Chemistry of inorganic compounds – Silicon or compound thereof – Elemental silicon
Reexamination Certificate
2000-02-18
2002-04-09
Griffin, Steven P. (Department: 1754)
Chemistry of inorganic compounds
Silicon or compound thereof
Elemental silicon
C423S342000, C423S348000, C423S349000
Reexamination Certificate
active
06368568
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates generally to the field of silicon purification, and more particularly to a method for improving the efficiency of such a silicon purification process
The production of high purity electronic grade silicon is the critical first step of the entire multi-billion dollar semi-conductor industry. The basic process consists of three steps; conversion of metallurgical grade silicon into a hydrohalosilane such as trichlorosilane, purification of this material by distillation and other means, and decomposition of the material back to silicon.
There are two established ways to produce the hydrohalosilane; a low temperature (300-400 C.) low pressure (1-5 atm) high yield (90%) process using a hydrohalide, such as hydrogen chloride; and a high temperature (400-500 C.) high pressure (30-40 atm) low yield (12-24% depending on catalyst and conditions) process using silicon plus hydrogen to hydrogenate a silicon tetrahalide such as silicon tetrachloride.
The hydrohalide process was invented by Siemens and is used by the majority of the silicon producers; the hydrogenation process was invented by Union Carbide and is used in two facilities as part of their silane process.
The following equations show the desired reactions for the two processes but as noted above reaction 1 has a much higher yield
Si+3HCl→SiHCl
3
+2H
2
Hydrohalide process 1
Si+3SiCl4+2H
2
→4SiHCl
3
Silicon tetrachloride hydrogenation 2
Purification is normally done by distillation, but reactive distillation is also used, as is adsorption. In most facilities there is extensive recycle and purification of hydrogen. Three facilities produce silicon hydride or silane, two from the hydrohalosilane by disproportionation (Union Carbide Process) and one from silicon tetrafluoride by reduction via aluminum hydride (Ethyl Process).
The decomposition reactors are all rod reactors except for fluid bed reactors operated on silane as part of the Ethyl Process. Fluid bed reactors have significant capital, operating and energy advantages but have proved difficult to implement. The only operating fluid bed units produce a dusty product contaminated with hydrogen that is not widely accepted.
There are two decomposition reactions for hydrohalosilanes; thermal decomposition and hydrogen reduction. (Trichlorosilane is used in the examples but bromine or iodine can be substituted for chlorine, fluorine cannot)
4SiHCl
3
→Si+3SiCl4+2H2 (thermal)
SiHCl
3
+2H
2
→Si+3HCl (hydrogen reduction)
All halosilane reactors incorporate both and consequently produce an effluent, which has a range of silicon hydrohalides and tetrahalides and hydrogen halides and hydrogen.
The essence of the process is impure silicon in, pure silicon out plus small impurity streams. To accomplish this there are large recycle streams of hydrogen, silicon and halide containing streams and is important not to produce low value by-products or waste streams.
The key problem is the silicon tetrahalide, which is difficult to convert to silicon, and thus causes a difficult problem in closing the plant silicon and chlorine balances without large waste streams. In the preferred hydrohalide reaction process the silicon tetrahalide is also produced albeit in small quantities (4-5%). Thus there is a net production of silicon tetrahalide as a byproduct. Most plants try to minimize this byproduct production and then convert what they have to fumed silica, which is not as valuable as electronic grade silicon but enables recovery of the hydrogen halide for reuse. The Union Carbide hydrogenation of silicon tetrahalide was invented to overcome this problem but is an expensive and dangerous solution (one accident and two fatalities have been reported to date). Another approach was taken by Wacker-Chemie as is shown in the U.S. Pat. No. 4,454,104 by Griesshammer where silicon tetrachloride is reacted with hydrogen in a reactor parallel to the deposition reactor. The effluent from both reactors are then mixed and compressed to 8 bar and cooled to −60 C. in order to force the hydrogen chloride into solution and allow separation of the hydrogen. No mention is made in this patent of the benefits of controlling the temperature of the effluent or of quenching the reaction.
In a recent patent U.S. Pat. No. 5,910,295 by de Luca a closed loop process is proposed which combines the Union Carbide approach of producing silane by disproportionation with the Siemens approach of producing trichlorosilane from hydrogen chloride at high yield. The solution is to react the excess silicon tetrachloride with hydrogen and oxygen. The overall silicon production reaction is simply the thermal decomposition of trichlorosilane to silicon and silicon tetrachloride. The excess silicon tetrachloride is then oxidized with hydrogen and oxygen which chemically is the same as reacting with water.
4SiHCl
3
→SiH4+3SiCl4
SiCl
4
+2H
2
O→SiO
2
+4HCl
Thus 3 moles of fumed silica are produced for every mole of silicon produced. Thus the trichlorosilane production reactors and purification processes must be four times larger than if all the silicon in the trichlorosilane were converted to silicon. The increased capacity of the above process is used to make fumed silica, which is a much less valuable product and does not require the high purification levels that the electronic grade silicon product does. Other silicon production processes make great efforts to promote the hydrogen reduction reaction because it produces more silicon from a mole of trichlorosilane.
SiHCl
3
+2H
2
→Si+3HCl
Such efforts typically include running the reactor at higher temperatures (1100 C.) than needed for thermal decomposition (850 C.) and recycling silicon tetrachloride and hydrogen to the reactor until the silicon tetrachloride is consumed. All practical plants also convert some product to fumed silica either as a means of disposing of contaminated material or as a byproduct for sale. For an overall optimum facility one would want to have the flexibility of producing the desired product slate of silicon and amorphous silica depending on market conditions. Such optima depend on the market demand and pricing for the products together with the marginal production cost and equipment capability and can be optimized using linear programming techniques, as is done in oil refining, providing the equipment has some flexibility.
There have been a number of patents for silicon deposition reactors; the key rod reactor patent is the U.S. Pat. No. 3,041,14 by Schering. U.S. Pat. No. 4,092,446 by Padovani describes an optimized system using a fluid bed and extensive recycle of materials. U.S. Pat. Nos. 5,798,137 and 5,810,934 by Lord describe a fluid bed capable of operating with or without recycle on a variety of feedstock. Various fluid bed patents describe methods of operating and of heating. U.S. Pat. No. 5,374,412 by Kim et al. describe use of two feed streams one of which is used to prevent wall deposition which would block the passage of the microwaves used for heating the beads.
All these systems take the effluent from the decomposition reactor as it is cooled down and removed from the reactor and then separate and recycle the components.
The primary deficiency in the prior technology is that it neglects the opportunities in the temperature regime between the deposition temperature which is typically between 750 and 1150 C. and the condensation temperature of the halosilanes in the effluent which are typically below room temperature. The effluent gases are allowed to cool and continue to react through this large temperature range thus producing more of the undesired silicon tetrahalide.
In this range the species in the effluent change composition with temperature and there is always an optimum temperature for recovery of the desired components which is typically 800-1000 C. At this temperature the desired hydrohalosilanes such as trichlorosilane and dichlorosilane are at
Griffin Steven P.
Johnson Edward M.
LandOfFree
Method for improving the efficiency of a silicon... 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 for improving the efficiency of a silicon..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for improving the efficiency of a silicon... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2859865