Band offsets engineering at CdxZn1−xS/Cu2ZnSnS4 heterointerface
Bao Wujisiguleng1, †, , Sachuronggui 2, Qiu Fang-Yuan1
College of New Energy, Bohai University, Jinzhou 121013, China
College of Engineering, Bohai University, Jinzhou 121013, China

 

† Corresponding author. E-mail: baowujisgl@bhu.edu.cn

Project supported by the Special Funds of the National Natural Science Foundation of China (Grant Nos. 11547226 and 11547180).

Abstract
Abstract

Cd1−xZnxS/Cu2ZnSnS4 (CZTS)-based thin film solar cells usually use CdS as a buffer layer, but due to its smaller band gap (2.4 eV), CdS film has been replaced with higher band gap materials. The cadmium zinc sulfide (CdZnS) ternary compound has a higher band gap than other compounds, which leads to a decrease in window absorption loss. In this paper, the band offsets at Cd1−xZnxS/Cu2ZnSnS4 (CZTS) heterointerface are calculated by the first-principles, density-functional and pseudopotential method. The band offsets at Cd1−xZnxS/CZTS heterointerface are tuned by controlling the composition of Zn in Cd1−xZnxS alloy, the calculated valence band offsets are small, which is consistent with the common-anion rule. The favorable heterointerface of type-I with a moderate barrier height (< 0.3 eV) can be obtained by controlling the composition of Zn in Cd1−xZnxS alloy between 0.25 and 0.375.

1. Introduction

Cu2ZnSnS4 (CZTS) solar cell material has an ideal band gap (Eg = 1.5 eV) and high absorption coefficient.[1,2] As a possible alternative to the commercially available Cu(In, Ga)Se2 (CIGS), CZTS does not contain indium or gallium. Therefore, it could be a cheaper and more sustainable solar cell material in the future. Recently, a conversion efficiency of 12.6% for selenium containing CZTS[3] has been reported. However more research is needed to understand what prevents the devices from reaching as high efficiency as that of CIGS, which has record efficiency over 20%.[4] As is well known, CZTS-based thin film solar cells usually employ CdS as a buffer layer. However, the development of CdS/CZTS thin film solar cells is restrained by its smaller band gap (2.4 eV). One of the reasons is an unfavorable alignment of the conduction band minimum (CBM) at the CdS/CZTS heterointerface, where CBM of CdS is lower than that of CZTS.[5] This interface leads to an increase in interface recombination and a loss in open circuit voltage, Voc.

Cadmium zinc sulfide (CdZnS) ternary compound has a higher band gap, which leads to a decrease in window absorption loss and thus an increase in the short-circuit current. CdZnS is potentially useful as a window material for p–n junctions without lattice mismatch in the device based on quaternary material like CuInxGa1−xSe2 or CuIn(SzSe1−z)2.[6] ZnCdS has a variable band gap energy of 2.4 eV–3.7 eV, primarily dependent on a relative ratio of Cd to Zn.[710] The optimum conduction band alignment for CZTS should lie between the CdS and ZnS values. It has been reported that the pure CdS is expected to give a negative conduction band offset with CZTS,[11,12] and ZnS should cause the current to block due to a high barrier. Irvine et al. noted that the control of the alloying of ternary Cd1−xZnxS and keeping the thickness of the CdS window layer constant can cause the band gap to broaden.[13]

As is well known, the band offset at buffer-absorber interface is one of the most important parameters, which is often used to assess some important interface effects, i.e., quantum confinement and carrier transport, in particular, for the design of solar cells and other optoelectronic devices. There are numerous reports on the electrical and optical properties influenced by Zn substitution in CdZnS material.[1418] The increasing of Zn composition in Cd1−xZnxS will enlarge the band gap, therefore, Cd1−xZnxS should lead to an increase of quantum efficiency (QE) in the shorter wavelength regime and also lead to favorable conduction band offset with CZTS. In this paper, we study the effects of Zn composition on the photovoltaic performance by calculating the band offsets for CdZnS/CZTS heterointerface with different Zn compositions. By changing the constituents of Cd1−xZnxS, the moderate barrier height at the Cd1−xZnxS/CZTS heterointerface is obtained.

2. Calculation

Alloying is a good approach to performing band-gap engineering to extend the available band gap. In bulk Cd1−xZnxS alloyed crystal, its composition (x)-dependent band gap energy Eg(x) can be expressed by Vegard’s law[19]

where Eg(ZnS) and Eg(CdS) are the band gap energies for bulk ZnS and CdS, respectively, and b is the bowing parameter and has a value 0.61.[19,20]

The calculation of density of states (DOS) was performed on the basis of the first-principles, density-functional and pseudopotential method by using the PHASE code developed by Institute of Industrial Science, University Tokyo.[21] We used the generalized gradient approximation (GGA) for the exchange-correlation interaction,[22] with valence electron configurations of S (3s2, 3p4), Cu (3d10, 4s1), Zn (3d10, 4s2), Cd (4d10, 5s2) and Sn (4d10, 5s2, 5p2). For a given atomic arrangement, the lattice constants and atom positions were optimized to minimize the total energy.

