J. MATER. CHEM., 1994, 4( l), 81 -85 Metal-Organic Chemical Vapour Deposition of YBCO using a New, Stable and Volatile Barium Precursor Sarkis H. Shamlian,” Michael L. Hitchman,*” Stephen L. Cookband Barbara C. Richardsb a Department of Pure and Applied Chemistry, University of Strathclyde 295 Cathedral Street, Glasgow, UK GI IXL The Associated Octel Co. Ltd., PO Box 17, Oil Sites Road, Ellesmere Port, South Wirral, UK L65 4HF A new highly volatile Ba-containing precursor which is thermally stable at I atm (101325 Pa) pressure is described. This precursor, which is a bis(P-diketonate) that has only C3F, as substituent groups and which is coordinated with the polyether tetraglyme, is shown to have reproducible carry-over rates and to give repeatable deposition rates for the chemical vapour deposition (CVD) of highly oriented crystalline BaF,.It has also been shown to be suitable for use in the preparation of superconducting YBCO films. Measurements of thermodynamic and kinetic parameters have been made and are compared with those obtained with other Ba-containing precursors. It is concluded that the new Ba-containing precursor is potentially a very promising material for the preparation of superconducting YBCO films. The deposition of inorganic compounds of barium is import- ant for the preparation of high-T, superconducting films or for coatings of, for example, BaF,. Currently, the best quality high-T, films are prepared by laser ablation, but as the demand for large-area samples, high-deposition throughput, good intra- and inter-sample uniformity, and conformality of layers on patterned structures increases, then CVD has the potential to become a powerful technique for preparing thin films of superconductors.However, although a significant number of papers have been published on the CVD of high-T, layers,’ there are still problems associated with reproducibility of growth and the control of layer stoichiometry. These problems often arise from the inadequacies of the precursors, particu- larly those of barium.2 The most commonly used precursors have been based on the fi-diketonate complexes with barium 1 where R and R‘ were initially alkyl groups, but more recently where fluorinated alkyl groups have been used. The main reasons for using fluorinated alkyl substituents is that they are sterically demanding relative to formula weight and that without fluorination intermolecular hydrogen bonding can occur, leading to oligomerisation and decreased volatility. The high electronegativity of fluorine also produces much lower van der Waals interactions.This strategy has been successful in reducing precursor sublimation temperature. For example, with R =R’ =CH, [Le. Ba(ACAC), where ACAC represents pentane-2,4-dionate, commonly known as acetyl- acetonate] it is not possible to sublime the precursor since it undergoes dissociation and only the volatile organic ligand leaves the evaporator,2 but with R =R’=CF, [i.e. Ba(HFA), where HFA = 1,1,1,5,5,5-hexafluoro pentane-2,4-dionate] sub- limation occurs3 at 240-255°C at 1 atm or at ca.165°C at a typical deposition pressure of ca. 10 T~rr.~ Unfortunately, considerable decomposition occurs at both pressures. We have recently rep~rted’.~ on an even more highly fluorinated barium precursor with R =R =C,F,. This compound [Ba(TDFND),.H,O] (where TDFND = 1,1,1,2,2,3,3,7,7,8, 8,9,9,9-tetradecafluorononane-4,6-dione)has an m.p. of 187 “C and, following loss of water, sublimes completely without decomposition at 1 atm. This was the first barium complex to show stable and complete volatilisation at ambient pressure. However, when used for CVD of BaF, it was found that6 whilst there was a very high initial deposition rate there was a marked drop in the rate when the same precursor material was used for a series of depositions.This growth rate variation with time was attributed to the loss of the water and sub- sequent slow structural changes in the molecule leading to stronger inter-monomer forces and decreased volatility. We have subsequently shown that anhydrous Ba(TDFND), gives stable and reproducible growth and that this material is a good precursor for the preparation of high-T, films.7 Anhydrous Ba(TDFND), still has, though, a relatively high melting point, and lower volatility, compared with the yttrium and copper precursors which are usually used for the CVD of YBCO. So to try to enhance the volatility of Ba(TDFND), even further we have coordinated it with a polyether, tetra- glyme (Le.2,5,8,11,14-pentaoxapentadecane).The approach using a polyether complex has been used previously with Ba( HFA),,%” but although the compounds can be completely sublimed at low pressure, at 1 atm decomposition occurs at a temperature near to the sublimation temperature. In this paper we report on the CVD of BaF, and YBCO using [Ba(TDFND), tetraglyme] as a precursor and we show that it gives stable and reproducible growth of high-quality crystal- line films of the former material and, after annealing, super- conducting films of the latter. Experimental Details of the synthesis and purification of the parent barium P-diketonate [Ba( T DFND),] have been given elsewhere.’ The preparation and characterisation of the tetraglyme com- plex have also been reported in detail,” but for convenience they are summarised here.The preparation was carried out under an atmosphere of dry nitrogen using standard Schlenk and syringe techniques. Barium hydride was used as supplied by Strem. Tetraglyme (Aldrich, 99 + %) was dried and stored over activated type 4A molecular sieve. Hexane (Prolab, Normapur grade) was dried by reflux over sodium/ benzophenone in the presence of tetraglyme and was freshly distilled before use. TDFND (5.0 cm3, 20mmol) was added dropwise with caution and starting at ambient temperature to ;I stirred suspension of barium hydride (1.39 g, 10 mmol) in hexane (20 cm3) and tetraglyme (2.2 cm3, 10mmol). An immediate effervescence was accompanied by a rise in temperature and dissolving of most of the solid.After the initial reaction had died down the suspension was heated to reflux in an oil bath for ca. 90 min and a turbid solution formed. The solution was filtered, yielding a clear, faintly yellow solution which, on refrigeration overnight to -20 "C, yielded 9.213 g (78%) of densely packed white crystals. Briefly the NMR analytical data obtained were as follows: 'H NMR dH 5.95 (s, 1 H, COCHCO), 3.73 and 3.53 (m, 8 H, CH,CH,) and 3.32 (s, OCH,). {'H} 13C NMR 6, 175.1 (m, CF,COCH), 117.6 (m, CF,CF,CF,), 109.4 (m, CF,CF,), 89.62 (s, COCHCO), 70.61, 70.22, 69.65 and 69.48 (s, CH2CH2), 58.38 (s, OCH,). Found vs (calculated): C, 28.80 (28.65); H, 1.97 (2.04); N,<0.3 (0); Ba, 11.5 (11.7) wt.%.Thermal analyses were carried out with a Stanton Redcroft STA 100 thermal analyser using Inconel crucibles at atmos- pheric pressure under a nitrogen flow of 40cm3 (standard) min-' (sccm) and at a heating rate of 20°C min-'. The CVD system was an impinging jet reactor (Archer Technicoat Ltd) with a resistive heated substrate platform (US Inc.) and standard gas-handling facilities. Stainless-steel precursor con- tainers were heated by circulating hot air in separate tempera- ture-controlled enclosures. Vapours generated in the containers were then carried by a flow of argon gas through stainless-steel gas lines. The gas lines were fitted with valves which allowed the precursors either to be vented to exhaust, permitting stabilisation of the precursor flows before the start of deposition, or to have direct entry to the reactor.All components of the system between the precursor containers and exhaust, as well as the walls of the reactor, were heated to a temperature of ca. 200°C to prevent condensation of the precursors. The precursors were mixed with 0, immediately prior to impinging on the heated substrate so minimising the possibility of reactions in the gas phase. Waste gases were passed through a cold trap to condense any unreacted precur- sors and were then removed from the system by a rotary vacuum pump. For the Ba precursor the precursor container temperature was either kept constant at 96°C for all the deposition experiments or it was varied in the range 80-107 "C for measurements of the effective enthalpy of vaporisation. The Y and Cu precursors used were Y(TMHD), and Cu( TMHD), (where TMHD =2,2,6,6-tetramethylheptane-3,5-dionate) and the precursor containers were kept at 108 and 101 "C, respectively.These temperatures were chosen to give an approximate ratio of 1 :2 :3 of the three elements in the YBCO films, as determined by EDAX. Measurements of each precursor temperature with a thermocouple probe touch- ing the precursor showed there was a negligible temperature difference between the container and its contents. Deposition of BaF, was performed at a heater platform temperature in the range 400-700 "C corresponding to a substrate tempera- ture range of 350-610°C. Deposition of films containing a mixture of Y,Ba and Cu were carried out at a substrate temperature of 660°C.The pressure in the reaction chamber and in all the gas lines was maintained at 10 Torr throughout for all depositions. The total argon gas flow for deposition was always 600 sccm with 200 sccm being used as carrier gas for each precursor. In the case of BaF, deposition 400 sccm were added just prior to the reactor entry. For an investigation of the order of reaction with respect to the precursor the precursor pot temperature was varied in the range 96-106 "C. The oxygen flow was kept at 200 sccm throughout, but for determining the reaction order with respect to oxygen the flow was varied from 30 to 200 sccm. Substrates for deposition were generally