Materials employed to harvest sunlight are commonly recognized to be at a premium when their optical absorption peaks in the visible, extends to the infra-red, is panchromatic, and matched to the solar spectrum. By contrast, natural photosynthetic absorbers such as chlorophylls and carotenoids display absorption spectra with narrow peaks for yet unknown evolutionary reasons. Beyond such general observations, a rigorous treatment of sunlight harvesting optimization is still lacking.
In this talk, we provide a quantitative analysis of optimal solar energy harvesting in materials. We show a procedure to derive optimal absorption spectra for sunlight harvesting as a function of absorber thickness, and elucidate the concept of solar-matched absorption and its applicability limits. We demonstrate that the shape of the optimal absorption spectrum depends on the thickness and geometry of the absorber, and provide optimal absorption spectra for all thickness regimes. In addition, we define a procedure to rank photovoltaic materials for sunlight harvesting. We close our talk with a possible explanation of why absorption in plants photosynthetic pigments occurs in narrow energy windows.

Novel, earth abundant materials are receiving increasing research interest as potential solution to the terawatt challenge. However, serious efforts to develop a photovoltaic technology based on alternative materials will require a solid proof of concept. The aim of the rapid development project is to accelerate the process from materials discovery, design, and device integration, by combining materials theory and combinatorial synthesis, characterization, and device fabrication. The role of theory is to (i) provide an initial assessment of basic photovoltaic properties of a candidate material, (ii) to explore novel materials structures and compositions, and (iii) to aid the device integration by addressing the issues related to the absorber/contact combination. This presentation will review recent computational studies in support of the rapid development, including Cu-Sn-S compounds, SnS-based alloys, and potential oxide photovoltaic materials systems.

The search for new materials for efficient solar energy conversion is one of the most active research areas globally. A material for such purpose should fulfill two basic requirements: it should be visible light sensitive and the carrier mobility should be high in the material. Many oxide materials with cations having a d0 or d10 electronic configuration can be rendered conductive with high carrier mobility but they do not absorb visible light due to large band gaps, e.g. transparent conducing oxides. In contrast, oxide materials with cations having partially filled d electrons can absorb visible light but the mobility of photo-generated carriers in these materials is rather low due to the localized character of d orbitals (flat bands). It is therefore desirable to combine the merits of the two categories of materials.
By hybrid density functional theory calculations, it is shown that materials with low band gaps and highly dispersive band edges can be realized in layered A2BB&’O6 perovskite oxides by utilizing the two dimensional character of the t2g conduction bands. La2MnTiO6, La2NiTiO6, Sr2FeNbO6 with alternating BO2 and B&’O2 layers along the [001] direction exhibit band gaps smaller than 2 eV. The valence band edge in each compound is composed of states from the element with partially filled d states, i.e. Mn2+, Ni2+ or Fe3+, whereas the conduction band edge is composed of highly dispersive unoccupied t2g states of the d0 element, i.e. Ti4+ or Nb5+. This means their band gap excitations are B → B&’ charge transfer transitions which are expected to result in spatially separated electron and hole pairs that migrate in separate layers. Furthermore, we predict that these layered A2BB&’O6 pervoskite can be synthesized under high pressure.
This work was partly supported by the Japan Science and Technology Agency (JST) Precursory Research for Embryonic Science and Technology (PRESTO) program.

In light of the dramatic price reduction of crystalline silicon photovoltaic modules and the pressing need to address anthropogenic emissions, there is a renewed emphasis on accelerating efficiency increases in candidate Earth-abundant PV materials. These novel materials may serve as top-cell materials for silicon-based tandems, or as single-layer thin-film absorbers.

During the past three years, we have investigated electrochemically deposited cuprous oxide (Cu2O) and PVD/pulsed-CVD-deposited tin sulfide (SnS) for photovoltaic applications. With the benefit of hindsight, we highlight the critical investigations that identified and successfully engineered performance-limiting mechanisms in Cu2O and SnS-based devices, resulting in efficiency improvements from ~1.3% range to ~4.5% and beyond. The parallel evolution of these two material systems suggests certain elements of a common framework to improve the device efficiencies of novel PV materials.

Bulk engineering: We discuss efforts to improve bulk collection length, including approaches for bulk carrier-collection length measurements, effects of deposition conditions and post-growth annealing on collected current, bulk impurity assessment, and efforts to simulate point-defect evolution using kinetic models.

Surface engineering: We review the effect of interface recombination on Voc, the role and properties of native oxides, and grain-orientation specific properties in polycrystalline materials.

Device engineering: Often underappreciated in materials investigations, it is often necessary to resolve baseline device loss mechanisms before "unmasking" more detailed materials science. We review the effects of series and shunt resistance, optical interference, efforts to improve device reproducibility, and the role of advanced device characterization techniques coupled with simulation.

Silicon is the second most abundant element in the earth&’s crust, existing naturally within various oxygen-rich minerals. In addition to its low cost and abundance, silicon maintains a native oxide layer and is readily dopable by other elements. These advantages help silicon-based technologies to maintain a significant market share of photovoltaic devices, despite fundamental light absorption limitations arising from the indirect nature of the band gap. We report on the synthesis of a new low-density, orthorhombic allotrope of silicon, Si24. Si24 was formed by thermal “degassing” of sodium from a Na4Si24 precursor that was formed at high pressure. Si24 contains nano-porous channels along the crystallographic a-axis that are formed from six- and eight-membered sp3 silicon rings, which facilitates the sodium removal when the material is subjected to thermal degassing. Powder X-ray diffraction and energy dispersive spectroscopy analyses confirmed the high purity of the new phase. Si24 is dynamically stable at ambient pressure and is energetically more favorable than known metastable silicon allotropes. Using Density Functional Theory and quasiparticle (G0W0) calculations we estimate a direct band gap of 1.34 eV and an indirect band gap 1.30 eV. These calculations were corroborated by electrical resistivity and optical reflectivity measurements. This quasidirect band gap near 1.3 eV is close to the theoretically suggested optimal value for solar-energy conversion applications, namely the Shockley-Queisser limit. Si24 therefore shows potential for cost-effective and high-efficiency solar energy-conversion technologies.

Heterostructure interfaces can play a decisive role in the optoelectronic performance of earth abundant semiconductor absorbers in thin film photovoltaic applications. I will discuss examples from cuprous oxide, zinc phosphide photovoltaic, and Zn-IV nitride semiconductor heterostructures. We examine the dramatic dependence of open circuit voltage on interface stoichiometry at zinc oxide/cuprous oxide interfaces, and the effect of heterostructure design on current transport interfaces between zinc phosphide and ZnS, ZnSe, ZnO and CdS. Approaches to and prospects for low recombination activity interfaces and carrier selective contacts in earth abundant materials will be discussed.

Thin-film solar cells using Earth-abundant materials are promising for renewable energy applications due to their low material cost and usage compatible with terawatt-level deployment, and inexpensive manufacturing potential. Cuprous oxide (Cu₂O) is a p-type Earth-abundant semiconductor with over 20 % single-junction theoretical maximum power conversion efficiency. In particular, its large bandgap of 2 eV is suitable for a top-cell in tandem solar cells. However, due to non-ideal band alignments and interfacial defects, the power conversion efficiency of Cu₂O-based heterojunction devices has remained low. The non-ideal heterojunction results in a significant energy loss by interface recombination, reducing the open circuit-voltage.
In this contribution, we increased the efficiency of Cu₂O-based thin film solar cells to over 4 % by atomic layer deposition (ALD) of Ga₂O₃ buffer layers. We designed and synthesized ~10-nm-thick buffer layers that mitigated the interfacial recombination problem by improving band-alignments across the heterojunction. The buffer layers were deposited by ALD on electrochemically-deposited Cu₂O layers prior to ALD of ZnO:Al. The buffer layer growth conditions were also engineered to minimize the density of Cu²⁺-related defects at the buffer layer/Cu₂O interface. The band-alignment and the heterojunction quality were characterized by x-ray photoelectron spectroscopy and capacitance-frequency (C-f) measurements, respectively. The Ga₂O₃ buffer layer showed reduced conduction-band offset relative to the Cu₂O layer, comparing with that of the amorphous zinc-tin-oxide buffer layer we used previously.[1] We studied the device impacts of the buffer layer by measuring the J-V characteristics and quantum efficiency. We report an enhanced efficiency by increased open-circuit voltage exceeding 0.9 V
[1] Y. S. Lee et al., Energy Environ. Sci., 6, 2112-2118 (2013)

Thin film solar cells consisting of earth-abundant and non-toxic materials were made from atomic layer deposition (ALD) of Zn(O,S) as the n-type buffer layer, and pulsed chemical vapor deposition (pulsed-CVD) of SnS as the p-type absorber layer using an inexpensive Sn precursor. A thin oxidized passivation layer at the SnS/Zn(O,S) interface significantly improves the open-circuit voltage above 300 mV and fill factor above 50%, and record power conversion efficiencies over 3% were achieved. Solar cell performance is further optimized by adjusting the oxygen to sulfur ratio of Zn(O,S) [1,2,3] and by in situ nitrogen doping using ammonium hydroxide to serve as both oxygen and nitrogen sources. [1] H. H. Park, R. Heasley, and R. G. Gordon, Appl. Phys. Lett.102, 132110 (2013). [2] P. Sinsermsuksakul, K. Hartman, S. Kim, J. Heo, L. Sun, H. H. Park, R. Chakraborty, T. Buonassisi, and R. G. Gordon, Appl. Phys. Lett.102, 053901 (2013). [3] L. Sun, R. Haight, P. Sinsermsuksakul, H. H. Park, and R. G. Gordon, Appl. Phys. Lett.103, 181904 (2013).

Thin film solar cells based on CdTe and CIGS have reached solar to power conversion efficiencies in the range of 20% after an extended development time. However, due to the criticality of some of the involved chemical elements there is a demand for alternative materials solution with similar conversion potentials. But so far most of the suggested alternatives are strongly limited in their performance. We will try with this contribution to deduce design principles for novel thin solar cells based on a systematic comparison of established and novel absorber materials.
Basis of this comparison are p-i-n structures of highly absorbing compound semiconductors which from a theoretical point of view are promising device structures for thin film solar cells. At first the device properties of CdTe and CIGS solar cells will be analysed with respect to the needs for obtaining high conversion efficiencies. As two probably most challenging duties we have identified the control of nucleation and growth also at low substrate temperatures as well as the engineering of interface properties. We will compare typical results which are considered to govern the success of CIGS and CdTe to the properties which have been obtained in the past for FeS2 and layered semiconductors like WS2. Recently we have also investigated novel absorber layers as e. g. SnS and previously studied layers of Cu2S. First results on the preparation and characterisation of thin films and simple devices of the materials will be presented and discussed.
It seems evident that in most cases the bulk properties of the deposited materials are already very promising when higher annealing temperatures can be applied avoiding a detrimental restructuring of the crystallites in the film forming pin-holes or grain boundary shunts which in principle could lead to reasonable conversion efficiencies. However, in most cases the contact materials are not appropriate and lead to only low built-in potentials due to the Fermi level pinning. For the electronic passivation of surface/interface states specifically engineered interfaces are needed as can be deduced from a comparion to successful devices.

A potentially low cost method for making Cu₂ZnSnS₄ (CZTS) or Cu₂ZnSnSe₄ (CZTSe) absorber layers for solar cells is annealing of films cast from colloidal dispersions of CZTS nanocrystals (NCs). Annealing such films in sulfur or selenium can transform them into polycrystalline thin films with micrometer size grains. Realization of inexpensive and high-efficiency CZTS and CZTSe solar cells via this approach requires a fundamental understanding of the effects of key annealing parameters on the evolution of film microstructure. To this end, we have studied the extent to which parameters such as S and Se vapor pressures, presence of alkali metal impurities, carbon content in the NC coating, annealing time, and annealing temperature affect the CZTS microstructure evolution. Films were cast from CZTS NC dispersions on substrates such as quartz, soda lime glass (SLG), Mo-coated quartz and Mo foil. NC films were than annealed in evacuated and sealed quartz ampoules containing known charges of solid sulfur and sometimes trace amounts of Na. The microstructure of annealed CZTS films is determined by three competing processes: densification by sintering of NCs, coarsening of NC grains by normal grain growth and surface energy driven abnormal grain growth. The NCs within the film grow and sinter, while abnormal grains nucleate on top of the nanocrystalline film and grow as large as 10 microns. Abnormal grain growth is driven by the high surface energy of the nanocrystals. The abnormal growth is enhanced at high temperatures, while the normal grain growth within the nanocrystalline film is more pronounced at high S vapor pressures and in the presence of alkali metals such as Na and K. Alkali metals exist as intentionally added impurities in SLG substrates and can diffuse into the CZTS film during annealing. Alternatively, controlled amounts of Na can be introduced into the annealing ampoule and eventually into the CZTS film via vapor phase transport. Either way, the presence of Na enhances grain growth. Larger grains are obtained on SLG than on quartz. The microstructure of the films annealed on SLG could be reproduced on quartz when Na is introduced into the CZTS films via the vapor phase. When the film is annealed in Se, its microstructure is different than that obtained when it is annealed in S. Annealing films in selenium can lead to the formation of a continuous layer of 2-5 µm size CZTSSe grains on quartz. The degree of Se incorporation depends on the annealing temperature and pressure. Annealing in Se also leads to the formation of a C and Se rich layer at the CZTS film-substrate interface. This segregation of C at the CZTSSe-substrate interface is commonly ascribed to the immediate formation of a capping/blocking layer of CZTSSe grains. However, we found that a continuous layer of CZTSSe grains is not necessary to observe C segregation to the film-substrate interface. In contrast, films annealed with S do not show such distinct C-rich layers.

Cu₂O is an earth abundant semiconductor that has been identified as a promising photovoltaic material due to its high absorption and long minority carrier diffusion length. Cu₂O is intrinsically p-type and thus requires a heterojunction partner to enable charge separation, which introduces the possibility of increased interface defects. Cu₂O devices are highly susceptible to deleterious interface reactions which impede high cell performance. The highest reported photovoltaic efficiency for a Cu₂O absorber device is 5.38%, whereas the detailed balance limit has been calculated to exceed 20%. One of the main reasons for low efficiency is instability of the Cu₂O surface with respect to oxidation or reduction.
ZnS/Cu₂O heterojunctions exhibit Type I band alignment and a clean stoichiometric interface can be achieved between Cu₂O and ZnS by sputtering. However difficulties in doping ZnS as well as a large conduction band offset measured for ZnS/Cu₂O band alignment preclude use of ZnS as an emitter, but make it a promising interfacial passivation layer for Cu₂O/ZnO heterojunctions, which have been studied extensively. It has been shown that the technique used to deposit the buffer layer and heterojunction partner has a significant effect on device performance and cleanest interfaces produce highest efficiency devices. One way to achieve a cleaner heterojunction interface is to deposit the interface in situ by plasma-assisted Molecular Beam Epitaxy (PA-MBE), which allows precise interface control in an ultra-high vacuum (UHV) environment. In this work we deposit phase-pure Cu₂O films on single-crystalline MgO substrates, followed by Zn-VI layers. The entire stack is deposited without breaking vacuum in order to achieve the cleanest possible interface. The heteroepitaxial orientation relationship is (110) Cu₂O on (110) MgO and (110) Cu₂O on (100) MgO, as determined by in situ Reflection High Energy Electron Diffraction (RHEED) and ex situ X-ray Diffraction (XRD). ZnO films exhibit (0001) texture. Phase purity of the interface is confirmed by X-ray Photoelectron Spectroscopy (XPS). The interface structure and defect landscape are further elucidated by cross-sectional Transmission Electron Microscopy (TEM).

Copper sulfide, an abundant, non-toxic p-type semiconductor material, has several stoichiometric and non-stoichiometric phases from Cu₂S to CuS. The tunable electrical and optical properties of this material make it a potential candidate for photovoltaic and plasmonic device applications, yet some phases have been reported as unstable, particularly as nanoscale materials. Material stability is an important requirement for device reliability. In this work, we investigate the electrical and optical properties, as well as the stability of M13 viral-templated copper sulfide nanowires synthesized under room temperature and pressure. The filamentous geometry of the M13 virus, approximately 880 nm in length and 6 nm in diameter, is particularly well-suited for templated nanowire synthesis. Electrostatic interactions between the phage template and aqueous-based chemical precursors were utilized to synthesize nanocrystalline Cu_1.8S along the length of the phage. Films of phage-templated Cu_1.8S were prepared by drop-casting the synthesis product onto Au/Ti contact pads. The film thickness was highly non-uniform due to phage agglomeration and bundle formation during synthesis. The measured film thickness ranged from 120 nm to 5.8 mu;m with a median of 1.1 mu;m. In thinner regions of the film (<500 nm in thickness) the morphology became porous and phage bundles were easily identifiable. To study the electrical response and stability of the synthesized material, electrical resistance was measured for freshly made and 3 day aged samples under two different conditions: (1) ambient and (2) low oxygen and low humidity (< 0.1 ppm) conditions. Samples exposed to ambient conditions showed nearly a factor of 40 decrease in resistance while the resistance of samples stored under low oxygen, low humidity conditions only decreased by a factor of 13. As synthesized, the UV/Vis/IR absorption spectrum showed strong absorption at wavelengths below 800 nm due to band-to-band transition and a small, but broad peak in the infrared (IR) region attributable to localized surface plasmon resonance (LSPR) caused by free carriers. With prolonged exposure to ambient conditions, the height of the LSPR peak grew corresponding to an increase in free carrier concentration, which is consistent with the decrease in resistance observed with electrical measurements. XPS analysis revealed that after 3 days of exposure to ambient conditions the surface of the synthesized material developed a mixture of Cu deficient compositions and various oxides. Cu ions within these Cu deficient phases were converted to higher oxidation states, increasing the concentration of free carriers, specifically holes. These studies are a first step towards understanding the electrical characteristics and stability of phage-templated Cu_1.8S material, both of which are critical to future device applications.

The solar-assisted hydrogen evolution from the water splitting can provide susceptible and green source of energy. Here, we report metal oxide core-shell nanowires as photoanodes for a water-splitting photoelectrochemical (PEC) cell. The core-shell nanowires were synthesized by chemical and heat treatment on metal oxide (WO_x) nanowires grown by chemical vapor deposition (CVD). The metal oxide core-shell nanowires showed enhanced PEC performance compared to original metal oxide nanowires due to increased light absorption in the visible. Also, the core-shell nanowires exhibit excellent chemical stability in a wide pH range and long-term stability against photocorrosion. Furthermore, plasmon-assisted photocurrent generation using the core-shell nanowires decorated with plasmonic nanoparticle will be discussed.

Thin films of Sb₂S₃(50 nm) and SnSe (200 nm) produced by thermal evaporation are integrated into a solar cell structure SnO₂:F/CdS(90 nm)/Sb₂S₃/SnSe/C-paint, which shows V_oc of 550 mV, J_sc of 3.6 mA/cm², FF of 0.38, and conversion efficiency of 0.75 %. Here, SnO₂:F is a commercial transparent conductive oxide (TCO) coating (TEC-15, Pilkington). A CdS thin film of 90 nm in thickness is deposited on TCO substrates using chemical bath deposition at 80 °C during 1 h giving a specularly reflective film of optical band gap (E_g) 2.4 eV and photoconductivity σ_p, 0.005 (Omega; cm)^-1. The Sb₂S₃ thin film is deposited by evaporating commercially available Sb₂S₃ powder (Sigma-Aldrich, 99%) at 10^-5 Torr on to the glass/SnO₂:F/CdS substrate at 400 °C. After such a substrate heating, the photoconductivity of CdS is found to increase to 0.1 (Omega; cm)^-1. The Sb₂S₃ film produced by the thermal evaporation has an E_g of 1.8 eV and σ_p of 10^-6 (Omega; cm)^-1. Subsequently, SnSe thin film is deposited by evaporating SnSe precipitate obtained in the laboratory from chemical bath containing SnCl₂.2H₂O, triethanolamine, sodium selenosulfate (Na₂SeSO₃), sodium hydroxide (NaOH), and polyvinyl pyrollidone (PVP). The optical absorption coefficient of the SnSe thin film in the visible region is > 10⁵ cm^-1 and the E_g, 1.19 eV. Thermoelectric measurement on the SnSe film shows p-type conductivity with a thermoelectric power of 450 mu;V/K that suggests a carrier concentration of asymp;10¹⁶ cm^-3. This film is photoconductive, with σ_p 10^-3 (Omega; cm)^-1. Optimization of the thickness as well as the thermal processing of the individual films is expected to lead to improved solar cell parameters.

Chemical deposition followed by thermal treatments to obtain thin films of crystalline silver antimony chalcogenides, AgSbS₂ and AgSbS_2-xSe_x, will be reported. Amorphous thin films are deposited on glass substrates kept for 1 h 15 min in a solution containing antimony trichloride (SbCl₃), sodium thiosulfate (Na₂S₂O₃) and silver nitrate (AgNO₃) maintained at 40 °C. By heating these films in nitrogen at 180-320 °C, we have obtained thin films of crystalline cubic-AgSbS₂. The thin film of AgSbS₂, thus produced, has an optical band gap E_g = 1.89 eV and photoconductivity σ_ph = 1.2x10^-5 Omega; ^-1 cm^-1. When the films are heated above 320 °C in a similar condition, decomposition occurs, and presence of elemental Ag and Sb is observed along with AgSbS₂ in such films. The heat treatment of the amorphous film in presence of Se-vapor results in the formation of silver antimony sulfide selenide films, AgSbS_2-xSe_x. XRD analysis confirms the formation of the solid solution AgSbS_1.25Se_0.75 or AgSbSe₂ depending on the extent of Se-vapor available during the heat treatment. The crystalline grain diameter is 15 nm in both the films. These films show optical absorption coefficients α, 10⁵ cm^-1 in the visible region; E_g of 1.46 eV for AgSbS_1.25Se_0.75 and of 1.1 eV for AgSbSe₂.The photoconductivity (p-type) σ_ph is 2.7x10^-2 Omega;^-1 cm^-1 of AgSbS_1.25Se_0.75 film. Thus the optical and electrical properties of the AgSbS_1.25Se_0.75 film are suitable for being a good absorber in solar cell. Presently, TCO(SnO₂:F)/CdS/AgSbS₂/C cells show open circuit voltage, V_oc, 600 mV; short circuit current density, J_sc, 0.8 mA/cm²; and fill factor, FF, 0.53. Having heated the cell with the C-electrode at 240 °C in nitrogen atmosphere to promote crystallization of the absorber layer, the characteristics reported here are stable. We will include results on TCO(SnO₂:F)/CdS/AgSbS_1.25Se_0.75/C cells as well, and illustrate a wide range of solar cell absorber materials for thin film and hybrid solar cell technology.

Semiconductor nanocrystals (NCs) are promising building blocks for next generation photovoltaic devices. FeS₂ (pyrite phase) NCs have been considered as a superior semiconducting material for solar energy conversion and photo-electrochemical applications since it possesses an appropriate band gap of ~1 eV and a very high absorption coefficient (α >10⁵ cm^-1). These characteristics coupled with the low cost, environmental compatibility, and abundant elements in the crust make pyrite to be a potential candidate for solar cell absorption layer materials in the form of ultrathin films (< 100 nm). NCs coating technique, which requires a well-dispersed NC ink, is simple and low-cost compared with the traditional vacuum processes for depositing pyrite thin films. However, as the effect of surface energy of NCs is significantly higher than that of larger particles, NCs will tend to agglomerate and disable the formation of continuous, smooth ultrathin films.
In this report, well-dispersed, stable pyrite NC inks were successfully produced by beads-milling technique and smooth pyrite thin films less than 50 nm were produced by spin coating the dispersed pyrite NC inks on various substrates. The absorption coefficient of coated thin films increased as annealing temperature increased. Hall measurements show that these films were P-type. The room-temperature resistivity decreased and hall mobility increased after 500 ^oC annealing. These films can be used for not only traditional photovoltaic devices but also flexible devices.
The grazing incidence X-ray diffraction patterns (GIXRD) show the pyrite thin films changed to amorphous phase when annealing temperature was less than 350 ^oC. However, the thin films maintained pyrite phase with increased crystallinity when annealed above 400 ^oC. Meanwhile, the small amount of marcasite phase coexisted with the pyrite phase (major phase). These phenomena were observed for the first time. A model related to the phenomena will be proposed in this report. Other analyses were also utilized to study the pyrite thin films in this report. Fourier transform infrared (FTIR) was used to determine if there is any organic residue after annealing. The band gap of pyrite thin films was obtained by Tauc plot extracted from UV-Vis spectrum. The surface morphology and thickness of pyrite thin films were acquired from field emission scanning electron microscopy (FESEM).

