CA1224013A - Silane purification process - Google Patents

Abstract

ABSTRACT

A process for the removal of and analysis of phosphine and arsine impurities in silane gas. Silane gas which normally contains the impurities of AsH3 and PH3 is contacted with a solution of NaAlH4 in dimeth-oxyethane to remove the impurities therefrom. The dimeth-oxyethane or other ether solution may then be hydrolyzed with water or alcohol to evolve hydrogen gas from the NaAlH4 and to re-evolve phosphine and arsine which may then be quantitatively determined by gas chromatography, atomic absorption, or other means.

Description

~LA~C IU~ O~IOY ~A~'~9 This invention relates in general to purification of gases and in particular to an advantageous method for the removal of phosphine and arsine from silane as well 5 as the quantitative analysis of phosphine and arsine in silane~
Silane gas is pyrophoric and those skilled in the art recognize that it reacts with many substances. The semiconductor industry uses large amounts of silicon 10 which are produced from gase~ including silane. Since phosphine (P~3) and arsine (AsH3) are common con-taminant~ in silane and are al o known to affect the electrical properties of ~emiconductor silicon, gases used for the preparation of such ~ilicon mu~t nece~sarily 15 be low in content of these impurities to reduce contami-nation of the silicon product.
Thus there e~ists a need for a process to remove arsine and phosphine from silicon while guarding against reaction of the ~ilane ga containing the impurities.
The semiconductor industry often require3 a silicon starting material having as little as a few part per billion or a few tenths of one part per billion of phosphorus or arsenic. This often may only be achieved by starting with a 3ilane gas precursor which contain~

equally small quantities of contaminants. Since suchtiny portions of contaminants are critical, it has been extremely difficult to devise such precise analy~ical techniques as will determine the amount of phosphine and 5 arsine in a sample of silane gas.
The present invention is directed to the problem of removing ar~ine and pho~phine impurities from silane gas while avoiding reac~ion with or contamination of the silane gas it~elf during purification.
The present invention offers the advantage of being able to determine very minute traces of arsine and phosphine impurities in a sample of silane gas. The advantage of the present invention is obtained by con-cen~rating the trace impurities from a large volume of 15 silane gas and then analyzing the concentrated portion in equipment with relatively high lower detection limit~ so as to calculate the contamination of the original silane ~ample.
Thus, the present invention i9 a proce3s for the 20 removal of phosphine or arsine from silane, said proces~
compri~ing intimately contacting silane ga~ containing phosphine or arsine with a liquid mixture consi~ting essentially of an innocuou~ reaction medium and an alkali metal aluminum tetrahydride and recovering the purified 25 9ilane gas.

The pre~ent invention is al50 a process for the analysi~ of phosphorus or arsenic content in silane, said process comprising intimately contacting silane gas con-taining phosphine or arsine with a liquid mixture con-5 taining alkali metal aluminum tetrahydride and thendetermining the amount of phosphorus or arsenic in the liquid mixture.
'Fhe present invention is also an analytical method for measuring the phosphine and arsine content of a 10 silane sample, said process comprising the steps of intimately contacting the sample with a solution of alkali metal aluminum tetrahydride to react the phosphine and arsine therewith so as to remove at least about 90%
of the phosphine and ar~ine from the ~ilane; hydrolyzing 15 the solution to evolve hydrogen ga~ from the alkali metal aluminum tetrahydride and re-evolve phosphina and/or arsine gase~ in a closed space, and determining the amount of arsine and phosphine conc2ntrated in the hydrogen gas.
Various alkali metal aluminum tetrahydride~ may be used according to the process of the invention. These compounds include sodium aluminum tetrahydride, lithium aluminum tetrahydride, and potassium aluminum tetra-hydride. Sodium aluminum tetrahydride i~ most preferred.

