Microcrystalline silicon emerged in the past decade as highly interesting material for application in efficient and stable thin film silicon solar cells. It consists of nanometer-sized crystallites embedded in a micrometer-sized columnar structure, which gradually evolves during the SiH4 based deposition process starting from an amorphous incubation layer. Understanding of and control over this transient and multi-scale growth process is essential in the route towards low-cost microcrystalline silicon solar cells. This thesis presents an experimental study on the technologically relevant high rate (5-10 °A s-1) parallel plate plasma deposition process of state-of-the-art microcrystalline silicon solar cells. The objective of the work was to explore and understand the physical limits of the plasma deposition process as well as to develop diagnostics suitable for process control in eventual solar cell production. Among the developed non-invasive process diagnostics were a pyrometer, an optical spectrometer, a mass spectrometer and a voltage probe. Complete thin film silicon solar cells and modules were deposited and characterized. It was established that under state-of-the-art high rate deposition conditions new challenges arise regarding temperature control since the high RF power dissipated in the plasma causes the substrate to heat up significantly during film growth. On the basis of experimental results a semi-empirical engineering model was developed that describes the magnitude of this plasma induced substrate heating for arbitrary reactor geometry and process settings. The experimental study revealed that plasma induced substrate heating leads to sub-optimal material quality and solar cell performance and it should be prevented by designing and incorporating a fast active substrate temperature control in deposition reactors. Another treated aspect of high rate deposition is the required high dilution of the SiH4 gas in H2, which is of importance to the on-going cost price reductions. It was established that under conditions of low H2 dilution transient depletion of the SiH4 source gas evolves through diffusion of SiH4 from the surrounding reactor volume back into the plasma and prevents successful nucleation of crystallites. A self-consistent analytical engineering model was developed for the general description of this transient depletion of source gases as the effect is considered to hold a generic importance to the materials processing field. A procedure of tailored SiH4 density was developed to prevent the transient depletion. Applying this procedure, highly efficient microcrystalline silicon solar cells were deposited without the admixture of H2, decreasing the needed total flow by about a factor 200 and the needed SiH4 flow by about a factor 2 with respect to the highly H2 diluted deposition regime. To explore the benefits of a controlled deposition process the developed diagnostics were simultaneously applied to a series of solar cell depositions. Process drifts acting on various time scales were recorded and attributed to effects like powder formation, transient SiH4 depletion and plasma induced substrate heating. An approach to process control was developed using profiling of process parameters like the SiH4 flow. An improved structural homogeneity in the growth direction was accomplished. Control over this structure evolution gives access to narrower process windows that are expected to become increasingly important as the technological development on microcrystalline silicon deposition advances. An example of a narrow process window is the deposition of microcrystalline silicon close to the transition to amorphous silicon. In this regime a variation of less than 10% in SiH4 flow can tune the crystalline volume fraction of the material to values between 0% and 60%. Applying process control in combination with a hot-wire deposited buffer layer, microcrystalline silicon solar cells were deposited with an open-circuit voltage breaking the 600 mV barrier. The thus obtained material showed an unusually low crystalline volume fraction of ~30%. The solar energy conversion efficiency was 9.8%. An attempt was made to apply the obtained knowledge on material growth and process stability to industrial production of thin film silicon solar cells on flexible substrate foils. Crucial was the recognition of the substrate dependence of microcrystalline silicon growth: The microcrystalline silicon solar cells only functioned well when they were grown as bottom cell in a tandem structure, which was attributed to the changed substrate surface morphology upon application of an amorphous silicon top cell. The best flexible modules showed an initial aperture area solar energy conversion efficiency of 9.4%. To conclude: by investigating the processes occurring beyond the glass windows of a microcrystalline silicon deposition reactor and by considering alternatives beyond the typically used glass substrates a deeper understanding of the microcrystalline silicon deposition process was gained and a step towards the realization of cost-effective thin film silicon solar cells was made.
|Kwalificatie||Doctor in de Filosofie|
|Datum van toekenning||11 dec 2006|
|Plaats van publicatie||Jülich|
|Status||Gepubliceerd - 2006|