PREPARATION AND PURIFICATION OF METAL ... - Science Direct

17 downloads 0 Views 375KB Size Report
and D.J. COLE-HAMILTON * ..... [7] DV. Shenai-Khatkhate, J.B. Mullin and D.J. Cole-Ham- ... [15] K.J. Irgolic, The Organic Chemistry ofTellurium (Gordon.
Journal of Crystal Growth 77 (1986) 27—31 North-Holland, Amsterdam

27

PREPARATION AND PURIFICATION OF METAL ALKYLS FOR USE IN THE MOCVD GROWTH OF Il/VI COMPOUND SEMICONDUCTORS D.V. SHENAI-KHATKHATE, E.D. ORRELL, J.B. MULLIN and D.J. COLE-HAMILTON *

~‘,

D.C. CUPERTINO

Chemistry Department, University of St. .4ndrews, St. Andrews Fife KYJ6 9ST, Scotland, UK

The preparation of adducts of Me

2Cd and Me2 Zn with a range of nitrogen donor ligands has been investigated. Those with two nitrogen atoms, which cannot for geometrical reasons bind to the same metal atom, are shown to give involatile polymeric adducts. which dissociate on heating at one atmosphere or in vacuo below the sublimation temperature of the adduct or the Lewis Base. These adducts may be suitable for the purification of Me2Cd or Me2 Zn. Those that dissociate Me2 Zn or Me2Cd below the melting point of the adduct may be suitable precursors for the delivery of these alkyls by the MEM method. A new preparation of Et2Te based on the two phase reactions of aqueous Na2Te with EtBr is also reported. This gives high yields of Et,Te which is polytellurium-free.

1. Infroduction In recent years, the growth of epitaxial layers of compound semiconductors by the technique of metalorganic chemical vapour deposition (MOCVD) has been extensively studied and is moving into the production phase for, e.g. gallium arsenide [1]. Extensive studies have shown that the purity of the semiconducting layer is often determined by the purity of the precursors (metal alkyls or hydrides). The problem of purity has been addressed by a number of academic and industrial laboratories and methods developed for removal of metal containing impurities to levels below 1 ppm. We [2] and others [3] have developed purification techniques for group III alkyls based on the preparation of invOlatile adducts with Lewis bases, which dissociate on heating either at one atmosphere or in vacuo. Volatile impurities can then be removed by evacuation at temperatures below the dissociation temperature of the adduct, whilst involatile impurities remain when the adduct is dissociated. Sometimes the adducting Lewis base can also be the solvent for the reaction [2] whilst in other cases it forms an adduct which can be *

Royal Signals and Radar Establishment, St. Andrews Road, Great Malvern, Worcs, WRI4 3PS. UK.

purified by crystallisation [3]. Using these techniques, both Me3Ga [2,3] and Me3 In [3] of electronic grade can be prepared routinely, although for Me3 In, there is some indication that the preparative method employed may be of considerable importance [4]. We have now turned our attention to the preparation of adducts which may be suitable for the purification of alkyls which are precursors for the MOCVD of Il/VI semiconductors such as ZnS, ZnSe, CdTe, Cd~Hg1 STe, etc. We now report preliminary results of studies on the preparation and/or purification of alkyls of cadmium, zinc, and tellurium. Full details have been [5], or will be, reported elsewhere and patents have been filed [6,7]. 2. Results and discussions 2.1. Dimethylcadmium Dimethylcadmium can readily be prepared from CdC12 and MeMgI in diethyl ether [8], and the use of 2,2’-bipyridyl to form a 1: 1 adduct from which Me2Cd can be released on heating has been known for some time [8]. Unfortunately, the adduct itself is slightly volatile [9]at the temperature required for dissociation of dimethyl cadmium so

0022-0248/86/$03.50 © Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)

28

D. V. Shenai-Khatkhate ci al.

/

Preparation and purification of metal a/ky/s

Cd /N

M~N©N~~2~N ©N~1 N-~ N

N

7—N N

N~ N

Cd Cd Me2 Me~ Fig. 1. Proposed polymeric Structure for the adduct (Me2Cd L = 4.4’-bipyridyl. 3.3’-bipyridyl or 4-dimethylaminopyridine.