The alkali metal aluminum tetrahydrides of the present invention are used in a liquid mixture so that the silane gas may be contacted with the mixture.
Preferably, the liquid mixture is a solution containing 5 the alkali metal aluminum tetrahydride. More preferably, the solvent~ used for dissolving the alkali metal aluminum tetrahydrides are ethers. Ethers usable according to the present invention include tetrahydro-~uran ~THF), dimethoxyethane (monoglyme), and the di-10 methylether of diethylene glycol (diglyme). Variousother ethers may also be used according to the invention. ~or purification, monoglyme is preferred because of its stability and good solubility of the various alkali metal aluminum tetrahydrides including 15 sodium aluminum tetrahydride. For silane analysis, a less volatile compound, such as diglyme is preferred.
Tertiary amines may also be used as a solvent according to the invention. Suitable tertiary amines include the trialkylamines such as triethylamine, tri-n-20 propylamine, tributylamine as well as mixed alkyl amines.
Various hydrocarbon mediums may also be usedaccording to the invention but usually form liquid mixtures which are not solution~ of the alkali metal aluminum tetrahydride.

~ 013 Since solid-ga~ reactions are known to proceed relatively slowly, a solution of the alkali metal aluminum tetrahydride is preferrea over a liquid mixture since the liquid mi~ture approximates the con~act between 5 a solid alkali metal ~luminum tetrahydride and a gas, silane. Thus, liquid mixtures of the alkali metal aluminum tetrahydride which are not solution~ are expected to react relatively slowly with the impurities o the silane gas.
Most notably, the invention is advantageous in that the silane gas to be purified and/or analy7ed, a commodity which is difficult and expensive to obtain, passes through the procedure and method of the present invention without detrimental effect to the silane itself.
We have found that silane ga~, even when evolved from alkali metal aluminum tetrahydride solutions during silane production is of insufficient purity for the manufactllre of silicon having an end use in semi-conductors or photovoltaics. We have also learned that 20 storage of silane in containers results in contamination with phosphine and arsine. Thus, the present invention i~ directed to a process for purifying as-produced silane or otherwise contaminated silane in preparation for its reduction to high purity silicon.
We have found that intimately contacting a liquid mixture of alkali metal aluminum tetrahydride with ~2~L3 contaminated ~ilane is effective for the removal of AsH3 and PH3, without reacting with the ~ilane gas containing these impurities. Furthermore, we have also devised a method for quantitatively determining the 5 amount of arsine and phosphine in a silane gas sample.
By intimately contacting we mean such contact be-tween the alXali metal aluminum tetrahydride liquid mixture and the ga~eous silarle as will enhance the factors of bubble submersion time, bubble size, mass 10 transport to bubble-liquia inter~ace, and the like.
Intimate contacting may be carried out by high shear agitation means including rapid ~tirring. Stirring is preferably carried out at a rate equivalent to at least 200 RPM's in laboratory glassware equipment. More 15 preferably, stirring i5 conducted in heavier equipment at as high a~ lO00 RPM's or higher where po~sible.
Superatmospheric pressure also tend~ to improve the efficiency of the inventive method. Pressures of only two atmosphere~ increa~e the rate of PH3/AsH3 20 removal during contact of silane and the tetrahydride liquid mixture7 Even higher pressure~ should further enhance the removal and are preferred.
Of course both agitation and pre3~ure have upper limits of practical operation and dimini~hing returns at 25 the maximum PH3/AsH3 removal rate from silane depending upon the particular equipment, alkali metal aluminum tetrahydride, and liquid used.

