CN1157259A

CN1157259A - Trichlorosilane production process

Info

Publication number: CN1157259A
Application number: CN96117960A
Authority: CN
Inventors: 若松智; 相本忠; 佐守广志; 小田开行
Original assignee: TOKUNOYAMA CO Ltd
Current assignee: TOKUNOYAMA CO Ltd; Tokuyama Corp
Priority date: 1995-12-25
Filing date: 1996-12-24
Publication date: 1997-08-20
Also published as: KR970074653A; DE19654154A1

Abstract

Process for preparation of trichlorosilane, includes preforming reaction in a fluidised bed from silicon particles, tetrachlorosilane and H2 at 400-700 DEG C in the presence of a copper silicide catalyst; and process for producing a catalyzer containing copper and silicon.

Description

Production method of trichlorosilane

The present invention relates to a method for producing trichlorosilane by a reaction between silicon particles, tetrachlorosilane, and hydrogen gas. And more particularly to a method for producing trichlorosilane, which can stably perform the above reaction in a fluidized bed at an extremely high reaction rate.

Trichlorosilane (SiHCl)₃) Are widely used as raw materials for producing high purity silicon. That is, when trichlorosilane is reacted with hydrogen ata temperature of 1000 ℃ or more, the following reaction may occur, thereby separating silicon.

In general, trichlorosilane can be produced by a reaction between silicon particles and hydrogen chloride. When tetrachlorosilane incidentally produced in the above-described production method of silicon is separated from the reaction gas and converted into trichlorosilane which is used as a raw material, high-purity silicon can be advantageously produced on an industrial scale.

A process for industrially converting tetrachlorosilane into trichlorosilane employs a reaction of hydrogenating tetrachlorosilane into trichlorosilane as represented by the following reaction formula.

The reaction is usually carried out in a fluidized bed formed in a fluidized bed reactor at a reaction temperature of 400 ℃ and 600 ℃ and a mixing molar ratio of hydrogen to tetrachlorosilane of about 2 to 5.

However, the above reaction involves problems such as extremely low reaction speed and extremely low productivity. In order to solve these problems, some means, such as increasing the size of the reactor, is required.

On the other hand, a method of increasing the reaction rate by using a catalyst containing copper or a compound thereof has been proposed.

Japanese laid-open patent application 56-73617 discloses a process for producing trichlorosilane using copper powder as a catalyst. This publication indicates a process by which trichlorosilane can be produced from silicon and hydrogen chloride and tetrachlorosilane can be converted into trichlorosilane by simultaneously reacting silicon particles, hydrogen chloride, tetrachlorosilane and hydrogen gas in a fluidized bed reactor at a temperature of 350-. It is indicated that copper particles are used as a catalyst in the above reaction.

Japanese laid-open patent application 60-36318 discloses a method for converting tetrachlorosilane to trichlorosilane by flowing hydrogen and tetrachlorosilane through silicon particles to react with the silicon particles at 500-. It is stated that cuprous chloride can be used as a catalyst in the above reaction.

Further, Japanese laid-open patent application 63-100015 discloses a method for reacting tetrachlorosilane with hydrogen gas or hydrogen gas and hydrogen chloride at a temperature of 150 ℃ or more in a flowing state. In the examples disclosed in this publication, the reaction is carried out in an autoclave at 260 ℃. It is pointed out that a catalyst containing metallic copper, metallic halides (including iron halides) and bromides and iodides of iron, aluminum or vanadium is used as the catalyst in the above reaction.

These known copper-based catalysts, i.e., catalysts containing metallic copper, copper chloride or the like, can be used as a particularly good catalyst when the reaction of silicon particles, hydrogen and tetrachlorosilane is carried out in a fixed bed. However, when the reaction is carried out at a temperature of 400 ℃ or more in a fluidized bed, which is most commonly used industrially, if these catalyst materials are directly introduced or introduced into the fluidized bed reactor in admixture with silicon particles during the reaction, the reaction is carried out, the copper-based catalyst or silicon particles are agglomerated to break the flow state, with the result that continuous stable operation is hindered or the reaction rate is lowered.

The inventors of the present invention have made continuous studies to solve the above-mentioned problems and found that the addition of a copper-based catalyst as copper silicide to a reaction system can prevent the copper-based catalyst or silicon particles from being agglomerated and is extremely high in reliability.

They have also found that a catalyst system consisting of a combination of the above copper silicide and an iron component, or a combination of copper silicide, an iron component and an aluminum component, can further improve the rate of the reaction for conversion to trichlorosilane.

It is therefore an object of the present invention to provide a process for producing trichlorosilane by reacting silicon particles, tetrachlorosilane and hydrogen in a fluidized bed at high temperature, wherein a novel catalyst system containing copper silicide is employed, whereby trichlorosilane can be produced more stably and at a higher reaction rate than a process using a known copper-based catalyst.

A second object of the present invention is to provide a method for producing trichlorosilane, which further improves the reaction speed of the conversion reaction into trichlorosilane by using a novel catalyst system in which the above copper silicide is combined with an iron component or with an iron component and an aluminum component.

