 25% efficiency and beyond: perovskite tandem solar cells. Molecular Systems Design & Engineering, 2016, 1(4): 370–376

[7] Todorov T K, Bishop D M, Lee Y S. Materials perspectives for next-generation low-cost tandem solar cells. Solar Energy Materials and Solar Cells, 2018, 180: 350–357

[8] Bakiyaraj G, Dhanasekaran R. Effect of annealing on the properties of chemical bath deposited nanorods of CdSe thin films. Crystal Research and Technology, 2012, 47(9): 960–966

[9] Bagheri B, Kottokkaran R, Poly L P, Sharikadze S, Dalal V. Efficient heterojunction thin film CdSe solar cells deposited using thermal evaporation. In: Proceedings of IEEE 46th Photovoltaic Specialists Conference (PVSC). Chicago: IEEE, 2019, 1822–1825

[10] Che S B, Nomura I, Kikuchi A, Shimomura K, Kishino K J P S S. Visible light emitting diode with ZnCdSe/BeZnTe superlattices as an active layer and MgSe/BeZnTe superlattices as a p-cladding layer. Physica Status Solidi (B): Basic Solid State Physics, 2002, 229(2): 1001–1004

[11] Jia S, Yun X, An Y, Li P, Xiao J. Fabrication and characterization of photo-detector based on CdSe_0.5S_0.5 quantum dots. Asia Communications & Photonics Conference, 2013, 2013: 978–981

[12] Mahawela P, Jeedigunta S, Vakkalanka S, Ferekides C S, Morel D L J T S F. Transparent high-performance CdSe thin-film solar cells. Thin Solid Films, 2005, 480–481(3): 466–470

[13] Wang C, Du X, Wang S, Deng H, Chen C, Niu G, Pang J, Li K, Lu S, Lin X, Song H, Tang J. Sb₂Se₃ film with grain size over 10 µm toward X-ray detection. Frontiers of Optoelectronics, 2020, doi:10.1007/s12200-020-1064-5

[14] Shin Y M, Lee C S, Shin D H, Kwon H S, Park B G, Ahn B T. Surface modification of CIGS film by annealing and its effect on the band structure and photovoltaic properties of CIGS solar cells. Current Applied Physics, 2015, 15(1): 18–24

[15] Song H, Zhan X, Li D, Zhou Y, Yang B, Zeng K, Zhong J, Miao X, Tang J. Rapid thermal evaporation of Bi₂S₃ layer for thin film photovoltaics. Solar Energy Materials and Solar Cells, 2016, 146: 1–7

[16] Xue D J, Liu S C, Dai C M, Chen S, He C, Zhao L, Hu J S, Wan L J. GeSe thin-film solar cells fabricated by self-regulated rapid thermal sublimation. Journal of the American Chemical Society, 2017, 139(2): 958–965

[17] Somorjai G A. Vapor pressure and solid-vapor equilibrium of CdSe (cadmium selenide). Journal of Physical Chemistry, 1961, 65(6): 1059–1061

[18] Li K, Kondrotas R, Chen C, Lu S, Wen X, Li D, Luo J, Zhao Y, Tang J. Improved efficiency by insertion of Zn_1−xMg_xO through sol-gel method in ZnO/Sb₂Se₃ solar cell. Solar Energy, 2018, 167: 10–17

[19] Major J D, Proskuryakov Y Y, Durose K. Impact of CdTe surface composition on doping and device performance in close space sublimation deposited CdTe solar cells. Progress in Photovoltaics, 2013, 21(4): 436–443

[20] Gnatenko Y P, Bukivskij P M, Faryna I O, Opanasyuk A S, Ivashchenko M M. Photoluminescence of high optical quality CdSe thin films deposited by close-spaced vacuum sublimation. Journal of Luminescence, 2014, 146: 174–177

[21] Tian L, Yang H, Ding J, Li Q, Mu Y, Zhang Y. Synthesis of the wheat-like CdSe/CdTe thin film heterojunction and their photovoltaic applications. Current Applied Physics, 2014, 14(6): 881–885

