Open Access
Add to BibSonomy
Abstract
We have demonstrated a reduction in the prevalence of surface defect states in FeS2 iron pyrite thin films by encapsulating them with ZnS. Thin films of FeS2 were grown and encapsulated, without exposure to atmosphere, by a variety of films. X-ray photoelectron spectroscopy (XPS) measurements, using a novel technique that permits us to selectively probe the FeS2-capping layer interface, show a reduction in the surface defect state characterized by the sulfur 2p doublet for ZnS-encapsulated films. We further present an atomistic density functional theory (DFT) model that explains this effect based on epitaxial bonding of FeS2 and ZnS across their interface.
1. Introduction
FeS2 has great promise as a next-generation photovoltaic (PV) material due to its favorable combination of earth abundancy, extremely high absorption coefficient, and a band gap suitable for either single or-multijunction applications [1]. In a recent survey of 23 known semiconductor systems with potential as PV absorbers, FeS2 ranked highest in potential annual electricity production based on known reserves and had the lowest extraction cost. Its bandgap is 0.95 eV, high enough to result in a potential solar to electricity conversion efficiency similar to that of Si, but unlike Si, it has an exceptionally high absorption coefficient of α = 6x105 cm−1 resulting in a required thickness of <40 nm for >90% absorption compared to typical thicknesses >100 μm for Si [2].
Despite these apparent advantages, FeS2 PV devices have not yet lived up to their potential. Efficiency has been limited to approximately 3% primarily due to open circuit voltage (VOC) of <200 mV, only ~20% of the bandgap [3–5]. These limits have been traced in part to surface-state defects related to the formation of unbound sulfur monomers at the surface of pyrite crystals, resulting in midgap defect states that lead to shorting and carrier recombination in devices [6]. While the exact nature of the surface states and electronic characteristics of pyrite are still under debate [7,8], many studies have confirmed the significance of the pyrite crystalline surface and its stark difference with the electrical properties of the bulk material, where surface effects dominate the electrical properties of thin films and nanostructures [9–12]. Additionally, theoretical studies [13,14] of FeS2 reveal that non-stoichiometric surfaces, with much-reduced band gaps, are thermodynamically more stable than the stoichiometric surface for certain ranges of the sulfur chemical potential, further suggesting that the FeS2 surface may require passivation to improve its properties for photovoltaic applications.
Therefore, we have studied the impact of various passivation layers on the nature of sulfur atoms at the pyrite interface. Previous studies of the FeS2 surface have revealed that there is a binding energy difference between the surface and bulk states of the sulfur 2p doublet [15]; we therefore use this difference as a useful metric to determine the effect of passivation layers on the surface. Experimental results indicate that surface defects in polycrystalline FeS2 films are indeed passivated by encapsulation in ZnS. We have also performed an initial DFT study supporting these experimental results that indicates that ZnS can create a defect-free interface with FeS2.
2. Method
Thin films of FeS2 were deposited on soda-lime glass substrates (Abrisa Glass) using RF magnetron sputtering. Prior to deposition, substrates were cleaned using successive steps of swabbed detergent solution (Liquinox, 1% solution), then sonicated in deionized water, acetone, isopropanol, with a final rinse in deionized water. They were then blown dry with N2 and transferred to a vacuum oven to fully dry.
Substrates were then transferred into a high-vacuum cluster tool (Kurt J. Lesker Octos) for film deposition. First, a 200 nm-thick bottom contact of aluminum was deposited using e-beam evaporation to provide a surface that allows for nucleation conditions that more closely mimic those that pyrite films will experience in real PV devices. Without exposure to atmosphere, samples were transferred to a separate chamber for RF magnetron sputtering through a high vacuum (~10−8 Torr) transfer chamber.
Films of pyrite were sputtered at room temperature from a 3” diameter compound FeS2 target (American Elements) at 200 W using an Ar plasma at 3 mT for a total film thickness of 400-500 nm. Extra sulfur vapor pressure (with a partial pressure of ~1x10−5 T) was added to the chamber during deposition by evaporation of elemental sulfur in a low-temperature thermal evaporation source; failure to increase the sulfur pressure within the chamber resulted in sulfur-poor films of pyrrhotite.
After FeS2 deposition, films were immediately transferred under high vacuum without exposure to atmosphere to another chamber for deposition of various capping/passivation layers. 40 nm thick layers of ZnS, ZnO, or SiO2 were deposited via e-beam evaporation. After this deposition step, films were removed from the cluster tool, and their properties were characterized using XRD, transmission electron microscopy, and XPS.
