Chemistry of inorganic compounds – Silicon or compound thereof – Elemental silicon
Reexamination Certificate
2000-06-06
2002-09-17
Silverman, Stanley S. (Department: 1754)
Chemistry of inorganic compounds
Silicon or compound thereof
Elemental silicon
C423S342000
Reexamination Certificate
active
06451277
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates generally to the field of deposition of silicon by chemical vapor deposition, and more particularly to methods for Heating a Fluidized Bed Silicon Deposition Apparatus which are more convenient, more efficient, of lower cost and provide better quality silicon beads than existing methods
High purity polycrystalline silicon is the basic raw material of the semiconductor and photovoltaic industries. It is currently produced by the decomposition of highly purified silicon containing gases onto a heated high purity silicon surface. This process is termed chemical vapor deposition. The standard industry technique uses high purity silicon rods as the heated surface. An alternative fluidized bed technology is also used on a limited scale. This latter technology provides a large surface area of heated silicon on the surface of beads fluidized by the silicon bearing gas and other diluents and offers the promise of reduced capital and operating costs and production of a more convenient form of silicon in the shape of beads. Many attempts have been made to implement the fluidized bed technology but all have suffered from severe operational and purity problems, which have prevented full commercialization.
The major use for the polycrystalline silicon is in production of single crystal silicon via melting and growth of single crystal silicon boules in Czochralski crystal pullers. Such pullers have specific requirements with regard to feeding the beads, contamination, ease of melting etc. which must be met in order to use silicon beads. Kajimoto et al documents some of these issues in U.S. Pat. No. 5,037,503.
The major operational problem is sintering of the beads in the reactor and the resultant plugging of the reactor, the major purity problems are metals, carbon, oxygen and hydrogen in the bulk and surface of the product and the major problem in feeding beads to the crystal puller is difficulty in controlling the bead flow due to variation in shape and size.
The sintering appears to be more prevalent as the temperature, deposition rate, silicon containing gas concentration and bead size increases and less prevalent as the fluidizing gas flow rate increases. Hence a violently fluidized bed will have a lower tendency to sinter but may tend to blow over more dust and will require more heat.
It has been accepted that it is important to have a reactor that does not contaminate the product and that the use of metal reactors is not feasible, see Ling U.S. Pat. No. 3,012,861 and Ingle U.S. Pat. No. 4,416,913 and hence metal contamination can be resolved by not contacting the beads with any metal parts.
Similarly contact with carbon or carbon containing materials leads to carbon contamination so graphite or silicon carbide parts are usually coated with silicon. Oxygen normally comes in through oxygen containing compounds such as water, carbon monoxide and carbon dioxide in the inlet gases and hence all such compounds are removed from the gas streams to as great a degree as is practicable. Oxygen containing materials such as silicon oxide (quartz) are frequently used as containment materials, see Ingle above, and can be used in contact with silicon although care must be taken to prevent erosion.
Hydrogen contamination is primarily caused during the deposition process when hydrogen remains trapped in the bead. This is a time, temperature and deposition rate dependent process which has been described by A. M. Beers et al “CVD Silicon Structures Formed by Amorphous and Crystalline Growth,” Journal of Crystal Growth, 64. (1983) 563-571. For rapid deposition rates of the order of 2-3 8 micron/minute which are desired in commercial reactors the silicon surface temperature must exceed 800 C. Typical rod reactors usually operate above this temperature as do halosilane based fluid bed reactors and thus such reactors do not suffer from this problem. The current silane based commercial fluid bed reactors operate below this temperature in at least part of the reactor and consequently have dusting problems see Gautreaux and Allen U.S. Pat. No. 4,784,840 and require a second dehydrogenation step as described by Allen in U.S. Pat. No. 5,242,671.
The problem of size and shape is not as important but most polycrystalline consumers would prefer large round beads because they flow better and have less surface area, thus less risk of contamination. Large beads require more gas flow to fluidize and hence more heat to bring said gas up to operating temperature.
One standard way to heat a fluidized bed is through the walls because the heat transfer from the wall to the particles is very good and wall heaters can be easily and cheaply built using electric heating coils. Another standard way is to preheat the gas reactants. A further standard approach is to recover heat from both the solid and gaseous effluent of the reactor by means of heat exchange. A yet further standard approach is to recycle unused reactant and or carrier gas. In a silicon deposition reactor there are problems facing all of these approaches. If the wall is heated then it is by definition hotter than the bed particles and hence more likely to be deposited on as the reaction rate is strongly influenced by temperature. The rule of thumb is that reaction rate doubles with each 10 degree Celsius rise in temperature. Hence a hot wall causes wall deposits which are a loss of product, increase the resistance to heat transfer through the wall and can cause breakage of the reactor on cool-down due to differential thermal expansion. There is also the problem that the heat load is localized to the inlet area where the incoming gases are heated up to reaction temperature. Thus hot beads may be present in the reactor but unable to circulate down to the inlet zone fast enough to provide sufficient heat. Heating the gas reactants is restricted by the thermal decomposition of the silicon bearing gases at around 350-400 C. Thus the gases cannot be heated above this temperature without depositing in the heater or in the inlet to the reactor. This problem is further compounded by heat conducted back into the inlet from hot beads located just above the inlet of the silicon bearing gases. The surface temperature of these beads should be over 800 C. to prevent hydrogen contamination, hence there is a high temperature gradient between the beads at 800 C. and the inlet which needs to be below the thermal decomposition temperature of the silicon containing gases which is 350-400 C. Recovery of heat is difficult because of the tendency of the silicon containing gas to form wall deposits which in turn means the wall temperature must be below 350 C. which is difficult when cooling gases or solids which are at 800 C. or greater. Recycle of unused reactants or carrier gas is also difficult for the same decomposition reason. The recycle gas must be cooled to below the thermal decomposition temperature of the silicon containing gases before mixing with them.
Thus the prior technology has attempted to deal with the heating issue in a variety of ways. Ingle, U.S. Pat. No. 4,416,913 noted the use of microwaves to heat the silicon beads directly through the quartz wall which itself is not heated by microwaves. Poong et al. in U.S. Pat. No. 4,900,411 advises using microwaves and notes the need to cool the wall and the distributor grid in order to prevent silicon deposits, which can then absorb the microwaves. Iya in U.S. Pat. No. 4,818,495 also suggests cooling the distributor grid and providing a heating zone above the reacting zone to compensate. Kim et al in U.S. Pat. No. 5,374,413 notes that cooling of the wall is not effective in preventing wall deposits and greatly increases power consumption and suggests a partition between the reacting and heating zone. Partitions have also been suggested by Ingle see above and Van Slooten in U.S. Pat. No. 4,992,245. Neither Iya in U.S. Pat. No. 4,818,495 nor Van Slooten in U.S. Pat. No. 4,992,245 provided means for the heated beads to travel down to the reacting area in sufficient quantity to he
Johnson Edward M.
Silverman Stanley S.
LandOfFree
Method of improving the efficiency of a silicon purification... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of improving the efficiency of a silicon purification..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of improving the efficiency of a silicon purification... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2825491