The above and other objects and advantages of the present invention will become more apparent from the following description.

According to the present invention, the first object and advantages of the present invention can be obtained by a method comprising reacting silicon particles, tetrachlorosilane and hydrogen in the presence of an added copper silicide containing catalyst at 400 ℃ -.

The second object and advantage of the present invention can be obtained by using a catalyst comprising copper silicide in combination with an iron component or copper silicide in combination with an iron component and an aluminum component.

FIG. 1 is a cross-sectional view of a fluidized bed reactor used in the process of the present invention.

In the present invention, the source of tetrachlorosilane as a raw material is not particularly limited, but industrially, it is preferable to use tetrachlorosilane that is incidentally produced by a deposition reaction of high purity silicon because the material is inexpensive. In the deposition reaction of high purity silicon, this by-produced tetrachlorosilane is obtained in a state that it contains unreacted trichlorosilane, hydrogen chloride and the like. In the present invention, the tetrachlorosilane can be used when it is or is substantially pure tetrachlorosilane from which the other components have been separated. It is preferred to use substantially pure tetrachlorosilane because it increases the reaction conversion.

In the present invention, the silicon particles are not particularly limited, but it is preferable to use silicon particles of metallurgical grade having a silicon content of 75% by weight or more, preferably 95% by weight or more. Preferably, the silicon particles have a large surface area to increase the reaction speed on the surface of the silicon particles in the reaction system.

The reaction of the present invention is usually carried out using a fluidized bed reactor as hereinafter described. In this case, the average particle diameter of the silicon particles is preferably 100-300 μm to obtain a good flow state.

In the present invention, hydrogen is not particularly limited, and it may be one produced by a known method or incidentally produced by other production methods or the like, without being limited to its source.

In the present invention, as a method for reacting silicon particles with a mixed gas of tetrachlorosilane and hydrogen, a method of reacting them with each other in a fluidized bed at a temperature of 400-. The reaction is usually carried out in a fluidized bed system, and the silicon particles, tetrachlorosilane and hydrogen gas may be supplied continuously or intermittently. Preferably, the silicon particles are intermittently supplied according to the amount of consumption thereof, and tetrachlorosilane and hydrogen gas are continuously supplied.

FIG. 1 is a cross-sectional view of a typical fluidized bed system reactor used in practicing the process of the present invention.

The reactor 1 comprises an ultrahigh (freeboard) section 2 and a fluidized bed section 3, which is provided at the bottom of the fluidized bed section with a conveying gas inlet pipe 5, a gas disperser 4 connected to the end of the pipe 5 and a particle conveying pipe 6 open at the upper end of the fluidized bed section. The open end of the particle discharge pipe 8 is open toward the middle portion of the fluidized bed portion 3 and the other end thereof is connected to a reaction gas discharge pipe 10 through the fine particle collecting cyclone 7. In this way, the gas containing the particles of the fluidized bed 11 (formed in the fluidized bed portion) separates fine particles in the fine particle collecting cyclone 7, and then the gas is discharged from the reaction gas discharge pipe 10. A silk screen valve 9 is provided at the beginning of the particle discharge tube 8.

In the above reactor, silicon particles are fed through the particle transport pipe 6. In which case the catalyst, which will be described in detail below, may be fed simultaneously through the particle duct 6. Meanwhile, tetrachlorosilane and hydrogen gas may be supplied from a conveying gas introduction pipe 5 through a gas disperser 4, thereby forming a fluidized bed 11.

Tetrachlorosilane and hydrogen, diluted with an inert gas which does not take part in the reaction, such as nitrogen or argon, may be fed in.

In the above reactor, the silicon particles may be introduced into the fine particle collecting cyclone together with the exhaust gas through the particle exhaust pipe 8 and discharged from the reaction gas exhaust pipe 10 as a gas substantially free of particles.

In the present invention, the amounts of tetrachlorosilane and hydrogen to be fed may be appropriately determined within a range that ensures a flow rate at which a fluidized bed can be formed. For the ratio of tetrachlorosilane to hydrogen, 1 mole of tetrachlorosilane is generally fed with 1 to 5 moles of hydrogen. However, since the total amount of trichlorosilane produced is the product of the degree of conversion of the conversion reaction into trichlorosilane and the amount of flow of tetrachlorosilane fed to the reactor, it is preferable that 1 mole of tetrachlorosilane is fed with 1 to 3 moles of hydrogen.

In the present invention, it is important to carry out the reaction of the silicon particles, tetrachlorosilane, and hydrogen gas as raw materials in the presence of a catalyst containing copper silicide.

By supplying a copper component as copper silicide to the reaction system for the catalyst, the agglomeration of the copper component or silicon particles can be effectively prevented, and the reaction can be stably performed without lowering or changing the reaction speed in the fluidized bed formed of these particles. Clogging of the reactor, in particular of the particle withdrawal pipe, due to agglomerated particles can advantageously be prevented.

Although copper silicide may be present in the reaction system independently of silicon particles, it is preferably present on the surface of the silicon particles in view of reactivity and easy handling.

