WO2011122959A1

WO2011122959A1 - Process for the production of disilane

Info

Publication number: WO2011122959A1
Application number: PCT/NO2011/000112
Authority: WO
Inventors: Vidar BJØRNSTAD; Dag Øistein ERIKSEN; Alexander Krivokapic
Original assignee: Polysilane As
Priority date: 2010-03-30
Filing date: 2011-03-30
Publication date: 2011-10-06
Also published as: NO334627B1; NO20121263A1; NO20100466A1

Abstract

Abstract A continuous process for the production of disilane is disclosed which comprises a transition metal catalyzed transformation of monosilane (monosilane). Hydrogen produced in said transformation is removed from the process by the use of a semi-permeable membrane or by an absorbent.

Description

Process for the production of disilane

Field of invention

The present invention relates to a novel and inventive process for the production of disilane. Background of the invention

In the production of electronics and photovoltaic devices today thin layers of silicon are deposited on substrates, most often as silicon wafers. Such layers are produced by use of CVD, Chemical Vapor Deposition, or MBE, Molecular Beam Epitaxy. The silicon is introduced to the reactor as a gas which is thermally decomposed at the substrate to be covered. The gases employed are monosilane, SiH₄, or TCS, Tri-Chloro Silane, SiHCl₃. The temperatures needed to decompose these gases are in the ranges of 800 and 1400 K, respectively. Thus, the substrates need to tolerate high temperatures. In addition, in CVD, which is the faster of the two processes, the gas is heated in the whole reaction chamber and a lot of silicon is deposited, i.e. lost, outside the substrate. If the deposition temperature could be lower, the loss would be diminished as the temperature gradient between ambient and reaction site is smaller and therefore easier to isolate thermally.

Disilane, Si₂H₆, represents an alternative to avoid the challenges of monosilane and TCS. Disilane is used for the deposition of amorphous silicon, epitaxial silicon and silicon based dielectrics via rapid low-temperature chemical vapor deposition (LTCVD) '. Disilane is also used in the epitaxial growth of SiGe films by molecular beam epitaxy (MBE) in conjunction with solid sources of germanium². Disilane is a precursor for the rapid, low temperature deposition of epitaxial silicon and silicon-based dielectrics .

Deposition of amorphous silicon, a-Si, on substrates using disilane can be performed at much lower temperatures compared to silane. It has also been shown that disilane is superior to silane in depositing polycrystalline thin films on various substrates. Chen et al.⁴ reports a 50% improvement in film thickness uniformity and 25% in surface roughness. Rogel et al.⁵ reports improved properties of thin film transistors produced on Corning glass substrates.

The thermal decomposition mechanism of disilane is not well determined, but the main steps are agreed upon. It is accepted that the mechanism of decomposition contains splitting of the disilane molecule into monosilane and SiH₂.⁶ The splitting of the Si-Si bond requires 340 kJ/mole compared to the Si-H bond which requires 393 kJ/mole⁷. Therefore, the temperature required to split the disilane molecule is lower than for braking up monosilane, but only half of the disilane molecule will deposit at this temperature. However, Niwano et al. propose a decomposition mechanism at 700K passing an intermediate step of H₂Si-SiH₂. If so, all the disilane will be transformed to silicon. Tonokura et al. studied hydrogenated silicon clusters as they were formed in time-of- flight mass spectrometry, but could not see any mono-silicon compounds.⁸ The same group in a later study determined the decomposition steps of SiH₄, SiH₂, H₃SiSiH, and production of higher silanes, i.e. tri- and tetrasilanes⁹.

Disilane for use in the methods described above has been produced by a variety of processes.

Disilane may be produced from acidic reactions with Mg₂Si¹⁰:

2 Mg₂Si + 8HC1 ¾ 4 MgCl₂ + Si₂H₆ + H₂ Eq. 1

Dilute hydrochloric acid (5%) can be used. The chemical yield of disilane compared to monosilane increases with increase in temperature. To operate the reaction at close to 100°C dilute sulfuric acid or phosphoric acid can be used.