[22] Chen C, Zhao Y, Lu S, Li K, Li Y, Yang B, Chen W, Wang L, Li D, Deng H, Yi F, Tang J. Accelerated optimization of TiO₂/Sb₂Se₃ thin film solar cells by high-throughput combinatorial approach. Advanced Energy Materials, 2017, 7(20): 1700866

[23] Parsons R B, Wardzynski W, Yoffe A D. The optical properties of single crystals of cadmium selenide. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1961, 262(1308): 120–131

[24] Rosly H N, Rahman K S, Harif M N, Doroody C, Isah M, Misran H, Amin N. Annealing temperature assisted microstructural and optoelectrical properties of CdSe thin film grown by RF magnetron sputtering. Superlattices and Microstructures, 2020, 148: 106716

[25] Ni Z H, Kong B, Zeng T X, Yang H, He Z Y. The effects of the intrinsic defects on the electronic, magnetic and optical properties for bulk and monolayer CdSe: first-principles GGA+ U investigations. Materials Research Express, 2019, 6(10): 105903

[26] Murali K R, Srinivasan K, Trivedi D C. Vacuum evaporated CdSe thin films and their characteristics. Materials Letters, 2005, 59(1): 15–18

[27] Murali K R, Srinivasan K, Trivedi D C. Structural and photoelectrochemical properties of CdSe thin films deposited by the vacuum evaporation technique. Materials Science and Engineering B, 2004, 111(1): 1–4

[28] Kosyak V, Opanasyuk A, Bukivskij P M, Gnatenko Y P. Study of the structural and photoluminescence properties of CdTe polycrystalline films deposited by close-spaced vacuum sublimation. Journal of Crystal Growth, 2010, 312(10): 1726–1730

[29] Hariskos D, Powalla M, Chevaldonnet N, Lincot D, Schindler A. Chemical bath deposition of CdS buffer layer: prospects of increasing materials yield and reducing waste. Thin Solid Films, 387(1–2): 179–181

[30] Leng M Y, Luo M, Chen C, Qin S K, Chen J, Zhong J, Tang J.Selenization of Sb₂Se₃ absorber layer: an efficient step to improve device performance of CdS/Sb₂Se₃ solar cells. Applied Physics Letters, 2014, 105(8): 083905

[31] Hu W D, Dall’Agnese C, Wang X F, Chen G, Li M Z, Song J X, Wei Y J, Miyasaka T. Copper iodide-PEDOT:PSS double hole transport layers for improved efficiency and stability in perovskite solar cells. Journal of Photochemistry and Photobiology A Chemistry, 2018, 357: 36–40

[32] Voswinckel S, Mikolajick T, Wesselak V. Influence of the active leakage current pathway on the potential induced degradation of CIGS thin-film solar modules. Solar Energy, 2020, 197: 455–461

[33] Wang L, Li D B, Li K, Chen C, Deng H X, Gao L, Zhao Y, Jiang F, Li L, Huang F, He Y, Song H, Niu G, Tang J. Stable 6%-efficient Sb₂Se₃ solar cells with a ZnO buffer layer. Nature Energy, 2017, 2(4): 17046–17053

Acknowledgments

This work was supported by the National Natural Science Foundation of China (Grant Nos. 61725401, 61904058, and 62050039), the National Key R&D Program of China (No. 2016YFA0204000), the Innovation Fund of WNLO, National Postdoctoral Program for Innovative Talent (No. BX20190127), the Graduates’ Innovation Fund of Huazhong University of Science and Technology (No. 2020yjsCXCY003), and China Postdoctoral Science Foundation Project (Nos. 2019M662623 and 2020M680101). The authors thank the Analytical and Testing Center of HUST and the facility support of the Center for Nanoscale Characterization and Devices (CNCD), WNLO-HUST.

Electronic Supplementary Material

Supplementary material is available in the online version of this article at https://doi.org/10.1007/s12200-021-1217-1 and is accessible for authorized users.

RIGHTS & PERMISSIONS

2021 Higher Education Press