In this case, the "surface of the silicon particle" represents a range that can be measured by EDS (energy dispersive spectrometer) of a scanning electron microscope. Since the signal from the EDS is indicative of the proportion of the component present on the surface of the particle, the composition on the surface of the particle can be known by analyzing the signal from the EDS. Further, after the acceleration transformer of the electron microscope was set to 20KV and the surface of the particle was seen, the magnification was set to 1000 times. The X-ray intensity of the EDS in a 10 micron square in the field of view was then determined to obtain the compositional ratio of the components on the surface from the intensity ratio.

As the copper silicide present on the surface of the silicon particles, particles having an alloy composition containing 85% by weight or less of copper are suitable. If the copper concentration on the surface exceeds 85 mole percent, the particles agglomerate at elevated temperatures. This is because the upper limit of the stable alloy composition ratio of copper to silicon is Cu₅The concentration of Si, i.e. copper, was 83.3 mol%. Under the condition that agglomeration of particles is liable to occur, for example, when small particles having an average particle diameter of 100 μm or less are used as the catalyst, the alloy composition ratio of copper to silicon is preferably Cu₄The concentration of Si, i.e., copper, is 80 mol% or less. The alloy composition ratio of copper to silicon is preferably Cu under conditions where agglomeration is more likely to occur₃The concentration of Si, i.e., copper, is 75% mole% or less. Although in practice the copper concentration will sometimes be above 85%, the essence of the invention is to prevent the copper containing an unstable silicide of excess copper and metallic copper that tends to agglomerate with other particles from being exposed to the surface of the particles.

Even when a small region having the above-mentioned high copper concentration exists, if the region having the high copper concentration occupies 10% or less of the total surface area of the particles, it does not cause a problem like deteriorating the flow state. Therefore, the silicon particles with copper silicide used in the present invention are preferably an alloy of copper and silicon having a copper content of 85 mol% or less, the alloy being present on 90% or more of the surface area of the silicon particles.

It is considered important that copper silicide be present in the vicinity of the surface of the particles participating in the reaction. According to the method for producing silicon particles having copper silicide on the surface provided by the present invention, at least 80% of copper silicide can be caused to exist at a depth of up to 10 μm (calculated from the surface), and therefore the amount of copper used can be reduced while maintaining the catalytic effect. The amount of copper silicide present at a depth of up to 10 microns (calculated from the surface) is determined by the method described below.

Several grams of the particle size-adjusted silicon particles were immersed in about 100 ml of a mixture solution containing 70% concentration of nitric acid and 50% concentration of hydrofluoric acid in a mixing ratio of 10 to 1 for 5 to 30 seconds with stirring, and then the solution was introduced into a large amount of water to terminate the reaction. After the particles were separated by filtration and rapidly dried, the weight of the particles was determined. From the weight reduction it is known how much silicon particles are dissolved in the solution and these conditions are repeated one or more times to dissolve the particles until the parts having an average depth (calculated from the surface) below 10 microns are removed. The copper content present in the depth (calculated from the surface) portion of 10 μm or less can be known from the difference between the amount of copper contained in the entire particle before the surface is dissolved and the copper content in the entire particle dissolved and removed to a depth of 10 μm (calculated from the surface) by the above-described method. The copper content in the whole particle can be determined by completely dissolving the particle in a mixed aqueous solution of nitric acid and hydrofluoric acid by ICP (inductively coupled plasma) spectrometry.

In the present invention, the amount of the catalyst (calculated on the basis of copper atoms) is preferably 0.1 to 30 parts by weight based on 100 parts by weight of silicon present in the reaction system of the trichlorosilane production reaction. In order to improve the degree of conversion and stabilize the flow state, it is more preferable to use 0.2 to 20 parts by weight of the catalyst.

In this case, since the silicon component is also contained in the silicon particles having copper silicide, the silicon particles can also be considered as silicon present in the reaction system.

Therefore, even if a catalyst having a copper content of 0.1 to 30 parts by weight, i.e., only silicon particles having copper silicide present in the reaction system, is prepared, the reaction can be satisfactorily performed.

Since copper contained in the added copper silicide is continuously converted into copper silicide by reacting with silicon in the reaction system and remains in the reaction system, it is sufficient to compensate for the reduction in the amount of copper silicide caused when the metallic silicon particles are removed to control the iron component (as described later) or the reduction in the amount of copper silicide crushed and diffused outside the reaction system once it is added to the reaction system, and it is almost unnecessary to newly add copper silicide.

The method for producing the silicon particles having copper silicide at least on the surface, which is advantageously used in the present invention, is not particularly limited, and for example, the silicon particles can be produced by the following method.

Heating silicon particles having an average particle diameter of 50 μm to 2 mm and cuprous chloride particles and/or copper chloride particles having an average particle diameter of 1 mm or less at a temperature of at least 250 ℃ under an atmosphere of a non-oxidizing gas.

