Project C2: Magnetic control in ribbon growth on substrate of silicon wafers

PI: Vladimir Galindo (HZDR)
Partner: HZDR, RGS Development

1. Scientific case for the project

1.1 Background

Photovoltaic silicon is today mainly produced by directional solidification of multi-crystalline silicon or by the Cz-growth of silicon crystals. A large amount of energy is needed in both processes. Wafers are then obtained by sawing those ingots. The unavoidable sawing losses are today still in the range of 40…50%. In terms of modern requests for energy and resource efficiency, this is not acceptable. There is an obvious need to look for technologies which have basically the potential to avoid those losses, thus significantly reducing photovoltaic electricity costs.
One solution for this problem consists in the so-called Ribbon Growth on Substrate (RGS) technology, see Fig. 12. It was suggested and developed during the last decade [1,2] and exists today on a pilot level. The process basically consists in melting the silicon in a casting frame and taking out the solidified Si foil on a cooled moving substrate. The efficiency of the process comes from a complete decoupling of the solidification and the casting velocities.
Fig. 12: Scheme of the Ribbon Growth on Substrate (RGS) process.
The principle of the RGS wafer growth process is that a “cold” (~1200oC) substrate is moved underneath a casting frame filled with liquid silicon (melting point 1410oC). Thus heat is extracted from the silicon melt, forcing a crystallization process of silicon from the substrate into the silicon melt. During the cool down process the silicon wafer separates from the substrate and can be removed from the process environment, while the substrate is re-used. The technical advantages of this technology compared to the state of the art, i.e. block crystallization and wafer cutting technology, are:
  • Silicon material yield is increased from about 40% to more than 90%.
  • Wafer cutting and cleaning steps are avoided.
  • Throughput per machine is increased from today's 2-10 MWp annual output per machine to 50 MWp with good possibilities for further up-scaling.
  • RGS silicon wafers are compatible with the currently available cut wafers and can therefore be processed to solar cells in standard solar cell manufacturing lines.
The RGS technique has mainly been developed by the company RGS Development B.V., located in Broek op Langedijk (The Netherlands). The company has built a full scale RGS wafer test production machine for the demonstration of the technology. A smaller but nevertheless real-scale test rig exists at Energy Research Centre of the Netherlands (ECN) in Petten.
For an eventual break-through of the technology, a number of remaining problems need to be addressed. The project, however, will only address generic aspects of those problems, but will not contribute directly to the technical solutions which have later to be developed for the industrial process.

1.2 Most important goals of the planned work

A leading problem consists in the occurrence of flow instabilities and meniscus oscillations at the open slits where the moving substrate enters and leaves the casting frame. The idea is to control those instabilities by means of tailored magnetic fields placed above and around the casting frame. The action of the related electromagnetic forces is to close in a contactless way the remaining slit at the exit of the foil out of the casting frame, as well as to reduce wave-like oscillations in the emerging Si film. Such kind of an “electromagnetic valve” (em-valve) is a known task in MHD, and several (mostly unsuccessful) attempts have been realized in the past, mainly for the case of a contact-free steel casting. An em-valve can only work if it acts in combination with the surface tension of the melt, i.e. for relatively small sizes of the slit of maximum a few millimeter. This, however, is exactly the case in the RGS technology. Recent activities on an em-valve for liquid silicon have been reported in [3]. The used electromagnetic fields are induction heaters of mid to high frequencies, the heating power of which must be incorporated into the overall heating scheme of the facility.
Once such an induction heater is used in the process, the following further questions arise:
  • As the induction heater frequency is a sensitive measure of the circuit load, can its measurement be used for a level sensing of the melt in the casting frame?
  • How does the induced current field depend on the presence of electrically conducting substrate materials? What is the influence of the substrate conductivity on the flow and on the Joule heat distribution in the melt?
  • How does the induction heater influence the flow instabilities in the model problem of a lid-driven cavity?
HZDR will perform numerical simulations for the action of induction heating AC fields on the silicon melt in the RGS process. HZDR will set-up a model experiment for the RGS process with a low melting liquid metal allowing to measure local velocities in the melt, thus characterizing experimentally the flow instabilities in the RGS process. HZDR will analyze numerically the possibility of a melt flow measurement in the RGS process by means of the CIFT technique.
The project will be done in direct cooperation with the industrial partner RGS Development B.V, who will not receive funding from the Alliance but will be involved in the definition of tasks.

2. Existing competencies and infrastructure

HZDR has a long-term experience in the numerical and experimental analysis and design of induction heating systems [4,5]. A very special, tailored induction heater has been developed for the purpose of a crucible-free melt extraction of thin metal fibres [5]. The experimental realization was based on numerical simulations of the electromagnetic fields and the resulting temperature distribution in the metal sheet. The melt flow in various induction heaters and its magnetic control has been studied in [6,7].
The generic case of a DC magnetic field influence on the flow and its stability in a lid-driven cavity has been investigated numerically [8].
HZDR has extensive capabilities in the fields of “cold” liquid metal model experiments and various techniques for the measurement of local velocity distributions in metal melts.
RGS Development B.V. is the leading company for the development of the RGS technology.

3. Resource planning and Budget Justification

The numerical and experimental works described above will be realized by a PhD student at HZDR.

Links: There are close relations to projects A3, A4, C1, and YIG. The experiments in project C1 on silicon wetting at various substrate materials will be of particular interest since the wetting behaviour is of crucial relevance in the RGS process both for the initial solidification on the substrate and for the occurrence of instabilities at the open slit where the silicon wafer leaves the casting frame.

References

[1] A. Schönecker, L.J. Geerlings, A. Müller, 2004, Casting technologies for solar silicon wafers. Block casting and ribon-growth-on-substrate. Solid State Phenomena, Vol. 95-96, 149-158.
[2] G. Hahn, A. Schönecker, 2004, New crystalline silicon ribbon materials for photovoltaics. J. Physics: Condens. Matter, Vol. 16, Topical Review, R1615-R1648.
[3] F. Santana, Y. Delannoy, A. Autruffe, 2009, Electromagnetic retention and stirring to study segregation in solar grade silicon. Int. Conf. EPM2009, Dresden, Proc. 891-894.
[4] A. Cramer, V. Galindo, G. Gerbeth, J. Priede, A. Bojarevics, Y. Gelfgat, O. Andersen, C. Kostmann, G. Stephani, 2009, Tailored Magnetic Fields in the Melt Extraction of Metallic Filaments. Metallurgical and Materials Transactions B, Vol. 40, 337-344.
[5] J.-S. Park, J. Pal, A. Cramer, G. Gerbeth, 2010, Optimization of induction heating for container-less melt extraction from a metallic sheet. Metallurgical and Materials Transactions B, Vol. 41, 1074-1083.
[6] V. Shatrov, G. Gerbeth, R. Hermann, 2008, Linear stability of an alternating-magnetic-field-driven flow in a spinning cylindrical container. Physical Review E, Vol. 77, 046307.
[7] A. Bojarevics, A. Cramer, Yu. Gelfgat, G. Gerbeth, 2006, Experiments on the magnetic damping of an inductively stirred liquid metal flow. Experiments in Fluids, Vol. 40, 257-266.
[8] V. Shatrov, G. Mutschke, G. Gerbeth, 2003, Three-dimensional linear stability analysis of lid-driven MHD cavity flow. Physics of Fluids, Vol. 15, No. 8, 2141-2151