We first obtained the energy level difference between the reference core levels (take the S 3s, Zn 3d, Cd 4d, and Sn 4d as reference core levels in the pseudopotential calculation method) and valence band maximum (VBM), from the band structures of Cd1−xZnxS and CZTS. Then, the core level difference was obtained from the band structure of the (001) Cd1−xZnxS/CZTS superlattice. The valence band offset EVBM and conduction band offset ECBM were obtained as[23,24]

where is the energy difference between the reference core levels and VBM for the CZTS (CdZnS) bulk, and is the difference between core levels in the Cd1−xZnxS/CZTS supercell. This equation is based on the idea that the energy difference between the core level and VBM in the respective bulk material is conserved in the heterostructure. ΔEg is the band gap difference between CdZnS and CZTS. For CdZnS/CZTS heterointerface, the lattice parameter of CdZnS is well matched to that of the CZTS layer: the lattice parameters are a = 0.541 nm for zinc-blende ZnS, a = 0.582 nm for zinc-blende CdS, and a = 0.551 nm and c = 1.123 nm for kesterite CZTS. Thus, the (001)-face lattice mismatch is smaller than 5.6% (with respect to CZTS) between CdZnS and CZTS layers, the effects of strain caused by lattice mismatch in the CdZnS/CZTS interface are negligible. So, in this work, we assume that the ΔEg for CdZnS/CZTS is the same as the difference between their bulk energy levels.

3. Results and discussion
3.1. Electronic structures and band gap

The total density of states for zinc-blend structures of ZnS and CdS are calculated as shown in Fig. 1. The zero point energy is taken as VBM. For the ZnS and CdS binary materials, cation d state and anion p state will repeal upward the VBM as shown in Fig. 1, Zn 3d states are shallower than the Cd 4d states. The calculated band gaps are 2.8 eV and 1.2 eV for ZnS and CdS, respectively. These calculated band gaps are smaller than the experimental values of 3.8 eV and 2.4 eV. It is due to the limitation of DFT, but it has less effect on the investigated electronic structure in the present work. As shown in Fig. 1, Cd-d states show two peaks and Zn-d states show a single peak, when a small amount of Zn content exists in Cd1−xZnxS. The Zn-d states change into two peaks while Cd-d states change into one single peak with the increase of Zn content. It means that the interaction of Zn-d and S-p states will play an important role in VBM nearby, when higher content Zn exists in Cd1−xZnxS. The calculated band gaps are 1.3, 1.5, and 1.8 for Cd1−xZnxS with the composition x values of 0.25, 0.5, and 0.75, respectively. As the Zn content increases, Cd1−xZnxS shifts toward the higher band gap; this trend is consistent with Vegard’s law[19] and other investigations.[25]

Fig. 1. Densities of states (DOS) for Cd1−xZnxS at composition x of 0, 0.25, 0.5, 0.75, and 1.
3.2. Band offsets for Cd1−xZnxS/CZTS heterointerfaces

The band offsets at Cd1−xZnxS/CZTS heterointerfaces are tuned by controlling the Zn composition (x) in Cd1−xZnxS (x = 0, 0.125, 0.25, 0.375, 0.5, 0.75, 1) alloy. Since the band gap is generally underestimated in the calculation based on GGA, the band gap energy of Cd1−xZnxS as a function of Zn composition (x) is determined using Eq. (1). The calculated values of band offsets at Cd1−xZnxS/CZTS heterointerfaces with different Zn content values are given in Table 1. The minus value means that the VBM (or CBM) of Cd1−xZnxS is lower than that of CZTS. The obtained valence band offsets are almost the same, i.e., about −1.0 eV∼ − 1.3 eV. This result is in consistence with the common-anion rule, where two semiconductors sharing the same anion would have a small valence-band offset.[26] The obtained conduction band offsets are −0.3∼ 1.0 eV, the CBM of Cd1−xZnxS increases with the increase of Zn composition in Cd1−xZnxS alloy. The conduction band offset becomes plus value when the composition content of Zn is higher than 0.25, that means the type of heterointerface is changed.

Table 1.

Calculated band offsets at Cd1−xZnxS/CZTS heterointerfaces with different values of Zn composition (x) in Cd1−xZnxS (x = 0, 0.125, 0.25, 0.375, 0.5, 0.75, 1).

.