In the above method, the silicon particles as the raw material are not particularly limited if they are metallurgical grade silicon having a silicon content of 75% or more. It may contain impurities such as iron and aluminum. Examples of the silicon particles include silicon I and silicon II described by JIS-G2312 and silicon iron I and silicon iron II described by JIS-G2302. Preferably, raw silicon particles produced by mechanically milling the above metallurgical grade silicon or by chemical treatment such as acid milling to control the particle diameter and particle size distribution thereof may be used, so that the reaction product may be added as a catalyst for the fluidized bed reaction to produce trichlorosilane. In general, when two kinds of particles different in average particle size are present in the fluidized bed, if their true densities are almost the same and their particle diameter ratio is 5 to 6 times, the homogeneous mixture state will disappear, and particles having a larger particle diameter will be accumulated in the lower portion of the fluidized bed, and this state is called "layered state".

As described above, since the average particle diameter of the silicon particles used as the raw material for the fluidized bed reaction, that is, the average particle diameter of the silicon particles in the reactor is 100-300. mu.m, the silicon particles of the additive added are preferably 20 μm or more in average particle diameter but 2 mm or less in average particle diameter so that they are uniformly mixed without occurrence of stratification in the fluidized bed. Copper silicide particles having a small particle diameter have a characteristic of easily forming agglomerates. If the particle diameter of the particles is 20 μm or less, a problem is easily caused when the catalyst is added during the reaction due to agglomeration which occurs immediately after the addition. For the reasons mentioned above, the average particle diameter should be between 30 micrometers and 2 mm, preferably between 50 micrometers and 1.5 mm, more preferably between 100 micrometers and 1.5 mm.

Preferably, the copper chloride used is not coarse particles having an average particle diameter of 1 mm or more. The reason for this is that since the average particle diameter of metallurgical grade silicon as a catalyst material is 30 μm to 2 mm, if the average particle diameter of copper chloride exceeds 1 mm, copper silicide containing an excessive amount of copper element is formed on the metallurgical grade silicon particles covered on the copper chloride due to the reduction of copper chloride. Since this copper silicide has extremely high viscosity, it is liable to cause further agglomeration between particles when it is used as a catalyst. This may cause some troubles.

The average particle diameter of copper chloride refers to the diameter of individual copper chloride particles rather than the diameter of agglomerates of the particles, since the individual particles are not separated during the formation of copper silicide and can be suitably separated and used normally by stirring the agglomerates of particles. Copper chloride may be copper (I) chloride or copper (II) chloride, and the purity thereof is not particularly limited.

In this reaction, metallurgical grade silicon particles and copper chloride are first uniformly mixed with each other, and the resulting mixture is maintained at a temperature of 250 ℃ or higher under an atmosphere of a non-oxidizing gas that does not produce undesired oxides or chlorides such as nitrogen, hydrogen, argon or a mixed gas thereof to form copper silicide. Although there is no problem when the mixture is in a static state, more uniform silicon particles can be obtained by stirring using a fluidized bed or a rotary drum. When the temperature of the product mixture is raised, copper silicide is formed on the surface of the silicon particles, while hydrogen chloride and an acidic component, such as tetrachlorosilane or trichlorosilane, are produced. The time required for heating varies to some extent depending on the heating state or kind of the ambient gas, but the completion of the copper silicide formation process can be estimated from the disappearance of the acidic component in the circulating ambient gas.

In the above reaction, the ratio of silicon particles to copper chloride particles is determined depending on the addition range of copper silicide.

Thus, silicon particles having copper silicide at least on the surface can be produced by the above method.

In the present invention, copper silicide has a catalytic effect even when used alone. However, the studies conducted by the inventors have revealed that the catalytic activity can be remarkably improved by the presence of an iron component or an iron component and an aluminum component in addition to copper silicide in the reaction system.

In the present invention, the components of iron and aluminum as the catalyst may be added to the reaction system in any manner, and iron and aluminum are not particularly limited as long as they can be added in a solid form, for example, a metal or a metal silicide thereof.

The amount of the catalyst composed of iron or iron and aluminum is not particularly limited. However, if it is too small, the reaction rate cannot be improved, but if it is too large, fine particles which are reactive substances themselves are covered with non-reactive substances.

Thus, for the amount of catalyst consisting of iron or iron and aluminum, the iron content should be 0.3 to 40% by weight, preferably 0.5 to 30% by weight (based on the silicon present in the reaction system), and the aluminum content should be 0.1 to 3% by weight, preferably 0.2 to 2% by weight, more preferably 0.2 to 2% by weight.

In the present invention, the ratio of each of copper, iron and aluminum in the reaction system can be known by maintaining a material balance in the reaction system and each component can be adjusted to fall within the above range according to the above ratio.

In order to conduct a more uniform reaction, it is more preferable to add iron and aluminum as the above-mentioned catalysts while they are present on the surface or inside of the silicon particles (as the raw material), even when they are added to the reaction system in the form of a metal or a metal silicide.

Therefore, as at least a part of copper, iron and aluminum, it is preferable to use silicon particles containing a large amount of these metal components or silicon particles having these metal components or the above-mentioned metal silicide adhered to the surface.