Energy generation from non-fossil fuels has been rapidly accelerating to meet the growing global energy demand. Solar-based technologies are one of the major contributors towards meeting the energy needs. As compared to conventional silicon-based solar cells, thin film solar cell panels are lightweight and flexible and are thus preferable for a number of applications. In particular, thin films solar cells fabricated using CdTe and CuInGaSe2 (CIGS) are already widely used. Despite the success, these materials are composed of toxic and less-abundant elements such as tellurium, indium and gallium. An attractive alternative being extensively investigated is Cu2ZnSnS4 (CZTS), consisting of non-toxic and relatively abundant elements, with a large absorption coefficient (~ 1 x 104 cm-1) for solar radiation. In view of identifying other materials composed of sustainable and non-toxic elements for thin film solar cells, we have been investigating Cu2FeSnS4, CuCdxZn1-xSnS4 and Cu-Sb-S systems. We have developed solution state syntheses methods for phase, size, and shape-selective growth of Cu2FeSnS4, CuCdxZn1-xSnS4 and Cu-Sb-S nanocrystals. The optical and electrical properties of the nanocrystals have been studied. In particular, the copper-antimony-sulfide system consisting of four major phases, namely CuSbS2 (Chalcostibite), Cu12Sb4S13 (Tetrahedrite), Cu3SbS3 (Skinnerite) and Cu3SbS4 (Femitinite), have been investigated in detail. All four phases are p-type semiconductors and their energy band gap lie between 0.5 and 2 eV, with large absorption coefficient values over 105 cm-1. We have developed simple colloidal hot-injection methods for the phase-pure synthesis of nanocrystals of all four phases (CuSbS2, Cu12Sb4S13, Cu3SbS3, Cu3SbS4) and carried out detailed electronic structure calculations in order to access their suitability for solar energy applications. The details of the syntheses methods, structural and optical characterizations and band structure calculation will be presented.

Thermal processing in presence of Se powder under nitrogen at 350 - 400 ^oC of chemically deposited thin film stack SnSe-ZnSe-Cu_2-xSe precursor produces thin film of Cu₂ZnSnSe₄ (CZTSe). To develop Cu₂ZnSnSe₄, we used chemically treated glass substrates to deposit compact and specularly reflective SnSe thin films of 300 nm from a bath containing tin(II) chloride, triethanolamine, sodium hydroxide, sodium selenosulfate, and a small quantity of polyvinylpyrrolidone. The deposition bath was maintained at 26 ^oC. As deposited, the SnSe thin films have orthorhombic crystal structure, an optical band gap E_g of 0.95 - 1.18 eV, and a p-type electrical conductivity of 1 #8486;^-1 cm^-1. Thin films of ZnSe and Cu_2-xSe are deposited on the SnSe films using chemical baths reported previously from our group [1, 2]. The films of CZTSe produced though heating the stack have tetragonal structure with lattice parameters of a = 5.68 Å and c = 11.92 Å and a crystallite diameter of 16 nm. The films are p-type, with electrical conductivity, 0.2 (Omega; cm)^-1and E_g of 1 eV. As preliminary results, in a CdS/SnSe thin film solar cell structure, we have observed an open circuit voltage, V_oc, of 225 mV; short circuit current density, J_sc, of 1.75 mA/cm² under an illumination of 850 W/m². The paper will deal with the morphology, composition, and other properties of the thin films of SnSe and CZTSe and results on solar cells using these two absorbers.
[1] V. M. García, M. T. S. Nair, and P. K. Nair, Semicond. Sci. Technol. 14 (1999) 366
[2] V. M. García, P. K. Nair, and M. T. S. Nair, J. Cryst. Growth, 203 (1999) 113

The stability of Cu₂ZnSnS₄ (CZTS) thin films during synthesis and annealing at high temperatures (400-600 ^oC) is an important issue as it affects the chemical and phase composition of the film. For example, if the appropriate vapor phase species are not present in sufficient pressure, CZTS may decompose into binary sulfides and volatile SnS may leave the film, changing the film stoichiometry. The vapor phase composition over the CZTS film also affects the microstructure of the film, as we recently showed that impurities like Na and K may be volatilized from soda lime glass substrates. Widely used open-system annealing and sulfidation processes further complicate these effects as the vapor species and pressures change over time, thus a closed-system synthesis approach is well suited for studying vapor-film interactions. To this end, we synthesized CZTS films via ex situ isothermal sulfidation of co-sputtered Cu-Zn-Sn thin films on various substrates in a sealed quartz ampoule. The precursor film composition is controlled by varying the sputtering power of Cu, Cu-Zn, and Cu-Sn targets. The sulfidation ampoule with the precursor film is charged with varying amounts of S, Na and K, evacuated to a 10^-6 Torr base pressure, sealed, and annealed at different temperatures and periods of time. The sulfidized films are characterized using a suite of methods including X-ray diffraction, confocal Raman spectroscopy and imaging, scanning electron microscopy, energy dispersive x-ray spectroscopy, and temperature dependent conductivity and Hall effect measurements. We explored a wide range of precursor metal compositions and found that the creation of SnS vapor was inevitable. While Sn loss through the formation of this phase has been well documented and is generally associated with CZTS decomposition, we found it to be essential during formation. During sulfidation, Cu_2-xS, ZnS, and SnS form first. The volatility of SnS results in the loss of the vast majority of the Sn initially present in the film at temperatures as low as 300 oC. However, the SnS that escapes into the vapor phase is reincorporated into the film by reacting with the solid Cu_2-xS remaining in the film to form the more stable Cu₂SnS₃ phase. Reincorporation of Sn is terminated when all of the Cu_2-xS is consumed, and The Cu₂SnS₃ is later converted to CZTS through ZnS incorporation. Any excess SnS remaining in the vapor eventually condenses on the ampoule walls during cooling. This mechanism results in the self-regulation of Sn concentration within the film, such that the Cu/Sn ratio is always 2 for a very wide range of precursor film compositions. If the precursor film is Sn deficient, Cu_2-xS remains in the final film, though the Sn deficiency can be corrected through the addition of elemental Sn in the sulfidation ampoule, creating additional SnS vapor. We find that these vapor-solid interactions are independent of substrate and whether or not Na and K are also present in the vapor phase.

Interest in Cu2ZnSn(S,Se)4 (CZTSSe) as a sustainable material for use in thin-film photovoltaics (PV) has increased rapidly in recent years. In addition, low-cost solution-based CZTSSe thin-film deposition methods that offer attractive scalability and handling characteristics have also attracted considerable attention. Of particular interest is the two-step process via selenization of nanoparticle precursors, which has been demonstrated for the fabrication of highly efficient PV devices by various groups. One of the successful nanoparticle routes is based on a mixture of binary and ternary chalcogenide nanoparticles, where efficiencies of 8.8% without antireflection coating have been achieved. Previously, we have shown that following high-temperature selenization, the binary and ternary nanoparticle precursor film transforms into an intriguing bilayer structure consisting of a large-grain CZTSSe layer on top of a fine-grain carbon-rich layer. Detailed investigations have shown that the large-grain CZTSSe layer is the active layer in the resulting device and that its properties control device performance. Improved understanding and control of the bulk and interface properties of the large-grain CZTSSe layer could lead to further improvements in efficiency. In order to devise strategies to improve the material properties of the large-grain CZTSSe layer, a clear understanding of the mechanism of formation of the bilayer structure during the selenization process is needed. In this paper, we will present results from detailed mechanistic studies on the selenization of binary and ternary nanoparticles. Selenization runs were interrupted at different stages of the annealing process, and samples were examined in detail by a combination of microscopy and spectroscopy techniques. Based on the experimental results, a growth model for the formation of the bilayer CZTSSe thin film from the binary and ternary nanoparticle precursor layer is proposed.

The ability of absorbing an incident solar spectrum efficiently over a range of angles into a planar dielectric film for absorption, conversion or optical transmission requires improvement from antireflective (AR) films with discrete layers to graded index films with near “ideal” AR performance. Such graded films can also enable transmission of optical energy to couple devices with mismatching numerical aperture (NA), size or shape. High transparency allows in film waveguide transport after AR coupling. We present the simulation, deposition and characterization of vacuum thin films with the required range of refractive index and low optical absorption for the design of an optical coupler that will concentrate the sunlight into a thin film wave guide or an optical fiber with minimal loss, by virtue of the ability of varying the material deposited both laterally and axially. Modeling results show an optimum design for such graded index profiles using pulsed DC reactive magnetron sputtering with substrate bias for the deposition of the films. Optical ellipsometry, Scanning Electron Microscopy (SEM) and Energy-Dispersive x-ray Spectroscopy (EDS) were used for their characterization. These capabilities provide sub-wavelength antireflective coating to maximize the coupled light with in film waveguide transport including an adiabatic mode converter to couple the light to a subsequent waveguide or optical fiber for transmission, conversion or thermal storage and use at high concentration.

Even though extraordinary progress has been made in the field of Si and Cu(In,Ga)(S,Se)2 (CIGS) solar cells/modules over the years and their performance significantly improved, fabrication/material costs and safety still remain major concerns. In terms of material design for compound semiconductors, the most cost-effective, safe, and easy-handling anions are sulfur and oxygen. Nowadays, to obtain compound semiconductors by reacting inexpensive metal cations (precursors) with gas-phase anions has become a significant issue. For realizing this, sulfurization and oxidation are appropriate physical processes because they are based on simple thermal diffusion of anions into a metal precursor. In fact, sulfurization/selenization techniques are adopted for commercially manufacturing CIGS solar modules because of the cost effectiveness, scalability, and uniformity of such techniques.
Tin monosulfide (SnS) has a direct energy bandgap of 1.3 eV and a high optical absorption coefficient of 104 cmminus;1. Therefore, SnS is perceived to be a promising candidate as a cost-effective and earth-abundant inorganic material for use in the fabrication of next-generation solar cells. Although the theoretical conversion efficiency of SnS-based solar cells is high, the demonstrated efficiencies of such cells are still low. Some of the reasons for low efficiencies are considered to be the poor crystal quality of a SnS layer, and the mismatch in the band diagram at a pn-heterojunction. In fact, most researchers simply refer to the CIGS solar cells structure. The growth mechanism of SnS or the appropriate band offset of SnS-based solar cells has not yet been clarified because it shows metastable phase.
Wide-bandgap transparent conducting oxide (TCO) films exhibit only n-type conductivity. On the other hand, NiO shows only p-type conductivity, and is composed of inexpensive and less toxic elements. Recently, NiO has also been used in visibly transparent solar cells, which are attractive because their optical transparency permits greater flexibility in terms of installation locations. In addition, NiO films can be easily obtained just by the oxidation of Ni. We have proposed a “NiO-based invisible sensor” that receives electrical power from a “NiO-based invisible solar cell”. This sensor can be fixed on a wall or ceiling without a wired connection. Moreover, the user will not be aware of these invisible sensors and will not experience the stress associated with being monitored. Therefore, there is a great market opportunity for various environments, and the application possibilities for invisible devices based on NiO-based solar cells are limited only by imagination.
In this presentation, we will introduce some earth-abundant solar absorbers (SnS, Cu2SnS3, Cu2O, and related compounds) and TCOs (NiO, CuAlO2, and related compounds) for use in photovoltaics, with focus on materials research and device development, in terms of the most recent advances and experimental results.

A great deal of interest is being focused on environmentally safe earth abundant materials including iron sulfide (FeS₂) in the pyrite phase. While promising, iron sulfide is presenting serious material science challenges. In addition the band gap of pyrite is much less than the optimal value of 1.36 eV. Another excellent candidate for an earth abundant absorber material is WSe₂ which can be directly grown as a p-type semiconductor with a band gap near 1.4 eV. In this work we present the structural, optical, and electrical properties of thin film WSe2 grown via the selenization of sputter deposited tungsten films. We will show that highly textured films with an optical band gap in range of 1.45 eV, and absorption coefficients greater than 105/cm across the visible spectrum can be easily achieved. In addition we will present Hall Effect and carrier density measurements as well, where will show densities in to 10¹⁶-10²⁰ cm^-3 range and p-type Hall mobilities in the 10 cm²/V-s range can be obtained. We employ these results to numerically simulate heterojunction solar cells based on this material, where we will show efficiencies greater than 20% are possible.

Tin monosulfide (SnS) has attracted considerable interest as a photovoltaic absorber material due to its potential for large-scale, cost-effective and environmentally friendly power generation. During recent years, thin-film solar cells based on SnS have shown substantial progress [1, 2]. Nevertheless, the latest published SnS-based champion cell still reveals a rather poor efficiency of 2.04% [2], far below the theoretical maximum.
We present a first direct comparison of two processing methods for solar cells using SnS as the absorber. We demonstrate thin-film devices based on thermal evaporation (TE) and pulsed chemical vapor deposition (CVD), following the previously developed substrate device stack (Mo/SnS/Zn(O,S)/ZnO/ITO/Ag) [2]. Upon device optimization, our best performing SnS-based solar cells yield new record power conversion efficiencies beyond 4%. We will present current-voltage, quantum efficiency and Suns VOC measurements to better identify loss mechanisms and opportunities for improvement in both methods. We analyze the structural, optical and electronic properties of as-deposited and post-annealed films to shed light onto differences in the device performances between TE and pulsed-CVD processed absorber layers.
In closing, we demonstrate high performing thin-film solar cells based on SnS as the absorber. While pulsed-CVD is a rather slow deposition method, TE enables high throughput processing allowing for industrial scale-up.
1. Ramakrishna Reddy, K., et al., Sol. Energy Mater. Sol. Cells 90, 3041-3046 (2006).
2. Sinsermsuksakul, P. et al., Appl. Phys. Lett. 102, 053901 (2013).

It is recognized that for a thin film solar cell technology to be viable, small area cell efficiency should be near 20% so that a module efficiency of 16% and projected large volume sales price of US$0.50/W are both feasible, which are currently available in commercial Si-solar cells. Among the least experimented alternate solar cell materials is antimony sulfide-selenide, the optical band gap of which is tunable in the 1.0 (Sb₂Se₃)-1.8 (Sb₂S₃) eV region, depending on the composition of the solid solution, Sb₂S_xSe_3-x which forms in the orthorhombic system. The name ‘anti-mony&’ signifies the presence of Sb-element in numerous minerals. The cost of Sb-ingots has remained below US$15/kg for most part of recent years. Native antimony metal, large single crystals of antimony sulfide (stibnite-crystalline Sb₂S₃) weighing in kg or as metastibnite (amorphous Sb₂S₃) are common; but large single crystals of antimony selenide (Sb₂Se₃ antimonselite) are less common.
In the 1990&’s chemically deposited Sb₂S₃ thin films have been reported in heterojunction or Schottky barrier solar cells with 5-6 % efficiency; in 2008 in thin film solar cells with 1%; in 2010 onward in dye sensitized solar cells with 3-6%; and in 2013 in CdS/Sb₂S_1.2Se_1.8/PbSe thin film cells with 2.5% conversion efficiency (eta;). In this work we present an overview of the chemical deposition methods to produce different solar cells: CdS/Sb₂S₃/C; CdS/Sb₂S₃/PbS; CdS/Sb₂S_xSe_3-x/PbS or PbSe, all with efficiency 1-2.5% range. We also report CdS/Sb₂S₃/C and CdS/Sb₂S₃/PbSe solar cells with eta;, 1.5-2.5%, in which case Sb₂S₃ films are obtained by thermal evaporation from either Sb₂S₃ prepared in our laboratory or from a commercial product. The basic feature of all these thin film solar cells is good thermal stability, which is essential to meet commercial prospects.
The limitation so far with our methods is the small crystalline diameters in the absorber film. At present, the best short circuit current density J_sc in these antimony sulfide-selenide cells is in the 10-15 mA/cm² range, but the higher J_sc is accompanied by a lowering of the open circuit voltage V_oc from 650 mV toward 450 mV. Typical cell parameters for isolated cell area of 0.5 cm² in the case of structures obtained by sequential chemical deposition on glass substrates coated with SnO₂:F are: V_oc ,640 mV; J_sc, 8.8 mA/cm²; FF 0.3; and eta;, 1.88% for CdS/Sb₂S₃/PbS-C; V_oc, 486 mV; J_sc, 12.1 mA/cm²; FF 0.44; eta;, 2.48% for CdS/Sb₂S_1.2Se_1.8/PbS-C; and V_oc , 454 mV; J_sc, 12.5 mA/cm²; FF 0.44; eta;, 2.5% for CdS/ Sb₂S_1.2Se_1.8/PbSe-C. With evaporated thin films of Sb₂S₃, we have observed in SnO₂:F/CdS/Sb₂S₃/C, V_oc,586 mV; J_sc, 6.3 mA/cm²; FF 0.42; and eta;, 1.5%, and in SnO₂:F/CdS/Sb₂S₃/PbSe-C, V_oc, 608 mV; J_sc, 10 mA/cm²; FF, 0.38; and eta;, 2.3%. We consider that the low melting point, 550 ^oC of Sb₂S₃ would permit improvement of crystalline grain diameter through post deposition thermal processing and lead to better cell parameters.

Numerical modeling of Hall effect data is used to demonstrate the existence of a conductive inversion layer at the surface of high-quality n-type single crystals of iron pyrite (cubic FeS2) grown by a flux technique. The presence of the inversion layer is corroborated by photoemission spectroscopy and Hall measurements as a function of crystal thickness. This hole-rich surface layer can explain both the low photovoltage of pyrite solar cells and the near-universal high p-type conductivity of polycrystalline pyrite thin films that have together perplexed researchers for the past thirty years. We find that the thickness and conductivity of the inversion layer can be modified by mechanical and chemical treatments of the pyrite surface, suggesting that it is possible to eliminate this hole-rich layer by passivating surface states and subsurface defects. Furthermore, modeling of the high-temperature electrical conductivity shows that the electronic band gap is 0.8 ± 0.05 eV at room temperature (compared to 0.94 eV according to optical transmission data), confirming that photovoltages of ~500 mV should be attainable from pyrite under solar illumination.

Although holding potential as an earth-abundant efficient photovoltaic material with a gap of approximately 1 eV, pyrite (FeS2 ) has so far exhibited in experiments an open-circuit voltage (OCV) of only 0.2 V, rendering it unsuitable for actual PV applications. Absorption experiments show large subgap absorption, commonly attributed to defects or structural disorder. In this talk, we report computations using density functional theory (DFT) with a semi-local functional approximation predict that the bottom of the conduction band consists of a very low intensity sulfur p-band that may be easily overlooked in experiments because of the high intensity onset that appears 0.5 eV higher in energy. The intensity of absorption into the sulfur p-band is found to be of the same magnitude as contributions from defects and disorder. Our findings suggest the need to re-examine the value of the fundamental bandgap of pyrite presently in use in the literature. If the contribution from the p-band has so far been overlooked, the substantially lowered bandgap would partly explain the discrepancy with the OCV. Furthermore, we show that more states appear on the surface within the low energy sulfur p-band, which suggests a mechanism of thermalization into those states that would further prevent extracting electrons at higher energy levels through the surface.
Reference: P Lazicacute; et al 2013 J. Phys.: Condens. Matter 25 465801 doi:10.1088/0953-8984/25/46/465801

FeS2 is a promising next generation Earth-abundant photovoltaic material with a 0.95 eV bandgap well suited to absorption of the solar spectrum and an extremely high absorption coefficient. Yet, its performance to date has been severely hampered by mid-gap electronic defect states that are created by variations in sulfur bonding states at the surface. Bulk FeS2 crystallizes in the cubic pyrite structure with sulfur atoms paired in a sulfur-sulfur dimer. The surface sulfur atoms have a different coordination due to the absence of additional sulfur atoms to form the S-S dimer. This sulfur coordination change results in electronic properties at the surface closer to those of iron monosulfide, which has a bandgap of only ~0.3 eV. In photovoltaic devices this causes excessively high dark current and extremely low open circuit voltage, severely limiting efficiency. We present here the potential of thin inorganic capping films as passivating layers, providing bulk-like coordination at the surface of FeS2 films. Thin films of sputtered FeS2 are capped with various inorganic thin films and compared with uncapped films using a variety of techniques, including x-ray diffraction, transmission electron microscopy, photoluminescence spectroscopy, electron paramagnetic resonance, and x-ray photoelectron spectroscopy (XPS). Notably, we find that ZnS-capped films show a significant decrease in iron monosulfide character as determined by examination of XPS peaks related to the bulk and surface coordinations of sulfur. The result is supported by a density functional theory model that exhibits nearly perfect surface passivation of FeS2 by ZnS. These results are a positive indication that iron pyrite can be a viable next-generation photovoltaic material.

Fe2SiS4 and Fe2GeS4 crystalline materials posses direct bandgaps of ~1.55 and ~1.4 eV respectively and an absorption coefficient larger than 10^5 cm-1. Recent literature reports indicate that the two materials hold theoretical solar cell efficiency potential comparable with CIGS (>20%). Furthermore, they can operate at a 0.1mu;m required thickness to accomplish similar performance with other thin film materials (CZTS, CIGS). However, no solar devices have been reported to date.
In the presented work, nanoprecursors to Fe2SiS4 and Fe2GeS4 have been fabricated and employed to build ultra-thin-film layers via spray coating and rod-coating methods. Temperature-dependent X-Ray diffraction analyses of nanoprecursor coatings show an unprecedented low temperature for forming crystalline Fe2SiS4 and Fe2GeS4.
We are currently working on fabricating ultra-thin-film photovoltaic devices utilizing Fe2SiS4 and Fe2GeS4 as solar absorber material, and results of this work will be presented. This will be a first report of solar devices fabricated with Fe2SiS4 and Fe2GeS4.

For long-term, economically competitive, sustainable solar energy to become a reality, absorber materials for photovoltaics should be selected for their earth abundance and low extraction costs. Iron pyrite, FeS2, has been proposed as an excellent candidate for these reasons, in addition to its reasonable band gap and large absorption coefficient. [1,2] Even though its photovoltaic properties were explored as early as the 1980&’s and a recent resurgence in pyrite research has produced a number of publications, experimental studies have produced only poor results with regards to the photovoltaic properties. Many have attempted to explain the problems plaguing the material, but no experimental data has shown improved performance above the record 3% efficiency. [3]
Recently, a class of materials has been proposed as an alternative to pyrite; Fe2MS4 (M = Si, Ge). Calculations have been used to predict nearly ideal photovoltaic properties (a band gap of 1.40 - 1.55 eV and a large absorption coefficient of 10^5 cm-1). [4] One hypothesis for the poor performance of pyrite is the possible decomposition of the desired phase to other iron-sulfur phases with very small band gaps. According to theory, there are no binary phases in the Fe2MS4 system that are calculated to be more stable than the ternary phase, suggesting it may be a significantly better candidate for solar cells because it may be more thermodynamically stable than the binary Fe-S pyrite phase. Historically, the bulk material has been studied for its interesting magnetic transitions as a function of temperature, but experimental reports of photovoltaic properties are limited. Herein, we report the synthesis of colloidal Fe2GeS4 nanocrystals for use in photovoltaic devices. The as-synthesized nanocrystals are phase-pure and form plate-like structures that show broad absorption in the visible region. While the nanocrystals exhibit poor stability under ambient conditions, methods of surface passivation for improved stability and enhanced photovoltaic properties will be discussed. These results pave the way towards a new earth-abundant material for use in low-cost photovoltaics.
[1] Wadia, C.; Alivisatos, A. P.; Kammen, D. M. Environ Sci Technol 2009, 43, 2072.
[2] Ennaoui, A.; Tributsch, H. Sol Cells 1984, 13, 197.
[3] Ennaoui, A.; Fiechter, S.; Pettenkofer, C.; Alonsovante, N.; Buker, K.; Bronold, M.; Hopfner, C.; Tributsch, H. Sol Energ Mat Sol C 1993, 29, 289.
[4] Yu, L.; Lany, S.; Kykyneshi, R.; Jieratum, V.; Ravichandran, R.; Pelatt, B.; Altschul, E.; Platt, H. A. S.; Wager, J. F.; Keszler, D. A.; Zunger, A. Adv. Energy Mater. 2011, 1, 748.