Japanese company Mitsui employs the transformation of magnesium silicide with addition of ammonium chloride:

2 MgaSi + 8 NH₄C1 ¾ 4 MgCl₂ + Si₂H₆ + 8 NH₃ + H₂ Eq.2 These processes require silicide of the alkaline earth metals. One way of producing such silicides is by jet-milling of metallurgical grade (MG) magnesium together with MG silicon. Another method of producing disilane is to start with monosilane and activate these molecules so that they can react. One way of activation is by electric discharge and another to heat the silane gas. These methods are used to produce small amounts of disilane by the monosilane producers. Another method believed to be less energy consuming and with easy process requirements is to use a catalyst. Japanese companies have filed patents on the subject, but they have not been implemented industrially. These patents use platinum group metal complexes¹¹ or lanthanide complexes as catalysts¹². However, Japanese company Showa

Denko has published results using the period 4 transition metals Ni, Mn and Fe as catalysts in a process where disilane was formed in a reactor at 470 - 570K, separated cryogenically at 200K in another unit and the unreacted monosilane was recycled for 10 hours in a batch operation. Within one cycle the gas had to be treated in two stages with a temperature difference of > 250K. Hydrogen, which is not an inert gas in this process, was used as a carrier gas.¹³

Showa Denko has also filed a patent for producing trisilane and higher ones by

disproportionation of mono- and disilanes¹⁴. In this patent, hydrogen is used as a carrier gas and further described as being inert. The above mentioned methods all have various problems to be solved before they are suitable for large-scale industrial production; the most serious problems concerning many of the catalysts are too low activity and too much bi-products. Concerning electric discharge and heating, these methods suffer from low yield, high energy consumption, and low selectivity for disilane compared to higher silanes. In addition, large amounts of elemental silicon as dust are lost in the latter processes. From the above it is clear that there are several methods that can be used to produce disilane. However, in addition to the yield and content of higher silanes, the chemical purity is imperative when it comes to use in photovoltaics and electronics. Because of the importance of di-, and higher silanes, new methods of making the compounds are constantly sought in order to provide improved or alternate methods of manufacture. The present invention is thus concerned with a novel and improved method for the production of disilanes.

Description of the invention

The present invention provides a novel and improved process for the manufacture of disilane. The present process avoids some of the disadvantages known in the prior art concerning the manufacture of disilane by being continuous, having a high energy and atom efficiency, and having a high per-cycle yield of product.

An important step in the present invention is a transition metal catalyzed reaction for the preparation of disilane utilizing monosilane as starting material:

M

2SiH4→ SiH₃SiH₃ + H₂ Eq. 3

Wherein M is a group 5 - 10 metal, or a compound thereof. As is also seen from Eq.3, hydrogen gas is produced in the reaction. Thermodynamically, it is known that many chemical reactions may reach an equilibrium where usually all components are present.

Whether a reaction will reach an equilibrium is, amongst other things, dependent on the reversibility of said reaction. The concentrations, i.e. activities, of the species are then governed by the equilibrium constant defined by the law of mass action. Using this law we deduced that removal of one (or more) of the products during the reaction could lead to further production of the desired product to maintain the equilibrium constant. This would lead to a more efficient process for the manufacture of disilane.

This hypothesis was confirmed in our laboratory by running test reactions with various amounts of hydrogen or an inert gas, i.e. argon. When hydrogen was added the disilane yield was reduced, in accordance with the above mentioned law.

The use of a carrier gas is not imperative for the process to run, but any gas which is inert under the process conditions, such as nitrogen, argon, and helium can be used in the present process as carrier gas. As was seen in the test reactions, hydrogen is not an inert gas in the present process. The more hydrogen to be removed from the process gas mixture, the higher the yield of disilane.

The hydrogen gas produced in the reaction of monosilane to produce disilane can be removed by using a separator. In the present application, the term separator is intended to mean a hydrogen selective membrane or absorbent, an absorbent being a compound capable of selectively absorbing hydrogen, or a metal capable of reversibly absorbing hydrogen. More generally, the separator does not require any condensation process to separate the hydrogen gas from the process stream.