Si( loo), but for investigations of supercon- ducting YBCO cleaved single-crystal MgO tiles (1 cm x 1cm) were used.For BaF, identification of crystalline phases and preferred J. MATER. CHEM., 1994, VOL. 4 orientation in deposited films was achieved using an X-ray powder diffractometer with Cu-Kcc radiation. Film-thickness profiles were measured by a surface profilometer (SLOAN DEKT AK 11) after films had been photolithographically patterned and etched with a suitable mineral acid. Thicknesses, as well as refractive indices, were also obtained by using a two-angle ellipsometer (Gaertner L116B). Films consisting of a mixture of Y and Cu oxides and BaF, were converted to YBCO by postdeposition annealing at atmospheric pressure first of all in a flow of water vapour and then in 0, using conditions and a temperature programme as suggested by Kirlin et a1.12 Measurements of resistance-- temperature characteristics for the YBCO films were made using a standard four-point probe method across the tempera- ture range 10-300 K.Determinations of the quantity of the barium precursor transported into the reactor were made by condensing the precursor vapour in a stainless-steel U-tube (weight ca. 135 g) attached to the line emerging from the precursor pot. The U-tube was cooled by ambient air. The amount of material transported (ca. 3-30mg in 60min) was monitored as a function of precursor temperature, carrier gas flow, and of the time the precursor was held at a given vaporisation temperature.Results and Discussion Fig. 1 shows the results for simultaneous thermal analysis (STA) (combined TG and DTA) for [Ba( TDFND),- tetraglyme], and Table 1 summarises data obtained from this figure together with comparable results for other barium fl-diketonate precursors. The DTA in Fig. 1 clearly shows the sharp melting point of the new barium complex and the TG shows that it sublimes completely without decomposition at a total pressure of 1 atm. Also given in Fig. 1 is an STA for a sample of the complex which had been used for a series of six deposition runs and again no residue remains after complete volatilisation, indicating the thermal stability of the complex. Also since no precautions were taken during handling of the compound to prevent contact with air or atmospheric moist- ure the material would appear to have good resistance to aerial oxidation and not to be particularly moisture sensitive. Samples have been found to have unchanged melting points on prolonged storage in sealed containers and following vacuum sublimation.The slight increase in the melting point of the used sample of the tetraglyme complex compared to that of the fresh sample may be associated with the removal of volatile contaminants, in particular excess tetraglyme, lead- ing to an effective purification of the material. The data summarised in Table 1 show that 1 1 0 50 100 150 200 250 360 3 7°C Fig. 1 Simultaneous thermal analysis of (a) freshly prepared sample of [Ba(T DFND), tetraglyme] and (b)sample after use for CVD.J. MATER. CHEM., 1994, VOL. 4 Table 1 Thermal analysis data for barium P-diketonate compounds water loss onset weightcompound T/"C loss (YO) melt T/"C [Ba( TMHD),H20] 50,90 3,7 -b [Ba( HFODj,H,O] 20,70 2 176 [Ba(DFHD)2H20] 80 2 200 [Ba(TDFND),H,O] 63 2 187 [Ba(TDFNDj21e --186 [Ba( TDFND),lf --196 [Ba(HFA),* tetraglyme] --149 [Ba( TDFND j2* tetraglyme] --70 volatilisation onset T/"C weight loss (%)a decomposition onset/"C total weight loss (YO) reference 200,320' 30 420,580,700' 84 240 38 300 91 220 88 350 91.5 200 99 b- >99 220 98 b- 98 220 99 b- 99.5 160 ca. 609 ca. 29Y 94 160 99 b- >99.5 TMHD, R =R' =C(CH,),; HFOD, R=C(CH,),; R' =C3F7; DFHD, R =CF,; R' =C3F7;TDFND, R =R' =C3F7; HFA, R =R' =CF,."Prior to onset of observable decomposition. bNot observed. 'Occurs in two stages. "Melt possibly detected in residue. 'Freshly prepared (after 0.5 h dehydration at 180°C in the STA). /Aged (after use for > 10 h in growth reactor -recovered sample). gUnder the STA conditions used these events overlap and values are therefore approximate. [Ba( TDFN-D), tetraglyme] has by far the lowest melting point of all the /?