Cu2ZnSnSxSe4-x (CZTSSe) has drawn world-wide attention due to its promising performance and earth-abundant components. So far, the state-of-the-art CZTSSe thin film solar cells have reached up to 12.0% power conversion efficiency via a hydrazine-assisted deposition approach. However, it is generally difficult to get efficiency above 12% and the PCE of CZTS solar cells is still far below the physical limit, known as the Shockley-Queisser (SQ) limit, of about 31% under terrestrial conditions. In this presentation, a new world record 12.6% CZTSSe thin film solar cell (performance independently certified) will be presented with extensive device characterization. The improvement of device performance is primarily due to the reduction of Voc deficit by ~50 mV. The depletion width xd and diffusion length Ld of the 12.6% CZTS solar cell, as deduced from C-V and voltage-dependence of IQE, are 0.15 ± 0.03 um and 0.75 ± 0.15 um, respectively, which are in good agreement with the 1-mu;m current-collection depth deduced from EBIC. The capacitance analysis indicates a substantial component of interfacial recombination within the record device. These results continue the recent trend of substantial increases in CZTSSe device performance improvements from 6.7% (2008) to 9.7% (2010) and 11.1% (2012), indicating the substantial promise of this new thin-film PV materials system.

Kesterite Cu2ZnSn(S,Se)4 (CZTSSe) thin films are attracting a lot of interest as an alternative system to Cu(In,Ga)Se2 (CIGS) thin films, owing to their majority carrier type (p-type), proper band gap energy (1.0-1.5 eV), and high optical absorption coefficient (> 10^4 cm-1). More promisingly, the CZTSSe is composed of earth-abundant (cf. In in CIGS), environmentally-friendly (cf. Cd in CdTe), and relatively cheap elements. We have recently reported that a highly efficient CZTSSe thin film solar cell can be fabricated via an electrochemical method involving the electrochemical deposition of metallic Cu-Zn-Sn (CZT) precursor thin films and the subsequent sulfurization/selenization process. In this study, we investigated the effect of the electrochemical deposition conditions on the physical and chemical properties of the metallic precursor films and the final CZTSSe thin films. In particular, the oxygen content in the precursor and CZTSSe thin films was found to influence the film morphology and the device performance. By carefully controlling the deposition condition, such as the bath temperature, solution composition, and the cell configuration, the oxygen content in the thin films could be successfully reduced. After some additional optimization processes including the annealing condition, an efficiency exceeding 8% was obtained, which is one of the highest efficiencies reported on the electrochemically deposited CZTSSe thin film solar cells.

Ever since several solution processes started to lead the breakthroughs of Cu2ZnSn(S,Se)4(kesterite) solar cells in the recent couple years, a thriving amount of research efforts have been spent on the exploration of more new solution-based methods to fabricate thin-film solar cells. Along the exploration, new problems are introduced because of the features of solution processes. The engagement of carbon and the requirement of high chalcogen pressure in the annealing treatments are two common features. Nanocrystal approach is one of the most promising routes to demonstrate high cell efficiency without involving major toxic or dangerous solvents. This study identified that the encounter of carbon-containing precursor films to chalcogen vapor has produced a vertical non-uniform composition distribution particularly at the top surface in the vicinity of the junction, which is detrimental for device performance. The observed spatial elemental distributions are likely shaped by the thermodynamic preference of selenium for Sn and Cu over Zn cations. To tackle the undesired composition deviation, we implemented a compensation approach through the use of nanocrystals with an high Zn/(Cu+Zn+Sn) and demonstrated a cell efficiency of 8.6%. The device was fabricated by all-solution process (solution process all the way from absorber to top electrodes), which to our best knowledge is the first report on CZTS solar cell without using vacuum deposition techniques to finish the device fabrication. This result shows a bright potential for kesterite to be manufactured with fully solution processes, in which the investment in vacuum budget is massively reduced.

We present the special device properties of Cu₂ZnSn(S_1-ySe_y)₄ (0le;yle;1) (CZTSSe) based solar cells under low illumination conditions, reporting an unexpected difference in behavior of the quaternary-like compounds (close to S and Se pure compositions), compared with those observed in the CZTSSe solid solutions with comparable S and Se contents. We study the optoelectronic properties of CZTSSe devices with efficiencies ranging from 3% to 6%, using illuminated I-V curves under different irradiance intensities (from 20 to 1000 W/m²). Furthermore, external quantum efficiency curves (EQE) were measured under different white bias light intensities, under different monochromatic bias lights wavelengths at approximately constant intensity (455 nm, 505 nm, 660 nm and 850 nm), as well as under different bias voltages. Complementary conductive atomic force microscopy (c-AFM) studies under different voltages and extra light illumination were also performed with the aim to analyze the role of grain boundaries and bulk properties in these effects. A remarkable increase of the corrected conversion efficiency from 5.6% and 3.3% (measured at 1000 W/m²), to 11.5% and 8.4% (measured at 20 W/m² and corrected to 1000 W/m²), for CZTSe and CZTS based devices respectively, were obtained. This gives a strong interest to these materials for the development of in-door photovoltaic applications related to the need for power supply in small autonomous systems, obtaining efficiency values that are comparable to those achieved in devices already at an industrial stage, and opens new interesting application fields for emerging kesterite based technologies.
The main changes observed in the devices at low illuminations intensity are associated to a strong increase of the shunt resistance. This could be related to the grain boundaries behavior under different light intensities. Conversely, CZTSSe solid solutions behave very different, and the corrected efficiency and the others optoelectronic parameters are almost constant with the illumination. Trying to understand this contrasting behavior, we analyze the EQE curves as a function of bias light intensity and wavelength, and bias voltage. The results confirm the previously observed trends, whereas the EQE for the CZTSSe based devices are almost unaffected for these parameters, those corresponding to pure and nearly-pure CZTSe and CZTS ones are strongly affected. In particular for this last device, EQE decrease can especially been observed when using long wavelength bias lights, whereas short wavelength bias light which is mainly absorbed in the buffer/absorber junction has less effect. This suggests that the light sensitivity is mainly due to the absorber bulk properties. Complementary c-AFM measurements under equivalent illumination conditions support the observed different roles of the grain boundaries. Potential origin of this behavior and the differences between CZTSe and CZTS, with CZTSSe based devices will be discussed.

Copper zinc tin sulfide (CZTS) is an attractive earth abundant and non-toxic material for solar cells, provided the efficiencies of the cells can be increased to beyond its current value of ~8.4%. Provided one is willing to accommodate some non-toxicity, by the addition of Se, solution-processing techniques have demonstrated efficiencies of around 12%. These numbers however need to be >15% or so in order to CZTS or CZTSSe to be viable candidates. Using examples from vacuum deposited material, we will describe some of the materials issues that limits current performance, the most significant among them being an open circuit voltage that is quite low. In order to identify the role of extended defects and grain boundaries on performance, we have also begun a study of epitaxial CZTS grown on near lattice matched silicon substrates by molecular beam epitaxy. It is possible to grow epitaxial films, though there is the formation of three domain variants that arise from a symmetry mismatch between epilayer and substrate. We will discuss the structural, optical and electrical results from the epitaxial films.

The U.S. Department of Energy SunShot Initiative is a collaborative national effort to reduce the price of solar energy to $1/W (5-6cent;/kWh) for utility-scale installations by 2020. This is estimated to enable solar energy to grow from less than 0.2% of the current U.S. electricity supply to roughly 14% by 2030 and 27% by 2050. To enable terawatt-scale deployment, the SunShot Initiative supports a diverse materials science R&D portfolio including research on materials comprised of earth-abundant elements. In this presentation, a selection of the materials science efforts funded by the SunShot Initiative will be highlighted. Furthermore, the future interests of the program related to materials development will be described.

Recently, we reported that a drastic improvement of conversion efficiency in Cu2O-based pn heterojunction solar cells was made possible by stabilizing the surface of polycrystalline p-Cu2O sheets as well as developing a low-temperature and low-damage deposition technology for applying thin films as an n-oxide semiconductor layer.1,2) In this paper, we describe the influence of the thin film/Cu2O sheet interface on the photovoltaic properties of transparent conducting oxide (TCO)/Cu2O heterojunction solar cells fabricated using polycrystalline p-Cu2O sheets. The Cu2O sheets (thickness of approximately 0.2 mm) that act as the active layer as well as the substrate were prepared by thermal oxidization of Cu sheets. Both Al-doped ZnO (AZO) and non-doped ZnO (ZO) thin films, used as the n-TCO thin film layer, were prepared by a pulsed laser deposition method that features an introduction of O2 gas into the deposition camber, either with or without the addition of O3 gas. It was found that the obtainable photovoltaic properties in the resulting n-TCO/p-Cu2O heterojunction solar cells are considerably affected by the pressure of the O2 introduced, either with or without any added O3 gas, during the deposition of the n-TCO thin films. For example, in transparent electrode AZO/AZO/Cu2O or AZO/ZO/Cu2O heterojunction solar cells fabricated using n-TCO thin films (thickness of 50 nm) deposited with various gas pressures, the obtained open circuit voltage (VOC) increased as the gas pressure was increased to 1.2-1.7 Pa, and then it decreased slightly as the pressure was increased further. The change of obtained VOC in the fabricated heterojunction solar cells from approximately 0.43 to 0.73 V was not correlated to that of the work functions (Fermi energy), resulting from a change in carrier concentration from 3×1019 to 5×1020 cm-3 in the deposited AZO and ZO thin films. It also was found that the gas pressure dependence of the parallel resistance measured on the fabricated heterojunction solar cells was correlated to that of the VOC, whereas the series resistance was independent of the gas pressure. This suggests that the obtainable photovoltaic properties in n-TCO/p-Cu2O heterojunction solar cells are considerably more affected by the surface condition of the p-Cu2O layer, i.e., the interface at the heterojunction, than the diffusion potential resulting from the difference of work functions between the p-Cu2O and n- oxide semiconductor layers.
References
[1] Y. Nishi, T. Miyata, J. Nomoto, and T. Minami, Thin Solid Films 520 (2012) 3819.
[2] T. Minami, Y. Nishi, and T. Miyata, Appl. Phys. Express 6 (2013) 044101.

Recent achievements for CIGS in the field of flexible and light-weight substrates, exceeding photovoltaic conversion efficiencies based on Soda Lime Glass (SLG), are encouraging the use of the same substrate configuration for CZTSSe based devices. Furthermore, the compatibility with high throughput processing techniques, such as Roll-to-Roll processes, and the broad application niche of light and flexible photovoltaic modules make this substrate choice highly appealing for the industry. In this work, we present preliminary results using 0.05 mu;m thick Austenitic Stainless Steel (SS) substrates. A two stage process based on DC sputtering of metallic precursors and vacuum annealing was developed to produce CZTSe absorbers and solar cells. Additionally, a Cr chemical barrier was used to impede Fe and other contaminants diffusion during the annealing. Film properties such as surface roughness, morphology and phase composition and crystalline quality of the different stacked layers trough AFM, SEM and XRD are presented, as well as depth resolved compositional analysis with Auger spectroscopy. Moreover, the impact of the temperature annealing (450 and 550 ^oC) was also studied. Solar cells were completed using a customized back contact configuration: SS/Cr/Mo/ZnO and an etching process was used to remove ZnSe from the absorber. The devices were characterized optoelectronically with illuminated AM1.5 I-V curves analysis and EQE measurements. The roughness obtained for the absorbers was very similar to compositionally analogue CZTSe films obtained on Mo coated SLG. Nevertheless, a difference in the absorber morphology surface was noticeable for both 450 and 550 ^oC annealed samples. A bi-layer structure with smaller grain size close to the CZTSe surface was observed. The absence of Na and difference in thermal properties between SLG substrates and SS and Cr could be in the origin of this morphology. Regarding phase composition and crystalline quality, polycrystalline CZTSe films with good crystalline quality were obtained along with Sn-Se secondary phases, most likely due to vapor condensation. The effectiveness of the back contact configuration to prevent the Fe diffusion to the absorber, even at 550 ^oC, is clearly demonstrated with the analysis of the depth resolved Auger profile for this element. Furthermore, different CZTSe metallic elements in-depth profile was observed for both 450 and 550 ^oC annealed samples. Finally first solar cells with 3.5% efficiency were obtained. Even if these values are lower than those achieved in standard glass substrates, they already show that the use of flexible SS substrates combined with a Cr diffusion barrier can be a promising route for CZTSe solar cells processing. We are currently introducing Na into the CZTSe absorber using Na doped Mo targets, MoNa, from which beneficial effects on absorber properties with the out diffusion of this alkali metal from the back contact are expected.

Cu2ZnSnS4 (CZTS) is a very promising material for Cu(In,Ga)Se2 (CIGS) absorber layers in thin film solar cells and is suitable for scale-up of production with cheap cost. To fabricate CZTS thin films, there are many methods, such as evaporation, sputtering, electro-deposition and solution synthesis with post-sulfurization. In CZTS fabrication methods, especially sulfurization of stacked metal or metal sulfide layers, have been studied and have led to cell efficiencies up to 6.7%. For high efficiency CZTS solar cells, CdS is widely used as a buffer layer deposited by chemical bath deposition (CBD) and it is important to make good p-n junction between CZTS and CdS. So, we report the study of interface states between CZTS and CdS. Here, CZTS thin film is deposited from co-sputtering with Cu, ZnS and SnS binary targets followed by sulfurization at 550C using elemental sulfur vapor atmosphere using two zone furnace. For CdS deposition, cadmium sulfate(CdSO4), Ammonium hydroxide(NH4OH) and Thiourea were used as precursors and CdS was deposited on CZTS/Mo/Soda-lime glass at 89C. Using CZTS/CdS samples, we report on the chemical and structural properties of interface between CZTS and CdS which is investigated by X-ray photoelectron spectroscopy (XPS) for chemical states, photoluminescence(PL) for interface defect analysis, scanning electron microscopy (SEM) for cross-section of two films and Raman spectroscopy. I-V characteristics of CZTS/CdS cell is also measured.

Cu2ZnSn(S,Se)4 (CZTSSe) is a promising alternative absorber material for thin-film photovoltaic applications because of its earth-abundant and non-toxic constituents, tunable band gap, and high optical absorption coefficient. Recently, we have reported a high efficiency CZTSSe PV device (~8.8%) employing CZTSSe thin film prepared by solution based chemical method. The printed CZTSSe films show an interesting double-layered microstructure consisting of a polycrystalline large-grain layer and an amorphous fine-grain layer underneath. In this talk, we will present our efforts on characterization of the fine-grain layer and understanding its impact to the PV device performances. The chemical compositions, electrical and optical properties of the fine-grain layer will be shown and based on the characterization results a device model will be developed to indicate the possible efficiency loss mechanisms.

Bi2S3 thin films can be prepared by chemical bath deposition using thioacetamide as the sulfur source [1]. Some recent works show that plasma treatment is an effective technique widely used to improve the optical and electrical properties of chemically deposited thin film of semiconductor materials. Thin films of are well known as n-type absorber layer that have been used in novel thin film solar cells [2] because of its characteristics of earth abundant and low toxicity material. The optical band gap of the as prepared Bi2S3 films is 1.59 - 1.68 eV, depending on the method of deposition. Furthermore, Bi2S3 thin films are important in the development of the ternary compound Cu3BiS3 [3] which offer excellent properties as a potential photovoltaic absorber material with high optical efficiency [4]. For this work the deposition of Bi2S3 thin films was carried out using chemical bath deposition technique from a reaction solution containing bismuth nitrate dissolved in triethanolamine as Bi3+ source and thioacetamide as S2- source. In this work we present a detailed analysis of the optical, electrical and morphological properties of the Bi2S3 thin films after the AC plasma treatment, produced in argon environment at a pressure of 3 Torr during 75 min with the discharge power supply maintained at an output of 12.5 V and 0.064 A, as well as, thermal annealing in air atmosphere at 250°C for 20 minutes. We observed that the Eg value decreases when film thickness increases from, 100 to 150 nm, as expected. Eg varies from 1.68 to 1.61 eV, 1.63 to 1.6 eV, and 1.61 to 1.59 eV, for the as prepared, heated in air, and plasma treated films, respectively. The thickness measurements showed a loss of 35% in thickness for the annealed samples compared with the as prepared films; however this loss in thickness can be avoided with the plasma treatment, because the films maintain their original thickness after plasma treatments. The XRD analysis showed that in both post-deposition treatments (thermal and plasma) a change in the crystalline phase from amorphous to crystalline with orthorhombic structure is produced. The electrical conductivity measurements showed an increase in the conductivity value for the films treated with plasma and thermal annealing. The conductivity values are σ= 5.8 (#8486; cm)-1, σ= 0.085 (#8486; cm)-1 and σ =1 - 2 x 10-3 (#8486; cm)-1 for the plasma treated, thermally annealed, and as-prepared films, respectively.
References:
[1] J.D. Desai, CD. Lokhande, Materials Chemistry and Physics 41 (1995) 98-10.
[2] Harumi Moreno-García, M.T.S. Nair, P.K. Nair, Thin Solid Films 519(2011) (21) 7364-7368.
[3] P. K. Nair, L. Huang, M. T. S. Nair, Hailin Hu, E. A. Meyers and R. A. Zingaro, Journal of Materials Research 12 (1997) (3) 651-655.
[4] Mukesh Kumar and Clas Persson, Applied Physics Letters 102 (2013) 062109.

A novel synthesis technique for the production of copper zinc tin sulfide (CZTS) nanocrystals has been developed using aerosol spray pyrolysis. CZTS is a quaternary semiconducting material that shows promise as a replacement to common semiconductors such as CdTe and CIGS for use in photovoltaic devices. CIGS is currently being commercialized in the photovoltaic industry, but rare and expensive indium and gallium components threaten its long term viability. CZTS looks to be one of the best alternatives to CIGS with all earth abundant and non-toxic materials and a band gap of 1.5 eV [1]. A number of synthesis techniques have been thoroughly studied and detailed previously. In our novel approach, we synthesis single phase 15 nm nanocrystals, starting with zinc, copper, and tin diethyldithiocarbamate precursors in a toluene solvent. The precursor solution is aerosolized using an ultrasonic nebulizer wherein the droplets are vacuumed through a tube furnace and nucleation occurs. We reproducibly synthesize kesterite, Cu2ZnSnS4, nanocrystals. This technique continuously converts the chemical precursor into high-purity nanopowder with a production rate of ~50 mg/hour for an un-optimized,
lab-scale reactor. Using the same precursor chemistry, we have also been able to deposit high-quality CZTS thin films directly onto arbitrary substrates using a spray pyrolysis technique. A discussion of process parameters on the stoichiometry of either the nanoparticles or the thin films will be presented. Results from extensive characterization via Raman spectroscopy, EDS, XRD, TEM and XPS will be presented. We are currently in the process of producing a printable ink to coat CZTS as the absorbing layer for use in photovoltaic devices.
[1] H. Wang. “Progress in Thin Film Solar Cells Based on Cu2ZnSnS4,” International Journal of
Photoenergy, 2011.

In the present work we synthesized single-phase and nearly stoichiometric crystals of quaternary Cu2ZnSnSe4 (CZTSe) via microwave synthesis using Oleylamine (OLA), Triethanolamine (TEA) and a mixture of TEA and water as solvent and complexing agents. The influence of reaction time and concentration of reactants was analyzed. The reaction was performed at 230 C and 60 psi. The characterization of its phase constituents, morphology, structure and optical properties were done using X-ray diffraction (XRD), Raman Spectroscopy, Scanning Electron Microscopy (SEM) and Energy Dispersive Spectrometry (EDS). The X-ray diffraction and Raman pattern showed that the CZTSe particles had single phase with a kesterite structure. The crystalline size was from 15 to 40 nm. On the other hand, EDS showed a nearly stoichiometric composition of Cu/(Zn Sn) varying in the range 0.8 - 1.1 and Zn/Sn 1.0 - 1.2 coinciding with theoretical values. Based on the powder synthetized we deposited films via evaporation process, the estimated band gap energy is in the range 1.0 to 1.5 eV, high optical absorption coefficient ( > 10^4 cmminus;1) and p-type electrical conductivity what makes them suitable for solar cell devices.

The high levelized energy cost of solar-energy technologies precludes them from competing head-on with conventional energy sources such as coal, oil, and gas. However, due to the importance of harvesting solar energy from an environmental and a national-security perspective there is continuing high interest in the development of cost competitive solar-to-electrical and chemical energy conversion solutions. In addition to this, current solar-to-electrical energy conversion technologies make it harder and harder for nascent technologies to compete on price alone.
Pyrite is a promising photovoltaic material including for incorporation into multi-junction cells. It&’s constituent elements are abundant, and solar-to-electrical energy conversion has been already been demonstrated using pyrite. However, efficient energy conversion has never been demonstrated, and this has been ascribed to deleterious bulk and surface chemistry but is not well enough understood to have been mitigated at this point in time.
The photoelectrochemistry of pyrite has been investigated for aqueous systems and has been correlated to near-surface crystal structure and surface chemistry. We are separating contributions from regenerative and non-regenerative processes to the observed current-voltage behavior. We are investigating the photoelectrochemistry in non-aqueous solutions and by systematically varying key properties of the solution redox couples thus working towards a more complete understanding of the photoelectrochemistry of pyrite electrodes.

Cu2ZnSnS4 (CZTS) attracts an increasing attention as a promising absorber material for thin film solar cells. However, CZTS is thermodynamically stable in only a narrow area of the phase diagram. The loss of sulphur and tin at high temperature makes controlling of its composition even more challenging. In this case, secondary phases, such as ZnS, Cu2S, and Cu2SnS3, could form depending on the processing conditions. Secondary phases have several effects on photovoltaic device performance. The formations of low band gap phases (like Cu2SnS3) can decrease the free carriers&’ life time and thus are detrimental to solar cell performance. Otherwise, secondary phases (like ZnS) can be regarded as an inactive high resistivity domain. The precipitated position of secondary phase is also important for the solar cell performance. Secondary phase precipitated within the space charge region should ideally be a high resistive barrier in order to increase the shunt resistance of the device. However, secondary phases within the quasi-neutral region should be low resistive barriers so that the series resistance is small and the detrimental impact on carrier separation is reduced. However, only a small portion of the secondary phases is challenging to characterize using neither X-ray diffraction due to overlap of the Bragg peaks with those of CZTS nor surface Raman spectroscopy when they are buried at the back contact. A cross-sectional transmission electron microscopy (TEM) equipped with energy dispersive spectroscopy (EDS) detector is an ideal tool to analyse secondary phases but TEM investigations of CZTS solar cells to date are scarce. We present micro-analysis results on our fully processed CZTS solar cells based on absorber layers deposited by sputtering of metallic-stack layers followed by a sulfurization annealing. The existence and distribution of secondary phases are identified by TEM and the role of these secondary phases on the solar cell performance is also analysed.

