Subscriber access provided by University of Sunderland
Article
Band gap tailoring and structure-composition relationship within the alloyed semiconductor CuBaGe SnSe 2
1-x
x
4
Garrett C. Wessler, Tong Zhu, Jon-Paul Sun, Alexis Harrell, William P. Huhn, Volker Blum, and David B. Mitzi Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.8b03380 • Publication Date (Web): 28 Aug 2018 Downloaded from http://pubs.acs.org on September 1, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 19 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 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
Band Gap Tailoring and Structure-Composition Relationship within the Alloyed Semiconductor Cu2BaGe1xSnxSe4 Garrett C. Wessler,† Tong Zhu,† Jon-Paul Sun,† Alexis Harrell,† William P. Huhn,† Volker Blum,*,†,‡ and David B. Mitzi*,†,‡ †
Department of Mechanical Engineering and Materials Science and ‡Department of Chemistry, Duke University, Durham, North Carolina 27708, United States
Abstract Recently, the I2-II-IV-VI4 (I = Cu, Ag; II = Ba, Sr; IV = Ge, Sn; VI = S, Se) materials family was identified as a promising source of potential new photovoltaic (PV) and photoelectrochemical (PEC) absorbers. These materials avoid the pitfalls of the successful photovoltaic semiconductors Cu(In,Ga)(S,Se)2 and CdTe, as they do not contain scarce (In, Te) or toxic (Cd) elements. Furthermore, ionic sizes and coordination preferences are very different for the I, II, and IV cations in the I2-II-IV-VI4 family, providing an intriguing avenue to avoid intrinsic antisite disordering that limits efficiency improvement in Cu2ZnSn(S,Se)4 (where Cu and Zn can easily substitute for one another). Here, we experimentally and computationally explore alloys Cu2BaGe1-xSnxSe4 (CBGTSe, 0 ≤ x ≤ 1) to fine-tune the structural, optical, and electronic properties for the relatively large band gap (Eg = 1.91(5) eV) unalloyed compound Cu2BaGeSe4 (CBGSe). We show that CBGTSe maintains the P31 crystal structure type of the parent CBGSe up to x ≤ 0.70. A minimum band gap value of 1.57(5) eV can be reached at x = 0.70 before the structure transforms to the Ama2 structure type. The experimental and theoretical investigations demonstrate the potential of CBGTSe for thin-film PV and PEC absorbers.
Introduction Thin-film photovoltaic (PV) technologies based on chalcogenide semiconductors, most importantly Cu(In,Ga)(S,Se)2 (CIGSSe) and CdTe, have had success in the research and industrial spaces, with lab-scale power conversion efficiencies exceeding 22%.1 However, CdTe and CIGSSe may ultimately be limited in their global deployment by the reliance of each material on scarce (i.e. In, Te) or toxic (i.e. Cd) elements. Thin-film solar cells based on the kesterite Cu2ZnSn(S,Se)4 (CZTSSe) absorber emerged from the desire to replace the toxic and scare elements with benign and earth-abundant elements Zn and Sn in these PV technologies. While there is a great interest in CZTSSe as an absorber material, the efficiency of research-scale devices has stagnated at 12.6% for the past several years.2,3 This efficiency stagnation stems at least in part from significant CuZn and ZnCu antisite disordering in CZTSSe, leading to potential fluctuations in the electronic band structure and band tailing, which in turn effectively limit the PV device open circuit voltage (Voc).4 Additionally, CuSn, SnCu, ZnSn, and SnZn antisite defects can contribute defect energy levels deep within the band gap of CZTSSe and increase recombination in CZTSSe PV devices, further hampering the power conversion efficiency.5
ACS Paragon Plus Environment
1
Chemistry of Materials 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 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 19