that contamination by nitrogen may be a problem. Our studies, some of which have already been reported [5], have shown that adducts with a number of nitrogen containing Lewis bases can be prepared, particularly if the Lewis base (L) contains two nitrogen atoms. If the nitrogen atoms are dispersed in such a way that an unstrained chelate ring can be formed, e.g. with L2-bis(dimethylamino)ethane, the adducts sublime without dissociation [5]; whereas, if the two nitrogen atoms cannot, for geometrical reasons, coordinate to the same metal atom, they bridge metal atoms and involatile, dissociable, adducts are formed. In most cases, e.g.. L 4,4’-bipyridyl [5], 3,3’-bipyridyl or 4-dimethylaminopyridine, the adducts have the stoichiometry (Me2Cd. L) and are believed to be polymeric *, with a structure similar to that shown in fig. 1. For L 1,4-bis(dimethylamino)benzene (TMPD) [5], on the other hand, an adduct of stoichiometry (Me7Cd)2L3 is formed, which may have the structure shown in fig. 2, or may possibly he polymeric (as structure 1) but with 0.5 moles of Lewis base of crystallisation. All of these adducts dissociate dimethylcadmium on heating in vacuo (see table 1) so may be suitable for the purification of dimethylcadmium. For L TMPD and 3,3’-bipyridyl. the adduct melts before or at the same temperature as dissociation occurs. For L 4,4’-bipyridyl or 4-dimethylaminopyridine on the other hand, dissociation occurs at temperatures below the melting point of the adduct. These last two adducts may, therefore, be suitable for delivery of dimethylcadmium using the modified entrainment method (MEM) [10]. The adducts are also considerably safer to handle than the parent alkyl, being only mildly air sensitive and not pyrophoric. =

=

=

=

*

)N ©NM~

The high dissociation entropy (153 J K’ mol~ ) is consistent with this interpretation,

Fig. 2. Proposed structure for the adduct of stoichiometry (Me2Cd)2 L3 L = I ,4-bis(dimethylamino)benzene.

The free Lewis bases as well as their adducts with metal alkyls are involatile at the corresponding dissociation temperatures of the adducts hence nitrogen contamination should not be a problem. 2.2. Dimethylzinc Dimethylzinc has normally been prepared by the reaction of a mixture of iodomethane and bromomethane with a zinc/copper couple [11]. Although this apparently gives relatively pure dimethyizinc, there is a worry that copper contamination may occur. Following our success [21 in developing routes to the preparation of ultrapure trimethylgallium using high boiling ethereal solvents, we have investigated the possibility of synthesising dimethylzinc in a similar manner from the reaction of ZnCl2 with MeMgI in high boiling ethers. The reaction proceeds smoothly and high yields of Me2 Zn can be achieved after distillation directly from the reaction pot at 180°C. Although dimethylzinc prepared in this way is relatively pure, the chemical similarities between zinc and cadmium have led us to investigate adducts of dimethylzinc with the same nitrogen donor ligands as we have employed for the purification of dimethylcadmium. In each case, crystalline adducts of stoichiometry [Me2Zn~LI, which we again assume to be polymeric, can be isolated (see table 1). Preliminary results suggest that these adducts dissociate on heating, but more detailed studies are in progress. 2.3. Diethyltellurium

The normal laboratory synthesis of Et2Te involves reaction of sodium with tellurium metal in liquid ammonia followed by reaction of the Na5Te so formed with, e.g., bromoethane [12]. We find that this reaction is satisfactory. although care must he taken to avoid the formation of po/ytel-

D. V Shenai-Khatkhate et 0/.

/

29

Preparation and purification of metal a/ky/s

Table I Adducts of dimethylcadmium and dimethylzinc with nitrogen donor Lewis bases Metal alkyl (M)

Lewis base (L)

solvent(s) employed

Dimethylcadmium

I,2-bis(dimethylamino)ethane 4-dimethylaminopyridine 2,2’-bipyridyl 3,3’-bipyridyl 4,4’-bipyridyl I,4-bis(dimethylamino)benzene 3,3’-bipyridyl 4,4’-bipyridyl 4-dimethylaminopyridine