As the silicon particles containing a large amount of at least one of copper, iron, and aluminum, silicon particles containing impurities such as iron or aluminum can be used. Examples of the silicon particles containing impurities such as iron or aluminum include silicon No. 1 and silicon No. 2 described by JIS-G2312 and silicon iron No. 1 and silicon iron No. 2 described by JIS-G2302, as previously given as a raw material of the silicon particles having copper silicide on the surface.

In addition to the above-mentioned feeding methods, copper, iron, and aluminum as the catalyst may be fed to the reaction system as a compound or mixture containing copper and iron or a compound or mixture containing copper, iron, and aluminum.

Although the added iron component remains in the reaction system like the copper component, when it is supplied as an impurity of silicon particles, the concentration of iron gradually increases as the silicon particles are consumed in the reaction system.

When the concentration of iron becomes particularly high, for example, the proportion of iron exceeds the upper limit of the above range by 40% by weight (based on silicon in the reaction system), the degree of conversion will decrease for the reasons described above. Therefore, it is necessary to remove the iron component from the reaction system or newly add the silicon particles after a lapse of a reaction time, thereby controlling the iron concentration within the above range. The method for removing the iron component is not particularly limited, and smaller particles can be removed from the system by using a cyclone for circulating the particles and periodically changing the classification efficiency of the cyclone. Since a large amount of iron silicide is contained in the smaller particles in this case, the iron component can be removed with high selectivity.

In addition, the aluminum component is converted into aluminum chloride in the reaction system by its reaction with tetrachlorosilane, which is different from copper and iron. Aluminum chloride is a gas at temperatures of 400 c or higher and can therefore be removed from the system. In order to allow aluminum to be always present in the reaction system as a catalyst, it is preferable to control the concentration of aluminum within the above range by including it in the silicon particles rather than adding it in the form of powder.

According to the present invention, clogging of the feed line and a decrease in the flow state caused by agglomeration caused when copper powder or copper chloride is added as a catalyst do not occur at all. In other words, trichlorosilane can be continuously and stably produced for a long time with a high degree of conversion while maintaining a flow state as it is when general metallurgical-grade silicon particles are charged.

The reaction of the silicon particles, tetrachlorosilane, and hydrogen gas may be performed in the presence of a catalyst containing an iron component or an iron component and an aluminum component in addition to copper silicide, so that the reaction speed may be increased and the reaction time may be shortened. Thus, the size of the reactor can be reduced by the process of the present invention employing such a catalyst.

The following examples and comparative examples serve to further illustrate the invention. However, it should be understood that the present invention is not limited to these examples.

In the following examples and comparative examples, the average particle diameter, the degree of conversion and the reaction rate were data obtained as described below. (average particle diameter)

The predetermined particles are classified with a classifying screen. The respective fractions were cumulatively added from the fraction having the smallest particle diameter, and when the cumulatively added value reached 50% by weight, the value was taken as the average particle diameter. (degree of conversion)

The gas concentrations before and after the raw materials were fed into the reactor were measured by gas chromatography, and when the mole number of tetrachlorosilane fed into the reaction system was determined as 100%, the percentage of conversion from tetrachlorosilane fed into trichlorosilane was obtained by the following equation. Degree of conversion (%) ([ mole of tetrachlorosilane converted to trichlorosilane]/[ mole of tetrachlorosilane fed]× 100 (reaction speed)

The degree of conversion is obtained from the following equation.

R＝1/tIn(c₀/c₀-c) in the formula

R: the reaction speed;

t: reaction time or, in the case of reaction with a fluid gas, average contact time between the silicon particles and the gas;

c₀: equilibrium conversion rate at reaction temperature (24% when the reaction is carried out at 500 ℃ and 0.7 MPaG); and

c: degree of conversion (%) at reaction time t. (composition of silicon particles used)

Two different compositions of silicon particles, shown as a and B in table 1 below, were used.

TABLE 1

	Fe	Al	Cu	Cr	Ni	Mn
	Fe	Al	Cu	Cr	Ni	Mn		(wt%)
A	0.12	0.51	＜0.01	0.44	＜0.01	0.02		(wt%)
A	0.12	0.51	＜0.01	0.44	＜0.01	0.02	B	0.09	0.05	＜0.01	0.01	＜0.01	0.01
	Ti	Ca	C	P	B	As	B	0.09	0.05	＜0.01	0.01	＜0.01	0.01
	Ti	Ca	C	P	B	As		(wt%)			(PPM)
A	0.01	0.03	0.02	20	20	＜10		(wt%)			(PPM)
A	0.01	0.03	0.02	20	20	＜10	B	＜0.01	0.01	0.01	＜10	＜10	＜10

Example 1 (copper silicide alone)

5 kg of silicon particles having an average particle diameter of 150 μm and a composition shown in A in Table 1 (purity: 98%) were mixed with 2 kg of monovalent copper (I) passed through a sieve having an opening diameter of 2 mm, and the resulting mixture was held at 300 ℃ for 12 hours while being gently blown in a fluidized bed with a mixture gas composed of nitrogen and hydrogen at a mixing ratio of 1: 1 to effect a reaction. After cooling, the reaction product was taken out and weighed, and its weight was 6.2 kg. Part of the reaction product was dissolved in a mixed aqueous solution of nitric acid and hydrofluoric acid, and the copper content in the reaction product was measured by ICP (inductively coupled plasma) spectrometry. The copper content was found to be about 20% by weight. The X-ray intensities at four different random points (points A, B, C and D) were measured by EDS to obtain the results shown in table 2.