~L~2~3 A very effective method of intimate contact i3 bubbling silane through a tall column of the liquid mixture or more preferably directing a ~tream of silane through a countercurrent flow of the liquid mix~ur~ as in 5 a column or conventional scrubber. Thi5 may be arranged to receive silane gas as it i~ produced, from a reactor.
The purification process of the invention is carried out in a liquid mixture consisting e~sentially of an innocuous reaction medium and an alkali metal aluminum 10 tetrahydride. By this we mean that the liquid mixture, while it may contain some small portions of impurities or the like, doe3 not contain a significant portion of ~ilane-producing reactant such a~ is present during the manufacture of silane from an alkali metal aluminum 15 tetrahydride and a ~ilicon tetrahalide (e.g., SiC14 and LiA1~4). Rather, the purification proces~ of the inven~ion i~ carried out in an environment that enhances contact of "contaminated" silane with a purifying liquid mixture that i~ primarily an innocuou~ reaction medium, 20 preferably a solvent, and alkali metal aluminum tetrahydride.
For ~;.lane purification, preferably, the liquid mixture~ and solutions of the invention are usually sufficiently free of other contaminant~ and sources of 25 pho~phorus or arsenic. For analysis the alkali metal aluminum tetrahydride is preferably purified by known means prior to use in order to reduce the analytical blank.
We have found that an agitated solution of sodium aluminum tetrahydride, for example, i~ u~ually effective 5 to remove about 90% of the phosphine and greater than 9O~
of the arsine present in a silane gas. These approximate numbers of 90% and greater than 90% seem to hold true whether the contamination of phosphine and arsine in the silane gas is of a relatively large amount like lO0 parts 10 per million or even if it i~ of a relatively miniscule amount of about one part per billion.
The ~ilane solution to be purified or analyzed according to the present invention may be contacted with a liquid mixture of alkali metal aluminum tetrahydride 15 according to the inven~ion by various means. One mean~
would be a countercurrent flow of silane gas through a liquid mixture of, for example, sodium aluminum tetra hydride in monoglyme. Various columnar and bedlike device~ may be u~ed for contacting the silane gas with 20 the liquid mixture according to the invention.
The important aspect of the analytical method of the present invention is that minute portions o pho~phine and arsine pre~ent in a silane stream may be gathered into a liquid mixture by an alkali' metal 25 aluminum tetrahydride and then re-evolved as phosphine or arsine or the like as a gas in a concentrated form thereby making quantitative analysis more practical.
This makes available for determination of ~hese trace impurities the various quantitative analysis techniques normally used where only large amounts of material are 5 found. For example, about 2,000 liters of a silane gas contaminated with trace impurities of phosphine and arsine at about the O.l parts per billion level may be contacted with a relatively small solution of sodium aluminum tetrahydride in order to gather about 9O~ of the 10 contaminant phosphine and arsine into the relatively small volume of hydride in ether. Although the level of contamination i5 small by normal analytical standards ~although large by high purity silicon stand~rds), the concentration of phosphine and arsine may be brought 15 above minimum detection limits for the equipment in question. This then enables an analyst to use a photo-ionization detector with a lO ppb threshhold or a molybdenum blue orthophosphate test.
According to the analytical method of the inven-20 tion the phosphine and arsine which have been reacted and gathered into the solution or liquid mi~ture of the invention may be quantitatively determined by various means. Pref~rably, the phosphorus and arsenic content in the liquia are first separated before the quantity is 25 determined. More prefe.ably, the phosphoru~ or arsenic values are liberated as PH3 or AsH3. Still more ~P~ 3 preferably, the PH3 and ASH3 are evolved into a closed ipace, such as the headspace above the liquid~ An expandable container is suitable for this purpose. Most preferably the ASH3 and PH3 are re evolved by 5 hydrolysisO Various agents may be used to re-evolve phosphine and arsine. Among these are water and methanol.
A small portion of water or other hydrolysis agent may be injected into the liquid mixture containing alkali metal aluminum tetrahydride in such a manner a~ to re-evolve 10 substantially all of the phosphine or arsine in the liquid mixture. The hydrolysis may be carried out in a closed expandable bag such that the phosp~ine and arsine are re-evolved with hydrogen gas from hydrolysis of the alkali metal aluminum tetrahydride into the head space 15 above the liquid mixture. Thereafter, the gas may be readily analyzed by such techniques as gas chromatography ~using a photoionization detector) or by atomic ab~orp-tion or other direct analysis techniques which have lower limits of detection not u~ually suitable for trace 20 impurities in silane.
Other suitable agents to hydrolyze the tetra-hydride ~nd the phosphoru3/arsenic value~ are any comp~unds with a source of active hydrogen. The~e include acids such as propionic, nitric, s~lfuric, acetic, and 25 HCl; primary and secondary amines such as n-propylamine, n-butylamine, diethylamine, mixed dialkylamines and l3 others alcohols such as methanol, ethanol, isopropanol, tert-butanol, and sec-butanol as well as phenols and alkylated phenols. Where the liquid mixture is a monoglyme or similar solution of alkali metal aluminum 5 tetrahydride, it is preferred to pump in a ten percent solution of water in diglyme from a syringe or the like.
Alternatively, the hydrogen gas containing phosphine and arsine in the head space above the liquid or liquid-solid mixture remaining after hydrolysis of the 10 alkali metal aluminum tetrahydride may be used in other analytical techniques. For example, the phosphine gas may be swept to an oxidizing solution such as a hypo-chlorite, preferably a sodium hypochlorite solution and then acidified and boiled to form orthophosphate (PO4 ).
15 In ~his manner, the phosphorus content may then be determined by colorimetric phosphate solution pro-cedures known to those skilled in the analytical arts.
Such a procedure i commonly referred to a~ a standard orthophosphate procedure or molybdenum blue orthophosphate 20 procedure. Notably, th~ sodium hypochlorite solution may not be used directly because silane reacts explosively therewith.
The amount of removal of phosphine and arsine gas along with the hydrogen gas evolved upon hydrolysis is dependent upon the technique used for contacting water or other hydrolysis agent with the alkali metal aluminum tetrahydride liquid mixture. A ga~ washing bottle with a side septum for injection of hydrolysis agent may be used in a manner where the gas liquid contact is high so as to 5 evolve all of the contained phosphorus and ar~enic values a5 PH3/A5H3-The major advantage of the pre~ent invention is that it is now possible to concentrate the relatively minute quantities of phosphine and arsine present in a 10 silane gas into a much smaller volume wherein the impur-ities may be analyzed with increased ~en itivity. A
better understanaing of the invention will be had by a review of the examples of the invention as given below.
We observed that the removal of phosphine from the 15 silane or mixed gas solution~ is more effectiv2 with high rates of agitation a~ shown in the table~ below. Variou~
methods o~ agitation may be used within the scope of the invention ~o long as the phosphine and arsine are reacted to be taken up in the scrubber solution and then option-20 ally re-evolYed for testing to determine purity of the silane stream so treated or to determine the purity of the incoming silane stream. According to the invention one may test a large volume of a ~ilane stream to determine the amount of phosphine or ar~ine contamination 25 to determine whether additional purification i~ required or whether the sil~ne stream may then be u~ed to form silicon in a high purity silicon manufacturing process w~erein the silicon is ultimately used for semiconductor or photovoltaic applicationsO One oE the mo~t advan-tageous features of the present invention is that the liquid mixtures for scrubbing a silane gas flow according 5 to the present invention do not react with silane and thus do not affect the ultimate u~e of the ~ilane, usually in the production of semiconductor or photo-voltaic grade silicon.
In a large scale purification operation, we 10 estimate that about a 15~ solution of sodium aluminum tetrahydride in monoglyme is the best form of purifica-tion according to the process of the present invention as now Xnown to U8u In the Examples below, a gas flow containing ASH3 15 and PH3 wa~ first sampled and analyzed. The levels of AsH3 and PH3 were set high enough for conventional detection. The gas flow was then passed through a liquid mixture and again immediately sampled and analyzed. The Examples were carried out to show the removal of PH3 or 20 AsH3 from silaneu During the course o~ the experiments we found that the rate~ of removal of PH3 and AsH3 from nitrogen/silane mixture~ and nitrogen gas alone are about the same as for silane alone. Of course, silane gas in practice requires handling and use of reagents 25 that will not react therewith as compared to inert nitrogen.