The photovoltaic absorber, Cu2ZnSn(S,Se)4 (CZTS) has the potential to cost compete with Cu(In,Ga)Se2 (CIGS) and CdTe thin film solar cells due to an ideally suited band-gap and abundance of these elements in the Earth&’s crust. In order to optimize the built-in potential and help CZTS based PV approach its theoretical maximum power conversion efficiency (PCE) of 32%,1 knowledge of surface and buffer interface electronic properties is essential. This contribution addresses the nature of CZTSe surface composition and band edge alignment across CZTSe/Zn(O,S) heterojunctions and how these properties relate to device performance. Polycrystalline kesterite CZTSe thin films deposited by thermal coevaporation and having demonstrated baseline PV efficiencies close to 9% (when employing CdS buffer layers), were used as platforms for study of n-type buffer layers based on Zn(O,S) grown by chemical bath deposition (CBD). Interconnected, in-situ growth and surface processing methods with UHV surface analysis such as X-ray and ultraviolet photoemission (XPS/UPS), and scanning tunneling spectroscopy (STS) were used to measure the composition and electronic structure at the CZTSe/Zn(O,S) interface. Cation exchange between CZTSe and Zn “partial electrolyte” (P.E.) solutions was assessed. This system exhibits slow reaction kinetics and the subsequent effects on the valence band maximum (VBM) are negligible for Zn-rich CZTSe surfaces. Lack of compositional changes from Zn P.E. treatment and the slow reaction kinetics form the basis for treating the CZTSe/Zn(O,S) as an abrupt heterojunction, unlike those utilizing CdS buffer layers. XPS valence band measurements place the CZTSe VBM 0.6 eV below the Fermi energy in the near surface region. Core level XPS measurements of layer-by-layer grown Zn(O,S) on CZTSe shows an additional downward surface band bending of 0.2 eV in CZTSe, accompanied by 0.3 eV downward band bending for the Zn(O,S) buffer layer. The CZTSe surface electronic gap was measured using STS. Our measurements indicate that the CZTSe near-surface region is slightly n-type. Derivation of the band diagram will be discussed and related to the performance of completed solar cells. These results, placed in the context of operational CdS buffer layers offer guidance for the function, optimization and processing of Zn(O,S) buffer layers applied to CZTS solar cells.
1. W. Ki and H. W. Hillhouse, Advanced Energy Materials 1 (5), 732-735 (2011).

A great deal of interests in the research of nontoxic semiconductors has been done in recent years, both from a fundamental as well as technological point of view. Several ternary sulfides of the system Cu-Sn-S system have been reported, generally synthesized by solid-state reactions over several days at elevated temperatures (850-1200 °C), with interesting properties and potential applications as suitable candidate for photovoltaic applications. In this work, thin films of SnS with two different crystalline structures were prepared by chemical bath deposition technique (CBD), with terminal thickness of 100-450 nm. These films were coated with a layer of thermally evaporated Cu, with different thickness. The SnS/Cu layers were heating at 300-400 °C for 1 h, in vacuum, nitrogen and argon atmospheres, resulting in the formation of Cu2SnS3 and Cu4SnS4, depending on the SnS crystal structure and/or the heating conditions. We present the results of structural analysis (X-ray Diffraction Technique, XRD), morphologies of the films (Scanning Electron Microscopy, SEM and Atomic Force Microscopy, AFM), optical and electrical properties (UV-Vis-NIR spectrophotometer and photo-response characterizations) as well as composition studies using X-ray Photoelectron Spectroscopy (XPS).

In this work, we investigate photoelectrochemical (PEC) properties of thermal oxide films grown on α-brass, which is an alloy of copper (85 atomic %) and zinc (15 atomic %). In contrast to pure metallic foils, alloys such as brass can be converted to either a p-type (copper oxide), n-type (zinc oxide) or a multiphase oxide by carefully controlling the thermal processing parameters. Additionally, the facile thermal synthesis of oxides grown on brass provide highly adherent oxide films which enables post processing integration and large scale electrode fabrication.
We studied the oxide composition and morphology formed on as-received brass at a series of different temperatures. Since for PEC applications the oxides of Cu (Cu₂O and CuO) have favorable position of their band edges with respect to the reduction potential of H₂O/H₂,^{1, 2} we modulated the surface composition to controllably obtain these oxides. Starting with α-brass, Ar gas anneals were performed at 300^omicron;C and higher to cause removal of surface Zn (dezincification) and obtain a porous copper rich surface layer³. This dezincified brass was oxidized to form a series of pure and mixed copper oxide phases. Characterization techniques including SEM analysis, Raman spectroscopy and x-ray diffraction confirmed the surface structure and chemical composition of the oxide layer. At 400^omicron;C, dense CuO nanowires with diameter 90 nm were formed on the oxide surface. As a result of the CuO-Cu₂O scale formation, ZnO phase development was delayed and seen only at relatively higher temperatures (> 500^oC) of oxidation. As expected, the oxide films were extremely well adherent and resistant to cracking.
Photoelectrochemical properties of the as-oxidized brass was compared to that of the oxidized samples after surface dezincification, under Xe lamp illumination. It was found that the oxide samples after dezincification had a higher photocurrent density and onset potential than the as-annealed ones. While further work is needed to understand the electronic and optoelectronic properties of this system, our preliminary data shows that oxidizing alloys into biphasic, metal oxide nanostructured films have potential to be engineered into viable solar energy harvesting devices.
1. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem Rev, 110, 6503 (2010).
2. A. Paracchino, V. Laporte, K. Sivula, M. Grätzel and E. Thimsen, Nat Mater, 10, 456 (2011).
3. R. Balluffi and B. Alexander, J Appl Phys, 23, 1237 (1952).

Hematite (α-iron oxide, Fe2O3) has been a popular material for photoelectrochemical conversion of sunlight to fuel. Its bandgap of around 2.1eV is optimal for water splitting, and it is also cheap, abundant, non-toxic, and stable. However, a major drawback of using hematite is its poor electric conductivity together with low optical absorptivity, leading to very low photocurrent. Nanostructuring, however, could effectively reconcile the different requirements of photon absorption and carrier extraction. In this work, we aim at maximizing sunlight absorption within a very thin layer of hematite by means of nanophotonic light management.
We choose 20nm as the operating thickness of hematite because, on one hand, the diffusion length of holes is at most 20nm, and on the other hand, the thickness needs to sufficient to build up a depletion region. We also rely on two considerations to maximize absorption. On one hand, antireflection is absolutely necessary due to impedance mismatch. Tapering provides a graded index for this purpose. On the other hand, we aim to pack as much hematite as allowed in the given volume, effectively orthogonalizing the carrier transport and photon absorption.
One structure that could meet the geometrical constraint is a square lattice of the core-shell nanocone structure. We smooth the bottom of the indium tin oxide (ITO) nanocone, and as a result, any location within hematite is less than 20nm away from the nearest hematite-water interface to guarantee the hole collection. Fortunately, both antireflection and orthogonalization strategies favor high aspect ratio cones. Later, we will see that the strategies are so powerful that the aspect ratio is far from extreme for near-complete absorption.
We simulate sunlight absorption in the nanostructures using the rigorous coupled-wave analysis (RCWA). Aiming to maximize the photocurrent, we adjust the parameters to search for a global optimum. The result shows that, as the height approaches 3mu;m, we see almost perfect absorption.
Although we focus on right circular cones to demonstrate our strategies, the same principles apply generally to other tapering structures, for example, pyramids or inverted pyramids. We indeed observe rich features with the nanopyramid structures, confirming the effectiveness of our light management strategies.
In all cases, the photocurrent generated by the nanostructured iron oxide photoanode could reach 12.5mA/cm2, as the height increases to a few microns while keeping the thickness to be 20nm.

An effective defect passivation route has been demonstrated in the rapidly growing Cu2ZnSn(S,Se)4 (CZTSSe) solar cell device system through using Cu2ZnSnS4:Na (CZTS:Na) nanocrystals precursors. CZTS:Na nanocrystals are prepared by sequentially preparing CZTS nanocrystals and surface decorating of Na species, while retaining the kesterite CZTS phase. The exclusive surface presence of amorphous Na species is proved by X-ray photoluminescence spectrum and transmission electron microscopy. With Na-free glasses as the substrate, CZTS:Na nanocrystals-based solar cell device shows 50% enhancement of device performance (~6%) than that of un-passivated CZTS nanocrystals-based device (~4%). The enhanced electrical performance is closely related to the increased carrier concentration and elongated minority carrier life time, induced by defect passivation. Solution incorporation of extrinsic additives into the nanocrystals and the corresponding film enables a facile, quantitative and versatile approach to tune the defect property of materials for future optoelectronic applications.

Thin film photovoltaic (PV) technologies based on Cu(In,Ga)Se₂ (GIGS) and CdTe combine moderate efficiencies with low cost. However, expanding to terawatt production brings concerns about material sustainability and hence, future costs. As such, there is a clear and present need for an alternative absorber, whose constituent elements are earth abundant, environmentally benign, and therefore, exhibit even lower projected materials and recycling costs. Copper sulfide (Cu₂S) has a rich history in thin film heterojunction PV due to its exceptional combination of these desired properties, as well as a suitable band gap of 1.2 eV, an absorption coefficient >10⁴, and a theoretical solar conversion efficiency of 28%. Despite such promising characteristics, a major hurdle preventing its practical utilization is the instability of Cu₂S toward the formation of the sub-stoichiometric Cu_1.96S, which is exacerbated in the presence of oxygen. The result is a highly degenerately doped (>10²⁰ cm^-3) semiconductor no longer useful as a PV absorber layer. Starting with clean chemistry, preventing surface oxidation, and limiting Cu diffusion into adjacent layers may avoid this deleterious conversion. Atomic layer deposition (ALD) is unique suited to accomplishing these tasks owing to its self-limiting surface chemistry, therefore enabling atomic-level thickness and composition control over large areas and textured surfaces. Here, we demonstrate that ALD is useful in the deposition of pure and stoichiometric Cu₂S thin films. Furthermore, the ALD of ultrathin PV-applicable barrier layers adjacent to Cu₂S is capable of preventing oxidation, mitigating Cu-diffusion, and thereby preserving the desirable semiconducting properties. These results represent the first key steps towards reviving Cu₂S photovoltaics.

Thin-film absorber layers for photovoltaics have attracted much attention for their potential for low cost per unit power generation, due both to reduced material consumption and to higher tolerance for defects such as grain boundaries. Cu2ZnGeSe4 (CZGSe) comprises one such material system which has a near-optimal direct band gap of 1.6 eV for absorption of the solar spectrum, and is made primarily from earth-abundant elements.
CZGSe thin films were fabricated from Cu, Zn, and Ge sputtered onto Mo-coated soda lime glass substrates. These metallic precursor films were selenized in a two-zone close-space sublimation furnace using elemental Se as the source, with temperatures in the range of 390 to 550 C, and at a variety of background pressures. Films approximately 1.5 microns thick were obtained with the expected stannite crystal structure. CdS and ZnO layers were deposited by chemical bath deposition and RF sputtering, respectively, to make devices.
Next, Cu2ZnSnSe4 (CZTSe), which has a direct band gap of 1.0 eV, was prepared in a similar manner and combined with CZGSe as either compositionally homogeneous or layered absorbers. The compositional uniformity of selenide absorbers made by selenizing compositionally homogeneous CuZnGeSn precursor layers was determined and the band gap as a function of composition was investigated in order to tune for optimum solar spectrum absorption. For layered CuZnGe/CuZnSn precursor films, the composition profile was measured before and after selenization to assess the stability of the layered structure, and its applicability for forming a band-gap-graded device for improved current collection.

In the past five years, there has been a steep interest in using Cu2ZnSnS4 (CZTS) as a material to replace conventional semiconductors in various applications like solar cells and Li-ion batteries.[1] This remarkable interest stems from the balance between good optoelectronic properties, availability of the constituent elements, toxicity and recyclability. However, there is a great deal of effort involved in developing low temperature and non-vacuum methods to produce CZTS.
The synthesis of CZTS in nanoparticulate form not only serves this purpose but also gives freedom to control and tune different properties at nanoscale level. One of the major challenges for nanoparticulate systems is to assure CZTS phase purity against by-products. Due to lower intensity and broadening of the signals, the characterisation of CZTS phase purity in nanoparticulate form poses additional challenges with respect to bulk materials. Therefore, there is an additional need to improve and develop characterisation techniques which can ensure purity of the CZTS nanoparticles as compared to bulk CZTS product.
In the present contribution we describe a facile synthesis of CZTS nanoparticles by hot injection method. In contrast to the state of the art, no inert atmosphere or vacuum was used during the synthesis.[2] Furthermore, the injection temperature influences the phase composition and crystalline quality of the final material. A detailed characterisation of nanoparticles was carried out using Raman mapping (MAPs) and UV-Vis spectroscopy, XRD, HRTEM, EDX. Combination of this wide range of techniques confirms the presence of CZTS phase and sheds light on the formation of crystal structures (kesterite, stannite, wurtzite) and by-products. The synthesised product was estimated to be more than 99.85% oxide free. Finally, the as-prepared films showed conductivity 10 times higher than the state of the art for hot injection method.[3] The paper will comprise the work recently published [2] and the characterization of crystallographic defects and secondary phases occurring during crystallization.
[1] Zhao, Y.; Burda, C., Development of plasmonic semiconductor nanomaterials with copper chalcogenides for a future with sustainable energy materials. Energy & Environmental Science 2012, 5 (2), 5564-5576.
[2] Ahmad, R.; Distaso, M.; Azimi, H.; Brabec, C.; Peukert, W., Facile synthesis and post-processing of eco-friendly, highly conductive copper zinc tin sulphide nanoparticles. J Nanopart Res 2013, 15 (9), 1-16.
[3] Lu, X.; Zhuang, Z.; Peng, Q.; Li, Y., Wurtzite Cu2ZnSnS4 nanocrystals: a novel quaternary semiconductor. Chem. Commun. 2011, 47 (11), 3141-3143.

This presentation will focus primarily on the conditions required to fabricate high quality CZTSSe p-type layers within Mo/CZTSSe/CdS/i-ZnO/ITO/Al solar cells. Optimization of the CZTS nanocrystal (NC) layer is key to realizing high quality devices. By optimizing the nanocrystal ink and thermal treatment steps, 7.5% CZTSSe solar cells have been fabricated.
Recent research has been directed toward the production of solution processable semiconducting colloids. Such colloids permit the active layers of nanocrystal-based devices to be literally printed at a comparatively low cost as compared to other methods. A particularly attractive candidate for thin film solar cells is CZTSSe. The use of CZTS nanocrystal (NC) derived active layers is favorable because the surface chemistry of the nanocrystals may be altered. This enables them to be dispersed in a variety of solvents and thus permits fully tailored semiconducting inks to be developed.
As prepared CZTS NC thin films have poor optoelectrical properties. This arises from the large number of grain boundaries within the NC films as well as the presence of organic ligands (inherent to NCs), both of which must be minimized in the final films to enhance device performance. The primary means to reduce organic (carbon) content is to choose an appropriate NC surface chemistry, typically consisting of short chain volatile ligands. To induce grain growth and further reduce impurities the CZTS NC films are typically heated at high temperatures (450-500 °C) in the presence of a high vapor pressure of selenium. The selenization process results in near-micron sized grains, which greatly enhances photoconductivity (up to an order of magnitude) in the case of CZTS films.
In order to improve overall device performance, it has been found that the CZTS NC composition, purification steps, NC surface chemistry, CZTSSe film thickness, selenization temperature, as well as the heating and cooling rates during selenization all need to be optimized.

To assemble the final device, the optimized CZTSSe NC film (on molybdenum) is coated with 60 nm of CdS via chemical bath deposition. Then i-ZnO is DC sputtered to a thickness of 50 nm followed by 150 nm of ITO via RF sputtering. Next an ~180 nm aluminium charge collection grid is thermally evaporated. Finally, the device is mechanically scribed. The active device area is accurately measured using an optical microscope and used to calculate the device performance.

We compare the photoluminescence (PL) characteristics of co-evaporated Cu2ZnSnS2 (CZTS) films with two types of microstructures: (1) polycrystalline films grown onto Mo-coated glass and annealed under sulfur overpressure (as in high-efficiency devices) and (2) large-grained (~10 mu;m diameter) epitaxial material grown on p+ Si substrates (device efficiency measurements underway). The epitaxial material is nearly stoichiometric in composition and a range of compositions of polycrystalline material have been characterized. In the epitaxial samples, intensity-dependent PL spectra at 4K contain a low-energy defect peak and a higher-energy shoulder. This is qualitatively similar to what is seen in polycrystalline CZTS samples. However, the lifetime of the defect recombination in epitaxial samples is ~tens of nanoseconds, compared to the ~10 microsecond lifetime of defect recombination in polycrystalline CZTS. We will discuss the reason for this difference in lifetime and compare the defect properties of epitaxial and polycrystalline CZTS.

We have used time-resolved terahertz spectroscopy (TRTS) to measure lifetimes and determine recombination mechanisms in Cu2ZnSn(S,Se)4 (CZTSSe) thin films fabricated from nanocrystal inks. TRTS is a valuable technique to probe photoconductivity on femtosecond to nanosecond time scales that are relevant for recombination in thin film photovoltaics. Terahertz frequencies (0.2-2.5 THz) correspond to typical scattering rates in semiconductors, enabling determination of carrier density and mobility. Ultrafast time resolution permits tracking the evolution of carrier density to determine recombination mechanisms. By manipulating the photoexcitation wavelength and fluence, we can tailor the generation profile of photoexcited carriers to distinguish between surface, Shockley-Read-Hall (SRH), and Auger recombination mechanisms and determine rate constants.
We will discuss our TRTS experiments and modeling to understand the role of recombination mechanisms and their contribution to CZTSSe photovoltaic performance. TRTS photoconductivity shows an instrument-limited onset within 1 ps, followed by a slow decay over nanoseconds. Photoconductivity was normalized to its maximum value, and decay kinetics were fit with a bi-exponential model with two time constants and a weight fraction. The short time constant is typically ~200 ps and roughly corresponds to diffusion to and recombination at the surface. The long time constant is typically ~2 ns and is attributed to SRH recombination. Assignment of these mechanisms is supported by the dependence of kinetics upon excitation fluence and wavelength. Normalized kinetics are independent of fluence over a range of 40x, indicating that no Auger recombination is occurring. Without Auger recombination, we can distinguish between surface and SRH rates by tuning the pump wavelength. As the excitation wavelength is shifted towards the blue, carriers are generated nearer to the front surface and the photoconductivity kinetics are sensitive to the surface recombination velocity. With blue excitation, we see that a larger fraction, ~0.5, of carriers recombine with a short time constant. With longer excitation wavelengths, the carriers are generated more evenly throughout the film and the kinetics are dominated by SRH recombination with the long time constant having a majority of the weight fraction, ~0.8. Despite the attractive simplicity of a bi-exponential model, it cannot fully capture the complexity of the carrier dynamics. We will present detailed simulations of the TRTS response obtained by solving the transient electron and hole continuity equations and Poisson&’s equation to track carrier transport and recombination. The combination of TRTS experiments and modeling provides a pathway to determine performance-limiting recombination mechanisms, measure key parameters like SRH lifetime and surface recombination velocity, and elucidate structure-property relationships for CZTSSe and other photovoltaic materials.

Cu₂ZnSnS₄ (CZTS) is considering as an important light absorber for next generation low-cost thin-film solar cells because of its optimal band gap energy for solar conversion efficiency and its high absorption coefficient [1-2]. The power conversion efficiency of CZTS-based solar cells, which currently obtained a maximum value of ~8.4% [3], is believed to relate essentially to the defect states existing intrinsically in quaternary compound CZTS [1,4-7]. Our recent results have pointed out the significant roles of the band tail states which cause picosecond-scaled localization of photocarriers as well as a sub-nanosecond multiple trapping processes in recombination dynamics of CZTS single crystals at room temperature (RT) [1]. In addition to knowledge obtained at RT, for some particularly practical uses of CZTS-based solar cells such as in aerospace industry, fundamental physical understandings of temperature dependent photocarrier dynamics in CZTS are also of importance.
In this study, we determined the temperature dependence of photocarrier recombination dynamics in CZTS singles crystals. The observed microsecond-scaled photoluminescence (PL) decay times at low temperatures indicate localization of photogenerated carriers at spatially separated band tails. A good consistency was observed for the temperature dependence of the decay rate and that of the PL intensity. As the temperature is increased, nonradiative recombination becomes dominant, leading to a considerable enhancement of the decay rate as well as a significant decrease of the PL intensity. A large change of about three orders in magnitude of the decay time, from microsecond to sub-nanosecond, indicates very efficient nonradiative recombination at high temperatures in CZTS, which might be one of the main reasons that limit the power conversion efficiency of the CTZS-based solar cells.
This work was supported by JST-CREST and the Sumitomo Electric Industries Group CSR Foundation.
References
[1]. L. Q. Phuong et al., Appl. Phys. Lett. 103 (2013) in press.
[2]. K. Ito et al., Jpn. J. Appl. Phys. 27, 2094 (1988).
[3]. B. Shin et al., Prog. Photovoltaics 21, 72 (2013).
[4]. K. Tanaka et al., Phys. Status Solidi A 203, 2891 (2006).
[5]. M. J. Romero et al., Phys. Rev. B 84, 165324 (2011).
[6]. S. Levcenko et al., Phys. Rev. B 86, 045206 (2012).
[7]. A. Nagaoka et al., Appl. Phys. Lett. 103, 112107 (2013).

The kesterite semiconductors Cu2ZnSnS4 and Cu2ZnSnSe4 are drawing intensive attention recently as the earth-abundant light-absorber layer in thin-film solar cells. The additional number of elements in these quaternary compounds, relative to binary and ternary semiconductors, results in increased flexibility in the material properties, but a large variety of intrinsic defects can also be formed, which have important influence on their optical and electrical properties, and hence their photovoltaic performance. Experimental identification of these defects is currently limited due to poor sample quality. Because of the limited knowledge about the defects, many experimental observations can not be understood, e.g., why the hole concentration in the stoichiometric samples is even higher than in non-stoichiometric ones, why the non-stoichiometric solar cells have higher efficiency than stoichiometric ones, whether there are recombination-center defects with a high concentration in these quaternary semiconductors.
Through the first-principles calculation study on the phase stability and defect properties, we gave a systematic explanation to these abnormal observations. We found that there are strong phase-competition between the kesterites and the coexisting secondary compounds, so it is highly possible that secondary compounds coexist in the synthesized samples and contribute to the significant non-stoichiometry. Furthermore, there are a series of intrinsic defects and defect complexes in Cu2ZnSnS4 and Cu2ZnSnSe4, among which the CuZn antisites and Cu vacancies are the dominant acceptor defects, determining the intrinsic p-type conductivity and the hole concentration. The compensation between low-energy acceptor and donor defects leads to the formation of fully or partially charge-compensated defect complexes, such as VCu+ZnCu, 2CuZn+SnZn, and CuZn+SnZn, which can have high concentration in samples and contribute to non-stoichiometry. 2CuZn+SnZn is an electron trapping center and CuZn+SnZn is an electron-hole recombination center. They can have high concentration in stoichiometric samples, which is detrimental to the solar cell performance. Based on the calculation, we explained the experimental observation that Cu poor and Zn rich conditions give the highest solar cell efficiency, as well as suggesting an efficiency limitation in Cu2ZnSn(S,Se)4 cells with high S composition.

