Skip to main content
European Commission logo print header

Development of SnS Based Solar Cells

Final Report Summary - SNS PV CELLS (Development of SnS Based Solar Cells)

Tin sulphide SnS is a IV-VI compound that has recently become of interest for application as an absorber material in low cost thin film solar cell devices. Similar to cadmium telluride CdTe and copper indium gallium diselenide CIGS, which are usually used to make commercial thin film solar cell devices and modules, SnS has a near optimum energy bandgap for photovoltaic solar energy conversion and a high optical absorption coefficient for photons with energies greater than the energy bandgap such that only a few microns of material are needed to absorb most of the incident light, minimising material costs. However unlike CdTe it consists of environmentally acceptable elements and unlike CIGS it consists of highly abundant elements. The overall objective of this research programme was to determine whether or not the properties of SnS and pn junctions formed with n-conductivity type wide energy bandgap buffer layers can be improved to make commercially viable solar cells.

Most of the SnS layers produced in this work were by the thermal evaporation of compound onto heated substrates in an oil pumped vacuum system operating in the 10-5 to 10-6 Torr range of vacuum pressures. The main growth variables were assumed to be the substrate temperature, the deposition rate and the film thickness. It was found that for a wide range of deposition parameters it was possible to produce SnS layers that were uniformly thick, free from pinholes, and highly adherent to the substrate surface. This was the case even for SnS film thicknesses > 20µm grown with high deposition rates e.g. > 2µm/min. Most of the layers studied had thicknesses in the range 3-5µm, i.e. thicknesses appropriate for using the SnS as an absorber layer material in thin film solar cells, and for deposition times of 3-5 minutes. A major influence on the composition of the layers was the substrate temperature, the layers tending to be stoichiometric and single phase SnS for layers grown in the substrate temperature range, 320 - 360°C. (This was confirmed by X-ray diffraction data and Raman studies). Optical spectroscopy data also indicated that these layers had a direct energy bandgap of 1.35 eV as expected for SnS. The layers grown for substrate temperatures < 320 oC and >360 oC were found to be non-stoichiometric.

To increase the grain size and crystallinity of the SnS it was decided to investigate the influence of post-deposition annealing of the layers. The annealing environments used initially included annealing in vacuum or in air. There was observed to be some sulphur loss for annealing in vacuum and some oxygen incorporation when annealing in air. In order to minimise the S-loss and to avoid O-incorporation, annealing was also carried out in an environment containing elemental S and an environment containing 5 % H2S in argon. The sulphur annealing was carried out at UNN whereas the H2S/argon annealing was carried out at Seoul Technological University (STU) in South Korea as they have a H2S annealing facility. In both cases by adjusting the annealing conditions (temperature and time used), it was possible to substantially increase the grain size, densify the layers and improve the crystallinity of the layers while maintaining stoichiometry of the SnS.

In an attempt to introduce excess Sn into Sn-deficient samples, the samples were dipped into a solution of SnCl4 in methanol and then annealed. In addition to the grain size being increased and the layers densified, it was found that this heat treatment type-converted the layers from p-conductivity type to n-conductivity type layers, the conductivity (and carrier concentration) in the layers being controlled mainly by the concentration of the solution used. This meant that a low concentration solution could be used to compensation dope the layers and an even lower concentration lightly dope the grain boundaries. Counter doping the grain boundaries (or possibly lightly oxidising the grain boundaries) should in principle passivate the grain boundaries in a similar way to that achieved with CdTe.

An alternative way of producing the SnS thin films was by depositing thin layers of Sn onto glass or Mo-coated glass substrates and then annealing the layers in an environment containing elemental sulphur or one containing H2S. As with the post-deposition annealing studies carried out on the thermally evaporated SnS, the S-annealing was carried out at UNN and the H2S/argon annealing at STU. It was found to be possible to produce thin films of SnS for Sn-thicknesses < 200 nm, an annealing temperature in the range 350 - 400°C and for an annealing time in the range 1-2 hr. However in particular for Sn thicknesses > 200 nm (on glass), the SnS layers formed were so highly stressed that they tended to peel off from the substrate used. Thicker layers could be used for the Sn layers deposited onto Mo, although the stress problems again became significant for SnS thicknesses > 0.5 µm. Because of the ease of success producing the layers using the thermal evaporation method the annealing route was not used to produce SnS devices.