Dimethylzinc



b) °

d)

[L]:[M] ratio in adduct h)

Class of adduct ~

1 1 1 2 3 1

1:1 1:1 1:1 1: 1 1:1 3:2

(i) (i) (ii) (ii) (ii) (ii)

2 2 I

1: I 1:1 I: I

(ii) (ii) (ii)

a)

Thermal behaviour d) (°C) m s d 62 70 75



138 35

70

— —



80

60 67 80 83 33

90 85—90 78

I = Adduct was prepared in diethylether and was purified by crystallisation at low temperatures from its diethylether solution. 2 = Adduct was prepared in diethylether from which it precipitated and was purified by recrystallisation at low temperatures from its tetrahydrofuran solution. 3 = Adduct was prepared in and purified from tetrahydrofuran solution. As suggested by microanalyses and H NMR studies on pure adducts. (i) Sublimable adduct formed. (ii) Adduct dissociates to give Me 2Cd or Me2 Zn at s 100°Cin vacuo. From differential scanning calorimetric studies; m = melts. s = sublimes ( 10—2 Torr). d = dissociates.

/uriums (Et2Te~,n ~ 2)

[13,14]. These polytelluriums break down on heating to give Et 2Te and tellurium metal but yields of Et 2Te are consequently lowered, For large scale synthesis of Et2Te in an industrial environment, the safe handling of large amounts of liquid can cause other severeroutes technical problems so ammonia we have developed which concentrate on safe straightforward procedures using readily available starting materials and which give as pure a product as possible. We find that this is possible using known methods for the production of Na 2Te in aqueous solution [15,16], followed by direct reaction with bromoethane. Similar approaches have been used in the past for the preparation of dialkylditellurium and of dimethyltellurium, but dialkyl sulphates or cosolvents [16,17] are generally required. Since these are not readily separated from Et2Te thisfor method has not previously been thought[15,17], suitable the preparation of Et 2Te. The Et2Te formed by our method is very pale yellow and can readily be extracted and purified. It is obtained in We highareyield and isassessing essentially polytellurium-free. currently the purity of the material and investigating possible chemical methods of purification,

3. Experimental All solvents were carefully dried by distillation from sodium diphenylketyl and were degassed prior to use. Reactions were carried out in an atmosphere of “white spot” nitrogen purified by 2~on silica. passing over a column consisting of Cr All greaseless joints and taps were employed and manipulations of the only moderately air-sensitive Lewis base adducts were carried out using standard Schlenk line and catheter tubing techniques. The purification, transfer and handling of base-free metal alkyls were carried out using high vacuum line techniques. All Lewis bases were reagent grade and were purified by crystallisation from an ethereal solution, or sublimation in vacuo. Microanalyses (C, H and N) were recorded on a Carlo-Erba 1106 hydrogen and were nitrogen t13Cd andcarbon, 125Te NMR spectra reanalyser. corded on a Bruker Associates WM250 nuclear magnetic resonance spectrometer operating in the Fourier transform mode with noise decou1H NMR spectra on a proton Perkin-Elmer pling spectrometer and R12B and a Bruker Associates WP-80 nuclear magnetic resonance spectrometer. Mass spectra were recorded on the Associated Electrical

30

D. V. Shenai-Khatkhate et a!.

/

Preparation and purification of metal a/ky/s

Industries M5902 double focussing mass spectrometer. Thermograms were measured on a Perkin-Elmer DSC 2 differential scanning calorimeter with a heating rate of 10°Cmin Melting points ~.

were determined on an electrothermal melting point apparatus in closed capillaries under argon and are uncorrected, 3.1. Preparation of dimethylcadmium and dimethylzinc

Me2Cd was prepared in diethylether by the reaction of MeMgI and CdCl2 and was purified from its adduct with 2,2’-bipyridyl by standard literature methods [81. Me2 Zn was prepared by the reaction of MeMgI, prepared by reacting magnesium g) 3) in highturnings boiling (32 ether, with cm g). Pure Me and iodomethane zinc chloride (78 (71.44 2Zn was obtained by the distillation of reaction mixture at 180°C(yield 85%).

