The nanostructure of hydrogenated amorphous silicon, examined by means of thermal annealing and light soaking

More Info
expand_more

Abstract

Photovoltaic energy is one of the key components of a sustainable energy future. While the market is currently dominated by crystalline silicon solar modules, thin-film silicon technology holds the promise of a cheap, resource-efficient and versatile alternative. The major drawback of thin-film silicon PV is its relatively low conversion efficiency, in part caused by the metastable defect phenomenon known as the Staebler-Wronski effect. To fully optimize the potential of thin-film silicon solar cells, a thorough understanding of hydrogenated amorphous silicon (a-Si:H) is required. To this end, an experiment is designed in which the effects of thermal annealing and light soaking on various material properties of a-Si:H are determined. Sets of p-i-n solar cells are deposited on Asahi VU substrates and sets of intrinsic a-Si:H films are deposited on Corning Eagle XG glass and on n-type c-Si wafers. The intrinsic layer of the solar cells and the films are deposited using hydrogen-to-silane flow rate ratios of 0, 2.5, 5, 7.5 and 10. The solar cells and films are annealed in vacuum for one hour at a time at temperatures increasing from 25?C to 500?C. In between annealing steps, measurementsare performed using the following techniques and methods: Fourier transform photocurrent spectroscopy (FTPS), Fourier transform infrared (FTIR) spectroscopy, Raman spectroscopy, reflectance/transmittance (RT), external quantum efficiency and solar cell external parameter determination. A separate set of solar cells and films is light soaked in a degradation chamber at 25?C with a light intensity of 1 kW/m^2 and an AM1.5 spectrum. At 0, 1, 10, 100 and 1000 hours of cumulative light soaking time, the same measurements as during the annealing experiment are performed (excluding FTIR and Raman spectroscopy). Additionally, the solar cells are light soaked in-situ in the FTPS setup using blue and white light with an intensity of 1 kW/m^2. Light soaking times increase exponentially and range from 0.5 s to approximately 200 h. FTPS measurements are performed after each step. Analysis of the FTIR results suggests that divacancies agglomerate into larger vacancies and nanosized voids during annealing, which is supported by the evolution of the bandgap obtained from RT measurements. Silicon-hydrogen bonds start breaking at around 300?C, at which temperature FTPS data show that the subgap absorption starts to increase, suggesting that defects are created in the form of un- or underpassivated vacancies and nanosized voids. Analysis of the FTPS data also reveals that the subgap absorption coefficient spectrum consists of at least 4 distributions, implying that the isolated dangling bond (which results in two defect distributions) cannot be the sole type of defect in a-Si:H. Analysis of the light soaking FTPS data reveals metastable defect creation with two distinct regimes: a fast regime (defect creation ? t^? with ? = 0.4 - 0.9) at timescales ranging from several hours to several days, and a slow regime (? = 0.1 - 0.2) at timescales ranging from days to weeks. This finding contradicts the commonly assumed single value of ? = 1/3, associated with the isolated dangling bond as sole type of metastable defect. The combined results strongly support a view of the nanostructure of a-Si:H, in which unpassivated divacancies and nanosized voids contribute significantly to metastable defect creation. Performance changes due to light soaking using different wavelengths of a-Si:H solar cells with absorber layers deposited using different hydrogen-to-silane gas flow rate ratios have been observed using FTPS. These observations are correlated to the external parameters of the devices. It is suggested that the fitted subgap distributions are linked spatially or energetically to particular defect entities. This highlights the potential of FTPS for monitoring metastability effects in solar cells.