Some layers of SnS were produced in India at SVU by spray pyrolysis and supplied to us at UNN for comparative studies and making solar cell devices. These layers were made using a solution containing tin chloride and selenourea in distilled water and sprayed onto Mo-coated glass using a substrate temperature of 350°C and at solution and gas flow rates of 6 ml/min and 8 l/min respectively. A major problem with these layers was that the maximum thickness that could be grown using this method was 0.6 µm and trying to deposit thicker layers by repeated depositions resulted in a very poor SnS/SnS interfaces. To make the pn junction device, a wide energy bandgap n-type buffer layer was deposited using either thermal evaporation (CdS or ZnS), sputtering (ZnO:Al) or using chemical bath deposition (CdS or ZnIn2Se4). Very good layers of CdS with steep optical absorption edges and low resistivities < 5.5 x 10-3 ?m were obtained for layers deposited using a substrate temperature of 180°C, a deposition rate of 0.1 µm/min. and for a film thickness of 1 µm. Likewise good layers of ZnS are produced using a substrate temperature of 300oC and with a deposition rate of 0.1 µm/min. However due to high resistivity of these layers the ZnS was co-evaporated with Al to extrinsically dope the layers and reduce their resistivities. The ZnO:Al was deposited using rf sputtering from a target containing ZnO and 2.0 % Al2O3, using argon as the sputtering gas at room temperature that showed an electrical resistivity of 2.0 x 10-3 ?m. Chemical bath deposition (CBD) is a standard way of producing the CdS using a aqueous solution containing a cadmium acetate and thiourea with few drops of ammonia as a complexing agent. This method results in very thin layers that are free from pinholes. The material deposited is actually Cd(S,OH) due to some oxygen incorporation into the layers. This results in layers with a wider energy bandgap than pure CdS. In this work we have also used CBD to deposit for the first time thin films of ZnIn2Se4 (patent application number: GB-1221098.4). This novel method uses a solution of ZnCl2, InCl3 and sodium selenite in water to supply the constituents to synthesise the compound. Our interest in this material was stimulated by a study in the US that showed that thermally evaporated ZnIn2Se4 was the best Cd-free buffer layer as an alternate to the CdS used in CIGS module manufacture.

Solar cells were initially made using the constituent layers in the 'superstrate' configuration. With this configuration the starting point was a tin oxide coated glass substrate. The n-type buffer layer, the p-type SnS absorber layer and finally the back contact were then deposited onto these substrates. Photovoltaic activity was observed for illuminated devices with the best efficiencies obtained for CdS/SnS junctions. These had efficiencies >1% with Voc = 220 mV, Jsc = 17 mAcm-2 and FF = 0.3 for illumination with an AM1.5 spectrum. The spectral response of all the devices exhibited the heterojunction window effect type behaviour with the main response between the energy bandgaps of the constituent materials. The long wavelength response was used to determine the minority carrier diffusion length, L in the SnS. The value of L obtained was >0.2 µm in the best devices, indicating that the quality of the SnS produced was promising. Capacitance-voltage measurements confirmed that the doping of the p-type SnS was in the range 1015 - 1016 cm-3 and capacitance-frequency data showed an interface state density of 3.06 x 1011 FC-1cm-2.

It was found that devices with improved efficiencies could be produced if they were made in the 'substrate configuration', i.e. they were fabricated by depositing the p-type SnS onto Mo-coated glass substrates, followed next by the n-type buffer layer, next by a ZnO:Al window layer and finally a top grid contact. The best devices used ZnIn2Se4 as the buffer layer. These had efficiency of 2.87 % with Voc = 472 mV, Jsc = 16 mAcm-2 and FF = 0.38. This is at present a world record efficiency for SnS-based devices. In the literature, two other research groups (Harvard, USA and Toyota, Japan) have recently reported efficiencies just over 2 % for SnS solar cells, Harvard for SnS/Zn(O,S) and Toyota for ZnMgO/SnS junctions. The recent improvements in efficiency of SnS devices has partly been achieved by replacing the CdS buffer with a wider energy bandgap Cd-free material that also forms a better interface with SnS. To study the latter effect, samples of few junctions have been investigated using X-ray photoelectron and photo yield spectroscopy to determine the band offsets at the junction. An energy and environmental analysis on the SnS based devices shows that they are favourable compared to both CdTe and CIGS cells with an energy payback time < 12 months and an energy ratio >20 for Southern Europe. Replacing the CdS with the ZnIn2Se4 and with further optimisation should lead to devices with efficiencies >10 %. Exceeding this target should result in devices that can sustainably generate power with costs comparable to those using fossil fuels and make the widespread use of PV the natural choice for power generation throughout the world.