† Corresponding author. E-mail:
Project supported by the National Natural Science Foundation of China (Grant Nos. 11164014 and 11364025) and the Gansu Science and Technology Pillar Program, China (Grant No. 1204GKCA057).
Using the first-principles plane-wave calculations within density functional theory, the perfect bi-layer and monolayer terminated WZ-CIS (100)/WZ-CdS (100) interfaces are investigated. After relaxation the atomic positions and the bond lengths change slightly on the two interfaces. The WZ-CIS/WZ-CdS interfaces can exist stably, when the interface bonding energies are −0.481 J/m2 (bi-layer terminated interface) and −0.677 J/m2 (monolayer terminated interface). Via analysis of the density of states, difference charge density and Bader charges, no interface state is found near the Fermi level. The stronger adhesion of the monolayer terminated interface is attributed to more electron transformations and orbital hybridizations, promoting stable interfacial bonds between atoms than those on a bi-layer terminated interface.
In recent years, the fabrication of pollution-free, low-cost, and high-efficiency photovoltaic cells has attracted successive attention. CuInS2-based thin film has been known as one of the most promising optical absorbers for high-efficiency solar cells.[1,2] It has garnered a great deal of interest due to its high optical absorption coefficient (> 105 cm−1) higher than conventional III–V semiconductors, desirable band gap (1.45 eV) close to the solar spectra, good thermal property, low cost, environmental, and electrical stability.[3–6] It is reported that CuInS2 (CIS) exists in three polymorphic modifications at different temperatures: the most common and thermodynamically stable chalcopyrite (CP-CIS), metastable zincblende (ZB-CIS), and wurtzite (WZ-CIS) structures.[7,8] It was found that the CIS with the wurtzite structure shows flexibility in stoichiometry, because the copper and indium atoms alternate on the cation site.[9] The wurtzite CuInS2 has been researched in a solar cell device because it provides the ability to tune the Fermi level energy over a wide range, which is beneficial for device fabrication.[10] It has been observed experimentally that this promising material has a high optical absorption coefficient and substantial photo-stability.[9] Because of its strong light absorption over the visible and near-infrared range and good size homogeneity, these nanocrystals were applied as the CIS absorber layer by spraycoating to fabricate working solar cells after the selenization process. The conversion efficiency of CIGS thin film soalr cells was beyond 20%.[11]
Cadmium sulfide (CdS) is a most important II–VI group wide-gap semiconductor with good stability and outstanding optical-electronic properties[12–14] due to its wide energy band gap (2.42 eV)[15] which happens to be in the visible range. CdS have two types of basic structures: the cubic sphalerite and the hexagonal wurtzite.[16] The wurtzite CdS shows novel electronic and optical properties as compared with the sphalerite CdS.[17] Many studies have also reported that the wurtzite CdS nanowires were synthesized in various nanocrystalline forms such as by a simple hydrothermal method.[18] Nowadays, CdS as an efficient n-type window material has been widely used in the CuInS2-based thin film solar cell devices.
Thin-film photovoltaic devices were fabricated with a conventional glass/Mo/CuInS2/CdS/ZnO/Al-doped/ZnO configuration.[19] CuInS2 with wurtzite structure is similar to CdS with wurtzite structure. They have the same space group and crystallize as a hexagonal structure. In the processes of preparing solar cells, choosing CIS and CdS with wurtzite structure at the same time, the absorption layer and window layer can make good lattice match, which may reduce the composite effect of the WZ-CIS/WZ-CdS interface. In the WZ-CIS thin film solar cells, the wurtzite CuInS2 nanocrystals were converted with a selenization process, and were applied as the p-type light-absorbing layer in the fabrication. The WZ-CdS window layer is generally deposited on the WZ-CIS absorption layer, which can form the WZ-CIS/WZ-CdS p–n hetero-junction. It is the core part of the WZ-CIS thin film solar cells. The interface between the light-absorbing layer and the window layer is especially important which has greatly influenced the conversion efficiency and performance of the solar cells. Therefore, investigating the interface between WZ-CIS and WZ-CdS gives a systematic understanding to improve CIS solar cell performance and may guide the experiments.