TABLE 2

	Si	Cu	Fe	Al
	Si	Cu	Fe	Al	Intensity ratio (A dot)	0.90	0.09	0.01	0.00
Intensity ratio (B point)	0.78	0.20	0.02	0.00	Intensity ratio (A dot)	0.90	0.09	0.01	0.00
Intensity ratio (B point)	0.78	0.20	0.02	0.00	Intensity ratio (C point)	0.45	0.45	0.07	0.03
Intensity ratio (D point)	0.14	0.81	0.03	0.02	Intensity ratio (C point)	0.45	0.45	0.07	0.03

Magnification: 1000 ×, measurement area: 10^-14mm²。

The intensity ratio is the ratio of the intensity of each atom to the sum of the total intensities of the above four atoms (taken as 1).

Thereafter, 35 kg of silicon particles having an average particle diameter of 150 μm and a composition shown in A in Table 1 (purity 98%) were charged into a fluidized bed reactor as shown in FIG. 1, and a fluidized bed was formed by using a mixed gas composed of hydrogen and tetrachlorosilane (molar ratio 2.5: 1) at a flow rate of 100Nm³In each case at 500 ℃ and a pressure of 0.7 MPaG.

A reactor as shown in figure 1 was used with the following characteristics. h1 (height from top of dispersion plate to bottom of fluidized bed portion): 650mmh2 (height of ramp portion): 150mmh3 (height of super high portion): 1100mmh4 (height of cyclone): 380mmh5 (height of upper part of cyclone): 150mmh6 (height of particle discharge tube): 1000mmd1 (internal diameter of fluidized bed portion): 298mmd2 (inside diameter of super high portion): 478mmd3 (inner diameter of cyclone upper part): 115mmd4 (inner diameter of particle discharge tube): 30mm

The degree of conversion increases gradually with the passage of time from the beginning of the reaction, but then becomes constant. The degree of conversion at this time is shown in Table 3.

After the conversion degree was constant, 6 kg of the silicon particles containing copper silicide prepared as described above was continuously added to supplement the decrease in the number of silicon particles caused by the fluidization by the reaction, thereby maintaining the height of the fluidized bed to a certain degree. The proportion of copper atoms based on silicon atoms was 6% by weight. The degree of conversion rapidly increased immediately upon introduction of a small amount of catalyst.

The degree of conversion at the completion of the 6 kg catalyst addition is shown in Table 3. During the introduction of the catalyst, clogging of the feed line did not occur at all, nor was the fluidization regime changed.

The reaction was continued for 60 days even after completely introducing 6 kg of silicon particles containing copper silicide as a catalyst while continuously adding silicon particles containing no copper silicide, thereby maintaining the height of the fluidized bed to some extent. The degree of conversion after 60 days is shown in Table 3.

TABLE 3

Determination of the degree of conversion

Degree of conversion of tetrachlorosilane (%)

After the reaction has been initiated with metallurgical grade silicon particles alone	9.0
	9.0	After the degree of conversion has stabilized	13.1
After adding 6 kg of silicon particles containing copper silicide	20.0	After the degree of conversion has stabilized	13.1
	20.0	60 days after the addition of 6 kg of catalyst	19.8

As shown in Table 3, even in the long-term reaction, the decrease in the degree of conversion was rarely seen, and the decrease in the fluidized state was not seen during the reaction. Further, after the reaction was forcibly stopped, the reactor was opened after cooling, and the inside of the fluidized bed reactor and the silicon particles taken out were checked. In either case, no large pieces of product or the like were seen.

1 hourafter the addition of the above copper silicide, the reaction rate was found to be 0.39 seconds^-1. Example 2

Silicon particles having different copper silicide contents as shown in table 4 were prepared by changing the conditions for preparing copper silicide in example 1 and added to the reactor to ensure that the copper contents as shown in table 4 were present in the fluidized bed, thereby carrying out the reaction.

In the above reaction, the degree of conversion and the reaction rate of tetrachlorosilane obtained in the same manner as in example 1 are also shown in table 4.

TABLE 4

	Degree of conversion (%) of copper atom content (%) and reaction speed (sec.)^-1)
					In copper silicide	In a fluidised bed	After the degree of conversion has stabilized	After 60 days
	1	5	0.3	16.0(0.24)	In copper silicide	In a fluidised bed	After the degree of conversion has stabilized	After 60 days	14.0(0.19)
2	1	5	0.3	16.0(0.24)	2	0.5	17.5(0.29)	15.5(0.23)	14.0(0.19)
2	3	10	0.5	17.5(0.29)	2	0.5	17.5(0.29)	15.5(0.23)	15.5(0.23)

The values are indicated by parentheses.