Figure 2 shows the band offsets at Cd1−xZnxS/CZTS heterointerfaces with different Zn content values. The VBM of ZnS is higher than that of CdS, it is due to the Zn 3d state being shallower than the Cd 4d state. The CBM of Cd1−xZnxS is higher than that of CZTS (type-I heterointerface), when the composition content of Zn is higher than 0.25 in Cd1−xZnxS alloy, the conduction band offset forms a barrier (spike) for the photoexcited electrons crossing the interface. Height of this barrier is larger than 0.4 eV; when the composition content of Zn is higher than 0.5 in Cd1−xZnxS alloy, this will considerably reduce the photocurrent.[27] On the other hand, if the composition content of Zn is smaller than 0.25 in Cd1−xZnxS alloy, the barrier against photogenerated electrons is not formed at the heterointerface, but the open-circuit voltage will be reduced by increasing the recombination rate for the majority carriers at the interface. Accordingly, it is generally consistent with the fact that the type-I interface with a moderate barrier height (< 0.3 eV) is the most favorable. Thus, it will obtain higher conversion efficiency by controlling the composition content of Zn in Cd1−xZnxS alloy to be between 0.25 and 0.375. Bhattacharya et al.[17] reported the 19.52%-efficient CIGS-based solar cells using a single-layer chemical bath deposited CdZnS buffer layer with a compositional ratio of Cd:Zn being 80:20. Song et al.[18] experimentally investigated the dependences of Cd1−xZnxS/CIGS device performance parameters on the Zn content in Cd1−xZnxS buffer layers with x = 0, 0.1, 0.2, 0.3, 0.4 and reported the higher conversion efficiency with x = 0.2 than other Zn content in Cd1−xZnxS, which is consistent with our prediction.

Fig. 2. Band offsets at Cd1−xZnxS/CZTS heterointerfaces with different Zn content values (in unit eV).
4. Conclusions

In the present paper, we calculated the band offsets at Cd1−xZnxS/CZTS heterointerface for different values of x, based on the first-principles, density-functional, and pseudopotential method. The band offsets at Cd1−xZnxS/CZTS heterointerface are tuned by controlling the composition content of Zn in Cd1−xZnxS alloy. Cd1−xZnxS/CZTS is a type I of heterointerface with a moderate barrier height smaller than 0.3 eV, when the composition content of Zn in Cd1−xZnxS alloy is between 0.25 and 0.375.

Reference
1Katagiri H2005Thin Solid Films480–481426
2Shin BGunawan OZhu YBojarczuk N AJay Chey SGuha S 2013 Prog. Photovoltaics 21 72
3Wang WWinkler MGunawan OGokmen TTodorov TZhu YMitzi D 2014 Adv. Energy Mater. 4 1301465
4Jackson PHariskos DWuerz RKiowski OBauer AFriedlmeier T M 2015 Phys. Status Solidi 9 28
5Bär MSchubert BMarsen BWilks RPookpanratana SBlum MKrause SUnold TYang WWeinhardt LHeske CSchock H W 2011 Appl. Phys. Lett. 99 222105
6Prem Kumar TSaravana kumar SSankaranarayanan K 2011 Appl. Surf. Sci. 257 1923
7Al Kuhaimi S ATulbah Z 2000 J. Electrochem. Soc. 147 214
8Chen PNicholls J EO’Neill MStirner THogg J H C 1998 J. Appl. Phys. 84 5621
9Dmitrienko A OBukesov S AMikhailova V V1993Inorg. Mater.29960
10Salem A M 2002 Appl. Phys. 74 205
11Barkhouse D A RHaight RSakai NHiroi HSugimoto HMitzi D B 2012 Appl. Phys. Lett. 100 193904
12Nagoya AAsahi RKresse G 2011 J. Phys.: Condes. Matter 23 404203
13Irvine S J CBarrioz VLamb DJones E WRowlands-Jones R L 2008 J. Cryst. Growth 310 5198
14Peter A JChang W L 2012 Chin. Phys. 21 087302
15Azizi SRezagholipour Dizaji HEhsani M H 2016 Optik-International Journal for Light and Electron Optics 127 7104
16Thahab S MOmran Alkhayat A HSaleh S M 2014 Mater. Sci. Semicond. Process. 26 49
17Bhattacharya R NContreras M AEgaas BNoufi R NKanevce ASites J R 2006 Appl. Phys. Lett. 89 253503
18Song JLi S SChen LNoufi R2006IEEE World Conference on Photovoltaic Energy Conference1534
19Singh J1996Optoelectronics: an introduction to materials and devicesNew DelhiMacgraw Hill
20Sze S M1969Physics of Semiconductor DevicesNew YorkWiley
21http://www.ciss.iis.u-tokyo.ac.jp/rss21/en/index.html
22Perdew J PBurke KErnzerholf M 1996 Phys. Rev. Lett. 77 3865
23Nelson A JSchwerdtfeger C RWei S HZunger ARioux DPatel RHochst H 1993 Appl. Phys. Lett. 62 2557
24Wei S HZunger A 1993 Appl. Phys. Lett. 63 2549
25Wei S HRen XHan ZLi TSu YWei LCong FZhang Y 2013 Mater. Lett. 102�?03 94
26Wei S HZunger A 1988 Phys. Rev. 37 8958
27Minemoto TMatsui TTakakura HHamakawa YNegami THashimoto YUenoyama TKitagawa M 2001 Sol. Energy Mater. Sol. Cells. 67 83