In the CZTSSe lattice, Cu, Zn, and Sn cations have a similar tetrahedral coordination and comparable ionic radii. These similarities contribute to small antisite defect formation energies and their prevalence in CZTSSe.5 One method to suppress the antisite disorder is to replace the Zn ion with a much larger alkaline earth cation such as Ba or Sr, introducing structural and ionic size diversity into the lattice. Cu2BaSnS4 (CBTS) and Cu2SrSnS4 (CSTS) structures adopt the trigonal space group P316,7 whereas the fully Se-substituted relatives Cu2BaSnSe4 (CBTSe) and Cu2SrSnSe4 (CSTSe) crystallize in the related orthorhombic Ama2 structure type (structural differences between these structure and CZTSSe are shown in Figure S1).8,9 These and other semiconductors within the larger family of I2-II-IV-VI4 (I = Cu, Ag; II = Ba, Sr; IV = Ge, Sn; VI = S, Se) materials form structures in which the II atom is eight-fold coordinated while the I and IV atoms maintain the tetrahedral coordination as in CZTSSe. The difference in coordination and ionic size between the II atom and the I/IV atom is predicted to hinder the formation of antisite defects and can decrease the amount of band tailing within films and devices based on these semiconductors.10,11 For example, Cu2BaSn(S,Se)4 (CBTSSe) has garnered interest as a potential alternative PV absorber to CZTSSe.11–17 Initial P31 space group-based CBTSSe solar cells and photoelectrochemical (PEC) devices have exceeded 5% power conversion efficiency and a photocurrent of 12 mA/cm2 at 0 V vs the reversible hydrogen electrode (RHE), respectively, with demonstrably reduced band tailing as compared to CZTSSe.13,18 The smallest band gap in the CBTSSe system is found at a Se:S ratio of 3:1 in the P31 structure (1.55 eV).12 At higher Se:S ratios, the structure type transitions to Ama2, associated with substantially higher band gaps. These tendencies indicate that both the P31 structure and high Se concentrations are associated with relatively low PV-friendly band gaps, closest to the ideal 1.0 – 1.6 eV range, within the I2II-IV-VI4 family.10,19 While CBTSSe illustrates a promising path for both PV and PEC16–18,20,21 applications, other absorber candidates, such as Cu2BaGeSe4 (CBGSe), have been identified with favorable optical and electronic properties.10 Within the I2-II-IV-VI4 family, the relatively unexplored CBGSe is the only Se-based compound to form in the P31 space group.22 CBGSe also exhibits a high absorption coefficient and relatively low electron (||a = 0.15m0, ||c = 0.18m0) and hole (||a = 0.26m0, ||c = 1.04m0) effective masses.10 Although CBGSe has been shown to have many properties that are attractive for PV and PEC applications, we have shown experimentally that the band gap (1.91 eV) of pure CBGSe is slightly above the ideal energy range for single junction PV, though it is well within the relevant range for multi-junction PV and PEC applications.10 To reduce the band gap of P31 CBGSe into a viable energy range for single junction PV and fine-tune the position of the valence and/or the conduction bands, it is natural to consider alloying one or more of its components. Given the already complete occupation of the anion site by Se, we turn to alloying one of the cation sites in this work. Cation alloying has been shown to be an effective method to tune the structural, optical and electronic properties for related materials CZTSSe and CBTSSe. Examples for CZTSSe include: targeted alloying of Ag23 or Li24,25 on the Cu site; Cd,26–29 Fe,30,31 Mn,32,33 Co,32 and Ni32 on the Zn site; and Ge34–38 and In39 on the Sn site. In a study over 40 years ago, the structural effect of alloying CBTS on the cation site (Ag for Cu) was also explored.6 In the present work, we initially consider several cation alloying approaches, including Ag
ACS Paragon Plus Environment
2
Page 3 of 19 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 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
for Cu, Sr for Ba, and Sn for Ge within CBGSe in both the P31 and Ama2 crystal structures using hybrid density-functional theory (DFT) calculations. We show that substitution of Sn for Ge holds the most promise to lower the CBGSe band gap in either of the prospective P31 or Ama2 space groups. Based on the preliminary direction suggested by the DFT-predicted variation of the alloyed band gaps, we experimentally and computationally probe the variation in structural and optical properties of the alloyed compound Cu2BaGe1-xSnxSe4 (CBGTSe), from the pure Ge (x = 0) compound to the pure Sn derivative (x = 1). As the Sn content (x) increases within CBGTSe, the material exhibits the expected phase transition from the P31 structure type to the Ama2 space group, which we find to occur just above x = 0.70, separated by a small apparent two-phase coexistence region. At a composition just below the phase transition, the band gap of P31 CBGTSe achieves a minimum value of 1.57(5) eV, within the relevant range for single and multi-junction solar cells and PEC devices.