Selective membranes, permeable only to the H₂ molecules, are known. One example of such membranes is described by Hsieh and Keller.¹⁵ This membrane is comprised of a separation layer of sulfonated-polysulfone on a support layer of polysulfone. As polysulfone they recommend polyarylethersulfone with at least one sulfonic acid group present on one of the aromatic rings, and as separation layer sulfonated bisphenol A polysulfone. By this means separation factors of hydrogen with respect to monosilane above 100 was achieved. Other membranes for selective hydrogen gas removal, preferably based on the size exclusion principle, are also suitable. Such membranes are typically made of metal such as Palladium alloys, as well as micro porous or dense ceramics based on for example silica, zirconia, SrCeO_x and BaCeO_x. Membranes of a more robust material for low temperature separation of hydrogen from silane can also be made of SiC. SiC-membranes have several advantages like high themal conductivity, thermal shock resistance, resistance in acidic and alkali

environments, chemical inertness, and high mecanical strength. Therefore they are useful in membrane reactors. The performance of SiC membranes has been tested at 473K by Dr Tsotsis' group at USC¹⁶ and they find separation factors between methane and hydrogen of 29 - 78 in favor of hydrogen. Although the performance is reduced at a lowered temperature, operation conditions at 400 - 470K are acceptable.

Compounds capable of selectively absorbing hydrogen may be found in the groups named zeolites, metal-organic-frameworks, covalent-organic frameworks, or active carbon. The group of metals capable of reversibly absorbing hydrogen comprises the hydrid forming metals like magnesium, titanium, nickel, palladium, and platinum, and alloys like lanthanum- nickel. Also, compounds like L1AIH₄ may be used for H₂ absorption. However, it is imperative that the absorbent absorb H₂ at one temperature, and desorb the gas at an enhanced temperature.

Thus, the present invention comprises the following main steps:

1. Monosilane gas, SiH₄, with or without an inert carrier gas, is passed through a

contactor with a catalyst.

The contactor may be a packed bed reactor, a fluidized bed reactor, a membrane reactor, a combination of these, e.g. packed bed catalytic membrane reactor (PBCMR) as well as other types of suitable reactors. The catalyst comprises a transition metal of group 5 to 10, or a compound thereof. Preferred metals are V, Cr, Mn, Fe, Co, and Ni. Commercial catalysts on substrates made of aluminate and silicates have shown good efficiency.

The catalyst should be heated to a temperature ranging from 290 to 520 K. In order to achieve a more energy efficient process, the catalyst is preferably heated to a temperature ranging from 330 K to 470 K, more preferred 400-470 and even more preferred 410-430 K. The exhausted gas from the contactor is subjected to (distillative) separation to separate non-reacted monosilane, as well as the optinional carrier gas and hydrogen generated in the contactor, if said hydrogen is not already separated in the contactor, as gas from disilane as liquid. The hydrogen gas generated in the condensation reaction in the contactor is separated from the non-reacted monosilane gas and the remaining components of the process stream by use of a separator suitable for separating hydrogen gas from monosilane. The separator is situated in the process stream as suitable. Preferably the separator is incorporated into the contactor/reactor in step 1 , or optionally after the distillative separation in step 2. The preferred operating temperature of the separator will depend on where in the process stream it is situated. When incorporated into the

contactor/reactor, the optimum operating temperature should be in the same range as for the catalyst, i.e. a temperature ranging from 330 K to 470 K, more preferred 400- 470 and even more preferred 410-430 K. When the separator is placed after the distillative separation step, a preferred operating temperature would be in the same range as the temperature used in said separation step, i.e. 190-350 K, preferably 250 - 320 K.

4. The non-reacted monosilane gas removed from formed disilane and depleted of

hydrogen is returned to a reservoir (make-up tank) for further reaction in the contactor.

Virgin monosilane can be added in each cycle in an amount equal to the one removed as disilane in the distillative separation, thus maintaining a constant amount of monosilane in the feed.