The defect physics of earth abundant quaternary compound Cu2ZnSnS4 (CZTS) and Cu2ZnSnSe4 (CZTSe) is extremely complex not only because of the many competing phases which considerably restrict the domain of stability of CZTS and CZTSe but also because of the many possible cationic substitutions. Previous theoretical studies have indicated that CZTS and CZTSe have a quite high hole concentration originating from intrinsic defects such as Cu vacancies and Cu-on-Zn antisites.[1] In this study, we have carried out state-of-the-art defect calculations including (i) thermochemical corrections to the phase diagram,[2] (ii) correction to the formation enthalpy of shallow defects due to the large band gap underestimation in DFT.[3] The latter was found to be a critical point in the analysis of intrinsic defects in both CZTS CZTSe. Indeed, and at variance with previous studies, our GW-corrected ab initio defect calculation reveals that both Cu-on-Zn and Zn-on-Cu antisites have comparable formation enthalpy, resulting in the pinning of the Fermi level in the mid-band gap region. Such discovery is in line with the disordered kesterite structure usually experimentally observed in CZTS.[4] Pinning of the Fermi level has two important consequences (i) the relatively low carrier concentration in CZTS or CZTSe samples and (ii) the limitation of the open circuit voltage. Such intrinsic limitation of CZTS could explain the large loss in open circuit voltage observed for CZTS-based solar cells. It is found that higher carrier concentrations are achievable under growth conditions where CZTS and CZTSe are only marginally stable and may decompose into binary or ternary competing phases.
[1] S. Chen et al., Phys. Rev. B, 81, 245204 (2010)
[2] V. Stevanovic et al., Phys. Rev. B, 85, 115104 (2012)
[3] H. Peng et al, Phys. Rev. B, 88, 115201 (2013)
[4] L. Choubrac et al., Phys. Chem. Chem. Phys., 15, 10722 (2013)

The increasing cost and the limited supply of elements like In and Te have led to a widespread interest in developing novel earth-abundant thin-film photovoltaic materials for future large-scale PV deployment. A popular strategy to lower material cost is to replace the expensive elements with a combination of cheaper alternatives such that atomic and thus electronic structures are not significantly changed. This strategy, as applied to chalcopyrite CIGS, led to development of kesterite Cu₂SnZn(S/Se)₄ (CZTSSe), both with tetrahedrally bonded structures related to zincblende ZnS. Although remarkable conversion efficiencies above 10% have been achieved with CZTSSe, the chemical complexity resulting from the combination of four difference valences (I2-II-IV-VI4) leads to challenges in the control of composition and electronic properties.
Here we demonstrate an alternative approach to design of such materials, evaluating candidates grouped by constituent elements rather than underlying crystal structures. As an example, we evaluate thermodynamic stability, electrical transport, electronic, optical and defect properties of Cu-Sn-S materials using complementary theory and experiment. Our evaluation shows that Cu₂SnS₃ avoids many issues associated with the properties of Cu₄SnS₄, Cu₄Sn₇S₁₆ and other Cu-Sn-S materials. Cu₄SnS₄ has too high experimental conductivities, and Cu₄Sn₇S₁₆ has low calculated absorption coefficient, poor hole transport and Fermi level pinning due to Cu vacancy defects. In contrast, Cu₂SnS₃ has a wide thermodynamic stability window, a high absorption coefficient and is resilient to the Fermi level pinning. Identification of Cu₂SnS₃ as the best candidate solar absorber within Cu-Sn-S family is consistent with a cell efficiency of 6% that has recently been achieved with Cu₂SnS₃ based absorber (Umehara et al., 2013).
This example demonstrates how this element-specific approach quickly identifies potential problems with less promising candidates and helps focusing on the more promising solar cell absorbers.
Refereces:
Umehara, M., Takeda, Y., Motohiro, T., Sakai, T., Awano, H., & Maekawa, R. (2013), 6, 15-17.

Use of fossil fuels to meet the power demands will continue to drive global climate change through the 21st century. Photovoltaic cells tap into the most abundant renewable and clean energy source on Earth - sunlight. The lack of inexpensive and high-efficiency solar cells, however, hinders large-scale deployment. Historically, the development of new solar materials has largely been a trial-and-error undertaking with little attention given to establishing fundamental efficiency limits on the basis of materials properties. In this contribution, I will discuss unique approaches for rational design of absorber materials with particular focus on compounds with rapid onset to high absorption near the semiconductor band gap as a means to realize high efficiencies. The absorption-merit based SLME computational tool is used to recognize best-of-class absorbers from a large pool of candidate ternary transition-metal chalcogenides. Identified, earth-abundant materials families offer promise for realizing efficient absorption of solar radiation in films much thinner than 1 mu;m. The use of such ultra-thin absorber layers also has important implications with respect to relaxing electrical-transport and defect-property requirements for a high-efficiency cell.

As the world&’s demand for energy grows, the search for cost competitive and earth abundant photovoltaic materials is becoming increasingly important. One promising material system for earth abundant photovoltaics is the Cu-Sn-S family, in which several ternary compounds, including Cu₄SnS₄, Cu₄Sn₇S₁₆, and Cu₂SnS₃ , have been suggested as PV absorbers in prior literature. Our work focuses on the Cu₂SnS₃ phase, which in our previous studies has been shown to avoid the numerous problems associated with Cu₄SnS₄ and Cu₄Sn₇S₁₆ compounds. Despite the relatively narrow band gap of Cu₂SnS₃ , we believe it to be the most promising absorber material in the Cu-Sn-S family (P. Zawadzki et al., submitted).
Although theoretical work suggests good absorber properties for Cu₂SnS₃, current device efficiencies remain quite low. For a device using a Cu₂SnS₃ absorber, the record efficiency (reported by Koike, et al. in 2012) is 2.8%. It is likely that one of the limiting factors for current device performance is poor electrical transport properties resulting from non-ideal absorber morphology. During the development of Cu(In,Ga)Se₂ absorbers, similar morphology challenges were addressed using multi-stage growth techniques. In CIGS growth, a Cu-rich deposition step is used to promote sintering and grain growth, resulting in films with large grain sizes. Then, the deposition is continued in a Cu-poor environment to achieve the desired film stoichiometry and to avoid Cu₂Se impurities.
In this work, we develop a multi-stage growth process for morphology control in the Cu₂SnS₃ absorber layer, with the goal of improving the Cu₂SnS₃ PV device efficiency. To synthesize Cu₂SnS₃ , we use combinatorial RF sputtering from Cu₂S and SnS₂ targets on stationary heated substrates. This lets us create and exploit a spatial gradient in composition for high-throughput studies. To explore the resulting films, we use spatially resolved atomic force microscopy, which allows us to assess the film roughness and grain size. Further compositional and structural characterization is performed using other spatially resolved techniques such as XRF, XRD, and Raman.
We find that a simple two-step Cu-rich/Cu-poor growth process provides a significant enhancement in the Cu₂SnS₃ grain size. Films grown in a single-stage process have grain sizes <100 nm; the two-stage process results in grain sizes between 200-500 nm. Currently, these grain sizes are limited by the film thickness, and are expected to be larger for thicker films, such as those used for absorber layers in PV devices. Initial studies of the effect of synthetic control over film morphology on solar cell device efficiencies are in progress and will be reported.
The project “Rapid Development of Earth-Abundant Thin Film Solar Cells” is supported as a part of the SunShot initiative by the U. S. Department of Energy, Office of Energy Efficiency and Renewable Energy under Contract No. DE-AC36-08GO28308 to NREL.

Copper iron sulfide materials are potentially useful earth-abundant solar light absorbers that incorporate the favorable qualities of copper sulfide with enhanced phase stability. Copper sulfide served as a component in some of the earliest commercial photovoltaics and retains promise as an earth-abundant solar absorber due to its optical properties and availability.1 Use of Cu2S in photovoltaics was discontinued due to phase instability and the related high copper conductivity. This phase instability resulted in deposition of metallic copper at interfaces and alteration in the band gap and minority carrier diffusion length. We have demonstrated that these issues are exacerbated in nanostructured Cu2S,2 an issue we sought to address by synthetic manipulation of copper sulfide chemistry. Solvatothermal synthesis of such nanoparticles enables synthetic control over phase and extent of iron incorporation. Notably, incorporation of even very small amounts of iron produced materials with optical properties similar to the copper sulfide phase sought after for thin film cells, but which resisted phase transformation over time.3 We have expanded on this finding to explore a range of copper iron sulfide nanoparticles. In this ternary system, an array of phases were accessed by variation of the amount of iron incorporated. Phases produced range from tetragonal Cu2S to CuFeS2. Solid solution formation allowed a linear relationship between ratio of Cu:Fe in solution and solid, resulting in gradual tuning of the optical behaviors and phase stability. The band gaps and plasmon band absorption of these copper iron sulfide nanoparticles were controllably altered by the extent of iron incorporation, yet the band gaps generally fall in the range optimal for a solar absorber. The presence of plasmon band absorption is one measure of the phase instability and resultant loss of copper, and is systematically eliminated as iron incorporation increases. This is supported by cyclic voltammetry and X-ray diffraction measurements. For species with low iron concentrations, the phase stability is greatly enhanced by surface passivation using metal sulfide ions.
(1) (a) Wadia, C.; Alivisatos, A. P.; Kammen, D. M. Environ. Sci. Technol. 2009, 43, 2072-2077; (b) Zhao, Y. X.; Burda, C. Energy Environ. Sci 2012, 5, 5564-5576.
(2) Lotfipour, M.; Machani, T.; Rossi, D. P.; Plass, K. E. Chem. Mater. 2011, 23, 3032-3038.
(3) Machani, T.; Rossi, D. P.; Golden, B. G.; Jones, E. C.; Lotfipour, M.; Plass, K. E. Chem. Mater. 2011, 23, 5491-5495.

Strict level control of stoichiometry for compound semiconductor materials is challenging but necessary for photovoltaic (PV) and other electronic applications, with rare exceptions (ODCs in CIGS and ZnS in CZTS). Stoichiometry control can be achieved by an overflux method if one component is in gas phase at the growth temperature, such as arsenic in GaAs or oxygen in ZnO. Here, we apply this overflux growth strategy to the thin films of CuSbS2 grown on heated substrates by RF sputtering from Cu2S and Sb2S3 targets. The target materials have large difference in vapor pressures which has been exploited to grow films with near stoichiometric cation ratios (Cu/Sb=1.0). During the synthesis, the overflux of vapor phase Sb2S3 is such that at ambient substrate temperature the bulk equilibrium composition would be CuSbS2 + Sb2S3 impurities. However, high substrate temperatures ensure that Sb2S3 remains in the vapor phase, thereby preventing precipitation of Sb2S3 impurities in the film. In this way, Sb2S3 is incorporated into the growing film only in the case that it reacts with Cu2S to form the ternary CuSbS2 compound.
The critical substrate temperature for the overflux growth, above which Sb2S3 remains in gas phase was identified using high-throughput combinatorial approach with spatially-resolved characterization techniques. Phase identification was carried out by X-ray diffraction (XRD), stoichiometry and thickness was verified by X-ray fluorescence (XRF) confirmed by other techniques, and conductivity was determined from four point probe sheet resistance measurements. Below the critical temperature, conductivity of the samples varied by 6 orders of magnitude across a narrow range of composition around the CuSbS2 due to presence of either Sb2S3 (n-type) or Cu12Sb4S13 (metallic) impurity phase controlled by relative Cu2S and Sb2S3 fluxes. Above the critical temperature, the resulting phase-pure CuSbS2 samples grown by Sb2S3 overflux approach had uniform and consistent hole concentration of 10^17 cm-3 and hole mobility of 0.2 cm2/V-s (as determined by Hall effect) and steep optical absorption onset at 1.5 eV with no signs of sub-gap absorption caused by metallic secondary phases (as determined by optical spectroscopy). Our progress towards integration of the resulting phase-pure high-quality CuSbS2 samples into solar cell prototypes will be reported.
In summary, we demonstrate the application of the overflux growth approach to growth of stoichiometric high-quality CuSbS2 thin films.This technique may also offer some control of grain orientation and dopant level off-stoichiometry in this materials. In addition, it may be applicable to other chemically similar Cu-S based ternary materials, such as Cu2SnS3.

The market for photovoltaic (PV) modules has shown nearly exponential growth over the last years. Though the PV sector is booming, PV generated electricity in most places can still not compete in price with conventionally generated electricity and is consequently dependent on subsidies. To reach grid parity further price reductions for PV systems are required, calling for novel PV cell concepts based on cheap materials and low cost deposition methods. Metal oxide (MO) semiconductors are very attractive to achieve that goal; they are chemically stable, many MOs are non-toxic, abundant, and fulfill the requirements for low cost manufacturing methods at ambient conditions. Consequently, devices made of MO semiconductors can be very inexpensive, stable and environmentally safe, which are besides the conversion efficiency the most important requirements for photovoltaics. From the optical point of view there are a number of MOs suitable for PV applications. However, the electronic properties of most known MOs, i.e. short lifetime of excited electronic states and low mobility of electronic charge carriers, prevented their use as active solar cell materials.
To bring all-oxide PV to a breakthrough we believe that new materials have to be investigated. The prospect of finding these unique materials lies in combinatorial material science which can produce novel MOs consisting of two, three, four or more elements. While most binary MOs are known, the number of unknown compositions is drastically increasing with the number of components; AkBlOn, AkBlCmOn, etc. with A, B, C being different metals and k,l,m,n being integers.
Combinatorial MO synthesis allows producing a large number of material compositions on a single substrate. The new materials are tested in combinatorial PV device libraries to investigate them under PV operating conditions. Data analysis, storage and handling are organized using a dedicated database. Tools for data mining are applied to reveal unexpected or complex correlations between the material properties and the device performance and to improve our understanding of all-oxide PV device operation.
The methodology, new photoactive MOs and all oxide photovoltaic cells will be reported.

Recently, record efficiencies of 5.4% and 4.4% have been achieved for both wafer-based and thin-film Cu₂O respectively. Efficiency enhancement in both classes of device has largely been driven by interface engineering which involves controlling the chemistry and band-alignment at the hetero-interface using buffer layers to improve open-circuit voltage (V_oc). However, short-circuit currents (J_sc) for both device architectures are still below the 14 mA/cm² theoretical entitlement for a 2.1 eV bandgap absorber. Quantum efficiency analysis indicates that increasing charge collection length by reducing bulk recombination is an important next step towards higher J_sc. Furthermore, reducing bulk recombination can also allow the electron quasi-Fermi level in the absorber be pushed closer to the conduction band, further increasing V_oc. To this end, the ability to engineer Cu₂O in ways that can mitigate the effects of deleterious bulk defects is crucial. In addition, using appropriate tools to characterize relevant bulk electronic properties such as defect levels and bulk minority carrier lifetime (tau;_b) is important so that the absorber can be systematically optimized for PV devices.
In this work, we measure a high tau;_b of up to 10 mu;s in thermally oxidized 100 mu;m Cu₂O bare wafers using the microwave reflection photoconductance decay technique. This is achieved by tailoring the oxygen partial pressure and temperature profile in the furnace during the post-annealing step. We use complementary characterization techniques to elucidate the relationship between the growth process, defect-structure and electronic properties of Cu₂O. Spectrally-resolved room-temperature photoluminescence (PL) reveals a broad mid-gap defect band centered at 1.2 eV and samples with lower lifetime show comparatively stronger defect-related PL emission. This suggests that radiative recombination from mid-gap states is detrimental and dominates recombination activity in lower lifetime samples. Spatial PL mapping is used to gain insights into the structural dependence of defect-assisted recombination and PL images show that grain boundaries are relatively more recombination active than the bulk, pointing towards the need of larger grain sizes for higher device efficiencies. We will also present PV device characteristics to compare the effects of post-annealing conditions. We conclude by discussing how the framework used in this study can be applied more broadly to advance other Earth-abundant materials.

Cuprous oxide (Cu₂O) is a 2.1 eV bandgap semiconductor that holds promise for thin-film photovoltaics because of the low cost and abundance of its component elements and its straightforward processing. Crystalline wafers of Cu₂O with bulk carrier transport properties appropriate for a solar absorber can be synthesized by thermal oxidation of Cu foils. Furthermore, Cu₂O has a homojunction solar cell detailed balance efficiency of ~20% and also the potential for an independently connected Cu₂O/Si tandem device with a detailed balance efficiency of ~43%. Despite favorable minority carrier properties, the record Cu₂O solar cell efficiency is 5.38%. There are several challenges to making a Cu₂O solar cell, including difficulty in doping, its relatively low chemical stability, and a lack of suitable heterojunction partners due to its small electron affinity.
We have focused our research on the low chemical stability of Cu₂O, which is due to its low enthalpy of formation (-168.6 kJ/mol). The low heat of formation means Cu₂O will be reduced when in contact with nearly any elemental material. The effect of interfacial Cu on device performance has been well characterized, especially for Cu₂O Schottky devices. However, Cu has another stable oxide, CuO, which may also form at the interface in oxidizing conditions. The specific impact of CuO on device performance has not been characterized.
We have used XPS to characterize chemical reactions at the heterojunction interface by analyzing shifts in the core level and Auger peak positions of Cu₂O covered with 0.5-1.5 nm of Zn-IV material. For example, by altering the partial pressure of O₂ in the atmosphere during ZnO sputtering, we can produce a Cu₂O/ZnO interface that is Cu-rich, stoichiometric, or O-rich. Furthermore, photovoltaic device performance and valence band offset is strongly correlated with interface species. We have found the ZnO/Cu₂O heterojunctions made with stoichiometric interfaces have the highest open circuit voltages, V_OC = 530.4 ±4 mV. Devices with O-rich interfaces have V_OC = 102.3 ±11 mV. Devices with Cu-rich interfaces have V_OC = 347.2 ±30 mV. We have also found the valence band offset between Cu₂O and ZnO shifts from ΔE_C = -2.4 eV for stoichiometric interfaces to ΔE_C = -2.0 eV for Cu₂O/ZnO interfaces with CuO present. Valence band shifts with interface species have not been analyzed before for Cu₂O, and could account for some of the variation in reported Cu₂O/ZnO valence band offsets. We have performed similar analysis for other Cu₂O/Zn-VI interfaces including ZnS/Cu₂O and ZnSe/Cu₂O and will report those findings as well.

Cuprous oxide (Cu2O), despite its band gap of 2.17 eV, is a promising material for photovoltaic applications, due to its high absorption coefficient, non-toxicity and the abundance of its composing elements. While recently more attention has been paid to heterojunctions, highest efficiencies were reached by employing copper sheets, which were oxidized and annealed at high temperatures. For technological applicability, a thin film deposition process with mass production capabilities, which provides decent film properties at low temperatures, has to be established. Such thin films however suffer from low carrier mobilities and lifetimes, due to their polycrystalline nature. It has been reported that post treatments with hydrogen can passivate grain boundaries. Copper oxide thin films were deposited from a metallic copper target via reactive radio frequency magnetron sputtering, utilizing gaseous argon and oxygen under the addition of hydrogen. The films were characterized and the influence of hydrogen was investigated via X-ray diffraction, X-ray photoelectron spectroscopy, Raman spectroscopy, photoluminescence and Hall effect.

Cuprous oxide (Cu₂O), a classical p-type semiconductor, is a candidate solar cell absorber material that has been spotlighted recently due to its low cost, earth abundant and non-toxic nature. Although the Shockley Queisser limit of Cu₂O based solar cells is 20%, the maximum efficiency reported to date is rather low (5.38%). The achieved low efficiency can in part be attributed to the combination of Cu₂O forbidden direct band gap (2.1 eV) with higher absorption threshold (2.6 eV). If the lowest possible energy transitions can be made allowed then the absorption threshold could be reduced to the band gap value, potentially leading to enhancement in solar cell performance. In this study, we experimentally alloyed Cu₂O with ZnO to modify the Cu₂O electronic band structure and reduce its absorption threshold.
Cu-Zn-O thin film alloys were synthesized on a glass substrate as a function of temperature and composition from Cu₂O and ZnO targets via combinatorial RF magnetron sputtering with temperature and composition gradient at fixed 20 mTorr Ar pressure. Through this high throughput study, a non-equilibrium Cu-Zn-O phase map was generated in the temperature range 100 - 400 °C and Zn cation fraction of 0 - 37 at%. Highly crystalline Cu-Zn-O alloys with Zn range of 0 to 25 at% were synthesized in Cu₂O structure without any second phases in the temperature range 200 - 270 °C. In this temperature range, the Cu-Zn-O alloy thin films were preferentially oriented along (200) direction at lower Zn at% (<= 7 at%), and along (111) direction at higher Zn composition. At lower temperatures (T<200 °C), either amorphous or poor crystalline Cu₂O structured alloys were observed, whereas at higher temperatures (T > 270 °C) and higher Zn composition (>25 at%), CuO or ZnO second phases were observed. A similar Cu-Zn-O non-equilibrium phase map was also generated via combinatorial-pulsed laser deposition and will be compared.
The optical reflection and transmission was measured in the energy range from near infrared to visible and the absorption coefficient was calculated. The absorption coefficient of all the Cu-Zn-O alloys was found to be higher than that of phase pure Cu₂O in wide energy range, and the absorption threshold was also reduced. For example, the absorption threshold of Cu-Zn-O alloy (with 10 at% Zn) reduced to 2.1 eV from 2.4 eV in Cu₂O at 2*10⁴ cm^-1 absorption coefficient. The electrical conductivity of all the synthesized Cu-Zn-O thin films alloys including baseline Cu₂O were measured to be in the range 3 - 6 mS/cm.
This research is supported by the U.S. Department of Energy, office of Energy Efficiency and Renewable Energy.

Zn₃P₂ is considered an ideal candidate for scalable photovoltaics, with a reported direct band gap of 1.5 eV and a long minority-carrier diffusion length (>5µm). However, much remains to be studied regarding the electronic properties of heterojunction devices incorporating Zn₃P₂ as a thin-film absorber. In this work, we have determined the energy-band alignment of epitaxial heterojunctions between Zn₃P₂ and n-type II-VI materials: ZnS, ZnSe, ZnO, and CdS as well as III-V materials: GaAs and GaP. The valence-band discontinuities were determined using high-resolution X-ray photoelectron spectroscopy measurements via the Kraut method. Amongst the band alignments measured, the ZnSe/Zn₃P₂ heterojunction demonstrated a large valence-band offset of -1.21 ± 0.11 eV and a negligible conduction-band offset of -0.03 ± 0.11 eV, indicating a nearly ideal alignment for a photovoltaic device. High-resolution transmission electron micrographs of the ZnSe/Zn₃P₂ interface showed that the morphological properties of the ZnSe epilayer were dominated by a thin (~1.5 nm) layer of amorphous zinc phosphide at the crystalline Zn₃P₂ surface. Various growth conditions were investigated in an attempt to remove this amorphous interfacial layer and improve the ZnSe crystallinity and the extent to which the films could be doped. Finally, the device properties of the ZnSe/Zn₃P₂ heterojunctions, including substrate and superstrate configurations grown epitaxially on GaAs, were characterized using current-voltage measurements performed under dark and simulated Air Mass (AM) 1.5, 1-Sun illumination. These results represent significant progress towards the realization of efficient, earth abundant Zn₃P₂ solar cells.

In order to spread solar cells widely, it is desirable that solar cell materials are made of low-cost and earth-abundant elements. In this light, we focus on the II-IV-V2 chalcopyrite semiconductor compound, ZnSnP2, which is a p-type semiconductor with a bandgap of 1.66 eV. ZnSnP2 is reported to have high absorption coefficient. Therefore, ZnSnP2 is a promising material for solar absorber. For the application to solar cells, the formation of thin film is required. In this work, the method to obtain thin film of ZnSnP2 is thus proposed and the properties are investigated.
As a technique to obtain ZnSnP2 thin film, we adopted the phosphidation of Zn-Sn thin film prepared by sputtering technique. In this method, an appropriate source of phosphorus should be selected. We thus considered to use samples with Sn / Sn4P3 dual phase as a phosphorus source, by which the phosphorus vapor pressure could be accurately controlled by temperature.
The phosphidation of Zn-Sn thin film was carried out at 500 C, under the phosphorus vapor pressure of 10 to -2 atm. In phosphidation for 30 min, the formation of ZnSnP2 was observed, while metal Zn and Sn were not identified. This indicates that phosphidation of Zn-Sn thin film was completed. The obtained ZnSnP2 thin film shows a p-type semiconductor, and the density and the mobility of holes were 10 to 16 - 10 to 18 cm-3 and about 1 cm2V-1s-1, respectively. These properties are attractive for a solar absorber. However, many protrusions, which might be obstacles for making photovoltaics, were observed on the surface of the thin films. To improve the surface morphology, we discussed the mechanism of the phosphidation. When the phophidation was carried out for various period, some particles with high Sn concentration were formed in the initial stage, and they may grow up to the protrusions of ZnSnP2. It was concluded that the formation of the particles might be caused by the compressive stress from phosphides, Zn3P2 and ZnSnP2, formed at the initial stage. Therefore, we consider to add elements, which make intermetallic compounds with Sn, to supress the phase separation. In the conference, the results of its improvement and the performance of photovoltaics will be presented.

ZnSnN2, a member of the class of “earth-abundant element semiconductors”, is examined with emphasis on evaluating its suitability for photovoltaic applications. It is predicted to crystallize in an orthorhombic lattice with an energy gap of about 2 eV. Instead, using molecular beam epitaxy to deposit high-purity, single crystal as well as highly textured polycrystalline thin films, only the monoclinic phase is observed experimentally. Far from being detrimental, cation sublattice disorder which inhibits the orthorhombic lattice is shown to have a profound effect on the energy gap, obviating the need for alloying to match the solar spectrum.
CdO, meanwhile, was the first transparent conductor to be identified over a century ago, but it has yet to realise its full potential. Here temperature-dependent optical absorption, Hall effect, and infrared reflectance measurements have been performed on as-grown and post-growth annealed CdO films grown by metal organic vapor phase epitaxy on sapphire substrates. The evolution of the absorption edge and conduction electron plasmon energy with temperature has been modeled, including the effects arising from the Burstein-Moss shift and band gap renormalization. The zero-temperature fundamental direct band gap and band edge effective mass have been determined to be 2.31 eV and 0.27m0, respectively. The associated Varshni parameters for the temperature dependence of the band gap are found to be alpha = 8 x 10^4 eV/K and beta = 260 K. Additionally, the electron mobility has been studied by comparing the "optical" intra-grain mobility with the transport inter-grain mobility. While ionized impurity scattering is dominant within the grains, grain boundary scattering is found to limit the overall mobility of the CdO films.