As shown in Table 4, even in the long-term reaction, the decrease in the degree of conversion was rarely seen and the decrease in the fluidized state was not seen during the reaction. Further, after the reaction was forcibly stopped, the reactor was opened after cooling, and the inside of the fluidized bed reactor and the silicon particles taken out were checked. In either case, no large pieces of product or the like were seen. Comparative example 1

6 kg of silicon particles having an average particle diameter of 150 μm and a composition shown in A in Table 1 (purity: 98%) were uniformly mixed with 470 g of cuprous chloride having an average particle diameter of 50 μm. The same fluidized bed reactor as in example 1 was used, and the reaction was started with silicon particles alone under the same conditions as in example 1. After the conversion degree was stabilized, 6 kg of the above mixture of cuprous chloride and silicon particles was introduced into the reactor in the same manner as in example 1. It can be seen that the pressure drop of the fluidized bed gradually shows abnormal fluctuation and the fluidized state becomes extremely poor. A large amount of flow particles are discharged from the reaction gas discharge pipe 10 located above the fine particle collecting cyclone 7 and the degree of conversion is lower than before the catalyst is added. The change in the degree of conversion of this comparative example 1 is shown in Table 5.

TABLE 5

Determination of the degree of conversion	Degree of conversion of tetrachlorosilane (%)
Determination of the degree of conversion	Degree of conversion of tetrachlorosilane (%)	After stabilizing the degree of conversion by adding metallurgical grade silicon particles	13.3
After addition of 6 kg of catalyst	9.0		13.3

When thereactor was opened after cooling and the granules were observed, agglomeration of copper powder and agglomeration of metallurgical grade silicon particles with copper powder were observed in the removed granules. Further, when the inside of the reactor was observed, it was seen that the inside of the particle discharge pipe 8 connected below the fine particle collecting cyclone 7 was partially clogged with the coarse product. Comparative example 2

6 kg of silicon particles having an average particle diameter of 150 μm and a composition shown in A in Table 1 (purity: 98%) were uniformly mixed with 300 g of electrolytic copper powder having an average particle diameter of 5 μm. The same fluidized bed reactor as in example 1 was used, and the reaction was started with silicon particles alone under the same conditions as in example 1. After the conversion degree was stabilized, 6 kg of the above mixture of the electrolytic copper powder and silicon particles was introduced into the reactor in the same manner as in example 1. It can be seen that the pressure drop of the fluidized bed gradually shows abnormal fluctuation and the fluidized state becomes extremely poor. A large amount of flow particles are discharged from the reaction gas discharge pipe 10 located above the fine particle collecting cyclone 7 and the degree of conversion is lower than before the catalyst is added. The change in the degree of conversion of this comparative example 1 is shown in Table 6.

TABLE 6

When the reactor was opened after cooling and the granules were observed, agglomeration of copper powder and agglomeration of metallurgical grade silicon particles with copper powder were observed in the removed granules. Further, when the inside of the reactor was observed, it was seen that the inside of the particle discharge pipe 8 connected below the fine particle collecting cyclone 7 was partially clogged with the coarse product. Example 3 (addition of iron and aluminum)

35 kg of silicon particles containing iron and aluminum as shown in A in Table 1 and having a purity of 98% and an average particle size of 150 μm were charged into a reactor in a fluidized bed as shown in FIG. 1, and the reaction was carried out at a temperature of 500 ℃ under a pressure of 0.7MPaG in the presence of a catalyst prepared by adding copper and iron to a composition as shown in Table 7 by blowing with a mixed gas composed of hydrogen and tetrachlorosilane at a molar ratio of 2.5: 1 and a flow rate of 100Nm³In terms of hours.

The above copper was provided by causing copper silicide to be present on the surface of the silicon particles in the same manner as in example 1.

The rate 1 hour after the start of the reaction is shown in table 7, and the degree of conversion which became stable after the start of the reaction and the degree of conversionafter the silicon particles were continuously fed while the reaction was continued for 60 days to maintain the height of the fluidized bed to some extent are shown in table 8. Comparative example 3

The same reaction test as in example 1 was conducted, except that no catalyst was added to the reaction system and only silicon shown in B in Table 1 was used. The reaction rate 1 hour after the start of the reaction is shown in table 7, and the degree of conversion which became stable after the start of the reaction and the degree of conversion after the silicon particles were continuously fed while the reaction was continued for 60 days to maintain the height of the fluidized bed to some extent are shown in table 8. Comparative examples 4 to 6 and examples 4 and 5

The same reaction test as in example 1 was conducted, except that only 1 or 2 different catalysts were added to the reaction system as shown in Table 7. The reaction rate 1 hour after the start of the reaction is shown in table 7, and the degree of conversion which became stable after the start of the reaction and the degree of conversion after the silicon particles were continuously fed while the reaction was continued for 60 days to maintain the height of the fluidized bed to some extent are shown in table 8. The above copper was provided by causing copper silicide to be present on the surface of the silicon particles in the same manner as in example 1.