_X~MP~E I
A 150 milliliter scrubber solution of Na~lH4 in monoglyme was charged to a 250 milliliter round bottom flask. The flask was equipped with a septum cover on an 5 opening near the top, a magnetic stirring bar operated at ahout 200 rpm. and a second opening near the top fitted with a dip leg in the stopcock for entry of the silane flow so as to be dispensed near the bottom of the liquid level in the round bottom flask in the area of high 10 agitation. The dip leg tube was fitted with a coarse porosity dispersion frit. The neck of the round bottom flasX contained a cooling water tube for the flow of cooling water in and out of the neck area to cool the effluent silane. The necX of the flask was fitted with a 15 two-way valve apparatus so that the effluent silane bubbled through the 150 milliliter scrubber solution could be sent to a burner a~ desired or to a 250 milliliter mass spectroscopy sample bottle. The entire flask and apparatu~ wa~ kept in a water bath on a stirring hot 20 plate. A small magnetic stirring bar was operated at 200 RPM's. Phosphine and arsine samples were available in nitrogen gas. The supply of phosphine gas contained about ppm PH3 in nitrogen. The supply of arsine gas con-tained about 8 ppm AsH3 in nitrogen. The nitrogen gas 25 supply was es3entially free of contaminants and the supply of ~ilane was Xnown to contain fewer than 10 parts per billion phosphine and arsine. A flow o gase~ wa~
~tarted with about 1.1 standara cubic feet per hour of the arsine source gas, 0.4 standard cubic feet per hour of the phosphine ~ource gas, 1.5 ~tandard cubic feet per 5 hour of the nitrogen ga~, and about 1.5 ~tandard cubic feet per hour of the silane gas. ~fter the equipment had been purged with nitrogen gas alone, the flow of gase~ a~
described above was directed to a 250 milliliter mas~
spectroscopy ~ample bottle. When the bottle was filled, 10 the sample was immediately analyzed with a photoioniza-tion detector and the number of integrater counts ~or the sample was noted. While the first sample was being analyzed, the flow of gases were directea through the ~parging tube and through the coar~e porosity dispersion 15 frit into the stirred solution in the round bottom flask.
The ~ilane gas wa~ evolved and collected in a second 250 milliliter mass spectro copy sample bottle and again immediately tested. The number of integrater counts in the sample which had pas~ed through the stirred solution 20 of sodium aluminum tetrahydride in monoglyme was recorded and thi~ value in conjunction with the value from the fir~t sample wa~ used to calculate the percent removal of phosphine and arsine from the gas flow~ The ~ample~ from each of the bottles were fir~t run through gas chroma-tography equipment so as to identify the phosphine andarsine which elute separately. ~otably, the arsine V~a3 removal could only be calculated as a "greater than"
percent figure since a small portion of some other compound was eluting at abou~ the same place on the G~Co For ea~e of calculation, the en~ire area of the peak for 5 arsine in the second sample was used as entirely arsine.
l'hus, this was the most ar~ine which could remain in the second sample.
Subsequent testing of ~amples which had set out for even ~hort periods of time evidenced the instability 10 of the phosphine and arsine samples by a rapid decrease in the total integrater counts obtained on such ~ubse-quent sample injections from the filled containers. This is a wall effect with the containers ~ince a qample tends to reach an equilibrium concentration of such impurities 15 in certain containers.
For this example and for ~xample II, the percent removal of phosphine or arsine were ea~ily calculated by atomic ab~orption and gas chromatography with a photo-ionization detector, according to the following formula.
20 ~ Removal =