The present process comprising the steps above provides a number of advantages. Use of a separator to deplete the carrier gas/monosilane flow of hydrogen gas significantly reduces the requirement of adding new inert gas in each cycle since only hydrogen gas is removed, as opposed to the known use of a condensator where parts of the inert carrier gas/hydrogen is removed from the monosilane gas by cooling. In addition, use of a condensator is not a feasible method for selectively removing hydrogen gas from the inert carrier gas. Further, the separator allows for a more energy efficient process by avoiding having to cool the

monosilane mixture, optionally containing an inert carrier gas, in a condensator. Further, and perhaps most surprisingly when considering the prior art, removal of hydrogen gas formed in the reactor, keeping the amount of hydrogen gas in the reaction process stream to a minimum, allows for a process with higher yield of disilane per cycle. Finally, separation of hydrogen and silane makes the gas mixture less susceptible to combustion if a leak would lead to contact with oxygen/air. Thus, the proposed process is safer than competing processes. Examples of catalysts used

Example 1 Nickel (Ni) as catalyst Commercial and non-commercial, purposely made catalysts were tested for catalytic efficiency by varying the temperature of the reactor, the flow rate and pressure of the feed, i.e. monosilane gas phase. The catalysts tested were Ni on substrates of aluminate, alumina-silica, and on the nano-material carbon cones. All worked satisfactory, but the best results were achieved with a reactor with a packed bed of a commercial catalyst with high Ni-loading. The yield was non-zero even at 300K. Optimum temperature was found to be below 450K. No trior higher silanes were formed.

Example 2 Platinum (Pt) as catalyst

Non-commercial, purposely made catalysts were tested for catalytic efficiency by varying the temperature of the reactor, the flow rate and pressure of the feed, i.e. monosilane gas phase. The catalyst tested was Pt on substrate of the nano-material carbon cones. It did not work satisfactorily. Example 3 Palladium (Pd) as catalyst

Commercial catalyst was tested for catalytic efficiency by varying the temperature of the reactor, the flow rate and pressure of the feed, i.e. monosilane gas phase. The catalyst tested was Pd on substrates of alumina-silica. It worked well, but the Pd-catalyst was not more efficient than Ni.

Brief description of the drawings

Figure 1 shows a flow chart of a process for the production of disilane, wherein the separator (32) is situated after the distillative separation step (23). In this case the separator is membrane based. Figure 2 shows a flow chart of a process for the production of disilane, wherein the separator is a membrane and is incorporated into the reactor (41). Figure 3 shows a flow chart of a process for the production of disilane, wherein the separators (32A) and (32B) are situated after the distillative separation step (23). These separators are run in opposing cycles, i.e. one separator is absorbing hydrogen, while the other is desorbing, and vice versa. Figure 4 shows a flow chart of a process for the production of disilane, wherein the separator is incorporated into the reactors (41 A) and (4 IB). These reactors are run in opposing cycles, i.e. one reactor is absorbing hydrogen and performing the catalytic reaction, while the other is desorbing hydrogen, and vice versa. Detailed description of four embodiments of the present invention

Embodiment 1: Process with a membrane separator (32) situated after the distillative separation step (23)

In Figure 1 , a process flow chart is shown. Tank (2) is a make-up tank for monosilane. This is the feed to reactor (22) where the catalyst is situated. The exiting gas mixture from reactor (22) enters separator (23) where disilane is separated from monosilane, optional inert carrier gas and hydrogen gas. Said disilane is liquefied and transferred as liquid to evacuated bottle (25), kept cool by a freezer (26). Said monosilane, optional inert carrier gas and hydrogen gas are compressed in compressor (33) to traverse membrane separator (32) where said hydrogen gas is separated from said monosilane and optional inert carrier gas. Said hydrogen gas is transferred to a pressure tank (3) where it is stored and used for other purposes, e.g.

regeneration of the catalyst, heating, or other

Said monosilane is transferred back to the make-up tank (2). In this way the monosilane is recycled until it is used entirely, or by having a reservoir of pure monosilane the extracted amount of disilane can be substituted by adding the corresponding amount of monosilane through tube (12).

