Production of polysilicon from silane pyrolysis in a fluidized bed

Abstract

The photovoltaic industry has experienced a tremendous growth over the last years. The backbone of the technology has so far been elemental silicon. Silicon is the second most abundant element on the earth, but in order to be utilized for solar panels, it needs to be purified. This purification process is very energy intensive. Behind a finished solar module, up to 30% of the energy goes into purification depending on route, Cucchiella and D´Adamo (2012). Reducing the energy payback time of solar panels is important, and focus on silicon production is crucial, since this is one of the most energy intensive parts. The production of polysilicon involves production of metallurgical silicon from quartz and further processing into polysilicon. The most common route for the latter step is the Siemens process. In this process, chlorosilane is produced from chlorination of the metallurgical silicon. Subsequently, trichlorosilane is reduced in a decomposition process after some additional refining. The decomposition reactor itself is where the energy consumption becomes large. The silicon containing reactant starts to decompose at a temperature of typically 350 - 480 °C depending on the process. However, in order to assure correct deposition, the deposition itself has to happen at temperatures as high as 1100°C for trichlorosilane, and about 650°C for monosilane. The layout of the Siemens process is simple. Electrical heating elements of silicon are distributed within a cooled dome. The heating elements are kept at typically 1100°C, while all other surfaces are kept at about 250°C. The process takes days and even weeks, depending on the process optimization, and the heat loss is therefore substantial. There is one other technology producing polysilicon to the market today, and this is the fluidized bed reactor. In a Fluidized Bed Reactor (FBR), the reactor vessel is filled with silicon particles. A gas is injected at the bottom of the reactor to fluidize the particles. Fluidizing the particles means the drag force on the individual particles is on the same scale as the weight of the particle. In this state, the bed of particles behaves like a liquid, and the flow of gas keeps the bed in continuous motion. The particles are heated to a temperature above the decomposition temperature and the reactant gas is inserted to the bed. Upon decomposition, the silicon deposit on the particles thus making them grow. After some dwell time, the particles have grown to a size suitable for extraction. The finished beads are then extracted, and new small seed particles are inserted to or produced within the bed. What complicates the picture is the decomposition of the reactant and how this influences growth. Challenges associated with FBR production of polysilicon involves dust (fines) production, unwanted depositions on surfaces other than the beads as well as inadequate quality due to porosity, amorphous inclusions and impurities. During the PhD project a state of the art fluidized bed has been designed, built and operated. This thesis discloses the design chosen as well as the background for the choices made. Further, the results of the experimental investigations are presented. The project has successfully achieved production of polysilicon. The results demonstrate two different types of silicon growth and how to control the process. The ability to alternate the nature of the produced material by process parameters has been demonstrated. High-density deposition and low fines production has been demonstrated. Low density and high fines production modes is also demonstrated and parameters leading to this mode is identified. Two different sources of fines formation are demonstrated. The first is nucleation and growth in gas phase. The other mechanism is the release of inadequately bound structures from the surface of the silicon beads post deposition. An ability to scavenge fines through optimization of process parameters is demonstrated. The process was first optimized for large dust production. The beads were then harvested and investigated before being reintroduced to the bed. The process was then optimized for dense depositions leading to a scavenging of the earlier deposited fines. The nature of the finished structures formed by this type of process is also presented in the thesis. The results of this work have been presented in four published journal papers and four conference proceedings. The design of the reactor is patent pending and the patent application is appended to this thesis. Further, a set of experiments was performed in a separate hot wall reactor also referred to as a free space reactor. It is known from the literature that silane based fluidized bed material may be associated with a characteristic periodic porous pattern. Through the presented free space experiments, the ability to grow a similar pattern in steady state has been demonstrated, thus indicating that the phenomenon is not directly linked to the fluid mechanics of a fluidized bed.