lOOX (Int. Counts, Sample 1 - Int~ Counts, Sa~ple 2) Int. Count~, Sample 1 The results of the above experiments using various solutions at variou~ operating temperatures are given in 25 Table I below.

TABLE I

Tests Conducted in Glass Reactor Scrubber Test Percent Removal For Solution Tem~., C. PH3 AsH3 0.53%1 24 go >9 NaAlH4 44 88 >90~
in THF 66 ~Reflux) 86 >902 10.4~ NaAlH4 24 58 ~99 in THF 50 44 >902 10 THF alone 24 None 50 10.6% NaAlH4 24 60 >952 in 50 60 ~952 Dimethoxyethane 75 65 ~952 8.1% 2~ 34 >902 15 NaAlH~ ~0 in Diglyme 15 48 > 902 1 All NaAlH4 value~ given are calculated from ga~
evolution analy~e~.

2 Since other components of small quantity were elut~ng at about the same peak a~ ar3ine, only a "greater than" percent removal can be given.

EXAMPLE II
The ~ame general procedure wa3 followed as in Example I except that ~ince we had learned that the 25 removal of phosphine and arsine is about the same according to t~i~ invention from silane as it is from nitrogen or mixtures of nitrogen and silane, some of the experiment~ in Example II were run using nitrogen only to demon~rate the workability of the invention. The same sources of arsine and phosphine in nitrogen solutions were supplied to a flow of either nitrogen or silane/
nitrogen mixture a~ for Example I. In this Example, however, the gas flow was provided to a one liter baffled 5 autoclave with a three bladed propeller stirrer. The autoclave was fitted with a dip tube feed of the gas supply below the agitator. The high pressure autoclave was used not for carrying out a high pressure reaction but for the purpose of conducting the invention at high 10 agitation without leakage. The effluent from the auto-clave reactor passed through a condenser and into a valve like that usea in example one so that the flow could be switched from a route to a burner for the silane to a route to be collected in a 500 milliliter sample bomb ~or 15 the hydrogen, pho~phine, and arsine mixture. In both this example and Example I, samples of the gas flow to the reactors were collected in a ~eparate bomb for comparison. This enabled us to calculate the percent removal based upon standard analytical equipment. A
20 sample 1 of the gas flow was first taken. After the gas flow had continued through a 526 gram sample of di-methoxyethane containing 8% by wPight ~odium aluminum tetrahydride, sample 2 was taken. ~gain the percent removal was calculated by the formula given in Example I
25 and the results are tabulated in Table II below.