Embodiment 2: Process with a packed bed catalytic membrane reactor (PBCMR), i.e. a process wherein the membrane separator is incorporated into the reactor (41).

In Figure 2, a process flow chart for a process where a packed bed catalytic membrane reactor (41) is used. The figure is similar to Figure 1 except that the membrane separator (32) is removed. In this process the hydrogen gas is removed by a membrane incorporated into the reactor. Said hydrogen gas is transferred to a pressure tank (3) where it is stored and used for other purposes, e.g. regeneration of the catalyst, heating, or other.

Embodiment 3: Process with two hydrogen separators operating in opposing cycles. While one separator is absorbing hydrogen, the other is desorbing.

Figure 3 shows a flow chart for a process where the hydrogen is removed by an absorbent or a metal/metal alloy. The figure is similar to Figure 1 except that the separator (32) is now duplicated to allow for running the separators in opposing cycles. During the desorbing of hydrogen from one of the separators, the other separator is absorbing hydrogen. In this way a continuous process is obtained. Some valves and compressors are not drawn to clarify the drawing. Embodiment 4: Process with two hydrogen separators incorporated in two catalytic reactors operating in opposing cycles. While one reactor is producing disilane and absorbing hydrogen, the other is desorbing.

In Figure 4 a process flow chart for a process where the hydrogen is removed by an absorbent or a metal/metal alloy mixed with the catalyst. The figure is similar to Figure 2 except that the reactor (41) is now duplicated, named (41 A) and (41B). Some valves and compressors are not drawn to clarify the drawing. In specific cases a metal or metal hydride may be used as both a catalyst for the production of disilane, and as an absorbent for the produced hydrogen. Said hydrogen is subsequently transferred to a pressure tank (3) by heating the reactor. During the desorbing of hydrogen from one of the reactors, the other reactor is catalyzing the reaction and absorbing hydrogen. In this way a continuous process is obtained.

Claims

1. A process for the production of disilane comprising the steps of:

a. contacting a monosilane flow, optionally including an inert carrier gas, with an immobilized transition metal catalyst from group 5 - 10, or derivatives of such metals, said catalyst being contained in a reactor; b. separating the formed disilane from the unreacted monosilane; and c. separating hydrogen gas, formed in the reactor, from the other components of the process stream by use of a separator.

2. A process according to claim 1 wherein the metal is V, Cr, Mn, Fe, Co, or Ni, preferably Ni.

3. A process according to claim 1 , wherein the temperature of the catalyst is in the range of 290 K to 520 K, preferably 330 K to 470 K, more preferred 400- 470 and even more preferred 410-430 K.

4. A process according to any of the claims 1—4, wherein the pressure in the reactor is in the range of 0.1 - 30 bar, preferentially 6-12 bar.

5. A process according to claim 1 , wherein the separation of disilane from

unreacted monosilane is done cryogenically.

6. A process according to claim 5, wherein the pressure in the disilane separation step is high enough to condensate the disilane gas at a temperature in the range of 300 - 350 K, preferentially 10 bar and 320K.

7. A process according to claim 1 , wherein the separator used in the separation of hydrogen gas from the process stream is a membrane based on the size exclusion principle.

8. A process according to claims 1 and 7, wherein the membrane used in the separation of hydrogen gas from the process stream is in the form of hollow fibers.

9. A process according to claims 1, 7, and 8 wherein the membrane used in the separation of hydrogen gas from the process stream is within the catalytic bed reactor itself.

10. A process according to any of claims 1-9, wherein the inert gas is argon,

nitrogen or helium.

11. A process according to any of claims 1-6, wherein the separator used in the separation of hydrogen gas from the process stream is a metal capable of absorbing hydrogen.

12. A process according to claim 11, wherein the metal of the separator also works as the catalyst.

13. A process according to claim 11 or 12, wherein the metal is nickel or an alloy containing nickel.

14. A process according to claim 1, wherein the separator used in the separation of hydrogen gas from the process stream is a compound from the group comprising zeolites, metal-organic-frameworks, covalent-organic frameworks, or active carbon, capable of absorbing hydrogen.