An easy and reliable method to extract the crystalline fractions in microcrystalline films is proposed. The method is shown to overcome, in a natural way, the inconsistencies that arise from the regular peak fitting routines. We subtract a scaled Raman spectrum that was obtained from an amorphous silicon film from the Raman spectrum of the microcrystalline silicon film. This subtraction leaves us with the Raman spectrum of the crystalline part of the microcrystalline film and the crystalline fraction can be determined. We apply this method to a series of samples covering the transition regime from amorphous to microcrystalline silicon. The crystalline fractions show good agreement with x-ray diffraction (XRD) results, in contrast to crystalline fractions obtained by the fitting of Gaussian line profiles applied to the same Raman spectra. The spectral line shape of the crystalline contribution to the Raman spectrum shows a clear asymmetry, an observation in agreement with model calculations reported previously. The varying width of this asymmetrical peak is shown to correlate with the mean crystallite size as determined from XRD spectra.

In silicon heterojunction solar cells, the main opportunities for efficiency gain lie in improvements of the front-contact layers. Therefore, the effect of transparent conductive oxides (TCOs) on the a-Si:H passivation performance has been investigated for Al-doped zinc oxide (ZnO:Al) layers made by atomic layer deposition (ALD). It is shown that the ALD process, as opposed to sputtering, does not impair the chemical passivation. However, the field-effect passivation is reduced by the ZnO:Al. The resulting decrease in low injection-level lifetime can be tuned by changing the ZnO:Al doping level (carrier density = 7 × 10 19 -7 × 10 20 cm −3 ), which is explained by a change in the TCO workfunction. Additionally, it is shown that a ∼10-15 nm ALD ZnO:Al layer is sufficient to mitigate damage to the a-Si:H by subsequent sputtering, which is correlated to ALD film closure at this thickness.

We have applied both sinusoidal and pulse-shaped rf substrate biasing techniques to the expanding thermal plasma deposition of hydrogenated amorphous silicon. Spectroscopic ellipsometry and Fourier transform infrared spectroscopy data demonstrate that both methods of substrate biasing can result in improved film properties at deposition rates ranging from 1.4 to 16 nm s −1 , at relatively low substrate temperatures (200 • C), as demonstrated by an increase in refractive index, and a decrease in the microstructure factor (R * ). An extra plasma forms in front of the substrate upon the application of both bias techniques, but optical emission spectroscopy data show that the emission intensities are significantly greater with the sinusoidal rf bias. Although the application of sinusoidal rf bias results in an initial improvement of the materials, the H content and R * values both increase at higher rf-induced substrate bias voltages (−V bias >∼ 70 V). The same bias conditions that result in increased H content and R * correspond to the conditions where the additional plasma in front of the substrate undergoes both a sharp increase in emission, and a decrease in measured ion current, suggesting an α to γ transition.

We have applied pulse-shaped biasing to the expanding thermal plasma deposition of hydrogenated amorphous silicon at substrate temperatures ∼200 °C and growth rates around 1 nm/s. Substrate voltage measurements and measurements with a retarding field energy analyzer demonstrate the achieved control over the ion energy distribution for deposition on conductive substrates and for deposition of conductive materials on nonconductive substrates. Presence of negative ions/particles in the Ar–H2–SiH4 plasma is deduced from a voltage offset during biasing. Densification of the material at low Urbach energies is observed at a deposited energy <4.8 eV/Si atom and attributed to an increase in surface mobility of mobile species as well as well as surface atom displacement. The subsequent increase in Urbach energy >4.8 eV/Si atom is attributed to bulk atom displacement in subsurface layers. We make the unique experimental abservation of a decreasing Tauc band gap at increasing total hydrogen concentration—this allows to directly relate the band gap of amorphous silicon to the presence of nanovoids in the material.

No abstract

No abstract

Expanding thermal plasma chemical-vapor deposition has been used to deposit microcrystalline silicon films. We studied the behavior of the refractive index, crystalline fraction, and growth rate as a function of the silane (SiH4) flow close to the transition from amorphous to microcrystalline silicon. It was found that the refractive index, a measure for film density, increases when the average sticking probability of the depositing radicals decreases. Furthermore, we studied the influence of the position at which SiH4 is injected in the expanding plasma on the film density. It was found that the film density becomes higher when the SiH4 is injected closer to the substrate. Both findings strongly suggest that the film density benefits from a high contribution of the SiH3 radical to the growth of microcrystalline silicon.

The surface roughness evolution of hydrogenated amorphous silicon (a-Si:H) films has been studied using in situ spectroscopic ellipsometry for a temperature range of 150–400 °C. The effect of external rf substrate biasing on the coalescence phase is discussed and a removal/densification of a hydrogen-rich layer is suggested to explain the observed roughness development in this phase. After coalescence we observe two distinct phases in the roughness evolution and highlight trends which are incompatible with the idea of dominant surface diffusion. Alternative, nonlocal mechanisms such as the re-emission effect are discussed, which can partly explain the observed incompatibilities.