~ 19 --Tests_Conducted in 1 Liter Autoclave TABLE II

Temp.Pressure Carrier ~ Removal For RPM C psig~ ~ PH3 AgH3 _ 1000 26 0 ~2 62 1400 22 0 N2 98> 99 1400 21 20 N2 99~99 1400 47 . ~2 96>9g 101400 26 o N2 94 450 20 0 ~/SiH4 68 1000 23 0 N2/SiH~ 94 1400 24 0 ~2/SiH4 95 , 1 Source of AsH3 exhausted - none used.
15 We noted that a slight pre~sure increased the removal of phosphine from a nitrogen ~olution and as a general rule we found that increased pres~ure increased the percent removal of phosphine from a ga~. We found that arsine i5 generally much more easily and completely removed than phosphine and we generally directed our eEforts to removal of phosphine at a high level which always resulted in the very efficient re~oval of ar~ine as well.
We also noticed that the u~e of the invention was ~5 effective for removal of a portion of the sulphur com-pounds including S02 which were found in qome sample~

of silane gas. Similarly, we found that fluoride compounds are fairly well removed by use of the invention.
Also diborane is removed. In particular, the contaminant sodium aluminum tetrafluoride present when the silane is S produced from 3ilicon tetrafluoride and sodium aluminum tetrahydride or the like, is fairly readily removed by using a countercurrent flow of reactant alkali metal aluminum tetrahydride in ether to the effluent silane from a silane reactorO Separate downstream scrubbers are 10 also effective.
In other experiments, phosphorus at the level o 100 parts per million was reduced to less than 20 parts per billion in a nitrogen stream and in a silane stream by passing each stream through an agitated solution of 15 -two percent purified NaAlH4 in diglyme at 90 C. The lower limit of PH3 detection i~ 20 ppb for the vapor pha~e chromatography eq1~ipment used. These e~periments indicate that higher temperatures may enhance PH3 removal.
In stlll another experiment, about 1 ml. of solution from the autoclave, after 15 minutes of contaminated gas flow (the gla~sware flow was operated for about ten minutes~, wa~ added to 149 mls. of a control solution of purified monoglyme containing 1.1 X
25 10 6 grams phosphorus and two percent purified NaAlH4O
The pho~phine and ars;ne gases were re-evolved from the solution along with H2 by slowly injecting (pumping from a ~yringe) a solution of about 6 mls~ distilled H20 in 44 mls. pure diglyme. ~he evolved (and re evolved) gases were swept to a pure solution of sodium 5 hypochlorite and colorimetric analysis revealed about 25 X 10 6 grams phosphorus--an increase of about 24 micrograms from one liter of the purifying solution.
It can readily be seen that internal standards can be established and the volume of silane analyzed or 10 purified along with the amount of liquid mixture can be recorded to analyze an unknown contaminated silane or to purify a silane of Xnown contamination level.
Additional experiments with other hydrolyzing agent~ such as methanol and other alco~ols confirm that 15 these agents may also be used according to the invention to hydrolyze the alkali metal aluminum tetrahydride and re-evolve the phosphine and arsine gases from the scrubber solution.

Claims (38)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A process for the removal of phosphine or arsine from previously synthesized silane, said process comprising intimately contacting silane gas containing phosphine or arsine with a liquid mixture consisting essentially of an innocuous reaction medium and an alkali metal aluminum tetrahydride and recovering the purified silane gas.

2. The process of Claim 1 wherein said innocuous reac-tion medium is a solvent for the alkali metal aluminum tetrahy-dride.

3. The process of Claim 2 wherein said solvent is a tertiary amine.

4. The process of Claim 2 wherein said solvent is an ether.

5. The process of Claim 4 wherein said ether is selected from the group consisting of tetrahydrofuran, diglyme, and monoglyme.

6. The process of Claim 4 wherein said ether is mono-glyme.

7. The process of Claim 3 wherein said tertiary amine is triethylamine.

8. The process of Claim 1 wherein intimately contacting comprises high shear agitation.

9. The process of Claim 8 wherein high shear agitation comprises rapid stirring at least about 200 RPM, so as to enhance the factors of bubble submersion time, bubble sire, and mass transport to bubble-liquid interphase.

