Manipulating the Hydrogenated Amorphous Silicon Growing Surface

Abstract

Our modern lifestyle is currently fueled by two billion years of accumulated energy reserves. For several years now there has been a strong rise in research interest and more recently also implementation of renewable energy sources in the European Union. Driving factors for these developments are the increasing awareness of global warming in our society, the limited nature of current fossil fuel sources, the need for energy security & independence and the opportunity to develop new technologies and business opportunities. Obtaining energy from solar radiation via solar cells based on cyrstalline silicon has seen continuous development since the demonstration of the first photovoltaic devices in the 1950’s. These types of solar cells are still the dominant technology today, but other solar cell technologies are about to make the transition from research labs into significant production volumes. One of these technologies is the hydrogenated amorphous silicon (a-Si:H) thin film solar cell, which promises to be a cheap alternative to the crystalline silicon solar cell, albeit at a reduced energy conversion efficiency. The deposition technique used to grow a-Si:H thin films in this work is the expanding thermal plasma chemical vapor deposition (ETP-CVD). Main benefit of ETP-CVD are the high growth rates of more than 1 nm/s that can be achieved, which is very beneficial for the growth of the thick intrinsic absorber layer in a solar cell. Substrate temperatures of >300ºC are required to obtain dense a-Si:H films suitable for solar cell application – temperatures that leads to degradation of previously deposited p-doped layers in a typical p-i-n solar cell. In order to reduce the temperature load during deposition, ion bombardment was employed with the goal to provide the growing film surface with kinetic energy from the bombarding ions, thereby allowing to deposit dense a-Si:H films at substrate temperatures around 200ºC. The results discussed in chapter 4 and 5 show that significant progress towards this goal has been achieved, as demonstrated by solar cells with > 7% initial efficiency at a growth rate of 1 nm/s and substrate temperatures between 200 – 180ºC for the intrinsic layer. To achieve this we have utilised two different types of substrate biasing, sinusoidal RF biasing and pulse-shaped biasing, with focus on the latter. Both biasing techniques result in ion bombardment of the film surface during deposition. Both biasing techniques enable the deposition of solar-grade intrinsic a-Si:H at substrate temperatures below the 350ºC required for unbiased deposition. In chapter 3 we demonstrate how RF substrate biasing leads to the growth of vacancy-dominated material up to 50ºC below the temperature required for unbiased deposition. This conclusion is based on FTIR analysis where a transition form vacancy- to void-dominated material is observed. We investigated the dependence of this transition on the reactor pressure. Ion bombardment at a DC bias voltage VDC of 14 V has hardly any effect on the transition, yet an increase in void fraction was observed for all temperatures and increasing pressures. Ion bombardment at VDC = 20 V for the high pressure series resulted in an increase of the transition from 9% total hydrogen concentration unbiased to 13% total hydrogen concentration biased. This is attributed to a reduced incorporation of ionic clusters and polymers into the film. The second type of substrate biasing utilized in this work is the pulseshaped biasing, PSB. As shown in chapter 4, with PSB we are able to obtain accurate control over the ion energy distribution in a range of 0 - 200 eV without the formation of a strong secondary plasma typically present for RF substrate biasing. Control over the ion energy was confirmed by retarding field energy analyzer measurements for conductive substrates, and non-conductive substrates covered with a conductive surface layer which is connected to the sample holder. For intrinsic a-Si:H deposited with PSB at a growth rate of 1 nm/s and substrate temperatures in the range of 180 - 200ºC we can distinguish roughly between two regions: region I < 4.8 eV of deposited energy per Si atom (eV/Si atom) and region II > 4.8 eV/Si atom. In region I we observe an increase in material density due to a decrease in nanovoid concentration as deduced from FTIR analysis. At the transition between region I and II around 4.8 eV/Si atom the densest material with low nanovoid concentration is obtained. The increase in material density and the reduction in surface roughness in region I are attributed to an increase in surface mobility of mobile species as well as surface atom displacement. Above 4.8 eV/Si atom we see an increase in Urbach energy which is related to bulk atom displacement in subsurface layers at higher ion energies. We report unique experimental evidence which indicates that the band gap is not correlated to the total hydrogen concentration, cH, as usually reported in literature, but instead to the presence of nanovoids in the film, asdetermined from the cHSM mode. Intrinsic a-Si:H deposited with PSB under the same conditions as the films from chapter 4 have also been implemented in p-i-n solar cells. These cells are discussed in chapter 5 and a reproducible record initial energy conversion efficiencies of 7.4% was obtained for cells grown by ETP-CVD at such high growth rates of 1 nm/s and low substrate temperatures of 200ºC. This efficiency was obtained for cells grown around deposited energies of around 1 eV/Si atom. The open-circuit voltage has a maximum of 0.82 V around 1 eV/Si atom and decreases at higher deposited energies per Si atom, which is attributed to the low band gap at higher deposited energy. The short-circuit current density reaches a maximum around 4.8 eV/Si atom and decreases at higher deposited energies, which is attributed to a reduced hole collection determined from external quantum efficiency measurements. The fill factor decreases above 1 eV/Si atom which we attribute to a lower mobility-lifetime product due to an increase in charge carrier recombination. This indicates defect formation at deposited energies above 1.7 eV/Si atom, significantly below the reported increase in Urbach energy around 4.8 eV/Si atom reported in chapter 4. Chapter 6 discusses the development of the a-Si:H surface roughness as function of several parameters like substrate temperature, RF substrate biasing and hydrogen dilution at growth rates of 0.1 nm/s. Important for this analysis was the spectroscopic ellipsometry technique, which allows to monitor the surface roughness development in-situ during film deposition. In the first section, from depositions with and without RF substrate biasing at different substrate temperatures the presence of a hydrogen-rich layer was suggested, which is removed/densified upon externally induced ion bombardment. In the second section of chapter 6, the surface roughness development was investigated as function of hydrogen dilution. We observe a discrepancy in the surface roughness development between spectroscopic ellipsometry and AFM measurements. We interpret this as another indication of the presence of a hydrogen rich/low density overlayer. At higher hydrogen dilutions, we obtain a thicker overlayer dominated by lower hydrides. From additional PSB experiments at two different dilutions we conclude that the hydrogen rich/low density layer is densified by the induced ion bombardment and/or excess hydrogen is removed.