So far, we have performed studies on the CIS solar cells interface: the bond characteristics, electronic structure and interfacial energetic of the Mo (110)/MoSe2 (100) interface,[20] and also for the WZ-CIS (100)/MoS2 (−100) interface[21] by the first principles calculations. In this work, we theoretically study the properties of bulk WZ-CIS, bulk WZ-CdS, their surfaces and interfaces. For the WZ-CIS/WZ-CdS interface, we investigated two different bond types, and discussed their lattice structures, electronic properties and the interface states at the atomic scale. Furthermore, our results also have been compared with others experimental and calculated results to well understand WZ-CIS solar cells.
All theoretical calculations were performed within the Vienna ab initio simulation package (VASP)[22–24] with the first principles of density functional theory (DFT). The projector augmented wave (PAW)[25,26] method and the generalized gradient approximations (GGA) functional with the Perdew–Burke–Ernzerhof (PBE)[27] were applied to describe the exchange-correlation energy. We used the pseudopotential for the electronic configurations are [Ar] 3d104s1, [Kr] 5s25p1, [Ne] 3s23p4, and [Kr] 4d105s2, for copper, indium, sulfur, and cadmium, respectively. Γ -centered k-point meshes were applied for bulk WZ-CuInS2, bulk CdS, WZ-CuInS2 surface, WZ-CdS surface, and WZ-CIS/WZ-CdS interface are 11 × 11 × 7, 11 × 11 × 7, 5 × 6 × 1, 5 × 6 × 1, and 5 × 6 × 1, respectively. We adopted the conjugate gradient (CG) method[28] to optimize atomic structure and selected a cutoff energy of 500 eV based on the plane wave basis. We apply the tetrahedron method with Blöchl corrections[29] for our band structure, density of states and total energy calculations in all systems. Density functional theory (DFT) always underestimates semiconductors’ band gaps. In our calculations we added a Hubbard U correction to the GGA energy functional. The values of effective U parameters employed for the 3d states of Cu atoms are 5.5 eV to solve the band gap problem presented in DFT calculations on WZ-CuInS2, which are consistent with other previous studies U-value corrections.[30–32]
To validate the accuracy of the computation methods, we firstly performed calculations on bulks. It is known that Wurtzite CuInS2 (WZ-CIS) belongs to the space group P63mc and crystallizes as a hexagonal wurtzite structure, as shown in Fig.
The band structure and density of states (DOS) were also calculated for bulk WZ-CIS.[21] Through DOS analysis, the WZ-CIS is a direct band gap semiconductor with the calculated band gap of 0.2 eV through the method of GGA + U (U − J = 5.5 eV), which is similar with another calculated value[9] but lower than the experimental band gaps (1.47 eV).[38] By analysis of partial density of states (PDOS), the bottom of the conduction band is attributed by 5s-orbital of In atoms, and the uppermost valence band is mainly attributed by 3p-orbital of S atoms.
In WZ-CuInS2, Cu and In atoms randomly locate on the same space group site. At the same time, WZ-CuInS2 has the sub-lattice structure: a tetrahedron containing an S atom at the center and four Cu/In atoms around the S atom. In every tetrahedron, In and Cu atoms may have different site occupancies. There must be five and only five possible occupancies: the S atom is surrounded by two Cu atoms and two In atoms, the S atom is surrounded by three/one In atom(s) and one/three Cu atom(s), the S atom is surrounded by four Cu/In atoms. The hybrid Hartree–Fock-like functional by Heyd, Scuseria, and Ernzerhof (HSE) can improve the calculation accuracy.[22–24] The first occupancy configuration is the generally used lattice model, which has a band gap 0.86 eV when Hartree–Fock screening parameter ω = 0.2 was employed in HSE06 functional. We also found that WZ-CuInS2 is a semiconductor that may be metallic for different local indium and copper atomic configurations. Metallic configurations have higher lattice energies while semiconductive configurations have lower lattice energies.