Zinc tin nitride (ZnSnN2) is a novel earth-abundant semiconductor. It is a member of the II-IV-Nitrides, that are analogous to the well-characterized III-Nitrides currently employed in light-emitting diodes and sensors. Hybrid DFT simulations of electronic and crystal structure have motivated fabrication of ZnSnN2, as its calculated direct band gap of 1.4eV lies in a permissible regime for solar energy conversion. Like the III-Nitrides, where alloying enables one to tune the band gap energy, we have demonstrated tunable alloys of Zn-IV-Nitrides. Thus, we are developing the Zn-IV-Nitride alloys for their potential applications in photovoltaics, light-emitting diodes, and sensors. Electronic characterization is important for device fabrication, and herein we report on the photoconductivity of ZnSnN2.
We fabricated zinc tin nitride thin films by RF reactive sputtering on sapphire and GaN templates on sapphire. By x-ray diffraction studies, polycrystalline thin films have grain sizes ranging from 5-50nm. Sample resistivities are in the 0.01 Ohm-cm range, exhibiting high n-type doping.
Devices constructed with selected solid state and liquid electrical contacts all display ohmic behavior. Reducing ZnSnN2 carrier concentration might allow better junction rectification. Ohmically contacted samples illuminated with white light have demonstrated samples are photoconductive. Spectral response of photoconductivity with chopped illumination reveals that photoresponse is due to intraband trap states. Additionally, solid state samples also exhibit a long-time transient conductivity on time scales greater than milliseconds. Photoconductivity measurements were performed in vacuum and in inert atmosphere to probe for molecular effects. It is hypothesized that the long-time response originates from photogenerated carriers of shallow defect states releasing adsorbed oxygen molecules, as in n-type ZnO. Presence of a photoresponse provides a positive outlook for development of ZnSnN2 for applications in photonic devices or gas sensors.

Cu3N is the parent binary of an exciting new class of materials based on the formula Cu-M-N, where M = Earth-abundant metal ion. This class of materials is predicted to display defect-tolerance, meaning that material properties such as conductivity and minority carrier lifetime are not significantly affected by the presence of point defects, high-angle grain boundaries, or surfaces. Cu3N has provided experimental evidence of this predicted defect-tolerance, in that it can be doped either n-type or p-type based solely on growth conditions. In this presentation, the control of bipolar doping behavior as a function of growth conditions in Cu3N is demonstrated, and hypotheses as to the underlying physics of this behavior are explored.
Thin films of Cu3N were deposited from a metallic copper target using reactive RF-magnetron sputtering and an atomic nitrogen source. In one set of experiments, growth temperature was varied combinatorially from 150°C to 50°C. All depositions in this set of experiments were performed with a target power density of 1.5 Wcm-2. In another set of experiments, target power density was varied from 1.0 Wcm-2 to 3.0 Wcm-2. In this set of experiments, all depositions were performed at a nominal temperature of 50°C, i.e. no active heating. For both experiment sets, Hall effect and Seebeck coefficient measurements were used to characterize carrier type with respect to deposition conditions. Crystallographic profiles of each sample were obtained via X-ray diffraction to confirm phase purity. Samples were also characterized for absorption onset and coefficient using UV-Vis spectrophotometry. Sheet resistance was determined using 4-point probe coupled with thickness measurements obtained via Dektak profilometry. Finally, Near Edge X-ray Absorption Fine Structure (NEXAFS) measurements were performed to investigate the possibility that bipolar doping in Cu3N arises from fundamental differences in structure brought on by varying growth conditions.
It was found that Cu3N grown under copper-rich conditions in which the activity of nitrogen was low (T > 120°C or T < 70°C, or target power density >1.5 Wcm-2) exhibited n-type conductivity with Seebeck coefficients on the order of -10 mu;VK-1. However, films grown under copper-poor conditions in which the activity of nitrogen was high exhibited p-type conductivity with Seebeck coefficients on the order of +300 mu;VK-1. NEXAFS measurements revealed the presence of mixed Cu valency (both Cu+1 and Cu+2), and this discovery helps to shed light on the underlying reasons behind Cu3N bipolar doping behavior.
This research is supported by the U.S. Department of Energy, office of Energy Efficiency and Renewable Energy.

Copper antimony sulfide (CuSbS₂) is one of the promising candidates for photovoltaic applications due to its ideal optical band gap (1.5 eV) and its composition of earth abundant precursor elements. In this work, CuSbS₂ thin films were prepared by heating glass/Sb₂S₃/Cu layered structures in low vacuum (10^-3 Torr). The Sb₂S₃ thin films of different thicknesses were deposited from a chemical bath containing SbCl₃ and Na₂S₂O₃ at 40 °C on well cleaned glass substrates. A copper thin film of different thicknesses was thermally evaporated onto glass/Sb₂S₃ and the layered structures were heated at different conditions. The thin films formed were characterized using different techniques. X-ray diffraction analysis showed the formation of chalcostibite structure. Morphology was analyzed by atomic force microscopy. Depth profile of composition of the films was determined by X-ray photoelectron spectroscopy. Optical and electrical properties of thin films were evaluated. Further, the CuSbS₂ thin films were incorporated in PV structures using chemical bath deposited CdS as window layer. The photovoltaic parameters were evaluated from the corresponding J-V curves measured under illumination using a solar simulator of AM1.5 radiation, yielding the values of Jsc, Voc and FF as 7.54 mA/cm², 405 mV and 0.32 respectively for the best device so far. A mini-module was fabricated by connecting various cells in series and parallel. The results will be presented in the meeting.

Earth-abundant photovoltaic absorbers cuprous oxide (Cu2O) and tin monosulfide (SnS) show promise as scalable PV materials. We have reported thin-film solar cells fabricated from these materials achieving efficiencies of 2.65% (Cu2O) [1] and 2.04% (SnS) [2], and recently exceeded 4.4% and 4.6% efficiencies, respectively [unpublished]. Cu2O and SnS films are paired with a high work-function metal as the back contact, and an n-type emitter layer, transparent conductor, and metal finger electrodes on the front.
Dominant losses in each set of cells are the low open-circuit voltages (VOC), well below the detailed-balance limits for their respective bandgaps. This is suggestive of low quasi-Fermi level separation, which may be limited by bulk or interface recombination, the Fermi level of either contact, or a high density of defects pinning the Fermi level at the heterojunction interface. Identifying the contribution of each of these mechanisms to VOC deficit is critical for device engineering.

Temperature-dependent J-V measurements are used to extrapolate the barrier height, or VOC at 0 K. Band bending and heterojunction band offsets are determined through capacitance-voltage (C-V) and x-ray photoelectron spectroscopy (XPS) measurements. The oxidation state of species at the interface is also determined through XPS. The VOC is shown to be correlated with heterojunction band offsets, as well as interface chemical composition determined by specific fabrication conditions.
Learning from this loss identification, state-of-the-art cells are fabricated exhibiting high VOCs in excess of 0.85 V and 0.35 V for Cu2O and SnS, respectively. This suggests a promising path towards enabling commercially viable efficiencies in these material systems.
[1] Y. S. Lee, et al. Energy Environ. Sci. (2013), DOI: 10.1039/C3EE24461J.
[2] P. Sinsermsuksakul et al. Applied Physics Letters 102, 053901 (2013).

2D materials such as the TDMC and graphene are called to be the basis of a revolution on the micro and nano-electronics. The recent performance records obtained on MoS2 transistors and phototransistors are suggesting the potential of this material to be used in the next generation of revolutionary optoelectronics application. Besides, MoS2 can be grown at a large scale in different commercial substrates enabling the material to be an alternative for cheap mass production application such as flexible electronics on tablets and cell phones or photodetection in general. The remarkable interaction of light and electricity found in the MoS2 monolayers and multilayers could also be potentially used in photovoltaic generation.
In this work, we present preliminary results on Schottky-based MoS2 junctions, and their possible use for solar to electric energy conversion. Large area MoS2 nanosheets were first achieved by direct sulfurization of thin Mo layers on Silicon wafer-scale substrates. The electrical behaviour of vertical MoS2/Si heterojunctions was then investigated at the micro and nanoscale. Thickness, surface roughness, morphology and crystalline quality of the MoS2 layers trough AFM, SEM, XRD and Raman are presented. The electrical properties at the nanoscale are further investigated by conductive AFM and allowed to identify that the conductivity of MoS2 is limited by the presence of grains and is thickness dependent. Grain conduction is revealed for thick (>30 nm) MoS2 layers while it becomes adjacent to the grain boundaries for thin (<15 nm) layers. Current-voltage measurements were then performed on large area (4×4 mm2) MoS2/Si heterojunction devices. The thickest device exhibited a p-type Schottky diode behavior (>104 on/off ratio), being conductive in forward bias and with a relatively low reverse leakage current. In contrast, when the MoS2 thickness is reduced, the devices exhibited higher current in reverse bias than in forward bias. This electrical response variation may be due to Si surface modifications during the sulfurization process. It is believed that thinner samples have a patch-contact topology leaving some Si regions exposed to the sulfur. This naturally makes the electrical response inhomogeneous in nature depending locally on the Si/MoS2 interface. S is well known to segregate at the interface acting as S donor-like traps or recombination centre, thus modifying the Schottky barrier height. The nanoscale mapping properties are therefore consistent with the large diode electrical response. Recent measurements done in light condition revealed that all the devices showed photosensitive behaviour, exhibiting relatively large Ilight/Idark ratio (superior to 10), suggesting that electron-hole pairs could be efficiently generated in the MoS2/Si Schottky-based junctions, and probably even applicable to solar cells. Other MoS2-based heterojunctions and Schottky diodes for solar cells applications are currently being investigated.

Chalcopyrite-type semiconductor CuInSe2 (CIS) is one of the most promising materials for low cost photovoltaic solar-cells. CIS has direct band bap which is suitable for absorption of sunlight and large light absorption coefficient. Moreover, due to the self-regeneration mechanism [1], it is known that the processing of CIS is rather easy compared to Si-based photovoltaic solar cells. However, from the point of resource security, high concentration of In in CIS is serious disadvantage. Recently, Cu2ZnSnS4 (CZTS) attracts much attention to overcome this disadvantage of CIS. In CZTS, all of In in CIS are replaced by Zn and Sn. This material has already been investigated as a photovoltaic material but the efficiency is no high enough [2]. In this paper, we propose how we can enhance the efficiency of CZTS by utilizing the self-organization phenomena caused by spinodal nano-decomposition of Cu & Cu-vacancy, S & Se, and Se & Oxygen. We will compare our design with the available experimental data of STEM-EDX, EELS, Atom Probe Tomography and Raman Scattering data.
Our calculations are based on the local density approximation (LDA) in the density functional theory. To correct the self-interaction in the LDA, we employ the self-interaction correction method originally proposed by Perdew and Zunger [3] and developed by Filippetti and Spaldin [4]. Substitutional disorder in co-doped CIS is treated by using the Korringa-Kohn-Rostoker coherent potential approximation (KKR-CPA) method. We use the MACHIKANEYAMA package for the present calculations [5].
In order to enhance the efficiency of solar energy conversion, we utilize spinodal decomposition. In CZTS, we introduce Se impurities at S sites. Mixing energy DE(x) = E(Cu2ZnSn(S1-xSex)4) - [(1-x)E(Cu2Zn SnS4)+xE(Cu2ZnSnSe4)] is calculated by using the KKR-CPA, where E is the total energy of the system. Positive DE indicates the tendency to phase separation and homogeneous mixing of S and Se is not favored in this system. In the calculated phase diagram, rather high spinodal temperature is predicted and we can expect spinodal nano-decomposition. According to our Monte Carlo simulations it is found that nano-structures with high concentration of impurities are self-organized under the layer-by-layer crystal growth condition. Since type-II band alignment is expected between Cu2ZnSnSe4 and Cu2ZnSnS4, we can expect efficient electron-hole separation at the interfaces between host and self-organized nanostructures [6].
References:
[1] S. B. Zhang et al., Phys. Rev. B57 (1998) 9642.
[2] D. A. R. Barkhouse et al., Prog. Photovolt: Res. Appl. 20 (2012) 6.
[3] J. P. Perdew and A. Zunger, Phys. Rev. B 23 (1981) 5048.
[4] A. Filippetti and N. A. Spaldin, Phys. Rev. B 67 (2003) 125109.
[5] H. Akai, http://sham.phys.sci.osaka-u.ac.jp/kkr/
[6] Y. Tani et al., Appl. Phys. Express 3 (2010) 101201. Jpn. J. Appl. Phys. 51 (2012) 050202.

ZnS thin films are ideal candidates for replacing CdS as buffer layers in CIGS, CZTS and CdTe photovoltaics because of their abundance, low toxicity, cost, and wider band gap. However, ZnS films deposited by chemical bath deposition (CBD) tend to incorporate oxygen in the form of oxide and hydroxide (thus called Zn(S,O,OH)). These impurities affect the stability and performance of the cells. In a batch CBD, bath composition varies with time, resulting in the film&’s properties and stoichiometry varying through its thickness. Understanding depth-dependent film properties is crucial, but measuring these properties is very challenging because the film thickness is small relative to the resolution of many characterization methods.
We report on CBD-Zn(S,O,OH) thin films using a continuous flow microreactor. The microreactor uses a sub-millimeter reaction channel, with the substrate acting as one reactor wall. The microreactor behaves like a plug flow reactor whereby the bath composition changes as the reaction proceeds with residence time down the channel, but the composition at any position is time-invariant. Transposing time-evolution into spatial-evolution results in a thin film with stoichiometry graded laterally over centimeters rather than through the ~50 nm thickness of a batch-deposited film. Spatially resolved characterization of the substrate enables rapid and direct correlation of material properties to growth conditions.
Laterally graded Zn(S,O,OH) films were grown from zinc salt, ammonia, and thiourea (TU) with 40 min residence time over a distance of 40 mm substrates. TU acts as a slow source of sulfur. Film morphology on glass is dominated by highly monodispersed nodules; small and frequent at initial positions, and large but sparse downstream. Deposition time- and spatial-dependent statistical analysis of nodules&’ size and number density provides a multi-dimensional view on the trajectory of nucleation and growth. Bath HS-/S2- concentration measured by ion-selective electrode (ISE) and bath pH evolution indicate that homogeneous nucleation dominates due to high OH- concentration at earlier positions, while growth is favored at downstream positions with high HS-/S2- concentration. Continuous thin films were grown on CdSe-coated substrates, and the films&’ spatially-dependent S/(S+O) and (OH)2/[O+(OH)2] ratios determined by XPS support the proposed mechanism. In contrast, thioacetamide provides faster S release kinetics than TU. Very fast deposition occurs at earlier positions with high nucleation and growth rate due to high HS-/S2- concentration and pH. Such fast deposition resulted in trapped water clusters in the film. We are investigating structural and optoelectronic properties to better understand the graded films and their potential for solar cells. The continuous flow microreactor offers unique insights into material formation trajectories and provides precise processing conditions for targeted material properties.

The Cu2ZnSnS4 (CZTS) and its related compounds semiconductors are the most promising absorber materials for the thin films solar cells because of their abundance, low toxicity, optical and electrical properties. The technological advance of CZTS via evaporation process and Cu2ZnSnSe4 (CZTSe) solar cells have reached efficiencies up 8.4 % and 11.1 %, respectively. Cu2ZnSn(SSe)4 (CZTSSe), also commonly referred to as kesterites because of their crystal structure, is of particular interest due to the achievement of 11.1 % efficient using nonvacuum, hydrazine based deposition process. These research results demonstrate CZTS based solar cells are currently the most promising materials to replace Cu(InGa)Se2 absorbers in photovoltaic device.
In this work, we demonstrate a simple route to fabricate CZTSSe films by using RF/DC sputtering from stacking orders of metal precursor. The approach for the stacking order of metal precursor has some advantages over simultaneous deposited precursor by using co-sputtering. Co-sputtering has the advantage of allowing an easy tuning and control of the percentage of composition ratio. However, this method is not suitable to fabricate large-scale and mass production. The stacking orders of metal precursor, on the other hand, is cost-effectively viable for large area production in the industrial field.
We recently submitted the paper about effect of precursor order on the properties of sulfurized Cu2ZnSnS4 thin films for solar cells. In this work, we found that the proper precursor sequence for the CZTS growth is Cu/SnS/ZnS/Mo. This structure allows for better control of the composition of the CZTS without Sn and Zn loss. We also used a compound material target such as thin sulfide (SnS) and zice sulfide (ZnS). They could lead to good morphological properties of the metal precursor films in the sputtering process. The CZTSSe absorber layer was simply fabricated for selenization annealing process of metal precursor film in the presence of S. The effects of the different annealing temperature in the metal precursor having S on the structural, morphological, electrical properties of CZTSSe absorber films were investigated. The best cell showed an open-circuit voltage of 0.42 V, a short-circuit current of 35.65 mA/cm2, a fill factor of 60.70 %, and a conversion efficiency of 9.1 %.

Cu2-xS is a direct p-type semiconductor with a band-gap that depends on its stoichiometry (1.1 -1.4 eV for x= 0 - 0.04; 1.5 eV for x= 0.2; 2.0 eV for x= 1). The combination of a suitable band gap, high absorption coefficient (104 cm-1), low cost, and low toxicity has made Cu2-xS a prime candidate for the large scale and sustainable deployment of photovoltaics. In fact, CdS/Cu2S heterojunctions were among the earliest thin-film solar cells to be investigated, but their further development was hindered by stability issues. Nevertheless, the possibility to make colloidal Cu2-xS NCs has created new ways to fabricate Cu2-xS based PV cells by using solution based techniques, leading to a renewed interest in them. Moreover, Cu2-xY (Y= S, Se, Te) NCs have been shown to possess the unique property of holding excitons and highly tunable localized surface plasmon resonances (LSPs) on demand. This makes Cu2-xY (Y= S, Se, Te) NCs promising materials for photovoltaics, photocatalysis, and nanoplasmonics.
The application of colloidal NCs in photovoltaic and nanoplasmonic devices requires strict control over the size, shape and polydispersity of the NC ensemble, since these characteristics are of crucial importance not only for the optoelectronic properties of the NCs themselves, but also for the quality of the NC thin films obtained by solution based deposition techniques. Many methods have been developed in recent years to control the size and shape of Cu2-xS NCs, but they lack flexibility, since different sets of physical-chemical parameters (concentrations, ligands, and reaction temperatures) have to be used for each different shape. In this work, a novel methodology for the size and shape control of colloidal Cu2-xS NCs was developed, which relies on changing just one single reaction variable: the concentration of Sn(IV)-compounds that are used to control the nucleation and growth rates of Cu2-xS NCs. Besides being simpler and more versatile than the other currently available strategies to control the size- and shape of colloidal Cu2-xS NCs, the method developed in our work also provides access to shapes not attainable by other methodologies. The improved size and shape control achieved in this work is beneficial not only to applications based on Cu2-xY NCs, but may also have a more general impact, since Cu+ ions in copper chalcogenides have been shown to be easily exchangeable by other cations. This opens up the possibility of using partial cation exchange reactions to convert Cu2-xY into other copper-based chalcogenides, such as CuInS2 (CIS) or Cu2ZnSnS4 (CZTS), with preservation of the size and shape of the colloidal NCs.

With the aim of exploring the anion arrangement on structural diversity and thermodynamic stability of Cu2ZnSnSxO4-x (CZTSO) and Cu2ZnSnSxSe4-x (CZTSSe) (0le;xle;4), the influence of oxidization and selenization of Cu2ZnSnS4 (CZTS) is studied by using first principle calculations. The pure Cu2ZnSnO4 (CZTO) was found to possess the lowest heat formation, followed by CZTS and finally Cu2ZnSnSe4 (CZTSe), which confirms the fact that oxidization process is inevitable, whereas selenization can only be accomplished under high temperature. The thermodynamic stability exploration reveals that ZnS/Se, CuS/Se, SnS/Se and Cu2SnS3/Cu2SnSe3 are the main secondary impure phases that limit the formation of pure CZTS /CZTSe. Also, it was found out that the pure CZTS is easier to fabricate than CZTSe, whereas selenization of CZTS is more difficulty than vulcanization of CZTSe, and vice versa. The effect of spatial anion distribution in the conventional cell is significantly different for oxygen and selenium atoms: the oxygen atoms have preferable sites and try to exclude with each other in the doped CZTSO system, while selenium atoms seem to have no preferable sites and are believed to distribute randomly in the CZTSSe.

Cu₂ZnSnS₄ (CZTS) is attracting considerable attentions as a less-toxic and low-cost absorber material for solar cells due to its optimal band-gap energy for solar energy conversion and its high absorption coefficients in near-infrared and visible spectral regions. Up to date, the best performance of CZTS-based solar cells has reached an energy conversion efficiency of ~8.4% [1], which is still considerably lower than the high conversion efficiency of 20.3% obtained for CuIn_1-xGa_xSe thin-film solar cells [2]. To enhance the conversion efficiency of CZTS-based solar cells, insightful understandings of fundamental optical properties of CZTS are important.
Because of the quaternary composition of CZTS, defect states are believed to take an important role in their electrical and optical responses [3,4]. Several photocarrier recombination models relating to the defect states have been proposed for CZTS based mainly on steady photoluminescence (PL) measurements [5-7]. To determine more exactly photocarrier recombination mechanisms in CZTS, time-resolved studies, for which there are only a few reports to date, might be essentially necessary.
In this study, we clarify the photocarrier localization and recombination dynamics in CZTS single crystals at room temperature (RT). The band-gap energies was identified consistently by various means, including PL excitation (PLE), photocurrent (PC), and transient reflectivity spectroscopy, and was estimated to be about 1.58 eV at RT. The large-density band tail states formed below the band edge determines the optical responses. The PL band, therefore, was assigned to radiative recombination of carriers localized in band tail states. We found that the photocarriers are rapidly localized to shallow tail states within a typical time constant of several picoseconds to a few tens of picoseconds and multiple trapping processes of carriers occur near the band edge.
This work was supported by JST-CREST and the Sumitomo Electric Industries Group CSR Foundation.
References
[1] B. Shin et al., Prog. Photovoltaics 21, 72 (2013).
[2] P. Jackson et al., Prog. Photovoltaics 19, 894 (2010).
[3] A. Nagaoka et al., Appl. Phys. Lett. 103, 112107 (2013).
[4] L. Q. Phuong et al., Appl. Phys. Lett. 103 (2013) in press.
[5] K. Tanaka et al., Phys. Status Solidi A 203, 2891 (2006).
[6] M. J. Romero et al., Phys. Rev. B 84, 165324 (2011).
[7] S. Levcenko et al., Phys. Rev. B 86, 045206 (2012).