10. The process of Claim 1 wherein intimately contacting comprises countercurrent flow of the silane with the liquid mixture.

11. The process of Claim 2 wherein a solution of alkali metal aluminum tetrahydride is flowed countercurrent to a silane gas containing arsine and phosphine.

12. The process of Claim 1 wherein said alkali metal aluminum tetrahydride is sodium aluminum tetrahydride.

13. The process of Claim 12 wherein said liquid mixture is an ether solution.

14. The process of Claim 13 wherein said ether is di-methoxyethane (monoglyme).

15. The process of Claim 1 carried out at superatmos-pheric pressure.

16. The process of Claim 1 carried out at a pressure of at least about two atmospheres.

17. A process for the analysis of phosphorus or arsenic content in previously synthesised silane, said process comprising intimately contacting silane gas containing phosphine or arsine with a liquid mixture containing alkali metal aluminum tetra-hydride and then determining the amount of phosphorus or arsenic in the liquid mixture.

18. The process of Claim 17 wherein said alkali metal aluminum tetrahydride is sodium aluminum tetrahydride.

19. The process of Claim 17 wherein said liquid mixture is an ether solution.

20. The process of Claim 17 wherein said liquid mixture is an ether solution containing alkali metal aluminum tetrahydride and said ether is selected from the group consisting of tetra-hydrofuran, dimethoxyethane (monoglyme), and the dimethylether of diethylene glycol (diglyme).

21. The process of Claim 20 wherein said ether is the dimethylether of diethylene glycol.

22. The process of Claim 21 wherein said alkali metal aluminum tetrahydride is sodium aluminum tetrahydride.

23. The process of Claim 19 wherein said determining the amount of phosphorus or arsenic in the liquid mixture after contacting the silane gas with the ether solution comprises hydrolyzing said solution to re-evolve phosphine or arsine along with hydrogen from the alkali metal aluminum tetrahydride.

24. The process of Claim 23 wherein the re-evolved phosphine or arsine are analyzed by chromatography, atomic absorption, or colorimetric procedures.

25. The process of Claim 23 wherein the re-evolved phosphine is reacted with a hypochlorite to form an orthophosphate for colorimetric analysis.

26. The process of Claim 24 wherein said hydrolyzing is caused by the addition of water or alcohol.

27. The process of Claim 17 wherein said intimately contacting comprises high shear agitation.

28. The process of Claim 27 wherein said high shear agitation comprises rapid stirring, at least about 200 RPM, so as to enhance the factors of bubble submersion time, bubble size, and mass transport to bubble-liquid interphase.

29. The process of Claim 17 wherein said intimately contacting comprises countercurrent flow of the silane with the liquid mixture.

30. An analytical method for measuring the phosphine and arsine content of a silane sample, said process comprising the steps of intimately contacting the sample with A solution of alkali metal aluminum tetrahydride to react the phosphine and arsine therewith so as to remove at least about 90% of the phosphine and arsine from the silane; hydrolyzing the solution to evolve hydrogen gas from the alkali metal aluminum tetrahydride and re-evolve phosphine and arsine gases in a closed space; and determining the amount of arsine and phosphine concentrated in the hydrogen gas.

31. The method of Claim 30 wherein said alkali metal aluminum tetrahydride is sodium aluminum tetrahydride.

32. The method of Claim 30 wherein said solution comprises an ether solution or a tertiary amine solution.

33. The method of Claim 30 wherein said solution comprises an ether solution.

34. The method of Claim 33 wherein the ether of said ether solution is selected from the group consisting of tetra-hydrofuran, dimethoxyethane (monoglyme) and the dimethylether of diethylene glycol (diglyme).

35. The method of Claim 34 wherein said ether is di-methylether of diethylene glycol.

36. The process of Claim 30 wherein the amount of arsine and phosphine concentrated in said hydrogen gas is determined by gas chromatography atomic absorption, or colorimeter procedures.

37. The method of Claim 30 wherein said solution is hydrolyzed by water, an alcohol, an acid, or a primary or secondary amine to evolve PH3 or AsH3.

38. The method of Claim 30 wherein said solution is hydrolyzed by water.