Figure
To study the properties of WZ-CuInS2/WZ-CdS interfaces, it is necessary to ensure that both sides of the interface models are thick enough to imitate the bulk. We firstly studied their surface structures. The WZ-CuInS2 (100) surface has two surface models: a bi-layer terminated surface and a monolayer terminated surface. Figures
The stability of a surface can be described via the surface energy. The surface energy γ [41] is obtained from a slab including two identical surfaces, through Eq. (
We have also obtained the electronic properties for the surfaces. Through analysis of the local density of states (LDOS) of WZ-CIS (100) and WZ-CdS (100) surfaces we found that no new electronic states appeared near the Fermi level, which means that there is no existence of surface states on the surfaces.
Building the WZ-CuInS2/WZ-CdS interface model for simulation needs good surface lattice matching. We find that WZ-CuInS2 (100) and WZ-CdS (100) faces can make a good lattice matching, and the lattice mismatch between two slabs is less than 5.6%, as shown in Fig.
Based on the above calculations, we build the WZ-CuInS2/WZ-CdS interface models for DFT computations and studied their local lattice structure and electronic properties through the interface models (Fig.
The atoms positions and bond lengths change before and after relaxation for bi-layer terminated and monolayer terminated WZ-CIS (100)/WZ-CdS (100) interfaces are shown in Fig.
Adhesion energy is the key factor to predict interfacial bonding stability.[42] We calculated the interface bonding energy to qualitatively analyze the interface combination stability of two different models. A negative bonding energy means a stable interface structure.[43] We use Eq. (
In order to further explore the bonding ability of the interface, we studied the local density of states (LDOS) and the partial density of states (PDOS), interfacial difference charge density distribution, Bader charges are calculated both for the bi-layer terminated (interface model A) and monolayer terminated (interface model B) WZ-CIS/WZ-CdS interfaces.
The local density of states (LDOS) and partial density of states (PDOS) of WZ-CIS/WZ-CdS interfaces were shown in Fig.
Figure
From Fig.
By comparing Fig.
We also calculated the difference charge density for two interface bonding styles of WZ-CIS/WZ-CdS to further clarify their interfacial bonding and electronic structure. The difference charge density ρ is defined in Eq. (
Figure
For the bi-layer terminated interface, seen from Figs.
Their Bader charges are also analyzed, to further clarify their electronic properties of bi-layer terminated and monolayer terminated interfaces. We marked the atomic charges near the interface in Figs.
From the above analysis, we can infer that the WZ-CIS/WZ-CdS interface has better stability, and no interface states appeared on the interface, which is beneficial to the conductivity of WZ-CIS based solar cells. Also, after investigating the bi-layer and monolayer terminated WZ-CIS (100)/WZ-CdS (100) interfaces, we found that the bonding style of the monolayer terminated interface has a more strong adhesion ability and stability than the bi-layer terminated interface. It is attributed to more electron transformations and orbital hybridizations on the interface promoting formation of the stable interfacial bonds, which is also beneficial to improving the conversion efficiency of solar cells.
In conclusion, we used the first-principles calculations to study WZ-CIS (100)/WZ-CdS (100) interfaces, including the interface structure, adhesion energy, interfacial bonding and electronic properties. Considering bi-layer terminated and monolayer terminated bonding styles, the main findings are as follows.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 | |
21 | |
22 | |
23 | |
24 | |
25 | |
26 | |
27 | |
28 | |
29 | |
30 | |
31 | |
32 | |
33 | |
34 | |
35 | |
36 | |
37 | |
38 | |
39 | |
40 | |
41 | |
42 | |
43 |