Cu2ZnSnS4 (CZTS) solar cell, which is reported to have a direct band gap of 1.4~1.5 eV and optical absorption coefficient of above 104 cm-1, has been extensively studied in recent years because it is known to be a candidate to replace CIGS for low cost thin film solar cell. In general CZTS is obtained by sulfurization of precursor prepared by both vacuum process (sputtering deposition, co-evaporation) and non-vacuum process (spray, electro-deposition and hydrazine-based solution process). In particular, sputtering process is favorable to obtain uniform thin film precursor with controllable composition. As for the sulfurization, the annealing with H2S gas has advantage for preparing CZTS solar cell under the controlled sulfur partial pressure. Therefore, sulfurization of metal precursor prepared by sputtering of pure Cu, Zn and Sn targets with H2S gas is suitable for fabrication of CZTS solar cell with large-scale at low cost. However, it has been reported that CZTS prepared by sulfurization of pure metal precursor with H2S gas showed low cell performance.
In this study, we investigated the influence of metal precursor structure on the microstructure, composition distribution and solar cell performance of CZTS thin film prepared by sulfurization with H2S gas. Cu-Zn-Sn thin films with stacked structure were deposited on the Mo coated glass substrate by magnetron sputtering of pure Cu, Zn and Sn metal targets. The stacked structures of precursors were (Cu,Zn,Sn)/Mo, (Cu,Sn)/Zn/Mo and Cu/(Zn,Sn)/Mo. The stacked precursors were sulfurized using a rapid thermal annealing(RTA) in the mixed N2+H2S atmosphere at 550°C. The microstructure and composition of CZTS thin films were investigated by X-ray diffraction(XRD), Raman spectroscopy, secondary electron microscopy(SEM), electron probe micro-analyzer(EPMA) and auger electron microscopy(AES). To fabricate CZTS solar cells, CdS buffer layer was deposited on CZTS thin film by chemical bath deposition and ZnO window layer was deposited by magnetron sputtering on CdS layer. Ni/Al grid was deposited by e-beam evaporation. The cell performance was measured by current-voltage measurements under standard AM1.5 illumination at 25°C. In contrast to the CZTS solar cell prepared by (Cu,Zn,Sn)/Mo precursor which showed very low conversion efficiency (~0.1%), the CZTS solar cell prepared by Cu/(Zn,Sn)/Mo and (Cu,Sn)/Zn/Mo precursors showed 1.2% and 2.6% cell efficiency, respectively. The cell efficiency of CZTS solar cell prepared with (Cu,Sn)/Zn/Mo precursor was greatly improved to be 4.3% by the incorporation of Se. The change in cell performance of CZTS thin films with different precursor structure will be discussed in terms of uniformity of microstructure and chemical composition through the formation of ZnS and Cu2SnS3 phases.

CZTS monograin layer solar cell technology is an alternative to the currently studied thin film technology. In this report the impact of preparation conditions on the properties of CZTS absorber material is presented. CZTS was synthesized in monograin powder form in molten CdI2 as a flux material. The influence of synthesis temperature and time on the chemical and phase composition and on optical properties of the absorber material was studied. Raman scattering, photoluminescence (PL), EDX and XPS analyses of powder samples are reported with the aim to understand the possible impact of preparation conditions on the phase composition and on the photoelectrical properties of CZTS monograin powders. The studies show that Cd incorporates from molten CdI2 into CZTS and the synthesis results in solid solution of Cu2Zn1-xCdxSnS4. There is found a strong correlation between the preparation conditions and the composition of the absorber compound which in turn impacts on the band gap value of the synthesised material.

Cu2ZnSnS4 (CZTS)-based solar cells represent a significant improvement because of a useful band-gap energy and a high absorption coefficient. In this study, we investigated two different effects of Na on CZTS solar cell characteristics.
We verified that the solar cell characteristics improve in direct proportion to the Na concentration in a Mo layer before CZTS deposition. The Na concentration in the Mo layer increased in proportion to the increase in Mo-layer annealing temperature. Secondary-ion mass spectrometry (SIMS) composition profiles show that Na is accumulated near the Mo-layer surface. The Na accumulated on the surface is rapidly diffused into the absorber layer at the initial stage of CZTS deposition. To assess the degree of defect within the absorber layer, we measured the voltage shortfall (Egminus;qVOC) as a function of the Mo-layer annealing temperature. As the Mo-layer annealing temperature increased, the voltage shortfall decreased. This indicates that the concentration of recombination centers in the absorber layer was reduced. The diffusion of a greater Na concentration into the absorber layer at the initial stage of CZTS deposition results in fewer defects in the absorber. In addition, the grain boundaries (GBs) in an absorber layer with a higher Na concentration are passivated more efficiently. A decrease in recombination-induced losses is observed in the samples that exhibit a more efficient passivation of GBs that act as a current flow channel. Such a correlation indicates that a CZTS absorber formed on a Mo layer with a high initial Na concentration contains fewer bulk defects, which correspond to a decrease in VOC losses and improvement of solar cell characteristics. Device with the Mo-layer annealed at 600°C improved from 494 to 585 mV in VOC, from 5.66 to 15.65 mA/cm2 in JSC and from 1.49 to 4.33 % in efficiency (Eff.).
Also we verified that the solar cell characteristics decline in direct proportion to the Na concentration in a CdS buffer layer. SLG/Mo/CZTS/CdS/transparent conducting oxide (TCO)-structured sample was annealed to control the Na concentration in a buffer layer. After CZTS absorber layer sulfurization process, Na is accumulated near the absorber layer surface. When a buffer layer was deposited on the absorber layer, Na near the absorber layer surface is diffused into a buffer layer. Transmission energy dispersive spectrometer (EDS) compositional profiles show that the Na concentration in a buffer layer decreased after annealing process. Na in a buffer layer acts as an ionized impurity and scatters the minority carrier; thus, the minority carrier path is deflected and the minority carrier mobility decreases. As the minority carrier mobility decreases, device series resistance (Rs) increases. So Na concentration in a buffer layer must be reduced. Device with TCO layer annealed at 300°C improved from 52.8 to 9.5 Omega; cm2 in Rs from 47.2 to 60.4 % in Fill Factor and from 4.33 to 6.48 % in Eff.

Polycrystalline Cu(In,Ga)Se2 (CIGS)-based solar cells are attracting attention because of their high absorption coefficient and optimal direct bandgap. However, some of the constituent elements in CIGS, such as In and Ga, are not abundant in nature. Additionally, Se is a toxic element. On the other hand, SnS is a promising candidate for a solar cell material because it has a high absorption coefficient of 1E04 cm-1 and direct bandgap energy of 1.3 eV. In addition, the constituent elements of binary SnS are relatively inexpensive and nontoxic.
Despite these promising properties, the present conversion efficiency of SnS solar cells is still low. Some reasons for this low conversion efficiency are the poor crystal quality and unknown physical properties of SnS. To date, there are only a few reports regarding the electrical and optical properties of SnS thin films and solar cells. In particular, the minority carrier lifetime and diffusion length tend to depend on the type and magnitude of the recombination processes inside the semiconductor. However, complex equipment are required to measure actual carrier lifetime.
Electrochemical impedance spectroscopy (EIS) analytic theory is increasingly being used as an analytical theory in chemical material research and has been shown to be invaluable for semiconductor studies of materials and devices such as CIGS solar cells. EIS is a simple and promising method for characterizing the heterogeneity near the pn-interface. In this presentation, we propose the carrier-transfer time-constant (quasi-lifetime) of SnS solar cells measured by EIS. We will describe the relationship between the quasi-lifetime that is reflected by carrier concentration or interface defects of SnS films and the electrical performance of SnS solar cell.

In the last years, interest in kesterite material Cu2ZnSn(S,Se)4 (CZTS) increased rapidly because it includes only non-toxic, earth-abundant and low-cost elements and because of the possible future difficulties on In and Te supply. The objective of this work is to understand what is the influence of the selenium partial pressure on the reaction path leading to the formation of Cu2ZnSnSe4, to determine the respective influence of kinetics and thermodynamics in the formation of secondary phases, and to use this understanding to synthesize Cu2ZnSnSe4 thin films with crystallinity and opto-electronic properties suitable high-efficiency solar cells.
In order to achieve that goal, we used cosputtered metallic precursors and we annealed them in a graphite box. In such conditions, for the sulfur compound, it was shown that because of sulfur loss, Cu2ZnSnS4 contains Sn in a (II) oxidation state [1]. The reduction of Sn from Sn(IV) to Sn(II) leads to the instability of CZTS and produces the SnS secondary phase. As SnS is a volatile element, its evaporation displaces the equilibrium of the reaction Cu2ZnSnS4 harr; (SnS + Cu2S + ZnS + ½ S2 towards the right member, leading to loss of Cu2ZnSnS4. The observations on Cu2ZnSnS4 are fully extrapolated on Cu2ZnSnSe4 because of the volatility of SnSe [2].
An Ellingham diagram of the binary selenides was built, showing that partial selenium pressure was a key parameter in reaction control during the annealing step. SnSe and MoSe2 appear for long annealing time or high temperatures, and pressure conditions favoring the stability of SnSe2 in the layer were found to be a way of minimizing CZTSe degradation. The existence of this compound in the layer was found to be dependant both on kinetics and thermodynamics, as well as the conversion of Mo into MoSe2
Nevertheless, even with SnSe2 inside the absorber layer, the films exhibit large grains with high crystalline quality. Glass / Mo / CZTSe / CdS / ZnO solar cells were synthesized, and conversion efficiency of 7.1 % were reached: at thiss level of performance, control of phase purity and MoSe2 thickness is not critical.
[1] J. J. Scragg, T. Ericson, T. Kubart, M. Edoff, and C. Platzer-Björkman, “Chemical Insights into the Instability of Cu2ZnSnS4 Films during Annealing,” Chem. Mater., vol. 23, no. 20, pp. 4625-4633, Oct. 2011.
[2] A. Redinger, D. M. Berg, P. J. Dale, and S. Siebentritt, “The Consequences of Kesterite Equilibria for Efficient Solar Cells,” J. Am. Chem. Soc., p. 156, 2011.

Visible-light transparent solar cells composed of NiO/NixZn1-xO/ZnO were fabricated. Leakage current tends to reduce by insertion of high-resistivity NixZn1-xO interlayer because of reduction it prevents abrupt band offsets and large lattice mismatch between NiO and ZnO. This result indicates that NixZn1-xO layer is effective for fabrication of NiO/ZnO oxide devices.
Transparent oxide semiconductors (TOSs) with wide bandgaps are mainly used in transparent conducting films of solar cells. In general, TOSs such as ZnO and indium tin oxide (ITO) exhibit n-type conductivity. Therefore, both p-type and n-type semiconductors having wide bandgaps are required to form transparent pn-junctions for visible light to pass through. NiO makes the visible light pass through it and is a promising candidate for a p-type TOS due to its wide bandgap energy (3.7 eV). NiO and n-type TOSs such as ZnO have been used in transistors, UV-visible light emitting diodes, and UV detectors. Therefore, NiO may be used in a “multifunctional window,” which comprises visible-light transparent solar cells, sensors, electronic circuits, and so on. Visible-light transparent devices are preferable because their optical transparency is not limited the installation locations.
There are several reports about abrupt band offsets and large lattice mismatch between NiO and ZnO. In general, the large band offsets cause a reduction in the threshold voltage as they lead to tunneling effects, which is an origin of leakage current. Moreover, the large lattice mismatch leads to a numerous defects and carrier recombination. In this presentation, we will investigate the use of NixZn1-xO (NZO) thin films as an interlayer or absorption layer for visible-light transparent solar cells, with the aim of preventing the abrupt band offsets of the NiO/ZnO heterojunction, while focusing on their band diagram, electrical, and optical properties. Each layer of the solar cells, including NZO, NiO, and ZnO has been deposited by sputtering, which is the most suitable method because it enables economical deposition of large-area films with well-controlled compositions. The critical importance of the Ni-Zn ratio in reducing the band offsets and lattice mismatch between NiO and ZnO will be addressed. In addition, the effect of the surface morphology of NZO thin films on the NiO/NZO/ZnO heterojunctions will be shown.
Polycrystalline NZO thin films with a thickness of approximately 200-400 nm were deposited by RF reactive co-sputtering using Ni metal and ZnO ceramics as targets and Ar-O2 as the sputtering gas. The valence band discontinuities were determined using the photoelectron yield spectroscopy measurements. The fabricated NZO-related solar cells exhibited photovoltaic effect under illumination. Moreover, optical transmittance of greater than 70% was obtained in the wavelength range of 400-800 nm. These results indicate that NZO is effective material for transparent electronic devices such as solar cells.

Single phase cupric oxide (CuO) thin films are deposited on SiO₂/Si substrates by reactive ion beam sputter deposition at 300 ~ 400^oC utilizing an anode layer ion source. Both argon and oxygen are passed simultaneously through the ion source to act as sputtering and reacting species, respectively. With an Ar:O₂ ratio of 2:1, single phase polycrystalline CuO thin films are obtained. As Ar:O₂ ratio reaches 3:1 ~ 4:1 , mixed CuO and Cu₂O are found. Further increasing Ar:O₂ ratio to 5:1 ~ 7:1 results in single phase polycrystalline Cu₂O. As Ar:O₂ reaches 8:1 and higher, the as-deposited sample shows mixed Cu₂O and Cu phases and are covered by Cu₂O nanorods. The bandgap of CuO and Cu₂O are 1.36 and 2.46 eV, respectively, both measured by optical transmission spectroscopy. CuO metal-semiconductor-metal (MSM) Schottky photodiodes (PD) are fabricated by depositing Cu interdigitated electrodes on CuO thin films at room temperature. Photosensing properties of the CuO MSM PD are characterized from 350 ~ 1300 nm and a maximum responsivity of 1 mA/W is found at 700 nm. The photocurrent transient properties are analyzed using a stretched exponential model and the decay time is found to be 1.0 ms. These findings may serve as valuable information for the optimization of CuO based phovovoltaic devices.

Photovoltaic (PV) devices usually exploit mid-range band-gap semiconductors which absorb in the visible range of the solar spectrum. However, much energy is lost in the IR and near-IR range. Efficient PV devices require fine tuning of the energy levels at interfaces between the absorber and the electrodes. Since IR absorbers possess a small band-gap, such tuning is difficult using common electrodes. While anodes are typically photosensitized by semiconductor deposition, cathode sensitization, requiring photo-induced hole injection, can expand the choice of materials for efficient PV energy conversion, and advance the basic understanding as well as device development for IR optoelectronics. In this work we study bulk-like PbS deposited on p-CuSCN to produce photocathodes for low cost photovoltaic devices utilizing IR and near-IR light.
Copper thiocyanate (CuSCN) is a promising cathode material with good hole conductivity, chemical stability, and transparency in the visible-IR range. Ivanova at el. [Electrochemistry communications 24, 1-4 (2012)] suggested the template-free electrodeposition of CuSCN nanowires, which can be used as a large surface area cathode for IR photovoltaics. An all-solid PV device was fabricated using PbS deposited on CuSCN nanowire array by facile, low cost, direct chemical bath deposition, with complete coverage of the complex nanowire morphology (crucial for preventing shorts), and evaporated Ag contacts as the anode. Photocurrent measurements under near-IR (784 nm) and mid-IR (1550 nm) illumination show photogenerated currents indicating light-induced hole transfer at the PbS/CuSCN interface. The ability to use large area p-type electrodes to extract charge carriers from a narrow band-gap semiconductor can advance the field towards high efficiency, low cost, IR and near-IR sensors and solar cells.

We use a variety of appropriate quantum mechanics methods to search for robust, efficient, and inexpensive materials for photo-catalytic electrodes (PCEs) that may convert sunlight, carbon dioxide, and water into fuels. Despite periodic media reports to the contrary, no efficient PCEs are available yet. I will discuss why it is so difficult to find effective PCE materials; in particular I will enumerate the nontrivial constraints that they must satisfy to achieve high efficiency. Limiting oneself to abundant elements further constrains the design space. As a result, we focus primarily on first row transition metal oxide parent materials. Key properties of conventional and novel materials, along with some new design principles, will be discussed. The work is revealing which dopants or mixed oxides are likely to provide efficient solar energy conversion materials.

In context of photoanodes for artificial photosynthesis, we describe combinatorial synthesis and characterization of earth abundant oxynitride photoabsorbers. Specifically, we focus on transition metal oxynitride ternary and quaternary phases with a band gap in the range 1.5-2.5 eV synthesized via a custom-built high-throughput combinatorial sputtering tool. Photoelectrochemical screening is used to ascertain the utility of these semiconductors for solar fuels applications.
While metal oxide semiconductors have been extensively studied as photoanodes, they tend to have a deep valence band (VB) dominated by O2p orbitals resulting in wide band-gap that is too large for efficient solar water splitting. Alternate anion chemistries such as nitrides and sulfides have been proposed to overcome this challenge since their VB position is usually higher in energy but these materials suffer from poor electrochemical stability. As a result, there is significant interest in identifying metal oxynitrides that afford tailoring material properties between those of metal oxides and metal nitrides. Our custom built high-throughput combinatorial sputtering tool equipped with six magnetron sputtering guns, varied substrate-source deposition geometries and µTorr level control of gas pressure allows us to explore multi-cation oxynitrides with independent degrees of freedom for the composition of transition metals and oxygen/nitrogen ratio. Synthesized materials are screened for band gap values within the range 1.5-2.5 eV using high-throughput diffuse reflectance UV-Vis measurements. The materials of interest are further characterized for incident photon conversion efficiency (IPCE) and phase information using high-throughput quantum efficiency and X-ray diffraction measurements. The application of more detailed characterization and interfacing with theory efforts will also be discussed.

Most low bandgap metal oxide materials with water splitting potential are not well explored with respect to charge carrier transport properties and carrier lifetimes. A variety of promising oxide based photoelectrocatalyst materials have been identified that satisfy energetic requirements for photocatalysis but almost all suffer from low mobility and small carrier diffusion lengths limiting the photocatalytic performance [1]. This substantially depends on the individual preparation techniques applied for thin film deposition of photoelectroactive layers. For an advanced development of metal oxide photoelectrodes it is crucial to understand how material composition is related to carrier transport and capture dynamics in a wide time range. We investigate bulk and interface reactions in metal oxides such as BiVO4 and Fe2O3 with time resolved spectroscopy from femtoseconds to milliseconds applying transient absorption, THz and microwave conductivity in transmission and reflection mode. The conductivity measurements reveal the photo-excited carrier lifetime in dependence of (oxygen)vacancies, interstitials, impurities or doping while optical absorption supplies information on the dynamics and energetics of bandgap states. Also, THz mobility spectra recorded at different times after photoexcitation yield profiles of the real and imaginary parts of the conductivity that allow for microscopic understanding of the transport mechanisms. Sub-picosecond THz photoconductivity is demonstrated for the first time for Fe2O3 prepared by pulsed laser deposition exhibiting ultrafast decay on the order of a picosecond in agreement with transient absorption measurements [2]. Non-Drude conductivity is demonstrated for BiVO4 doped/undoped spray pyrolysed film and single crystal samples. The spectra suggest weakly-bound carriers in agreement with recent reports [3].
[1] A. J. Cowan, J. R. Durrant, Chem. Soc. Rev. 42, 2281 (2013)
[2] A. G. Joly, J. R. Williams, S. A. Chambers, G. Xiong, W. P. Hess, D. M. Laman, J. Appl. Phys. 99, 053521 (2006)
[3] F. F. Abdi, T. J. Savenije, M. May, B. Dam, R. van de Krol, J. Phys. Chem. Lett. 4, 2752 (2013)

The drive to develop cost effective TW scale photovoltaics has generated renewed interest in the development of earth abundant transparent conducting oxides (TCOs). In short, indium-free TCOs with performance similar to indium-tin-oxide (ITO) are strongly desired. Further, material properties beyond just the conventional metrics of high transparency and low sheet resistance must be considered in developing TCO materials for next generation PV application. In particular, tunable band gaps and tunable band edge energies are necessary to optimize electrical contact to both current and emerging earth abundant absorbers. Often, the use of an additional thin interfacial (or charge transport) layer based on TCO-like materials improves PV performance as well. Accordingly, band edge energies, work function, dopability and morphology should also be considered in selecting TCOs for any specific PV absorber. We will demonstrate this through examples taken largely from our prior and ongoing research, a few of which are summarized next. Amorphous Zinc-Tin-Oxide (a-ZTO) is a work function tunable, but low conductivity material, that has been demonstrated utility as a charge transport layer. Crystalline Ga-doped (Zn,Mg)O is a band gap tunable TCO where the conduction band minimum (CBM) energy can be tuned over a 0.3 eV range by varying the Mg content from 0 to 30%, albeit with reduced conductivity as the Mg content increases. Sputtered SnO₂ TCOs are being developed with the goal of making inexpensive and chemically robust SnO₂ a viable TCO for device applications that can not tolerate the high temperatures used in making F-doped SnO₂ by pyrolytic decomposition processes. Finally, Nb-doped anatase TiO₂ provides a potential high index of refraction TCO. Finally, our approach to TCO materials development generally begins with a broad exploration of the relevant materials and synthesis space. This is most often done using combinatorial composition gradient samples (libraries) grown by co-sputtering. The benefits and challenges of such high-throughput approaches will also be discussed.

Transparent conducting oxides (TCOs) are essential to many technologies from solar cell to transparent electronics. While n-type TCOs (using electrons as carriers) are widespread in current applications (e.g., indium tin oxides or ITO), their p-type counterparts have been much more challenging to develop and still exhibit carrier mobilities an order of magnitude lower.
The difficulties in developing high mobility p-type TCOs can be related to the intrinsically high effective masses of holes in oxides. In this talk, we will report on a high-throughput computational search for oxides with low hole effective mass, wide band gap and p-type dopability. Screening thousands of binary and ternary oxides in the Materials Project Database using state of the art ab initio techniques, we will present several unsuspected compounds with promising electronic structures. Beyond the description of those novel TCOs candidates, we will discuss and chemically rationalize our findings, highlighting several design strategies towards the development of future high mobility p-type TCOs.

ZnM₂O₄ (M=Co, Rh, Ir) spinels are considered as a class of potential p-type transparent conducting oxides (TCO). Experimentally one has shown that polycrystalline samples of ZnM₂O₄ spinels exhibit p-type conductivity. We report the formation energy of acceptor-like defects using first principles calculations with an advanced hybrid exchange-correlation functional (HSE06) within density functional theory (DFT). Due to the discrepancies between the theoretically obtained band gaps with this hybrid functional and the -scattered- experimental results, we also perform GW calculations to support the validity of the description of these spinels with the HSE06 functional. The considered defects are the cation vacancy and antisite defects, which are supposed to be the leading source of disorder in the spinel structures. We also discuss the band alignments in these spinels. The calculated formation energies indicate that the antisite defects Zn_M (Zn replacing M, M=Co, Rh, Ir) and V_Zn act as shallow acceptors in ZnCo₂O₄, ZnRh₂O₄ and ZnIr₂O₄ , which explains the experimentally observed p-type conductivity in those systems. Moreover, our systematic study indicates that the Zn_Ir antisite defect has the lowest formation energy in the group and it corroborates the highest p-type conductivity reported for ZnIr₂O₄ among the group of ZnM₂O₄ spinels. To gain further insight into factors affecting the p-type conductivity, we have also investigated the formation of localized small polarons by calculating the self-trapping energy of the holes.

Recent experimental results have shown than doped n-type BaSnO3 may be a useful high performance transparent conductor based on abundant elements. A key difference from ITO is the perovskite structure, which allows considerable flexibility in strain tuning the properties, both through substitution or alloying on the Ba site and through epitaxy. We report first principles calculations examining the effect of different strain conditions and substitution on the band gap, optical and transport properties. As expected, strain couples to the perovskite tilt systems. It also strongly affects the band gap, though not in the same way as in transition metal perovskites like SrTiO3. Specifically, the s-electron nature of the conduction band leads to a relatively greater volumetric sensitivity of the gap and optical properties relative to the tilt dependence. The band width and inferred conductivity also show substantial dependence on strain. We find that these perovskite stannates are remarkably flexible in their electronic properties in spite of the simple s-electron nature of the conduction band. This work was supported by the Department of Energy, Basic Energy Sciences, Materials Sciences and Engineering Division.