TABLE 7

	Catalyst and process for preparing same	Example of the comparative example to System ( Silicon atom weight%)			Reaction rate (second)^-1)
						Cu	Fe	Al
		Example 3	Adding Cu, Fe and Al	1.0		Cu	Fe	Al	2.0	0.51	0.97
Comparative example 3	Without addition of catalyst component	Example 3	Adding Cu, Fe and Al	1.0	0	0.09	0.05	0.07	2.0	0.51	0.97
Comparative example 3	Without addition of catalyst component	Comparative example 4	Addition of Fe only	0	0	0.09	0.05	0.07	1.6	0.05	0.19
Comparative example 5	Addition of Al only	Comparative example 4	Addition of Fe only	0	0	0.09	0.51	0.16	1.6	0.05	0.19
Comparative example 5	Addition of Al only	Comparative example 6	Adding Fe and Al	0	0	0.09	0.51	0.16	2.0	0.51	0.25
Example 4	Adding Cu and Al	Comparative example 6	Adding Fe and Al	0	1.0	0.09	0.51	0.42	2.0	0.51	0.25
Example 4	Adding Cu and Al	Example 5	Adding Cu and Al	1.0	1.0	0.09	0.51	0.42	2.0	0.05	0.59

TABLE 8

	Degree of conversion (%)
	Degree of conversion (%)		Is stable at the degree of transformation Rear end	After 60 days
	Example 3	22.7	Is stable at the degree of transformation Rear end	After 60 days	21.0
Comparative example 3	Example 3	22.7	6.6	13.1	21.0

Comparative example 4	13.9	14.5
Comparative example 4	13.9	14.5	Comparative example 5	12.4	12.0
Contrast experimentExample 6	16.3	15.0	Comparative example 5	12.4	12.0
Contrast experimentExample 6	16.3	15.0	Example 4	20.5	20.2
Example 5	22.4	21.0	Example 4	20.5	20.2

Examples 6 to 9

The same reaction test as in example 5 was conducted by changing the concentration of the catalyst in the reaction system as shown in table 9. The reaction rate 1 hour after the start of the reaction is shown in table 9, and the degree of conversion which became stable after the start of the reaction and the degree of conversion after the silicon particles were continuously fed while the reaction was continued for 60 days to maintain the height of the fluidized bed to some extent are shown in table 10.

TABLE 9

	Proportion of silicon atoms in the System (% by weight)				Reaction rate (second)^-1)
	Proportion of silicon atoms in the System (% by weight)					Cu	Fe	Al	Catalyst and process for preparing same
	Example 6	1.0	0.4	0.51		Cu	Fe	Al	Catalyst and process for preparing same	1.9	0.59
Example 7	Example 6	1.0	0.4	0.51	1.0	23	0.51	24.5	0.87	1.9	0.59
Example 7	Example 8	1.0	13	0.51	1.0	23	0.51	24.5	0.87	14.5	1.12
Example 9	Example 8	1.0	13	0.51	1.0	2.0	2.8	5.8	1.72	14.5	1.12

Watch 10

	Degree of conversion (%)
	Degree of conversion (%)		Is stable at the degree of transformation Rear end	After 60 days
	Example 6	22.4	Is stable at the degree of transformation Rear end	After 60 days	20.5
Example 7	Example 6	22.4	23.5	21.5	20.5
Example 7	Example 8	24.0	23.5	21.5	22.5
Example 9	Example 8	24.0	24.0	22.5	22.5

Claims

1. A process for producing trichlorosilane, which comprises reacting silicon particles, tetrachlorosilane and hydrogen in the presence of an added catalyst comprising copper silicide in a fluidized bed at 400-700 ℃.

2. The method of claim 1, wherein the catalyst comprises copper silicide in combination with an iron component, or copper silicide in combination with an iron component and an aluminum component.

3. The process of claim 2, wherein the amount of the combination of copper silicide and an iron component or the combination of copper silicide, an iron component and an aluminum component present in the reaction system is adjusted to ensure that the proportionof copper atoms is 0.1 to 25% by weight, the proportion of iron atoms is 0.3 to 40% by weight and the proportion of aluminum atoms is 0.1 to 3% by weight (based on silicon atoms).

4. The process of claim 2, wherein metallurgical grade silicon particles containing 0.3% by weight or more of iron atoms (based on silicon atoms) are added at the beginning of the reaction.

5. The process of claim 2 wherein the aluminum is added to the reaction system as aluminum-containing silicon particles.

6. The method of claim 2, wherein the iron is added to the reaction system as iron-containing silicon particles.

7. The method of claim 1, wherein the copper silicide is added to the reaction system as copper-containing silicon particles having copper silicide present at least on the surface of the particles.

8. A process for producing the copper-containing silicon particles as defined in claim 7, which comprises heating silicon particles having an average particle diameter of 50 μm to 2 mm and cuprous chloride particles and/or copper chloride particles having an average particle diameter of 1 mm or less in an atmosphere of a non-oxidizing gas which does not generate oxides or chlorides at a temperature of at least 250 ℃.

9. The process of claim 8 wherein the amount of cuprous chloride and/or cupric chloride used is 30 parts by weight (based on 100 parts by weight of silicon particles) calculated on the basis of copper atoms.

10. The method of claim 8, wherein silicon particles are produced having 80% copper silicide present in portions (calculated from the surface) below 10 microns in depth.