Due to the energetic preference for defect formation in a free surface of FeS2, extra care must be taken during XPS measurements. The FeS2-capping layer interface must not be exposed; i.e. some finite thickness of capping layer must be present in order to prevent reconstruction of the FeS2 surface. For the same reason, we cannot perform measurements on an uncapped FeS2 film (even though comparing an uncapped film to capped films would be of interest). Surface reconstruction of the bare FeS2 surface, both in air and under vacuum in the XPS system, would prevent a valid comparison.
Samples were placed in the XPS analysis chamber (Thermo Scientific K-alpha, using an Al kα anode 1486.7 eV) and the capping layer was slowly removed by etching with argon ions until both Fe and capping layer peaks were present in the XPS spectra. This process is detailed in Fig. 1. Because XPS is a highly surface sensitive technique with a measurement depth of <10 nm, etching until both sets of peaks are present ensures that the FeS2 surface has not been exposed and has therefore not undergone any form of reconstruction while in the analysis chamber. For an underetched film, (Fig. 1(a)), insufficient signal is obtained from the pyrite surface for analysis, as the XPS probe volume is contained within the capping layer. The XPS signal in a correctly etched film (Fig. 1(b)) exhibits peaks of both the capping layer and pyrite with sufficient signal for proper analysis, with a capping layer thickness on the order of a few nm. An overetched film (Fig. 1(c)) no longer retains any XPS signature of the capping layer, creating a fully exposed pyrite surface.
Fig. 1 Schematic depiction of the experimental technique employed for XPS studies, with sample cross sections (left) and relative XPS signal contributions from each layer (right) for three different etching conditions. The underetched sample (a) primarily shows peaks from the capping layer; the correctly etched sample (b) retains a few nm of capping layer, with observable XPS signal from both the capping layer and pyrite; and the overetched sample (c) shows peaks from only the FeS2.
First-principles calculations were used to investigate the geometry and electronic structure of the FeS2-ZnS interface. Total energies and forces were calculated within the generalized-gradient approximation of Perdew, Burke and Ernzerhof [16] (PBE) to density-functional theory (DFT), using projector-augmented-wave (PAW) potentials and a standard planewave basis as implemented in VASP [17]. In order to describe accurately the electronic structure of FeS2, we used the DFT + U method of Dudarev [18] with U set to 2 eV for the Fe 3d orbitals; this is the same methodology used recently by Zhang, et al. [13] The interface calculations were performed in a slab geometry with four triple-layers of FeS2 and seven atomic layers of ZnS. All atomic positions were relaxed except for the innermost layers of each material. The plane-wave cutoff was 280 eV, and the sampling of the surface Brillouin zone was carried out with a 4 × 4 Monkhorst–Pack mesh. The band lineup across the heterojunction was calculated using standard techniques [19] in which the locally averaged electrostatic potential around atoms far from the interface was used to provide the reference energy levels required to determine the alignment of the valence and conduction band edges of the two bulk materials, which were determined in separate bulk calculations.
3. Results and discussion
Thin films of pyrite were grown using RF magnetron sputtering as detailed in the Methods section. Confirmation of pyrite growth was obtained by both X-ray diffraction (XRD) and transmission electron diffraction (Fig. 2 (a) and 2(b)). Films are polycrystalline in nature, with an average crystallite size of ~100 nm (Fig. 2(c)).
Fig. 2 (a) X-ray diffraction pattern with pyrite peaks identified; (b) transmission electron microscopy diffraction pattern showing FeS2 lattice spacing; (c) cross-section scanning electron microscopy showing dense pyrite film.
To analyze the impact of the capping layer on surface sulfur properties, films were characterized using XPS; three different capping layers were considered – ZnS, ZnO, and SiO2. FeS2 and ZnS have a nearly perfect lattice match, with lattice spacings of 5.417 Å and 5.411 Å, respectively. Further, ZnS has an identical cation coordination to that of bulk FeS2. ZnO is not epitaxially matched to FeS2, but retains a favorable cation coordination, and SiO2 has neither an epitaxial match nor similar coordination. Deconvolved spectra of the S 2p doublet for various capping/passivation layers are shown in Fig. 3(a)-3(c). Yellow areas represent deconvolutions of the bulk S 2p doublet (with a binding energy near −162 eV), while blue areas are surface 2p doublets (binding energy near −161 eV). Deconvolution was performed using a Pearson VII fitting routine in Thermo Scientific Avantage analysis software. The 2p3/2 and 2p1/2 peaks for each species doublet was constrained to a 2:1 peak area ratio with a 1.5 eV FWHM and a separation of 1 eV.[20]
Fig. 3 Deconvolved peak fitting of XPS results. Sulfur 2p doublets for surface (blue) and bulk (yellow) states are shaded for clarity.