Sulfide compounds and mixed anion oxide-sulfide materials have potential as solar absorbers or transparent contacts. Improved techniques for depositing these materials in thin-film form are necessary to obtain greater compositional and phase control. Controlling the metal-sulfur ratio in sulfides and the oxygen-sulfur ratio in oxide-sulfides is a dominant challenge to thin film growth of these material systems. Here, we report a deposition method with improved control over the sulfur content in thin-films through the addition of a radio frequency (RF) solids atom source (cracker) to a multiple-source sputtering system. This technique has enabled combinatorial growth of both sulfide and oxide-sulfide materials in thin film form.
The growth conditions for sulfur containing compounds can quickly be refined using this system for combinatorial synthesis. Co-sputtering from one or two targets provides a compositional gradient across a stationary substrate. In addition, a temperature gradient that is orthogonal to the composition gradient is induced across the substrate. An RF solids cracker is used to provide controllable amounts of activated sulfur across the entire substrate during the deposition. Typically, RF solids crackers are used in molecular beam epitaxy systems where the usual operating pressure is 10^-5 to 10^-6 Torr. Here, we employ a RF solids cracker as an addition to our sputtering system where the typical operating pressure is 3 mTorr. Together, the composition gradient, orthogonal temperature gradient and activated sulfur source can be used concurrently to control the composition and phase of the deposited thin films. For this work, all films were deposited on 2”x2” glass substrates at a chamber pressure of 3 mTorr with only argon flowing as a process gas. The temperature gradient was 485 °C to 375 °C across 2”. The film composition was measured using both Rutherford backscattering spectrometry and x-ray fluorescence. X-ray diffraction was used for structural and phase determination.
The growth of sulfides is demonstrated using Cu₂S and the growth of oxide-sulfides is shown with the Bi_xO_yS_z system. In particular, Cu₂S has been grown from both metallic Cu and ceramic Cu₂O targets. Bi_xO_yS_z films with tunable oxygen to sulfur ratios were grown from a Bi₂O₃ target. Further, the independent tuning of anion and cation ratios with this system is demonstrated by the growth of BiCuOS. The successful growth of both sulfide and oxide-sulfide compounds demonstrates the viability of this hybrid approach. These results also suggest that similar approaches with phosphides, oxide-phosphides and phosphide-sulfides would be achievable with hybrid deposition systems.
This work is supported by the U.S. Department of Energy, Office of Science, BES, under Contract No. DE-AC36-08GO28308 to NREL as part of the DOE Energy Frontier Research Center "Center for Inverse Design"

Solution-based nanoparticle inks offer an inexpensive route for the production of thin film solar cells. These methods avoid the requirement of costly and inefficient vacuum methods and allow the possibility of roll-to-roll processing, an attractive option for high-volume manufacturing. Copper zinc tin sulfide (CZTS) absorber layers have attracted significant interest in thin film solar cell technologies due to the significantly improved efficiencies recently reported and the earth-abundant low cost materials that are being used. Our efforts focus on microwave-based continuous flow synthesis of quaternary CZTS colloidal nanoparticles, and advanced annealing procedures for CZTS nanoparticle thin films.
We have found that integrating microwave-based methods with continuous flow reactors offers significant advantages over standard continuous flow or batch methods for preparing CZTS nanocrystals. This is due to the rapid, uniform nucleation of the nanoparticle precursors via microwave heating, and due to the large surface area to volume ratios and the enhanced mixing via continuous flow methods. We have found that microwave-based continuous flow nanoparticle synthesis methods allow significant control in composition, crystalline structure, and electronic properties of the nanoparticles, where we can modify the metal salt precursors, residence time, nucleation and growth temperatures, and sulfur sources.
To enable the fabrication of CZTS solar cells on low temperature (i.e., polymeric) substrates, we have begun annealing studies of the CZTS nanoparticle films using pulsed thermal processing (PTP). This method uses a 750 kW radiant plasma arc lamp to rapidly heat CZTS nanoparticle thin films to improve crystallinity and densify the films without providing the time necessary for unwanted diffusion of ions and phase separation. PTP can anneal large areas simultaneously providing a path to roll-to-roll scale up.
In order to find the optimal CZTS nanoparticle ink composition and PTP annealing conditions we have performed detailed characterization of the nanoparticle materials and associated films using: X-ray diffraction, X-ray photoelectron spectroscopy, scanning electron microscopy, transmission electron microscopy, inductively coupled plasma atomic emission spectroscopy, thermal gravimetry, secondary ion mass spectroscopy, Raman spectroscopy, and UV-vis spectroscopy. These results will be discussed in regards to scalable synthesis of CZTS inks and thermal processing.

Solar cells based on Cu₂ZnSn(Se,S)₄ (CZTS) absorber layers are appealing due to the earth-abundant constituent elements and tunable optical band gap near the optimal energy for single junction devices. However, the native point-defect chemistry of CZTS is complex (1), and there exist numerous competing crystalline phases (2), both of which factors present challenges for the growth of high-quality films with the properties required for high-efficiency solar cells. Additionally, the role of grain boundaries in CZTS solar cells is not yet well understood. Here we report on efforts to grow epitaxial CZTS, a model system that will enable fundamental studies relating to the native defect chemistry of CZTS, and to the role of grain boundaries in polycrystalline CZTS solar cells. The CZTS films in the study were grown by molecular beam epitaxy (MBE) on Si substrates, which has a nearly identical lattice constant at room temperature to that of CZTS. Combinatorially graded films, where the cation composition is intentionally varied across the substrate, were grown to determine the composition window in which epitaxial growth could be achieved, and to assess the effect of film composition on properties that relate to the presence of point defects, for example photoluminescence. Films crystallinity and morphology were characterized with scanning electron microscopy (SEM), scanning Auger microscopy (SAM), electron back-scatter diffraction (EBSD) and transmission electron microscopy (TEM). ZnS/CZTS/ZnS double heterostructure devices were fabricated to test the hypothesis that the ZnS/CZTS interface is passivated and therefore benign, an important issue since ZnS is a commonly observed secondary phase in device-quality CZTS films.
(1) S. Chen, J.-H. Yang, X.G. Gong, A. Walsh, and S.-H. Wei, “Intrinsic point defects and complexes in the quaternary kesterite semiconductor Cu₂ZnSnS4₄,” Physical Review B 81, 245204 (2010).
(2) I.D. Olekseyuk, I.V. Dudchak, and L.V. Piskach, “Phase Equilibria in the Cu₂S-ZnS-SnS₂ System,” Journal of Alloys and Compounds 368, 135-143 (2004).

Cu₂ZnSn(S,Se)₄ (CZTS) is much studied as a substitute for CuInGa(S,Se)₂ (CIGS) in thin film photovoltaics. CZTS is derived from CIGS by substitution of rare and expensive InGa by abundant and inexpensive ZnSn, yielding a material with comparable optical and electronic properties but with more scalable constituent elements. CZTS solar cell performance continues to improve and the record efficiency is now 12.0% [1]. However, the CZTS phase has a narrow composition window of thermodynamic stability, and CZTS solar cells are thought to suffer from the effects of phase segregation. Further improvements in device efficiency may hinge on using on kinetic stabilization to inhibit phase decomposition. This is an approach that applies generally to materials with complex stoichiometry.
We demonstrate the growth and detailed structural characterization of disordered (CuZnSn)S₄ alloys that are stabilized at low temperature (T). By growing films at room T we achieve an alloy with an expanded solid solution window in the pseudo-ternary CuS - ZnS - SnS phase diagram. Of particular interest for photovoltaics, this alloy allows independent tuning of the bandgap and carrier concentration. Here we will focus on the structural characterization of this new phase. We study the short- and long-range order using x-ray absorption spectroscopy (XAS) and x-ray diffraction, respectively. We use extended x-ray absorption fine structure (EXAFS) to quantify short range order, and x-ray absorption near edge structure (XANES) to quantify phase segregation. We complement our x-ray studies with atomic force and transmission electron microscopy. Our results point to the formation of a new, metastable amorphous phase with an expanded range of solid solubility and a tunable bandgap. We study the thermal stability of this new phase by studying a series of samples with growth temperatures between room T and 450 °C, and comparing to samples that were grown at room T and then annealed. For both series we see a continuous evolution towards the crystalline CZTS phase that is nearly complete at 450 °C.
Our results are a detailed materials characterization of a new semiconductor phase that holds promise as a photovoltaic absorber. Our results also inform the fabrication of conventional CZTS solar cells by establishing the temperature range over which thin films transform from a kinetically stabilized, metastable phase to a thermodynamically stabilized, crystalline phase.
[1] Winkler et al., Energ. Environ. Sci. (2013). DOI: 10.1039/C3EE42541J

Earth-abundant absorbers become even more cost-effective when used in an extremely thin absorber solar cell. Atomic layer deposition (ALD) and chemical vapor deposition (CVD) deposit highly conformal films and hence are important tools for developing nanostructured solar cells with scalability.
We present local structure studies of ZnS/Cu₂S, Cu₂SnS₃ and Cu₂ZnSnS₄ composite films prepared with ALD and CVD, using extended x-ray absorption fine structure (EXAFS) technique. The EXAFS technique has the ability to probe the local environment about different atoms, and can also give very precise ratios of elements using their fluorescence peaks. Previous work has shown that individual thin films of Zinc Sulfide (ZnS) and Copper (I) Sulfide (Cu₂S) resemble bulk structure. Yet multi-layer films of ZnS/Cu₂S, prepared using a wide range of parameters, produce films that are predominantly either ZnS or Cu₂S, with the other material being highly disordered. This can be attributed to the crystal structure mismatch of ZnS and Cu_xS, making ALD with these precursors unsuitable for a CuZnS alloy.
Another copper-based material, Cu₂SnS₃ (CTS), has a stable structure with good electrical and optical absorption properties. Composite films of CTS were made using CVD layers of Cu₂S and Tin (IV) Sulfide (SnS₂), with an anneal step. The metal precursors were Tin (IV) Acetate for SnS₂ and KI5 (a direct descendant of CupraSelect from Air Products) for Cu₂S. Of special note, the metal precursors are all non-pyrophoric, which is safer. The sulfur source was H₂S, generated in situ via a reaction between aluminum sulfide powder (Al₂S₃) and water. Cu₂SnS₃ also has the same structure as ZnS, which allows for the formation of the quaternary Cu₂ZnSnS₄ (CZTS). CZTS composite films were also made, using Zn(TMHD)₂ for the ZnS, which deposited on top of the Cu₂S and SnS₂ layers determined for CTS. Stoichiometric control was established by varying the deposition times of the binary compounds, and was measured using energy-dispersive x-ray spectroscopy (EDX), x-ray diffraction (XRD), and EXAFS techniques. Optical absorption results are promising for forming a photovoltaic device with copper-based ternary and quaternary materials as the absorber.

Annealing of reactively-sputtered precursors [1] is a promising route to production of Cu₂ZnSnS₄ (CZTS) kesterite thin-film solar cell absorbers on an industrial scale. The thermally-activated processes involved must be thoroughly understood if the composition and heterostructure of the corresponding devices are to be optimized. Element loss, such as the evaporation of S and SnS during thermal processing [2], can prevent the formation of kesterite-phase-pure CZTS material. It has also been shown that the standard back contact interface (CZTS/Mo) is unstable at the temperatures used to process the precursor into kesterite-phase absorbers [3].
To study the chemical processes at the interface between the CZTS precursor and the Mo back contact during annealing, we have used soft X-ray emission spectroscopy (XES) to first investigate the reverse interface - a thin Mo layer on the precursor. XES probes the local partial density of valence band states and was employed to monitor the buried interface at several temperatures. We find that MoS₂ is formed - even at temperatures below those used during absorber formation - and that this reaction does not require additional S in order to occur, confirming previously suggested chemical reaction pathways [3]. The possibility that the MoS₂ formation is influenced or enhanced by the disorder of the precursor phase was investigated by comparing it to similar experiments in which a Mo capping layer is deposited onto nominally phase-pure CZTS ‘standard&’ absorbers.
We also find, using XES, that after annealing in ultra high vacuum, a ZnS-like layer is formed at the surface of both precursor and ‘standard&’ CZTS. The loss of tin (via SnS) from the CZTS surface at elevated temperature has been previously observed [2]; possible explanations for the absence of Cu will be discussed. Hard X-ray photoemission spectroscopy (HAXPES) was used to characterize the composition of both the front- and back-sides of the absorber precursors annealed at four different temperatures. This analysis shows that, even after standard processing steps, the back side of the absorber is not single-phase CZTS, but consists of a mixture of phases and/or a disordered CZTS, whereas the front side data resembles (nearly) single-phase CZTS after the thermal treatment. HAXPES measurements of a CZTS precursor layer below a sputtered Mo capping layer confirm that MoS₂ does not form during sputtering, but only during subsequent annealing, in agreement with the XES results and with the previously-suggested reaction pathway [3]. The instability of the CZTS/Mo interface under standard processing conditions suggests that an optimization of the absorber/back contact structure may lead to significant gains in device efficiency.
[1] J.J. Scragg et al. Prog. Photovolt.: Res. Appl. (in press 2012).
[2] A. Weber et al. J. Appl. Phys. 107 013516 (2010).
[3] J.J. Scragg et al. J. Am. Chem. Soc., 134 19330 (2012).

The earth-abundant, less environmental impact Cu2ZnSnS4 (CZTS) thin film solar cells are fabricated by using a sulfurization technique. The CZTS film possesses promising characteristic optical properties; band-gap energy of about 1.5 eV and large absorption coefficient in the order of 104 cm-1. All constituents of this CZTS film, which are abundant in the crust of the earth, are non-toxic. Therefore, by using CZTS film practically as the absorber of thin film solar cells, we will be free from both of the resource saving problem and the environmental pollution. We believe that the key words of the after next generation have to be abundant and non-toxic. CZTS is one of the most promising materials for thin film solar cells.
In this study, CZTS absorber layers were synthesized by the sulfurization technique of sputtered precursors. In our previous report, we have clarified that the off-stoichiometry composition of Cu-poor and Zn-rich is desirable to achieve high conversion efficiency. By using CZTS compound target that provide such active composition, we could use a simple single sputtering method to prepare CZTS absorber.
In our laboratory, a two stages process of precursor preparation followed by sulfurization is a major fabrication method from the start of this study. We think that this method is suitable for a mass production. Therefore, the sulfurization process as well as the composition control is a quite important issue. In this paper, TG/DTA system available in the H2S atmosphere is introduced to optimize the sulfurization condition. When we examined the sulfurization condition, we used both the results of XRF composition measurements and TG/DTA thermal analysis. It was confirmed that the raising rate of the substrate temperature within the certain temperature region affected significantly to the film properties.

Cu2ZnSn(S,Se)4, or ‘CZTS&’, is being widely developed as an earth-abundant solar absorber. This multicomponent system presents several challenges in developing the desired properties in the bulk and at interfaces. X-ray absorption spectroscopy can be a powerful tool to understand both electronic and physical structure of a material. We report on the near edge X-ray absorption fine structure (NEXAFS) of the Cu, Zn, and Se 2p transitions for CZTS in a completed solar cell. The data collection employed scanning transmission X-ray microscopy (STXM) that can be used to map the NEXAFS throughout the device layers. Data was obtained from the bulk CZTS and near the front and back contact interfaces. Depletion layers near the front contact layers are observed in the STXM image analysis, which enable a comparison of the bulk and interface NEXAFS. The NEXAFS data is interpreted with the benefit of calculations of the X-ray transitions using density functional theory. The Zn and Se 2p NEXAFS exhibit the expected structure of a filled 3d shell. The Cu 2p however shows additional fine structure not predicted from the calculations. These results are discussed with respect to possible charge transfer effects and point defects within the system.

The occurrence of secondary phases in Cu₂ZnSn(S,Se)₄ (CZTSSe) is highly expected due to the off-stoichiometry conditions required for the preparation of high efficiency devices. A strong attention has been given to minimization of the occurrence and selective removal of ternary Cu-Zn-(S,Se) and binary Zn(S,Se) and Cu-(S,Se). However, so far very little attention has been given to the role of other potential phases as Sn(S,Se). The selective removal of this phase is important because it can occur either by stoichiometric deviation or also from condensation from the annealing atmosphere, which usually contains Sn and the chalcogen. As processes are further optimized to achieve higher efficiency devices, it is important to increase the level of control on the presence of these potential detrimental low band gap phases.
In this work we present a simple chemical route for the selective removal of Sn(S,Se) by using (NH4)₂S solutions as an etchant in the concentration range of 0.5 M to 3.0 M. The analysis of the etch rate performed on the different secondary phases as well as on CZTSSe, both in powder forms, allow us to conclude that a 3.0 M solution of (NH₄)₂S is highly selective for Sn(S,Se) removal. Using SEM, EDX and XRD we show that there are two types of Sn(S,Se) aggregates on the CZTSSe surface: those related to stoichiometric deviations typically encrusted on the surface, and those related to the condensation from the annealing atmosphere during the cooling down process, typically observed as surface over-growths. Whereas a 0.5 M (NH₄)₂S solution does not remove the Sn(S,Se) aggregates, a 1.0 M solution seems to slightly affect the aggregates and the 3.0 M solution is very effective at removing both types of aggregates. Using XRF we observe that both, the Cu/(Zn+Sn) and Zn/Sn ratios slightly increase, in agreement with XRD.
The impact of the presence of these phases on the characteristics of the solar cells has been assessed by the analysis of devices produced both with unetched and (NH₄)₂S etched absorbers. These data confirm that the etching with 0.5 M and 1.0 M solutions are only somewhat useful to remove Sn(S,Se) secondary phases, obtaining devices with almost the same optoelectronic parameters as those measured on the unetched samples. Conversely, devices obtained with layers etched in a 3.0 M solution exhibit a remarkable increase of the efficiency from typically 3.0% for the unetched sample, up to more than 5.0% for the etched ones. This is mainly related to an increase on the V_oc and F.F. of the solar cells, while the J_sc is almost unaffected. Using test samples with Sn(S,Se) phases intentionally grown onto the surface, we corroborate that this feature is not only related to the removal of the low band-gap Sn(S,Se) phases which explain the increase of the V_oc, but also to the passivation of the surface. Finally, the impact of this chemical route in the further development of CZTSSe is presented.

Thin-film Cu₂ZnSn(S,Se)₄ (CZTSSe) is a promising absorber material for low-cost, scalable photovoltaic applications. There has been substantial progress recently in the performance of CZTSSe devices, in spite of a complex intrinsic point defect chemistry and poorly understood kinetic processes during film growth and processing. We have employed in situ electrical characterization methods to investigate the defect chemistry and diffusion in CZTSSe.
Recent first-principles calculations predict that chalcogen vacancies form deep levels in CTZSSe materials that could negatively impact critical opto-electronic properties, including minority-carrier lifetime and electron and hole mobilities.[1] As a consequence, CZTSSe materials synthesized under low or high vacuum conditions, or via liquid-phase precursors, are often subjected to a post-deposition annealing process under relatively high chalcogen partial pressures. Chalcogen diffusion in CZTSe during chalcogen-ambient annealing was investigated via in situ monitoring of the molybdenum back-contact resistance during a selenization treatment. Diffusion of Se through the CZTSe layer results in conversion of the Mo layer to MoSe₂, increasing the sheet resistance of the film stack. By monitoring the rate of MoSe₂ formation as a function of annealing temperature, an activation energy of 0.5 eV has been measured for Se diffusion in CZTSe. The partial pressure dependence of chalcogen diffusion suggests that chalcogen vacancies are not the defect controlling chalcogen diffusion in CZTSe.
Cation diffusion in CZTS was investigated via in situ electrical measurements during film growth by elemental co-evaporation. In these experiments, CZTS sheet-resistance was monitored during film growth, and transients were observed when film growth was temporarily arrested. We interpret these transients as arising from re-equilibration of cation species associated with chemical potential variations during the growth process. By monitoring resistivity equilibration times at various processing temperatures, an activation energy for cation diffusion of 2.0 eV has been extracted. For both techniques, fundamental connections between the results and the overall defect chemistry within CZTS and CZTSe will be discussed.
1.) Chen, S., et al. Physical Review B 81(24) (2010).

Higher efficiency Cu₂ZnSnS₄ (CZTS) based solar cells have been based on absorbers synthesized under Cu-poor and Zn-rich compositions in order to avoid the formation of Cu-S and Cu-Sn-S secondary phases, and favor V_Cu vacancy defects. Raman spectroscopy has demonstrated its potential as a characterization tool for CZTS, allowing the detection of the main secondary phases expected in these conditions. However, these non-stoichiometric conditions also have a strong impact on the crystalline quality of the CZTS films, and can lead to structural disorder. This in turn has a potentially significant impact on the spectral features of the main peaks in the Raman spectra. However, evaluation of these effects in CZTS requires a deeper knowledge of the vibrational properties of this material. In order to deepen in the understanding of the CZTS vibrational properties, this work presents a detailed analysis of all Raman active modes of CZTS prepared by sulfurization of metallic precursors deposited by DC-magnetron sputtering onto soda-lime glass substrates. This process has led to device efficiencies around 5.5%. The surface Raman spectra have been measured using six different excitation wavelengths from NIR to UV, which include non-resonant and resonant conditions with Γ₁ and Γ₂ points of the electronic band diagram. Combining the information of all spectra, 18 peaks have been resolved, 5 of which were not observed previously, and have been assigned with the 27 optical modes theoretically expected for this crystalline structure allowing the complete characterization of the vibrational properties of the CZTS kesterite structure.
To analyze the impact of the defects in the Raman spectra CZTS samples of different crystalline quality have been prepared using the same methodology as for the reference samples, but with different annealing times (from 0 to 300 minutes). Variation of the Raman shift and shape of the two main peaks at 287 and 338 cm^-1 assigned both to A symmetry modes has been observed. These variations are assigned to confinement effects due to the crystalline lattice periodicity breaking determined by the presences of the defects in the crystalline structure, which has been modeled and quantified using a phonon confinement model. Simultaneous fitting of the two dominant peaks in the spectra is used to determine both the correlation length and stress induced shift components, that have been correlated with the grain size and crystalline quality assessed by complementary techniques (XRD and SEM). The estimated correlation length is shown to constitute a suitable quantitative indicator of the crystalline quality for these materials. Finally, a simple experimental methodology based on these non-destructive optical measurements is proposed for the structural assessment of these films. This methodology is compatible with micro-scale analysis configurations and can also be extended to Cu₂ZnSnSe₄ compounds.

The photovoltaic material Cu₂ZnSnS(e)₄ (CZTS) may be a game changing absorber in thin film photovoltaics because it is synthesized from elements abundant in the Earth&’s crust and because of its rapid development in devices over recent years. However, wide-spread usage of CZTS-based solar cells awaits further improvement of the material and a better understanding of the relationships among growth conditions, defects, surface, and interface properties. For example, it has been found empirically that the manner in which one finishes the growth by physical vapor deposition of a thin film of CZTS is critical to the functioning of the resulting solar cell.¹ The effects of and range of possibilities for beneficial post-growth surface modifications are in their initial stages of understanding. Aqueous processes are of particular interest. Simply exposing CZTS to pure water has been found to enhance the conversion efficiency of CZTS solar cells.² Also, aqueous chemical methods are typically used to deposit n-type buffer layers such as CdS on CZTS. In these examples and in many other cases, little is known about the microscopic details of the CZTS surface both before and after modification. In this contribution we report results from electron spectroscopic investigations on ex-situ grown CZTSe polycrystalline thin films. From X-ray photoelectron core level spectra (XPS) we find that similar to the case of Cu(In,Ga)Se₂, surfaces of high-performance CZTS films are copper poor relative to the bulk composition. Unlike Cu(In,Ga)Se₂, argon ion sputtering of CZTS at ~ 1 kV does not cause metallization of the surface as observed by valence band photoemission spectra. We find a strong correlation between CZTS near-surface composition and the Fermi-energy-referenced valence band maxima, indicating that substantial band bending exists at the surface of the as-grown absorber. The observed band bending increases with exposure of these surfaces to Cd_(aq). These results indicate that strong n-type doping exists at the CdS/CZTS interface, consistent with Zn_Cu antisite defects and Cd_Cu sites in CZTS acting as shallow donors. These experimental observations are used in conjunction with device-modeling results to construct a detailed composition dependent band diagram for the CdS/CZTSe interface, and implications for device performance are discussed. Finally we discuss our plans for investigating epitaxial films of the sulfide material Cu₂ZnSnS₄ grown in-situ in our laboratory&’s cluster tool.
References
1. Repins, I.; Beall, C.; Vora, N.; DeHart, C.; Kuciauskas, D.; Dippo, P.; To, B.; Mann, J.; Hsu, W. C.; Goodrich, A.; Noufi, R., Co-evaporated Cu₂ZnSnSe₄ films and devices. Sol Energ Mat Sol C101 (2012) 154-159.
2. Katagiri, H.; Jimbo, K.; Yamada, S.; Kamimura, T.; Maw, W. S.; Fukano, T.; Ito, T.; Motohiro, T., Enhanced conversion efficiencies of Cu₂ZnSnS₄-based thin film solar cells by using preferential etching technique. Appl Phys Express1 (2008) 041201.