van Racmm René Swaaij

Determining the material structure of microcrystalline silicon from Raman spectra

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 previo...

Influence of transparent conductive oxides on passivation of a-Si:H/c-Si heterojunctions as studied by atomic layer deposited Al-doped ZnO

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 × 1019–7 × 1020 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.

Fast deposition of microcrystalline silicon with an expanding thermal plasma

Abstract Microcrystalline silicon has been deposited using an expanding thermal plasma. High deposition rates are achieved, which is attractive for solar cell production. A first survey of the influence of the deposition parameters on the optical, electrical and structural material properties is performed. SEM analyses show columnar growth and infrared absorption shows varying oxygen content. For some of the deposition conditions the dark and photoconductivities approximate 10−7 and 10−5 S/cm, respectively. Also the deposition plasma has been studied by means of mass spectrometry. It is concluded that the expanding thermal plasma is well suited for the deposition of microcrystalline silicon at growth rates up to 3.7 nm/s.

The relation between the band gap and the anisotropic nature of hydrogenated amorphous silicon

The bandgap of hydrogenated amorphous silicon (a-Si:H) is studied using a unique set of a-Si:H films deposited by means of three different processing techniques. Using this large collection of a-Si:H films with a wide variety of nanostructures, it is demonstrated that the bandgap has a clear scaling with the density of both hydrogenated divacancies (DVs) and nanosized voids (NVs). The presence of DVs in a dense a-Si:H network results in an anisotropy in the silicon bond-length distribution of the disordered silicon matrix. This anisotropy induces zones of volumetric compressed disordered silicon (larger fraction of shorter than longer bonds in reference to the crystalline lattice) with typical sizes of ~0.8 up to ~2 nm. The extent of the volumetric compression in these anisotropic disordered silicon zones determines the bandgap of the a-Si:H network. As a consequence, the bandgap is determined by the density of DVs and NVs in the a-Si:H network.The network and nature of hydrogenated amorphous silicon (a-Si:H) are conventionally interpreted in terms of a continuous random network (CRN) of Si-Si bonds, weak Si-Si, Si-H bond and dangling bonds. A CRN requires that the smallest anisotropic features like dangling bonds and bonded hydrogen are randomly distributed and reside as isolated configurations in the network. However, in recent years more and more theoretical and experimental evidence have been found that both the isolated dangling bond and the isolated hydrogen are not present in the a-Si:H network. To the contrary, all studies come to the conclusion that the real nature of the a-Si:H is to contain more local structural order than expected from a CRN. These insights offer new opportunities to revisit the origin of several properties of a-Si:H, which are up to now explained within the framework of the CRN model. In this contribution we will discuss that many diagnostics like nuclear magnetic resonance, positron annihilation, small angle x-ray spectroscopy, density analysis and infrared spectroscopy on a-Si:H consistently demonstrate that a-Si:H exhibits an anisotropic network. In dense disordered networks the hydrogen predominantly resides in hydrogenated divacancies, whereas for less dense networks the hydrogen predominantly resides in poly-vacancies up to nanosized voids. We will discuss that hydrogenated divacancies in a disordered network contribute to the amorphous nature of a-Si:H and its electronic structure like the band gap, gap tails and the defect gap states.

Hydrogenated amorphous silicon deposited under accurately controlled ion bombardment using pulse-shaped substrate biasing

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 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.

On the surface roughness development of hydrogenated amorphous silicon deposited at low growth rates

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.

The role of the silyl radical in plasma deposition of microcrystalline silicon

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.

Importance of defect density near the p-i interface for a-Si:H solar cell performance

A cascaded arc expanding thermal plasma is used to deposit intrinsic hydrogenated amorphous silicon at growth rates larger than 2 A/s. Implementation into a single junction p-i-n solar cell resulted in initial efficiencies of ∼7%, although all the optical and initial electrical properties of the individual layers are comparable with RF-PECVD deposited films. The somewhat lower efficiency is due to a smaller fill factor. Spectral response measurements, illuminated J , V - measurements, and simulations indicate that a higher local defect density in the region near the p-i interface might be responsible for the smaller fill factor in comparison with conventional low- rate RF-PECVD. The higher defect density is most likely caused by the initial growth in the first 10 to 50 nm. Therefore, controlled initial growth of the intrinsic layer is suggested for good solar cell performance.

High-rate microcrystalline silicon for solar cells

In order to produce thin silicon films for solar cells at high growth rates we deposited films with a cascaded arc expanding thermal plasma. We demonstrate the power of this technique by applying amorphous films deposited at rates up to 1.4 nm/s in solar cells. We used the same deposition technique to produce microcrystalline silicon films. Growth rates up to 3.7 nm/s are achieved. The material structure is analyzed using Raman spectroscopy and XRD. We see that the crystalline fraction increases with the H/sub 2/ flow, whereas the amorphous and the void fraction decrease.

Temperature dependence at various intrinsic a-Si:H growth rates of p-i-n deposited solar cells

With a cascaded arc expanding thermal plasma, intrinsic solar grade amorphous silicon can be deposited at growth rates varying from 2 to 100 /spl Aring//s. The temperature above which good material is obtained becomes higher for higher growth rates. Higher deposition temperatures affect the p-layer within p-i-n grown solar cells, which will result in other optimum deposition temperatures of the i-layer. In this paper, the authors address the dependence of the p-i-n solar cell performance on the deposition rate and deposition temperature.