Experimental and computational Bulk sample preparation: Pellets (~200 mg) of CBGTSe were prepared by grinding stoichiometric amounts of BaSe (Materion, 99.5%), CuSe (Alfa Aesar, 99.5%), Ge (Alfa Aesar, 99.999%), Sn (Alfa Aesar, 99.999%) and Se (Alfa Aesar, 99.999%) in an agate mortar and pestle and cold-pressing into a pellet inside a nitrogen-filled glovebox. Prior to weighing, the BaSe and CuSe precursor powders were first baked on a hotplate, in an inert atmosphere, at 175 °C to remove slight SeO2 impurities. The pellets were loaded into prebaked quartz tubes, evacuated to ~ 5x10-6 torr and flame-sealed under dynamic vacuum using a hydrogen-oxygen torch. The samples were then heated to 600 oC or 650 oC (x = 1.00) with a ramping rate of 100 oC/hr and held at these temperatures for 24-60 hours. The grinding, pressing, and heating processes were repeated 1-2 times for each sample without exposure to air in order to improve phase purity. During the first heating, the pellets expanded and became fragile as the initial reaction proceeded, with an average mass loss of 6.9 mg noted. After the first heating, the pellets and powders were dark gray. During subsequent heating steps, the pellets did not deform and < 1 mg mass loss was observed. Characterization: Powder X-ray diffraction (PXRD) patterns were measured using a PANalytical Empyrean diffractometer with Cu Kα radiation under ambient conditions. These patterns were fit by a Pawley phase-fitting method using the PANalytical HighScore Plus software. The band gaps of the bulk powder samples were measured by diffuse reflectance using an Enlitech QE-R Quantum Efficiency/Reflectivity system. Diffuse reflectance analysis was conducted using the Kubelka-Munk function, F(R), defined as = 1 − ⁄2 , where R is the diffuse reflectance of the powder samples.40 Extraction of the direct band gap was done using a Tauc plot by plotting ℎ versus ℎ . The intersection of the linear fit of the linear region of the absorption edge and the minimum of the data was taken as the band gap of the powder sample (e.g., see Figure S2). Photoluminescence emission spectra were measured using a Horiba Jobin Yvon LabRam ARAMIS system with a 442 nm He-Cd excitation laser. Computational methods: To analyze the structural, electronic, and optical properties of CBGTSe alloy systems, the FHI-aims41–45 electronic structure code was used with "tight" numerical settings for all calculations reported in this work. Basis sets for Cu, Ag, Ba, Sr, Sn, Ge,
ACS Paragon Plus Environment
3
Chemistry of Materials 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 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 19
Se are given in Table S1. Unless otherwise noted, the screened hybrid Heyd-Scuseria-Ernzerhof (HSE) functional46,47 with a fixed screening parameter (ω = 0.2 Å-1) and a fixed exchange mixing parameter (α = 0.25) in a dense Γ-point centered 12 × 12 × 6 k-point grid are used in all calculations (we refer to these choices as the "standard HSE06 functional" in the remainder of this work).48 We note that we found this choice to lead to a systematic underestimation of predicted band gaps by approximately 0.3 eV in our past work on similar systems.10,12,13 In the literature, it is known that more careful parameterizations of hybrid functionals may yield somewhat reduced band gap prediction errors.49 However, even such more refined parameterized hybrid functionals still yield mean absolute errors of few tenths of an eV49 and are neither fully equivalent to a hypothetical exact generalized Kohn-Sham density functional, nor would they include other effects reflected in experiment, such as electron-phonon renormalization. We therefore opt to retain the shape of the standard HSE06 functional for its relative simplicity and for compatibility with past work of our own, relying on experimental validation of the key bandgap related findings as an additional safeguard in our work. Spin-orbit coupling (SOC) in a second-variational approach50 is included in all calculated band structures, densities of states (DOS) and optical properties. Unless otherwise noted, all unit cell parameters and cell-internal coordinates used in the computation of total energies, band structures, DOS and optical properties correspond to local minima of the HSE06 potential energy surface. Residual forces (stresses) of computed equilibrium structures are below 5 × 10-3 eV/ Å1 (5 × 10-3 eV/ Å3). Normal-incidence absorption coefficients α(ω) were calculated based on the real and imaginary parts of the dielectric function,51–53 obtained from the HSE06 functional in the independentparticle (i.e., random-phase) approximation including spin-orbit coupling (SOC), and based on dense Γ-point-centered 12 × 12 × 6 k-point grids using the same method as in our previous paper.10 A Gaussian broadening of 0.1 eV was used to obtain smooth α(ω) curves. For the HSE06 computational results pertaining to alloyed compounds (0 < x