It is readily apparent that both the ZnO and SiO2 cases have a much greater surface state component than that of ZnS. To further quantify this, the surface/bulk ratio Asurface/Abulk was calculated for each sample by taking a ratio of the areas under the respective peaks. ZnS has a ratio of 0.80, while both ZnO and SiO2 have ratios in excess of 2.0 (2.38 ZnO, 2.01 SiO2). The ZnS capping layer, therefore, results in better surface passivation than does either of the other capping materials.
While we do see a significant reduction in the surface/bulk ratio, we do not see complete elimination of the surface state character. This can be due to several factors, including unbound sulfur atoms near grain boundaries and non-epitaxial growth of ZnS. The former may appear in the measurement results as a result of the fact that the probe volume in the XPS measurement may include grain boundaries within the FeS2 film beneath the FeS2-capping layer interface. Thus, in order to have a complete passivation effect, we surmise that a film composed of FeS2 nanocrystals with their surfaces passivated by a thin ZnS layer may be the most promising route for FeS2 photovoltaics.
To better understand the interplay between ZnS and FeS2 electronic states, a preliminary atomistic model of the interface was developed. A pyrite core with exterior sulfur atoms was surrounded by sphalerite ZnS and allowed to equilibrate. The resulting three-dimensional geometry, shown in Fig. 4(a), establishes that sphalerite ZnS is in principle able to successfully passivate the FeS2 surface by chemically bonding to surface sulfur atoms, leading to a nearly perfect epitaxial interface. After relaxation, interfacial atoms maintain (or are close to) their bulk coordination values: all interfacial Zn and S atoms have their bulk coordination (4-fold and 3-fold, respectively) while interfacial Fe atoms are 5-fold coordinated (compared to 6-fold in bulk FeS2). Interestingly, this local passivation can follow atomic-scale corners and edges, suggesting that it will be able to passivate realistic crystalline surfaces.
Fig. 4 (a) Images of DFT modeling results showing passivation of surface sulfur atoms in an FeS2 crystal by ZnS. Closeup image of passivation around atomic-scale features is also shown, indicating that realistic crystallite shapes can be passivated in this manner; (b) Band alignment at an FeS2/ZnS/FeS2 interface, with offsets indicated.
In an optoelectronic application such as photovoltaics, band alignment for proper charge transport is crucial. Calculated band offsets (Fig. 4(b)) at the interface show a type I heterojunction, with an 0.8 eV valence band offset, suggesting that a hopping-mode transport between FeS2 domains across the ZnS layers epitaxial layers will be preferable. For sufficiently thin ZnS layers, high mobilities should still be achievable; this must be further explored experimentally in future work.
4. Conclusions
In summary, we have experimentally demonstrated a reduction in the undesirable sulfur surface state of FeS2 by using a ZnS passivation layer. Other passivation layers tried do not yield as large a reduction in FeS2 surface states, suggesting that there is a beneficial interaction between ZnS and FeS2. Theoretical modeling supports this by showing that ZnS is able to perfectly passivate FeS2 at an epitaxial interface through Zn-S bonding across the interface. DFT modeling of band offsets also predicts a type I heterojunction between the two materials. Successful passivation of FeS2 surface defects is an important step towards the realization of efficient pyrite-based PV.