3.4. Preparation of dialkylte/luriums

Bromoethane (52 cm3) was added to a solution of Na 2Te (55 g) in alkaline water. An exothermic reaction was found to occur and a pale yellow oil was obtained. The reaction mixture was heated under reflux on a water bath maintained at a steady temperature of 60°C for 5 h. The yellow oily layer was isolated from the aqueous layer by extraction with diethyl ether. The ethereal layer was dried over anhydrous calcium chloride and after filtration, diethylether and excess ethylbromide were removed in vacuo. The diethyltellurium was distilled at I atm. The resulting yellow oil was purified by vacuum distillation. Overall yield 60%. Diisopropyltellurium (b.pt. using 157—158°C, 52% yield) was similarly prepared 2-bromopro3). pane (65.7 cm =

=

Acknowledgements 3.2. Preparation of Lewis base adducts of dimethylcadmium and dimethyizinc

The adducts of various Lewis bases with the metal organics were prepared by addition of excess pure metal alkyl to the solution of free Lewis base in vacuo. The products were obtained by concentration of the solution and cooling. After filtration they were recrystallised from the solvents noted in table 1 (footnote a). 3.3. Preparation of ultrapure dimethylcadmium and dimethy/zinc from their adducts

The adducts of dimethylcadmium and dimethylzinc were heated in vacuo at the dissociation temperature of the adducts (see table 1), to yield pure metal alkyls, which were collected in a cold trap maintained at a liquid nitrogen temperature. Nuclear magnetic resonance spectroscopic studies on metal alkyls obtained in such a way showed the absence of either the free Lewis base or the solvent employed for the synthesis of metal alkyls (viz. diethylether).

We wish to thank The Ministry of Defence (UK) for financing this research.

References [1] See, e.g., Proc. 2nd Intern. Conf. On Metalorganic Vapour Phase Epitaxy, Sheffield, 1984 [J. Crystal Growth 68 (1984) 1—502]. [2] AC. Jones, A.K. Holliday, D.J. Cole-Hamilton, MM. Ahmad and N.D. Gerrard, J. Crystal Growth 68 (1984) 1. [3] D.C. Bradley, M.M. Faktor and S.J. Watts, Intern. Sym~ on Uses of Metal Organic Compounds, Oxford, April 1985. [4] AC. Jones, Epichem Ltd., private communication, 1986. [5] P.R. Jacobs, ED. Orrell, DV. Shenai-Khatkhate, J.B. Mullin and D.J. Cole-Hamilton, Chemtronics 1(1986)15. [6] P.R. Jacobs, J.B. Mullin, DV. Shenai-Khatkhate, ED. Orrell and D.J. Cole-Hamilton, UK Patent AppI. 1985, GB 8509055. [7] DV. Shenai-Khatkhate, J.B. Mullin and D.J. Cole-Hamilton, to be published. [8] G.E. Coates and S.I.E. Green, J. Chem. Soc. (1962) 3340. [9] K.H. Thiele, Z. Anorg. AlIg. Chem. 330 (1964) 8. [10] D. Battat, MM. Faktor, I. Garrett and RH. Moss, J. Chem. Soc., Faraday Trans. 1(1974) 2267. [11] CR. Noller, Organic Syntheses, CoIl. Vol. Il (Wiley, New York, 1943) 184.

D. V. Shenai-Khatkhate et a/.

/ Preparation

[12] L. Brandsma and HE. Wijers, Rec. Tray. Chim. Pays-Bas. 82 (1963) 68. [13] E. Zintl, J. Goubeau and W. Dullenkopf, Z. Physik. Chem. (Leipzig) A154 (1931) 1. [14] C.A. Kraus and J.A. Ridderhof, J. Am. Chem. Soc. 56 (1934) 79.

and purification of meta/ a/ky/s

31

[15] K.J. Irgolic, The Organic Chemistry of Tellurium (Gordon and Breach, New York, 1974). [16] L. Tschugaeff and W. Chiopin. Chem. Ber. 47 (1914) 1269. [17] F. Woehler, Ann. Chem. 35 (1840) 111.