Funding
US Naval Research Laboratory (NRL)
Acknowledgements
The authors acknowledge Mr. Michael Hunt and Dr. Vinh Nguyen for assistance with film processing.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. C. Wadia, A. P. Alivisatos, and D. M. Kammen, “Materials availability expands the opportunity for large-scale photovoltaics deployment,” Environ. Sci. Technol. 43(6), 2072–2077 (2009). [CrossRef] [PubMed]
2. A. Ennaoui, S. Fiechter, C. Pettenkofer, N. Alonso-Vante, K. Büker, M. Bronold, C. Höpfner, and H. Tributsch, “Iron disulfide for solar energy conversion,” Sol. Energy Mater. Sol. Cells 29(4), 289–370 (1993). [CrossRef]
3. M. Alam Khan, J. C. Sarker, S. Lee, S. C. Mangham, and M. O. Manasreh, “Synthesis, characterization and processing of cubic iron pyrite nanocrystals in a photovoltaic cell,” Mater. Chem. Phys. 148(3), 1022–1028 (2014). [CrossRef]
4. A. Ennaoui, S. Fiechter, H. Tributsch, M. Giersig, R. Vogel, and H. Weller, “Photoelectrochemical energy conversion obtained with ultrathin organo‐metallic‐chemical‐vapor‐deposition Layer of FeS2 (pyrite) on TiO2,” J. Electrochem. Soc. 139(9), 2514–2518 (1992). [CrossRef]
5. A. Ennaoui, S. Fiechter, W. Jaegermann, and H. Tributsch, “Photoelectrochemistry of highly quantum efficient single‐crystalline n ‐ FeS2 (pyrite),” J. Electrochem. Soc. 133(1), 97–106 (1986). [CrossRef]
6. M. Bronold, Y. Tomm, and W. Jaegermann, “Surface states on cubic d-band semiconductor pyrite (FeS2),” Surf. Sci. 314(3), L931–L936 (1994). [CrossRef]
7. R. Sun, M. K. Y. Chan, and G. Ceder, “First-principles electronic structure and relative stability of pyrite and marcasite: Implications for photovoltaic performance,” Phys. Rev. B 83(23), 235311 (2011). [CrossRef]
8. X. Zhang, M. Li, J. Walter, L. O’Brien, M. A. Manno, B. Voigt, F. Mork, S. V. Baryshev, J. Kakalios, E. S. Aydil, and C. Leighton, “Potential resolution to the doping puzzle in iron pyrite: Carrier type determination by Hall effect and thermopower,” Phys. Rev. Mater. 1(1), 015402 (2017). [CrossRef]
9. R. Murphy and D. R. Strongin, “Surface reactivity of pyrite and related sulfides,” Surf. Sci. Rep. 64(1), 1–45 (2009). [CrossRef]
10. J. Puthussery, S. Seefeld, N. Berry, M. Gibbs, and M. Law, “Colloidal iron pyrite (FeS2) nanocrystal inks for thin-film photovoltaics,” J. Am. Chem. Soc. 133(4), 716–719 (2011). [CrossRef] [PubMed]
11. D. Liang, M. Cabán-Acevedo, N. S. Kaiser, and S. Jin, “Gated Hall effect of nanoplate devices reveals surface-state-induced surface inversion in iron pyrite semiconductor,” Nano Lett. 14(12), 6754–6760 (2014). [CrossRef] [PubMed]
12. M. Limpinsel, N. Farhi, N. Berry, J. Lindemuth, C. L. Perkins, Q. Lin, and M. Law, “An inversion layer at the surface of n-type iron pyrite,” Energy Environ. Sci. 7(6), 1974–1989 (2014). [CrossRef]
13. Y. N. Zhang, J. Hu, M. Law, and R. Q. Wu, “Effect of surface stoichiometry on the band gap of the pyrite FeS2 (100) surface,” Phys. Rev. B 85(8), 085314 (2012). [CrossRef]
14. I. N. Yakovkin and N. V. Petrova, “Influence of the thickness and surface composition on the electronic structure of FeS2 layers,” Appl. Surf. Sci. 377, 184–190 (2016). [CrossRef]
15. K. Laajalehto, I. Kartio, and P. Nowak, “XPS study of clean metal sulfide surfaces,” Appl. Surf. Sci. 81(1), 11–15 (1994). [CrossRef]
16. J. P. Perdew, K. Burke, and M. Ernzerhof, “Generalized gradient approximation made simple,” Phys. Rev. Lett. 77(18), 3865–3868 (1996). [CrossRef] [PubMed]
17. G. Kresse and J. Furthmüller, “Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set,” Phys. Rev. B Condens. Matter 54(16), 11169–11186 (1996). [CrossRef] [PubMed]
18. S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, and A. P. Sutton, “Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study,” Phys. Rev. B 57(3), 1505–1509 (1998). [CrossRef]
19. C. G. Van de Walle and R. M. Martin, “Theoretical study of band offsets at semiconductor interfaces,” Phys. Rev. B 35(15), 8154–8165 (1987).
20. J. M. Guevremont, A. R. Elseinow, D. R. Strongin, J. Bebie, and M. A. A. Schoonen, “Structure sensitivity of pyrite oxidation; comparison of the (100) and (111) planes,” Am. Mineral. 83(11–12), 1353–1356 (1998).