-diketonates listed in the table. In addition, it is stable under conditions of ambient pressure, whereas the compound with the next lowest melting point, [Ba( HFA), tetraglyme], is only stable at reduced pressure. Using a higher polyether to complex with Ba(HFA), to form [Ba( HFA),.hexaglyme] (where hexaglyme =2,5,8,11,14,17,20-heptaoxaheneicosane) produces a compound with a lower melting point of 70-72 "C,* but again it is not thermally stable at ambient pressure.Preliminary studies have shown that the hexaglyme complex of Ba(TDFND), has an even lower melting point of 40°C and that it is also thermally stable at 1 atm, but its use as a CVD precursor has not yet been investigated further. An important feature of the polyethers of Ba(TDFND), is that they can conveniently be used in liquid form in a precursor container, and this is highly desirable from the point of view of reproducibility of carry- over rates. The structural form of a solid and the geometry of the container can markedly affect the rate of precursor evaporation., For example, if any sintering of a powder occurs during a deposition run then subsequent runs may give very different results due to changes in the carry-over rate of the precursor.A series of six measurements of carry-over rate made in between deposition experiments gave for a precursor pot temperature of 96 "Cand a pressure of 10 Torr a mean value of (367+38) pg min-' where the error is at the 95% confi-dence level. There was no systematic variation of carry-over rate with time and a significant cause of the size of the error was probably small variations in the temperature control of the precursor pot (see discussion below). The variation of carry-over rate as a function of temperature gave a linear plot of the logarithm of the carry-over rate against reciprocal thermodynamic temperature (Fig.2), and from the slope of this plot a value of the effective enthalpy of vaporisation (AHeff,,)of (99.2+ 11.2) kJ mol-' was calculated. It is import- ant to note that this value of AH,,, will not be a true thermodynamic value since it was determined under non-equilibrium conditions. It may, for example, contain a contri- bution from the energy of activation associated with mass transport and, as such, be dependent on precursor pot geometry. This effect can be shown quite clearly from, for example, the work of Fitzer et aL2 who measured the tempera- ture dependence of evaporation of Ba(TMHD), with two very different geometries for the precursor containers.Values of AHeff,,calculated from their data for the evaporation rates obtained with the two different containers differ by more than 30%. Similarly, literature values of AH,,, for Cu(TMHD), range from ca. 82kJ mol-',13 to 124kJ m~l-','~ again a precursor temperature/'C 115 105 95 a5 757 I I I I 6 31. ' . * ' . -. ' ' 2.5 2.6 2.7. . . ' '. . . .I2.8 2.9 lo3 KIT Fig.2 Temperature dependence of the carry-over rate of [Ba(TDFND)2 tetraglyme]. Total pressure = 10 Torr; carrier gas flow =200 sccm. variation of> 30%, while for Y(TMHD), the variation is >60%.15 Clearly, the AH,,, for [Ba(TDFND), tetraglyme] determined here cannot be meaningfully compared with values reported in the literature for other barium complexes.Indeed, literature values of 'enthalpies of vaporisation' should be treated with great care since the majority of them are far from true thermodynamic values and they should not be quoted as such. However, the relatively high value of AH,*,, does show the need for careful temperature control of the precursor pot; a variation of 1°C would give rise to ca. 10% variation in the carry over rate. Also it can be usefully compared to AH,,, for Ba(TDFND), determined under identical conditions of flow, pressure and precursor pot geometry. For Ba(TDFND),, the value of AHeE,, was found to be (211.8f22.8) kJ mol-'. This significantly higher value than that for the tetraglyme complex is probably indicative of stronger intermolecular bonding in Ba(TDFND),.At a constant temperature of 610°C an average value of (4.69 f0.30) nm min-' was obtained for the growth rate of BaF, from six deposition runs. There was no systematic variation of the growth rate with time. Fig. 3 shows the variation of the growth rate of BaF, with temperature. The presence of kinetic- and transport-controlled regions is evi- dent. However, as has been pointed out elsewhere,I6 there is J. MATER. CHEM., 1994, VOL. 4 : 0 2 1.01 II I 0 -0.5 0 0.01 . " ' 'I. ' " " 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 lo3 WT Fig. 3 Temperature dependence of the deposition rate of BaF,. Total pressure = 10 Torr; precursor temperature =96 "C;carrier gas flow = 200 sccm. not a sharp transition between these two regions and a true kinetic activation energy can only be obtained by allowing for the contribution of transport-controlled growth at low temperatures. This is done using the simple relationship l/jT= l/k + l/j, where j, is the overall or total growth rate, j, is the kinetically controlled growth rate, and j, is the transport-limited con-trolled growth.Taking an average value of j, of 4.86nm min-' from Fig. 3, the kinetic growth rates for T,<43O"C can be calculated. An Arrhenius plot of these j, values gives a kinetic activation energy (E,) of ca. 200 kJ mol-l. This is noticeably higher than the value of E, z 150 kJ mol-' deter-mined for the deposition of BaF, from Ba(TDFND),.7 The higher volatility of the tetraglyme complex probably indicates less intermolecular binding which could be associated with the polyether compound having reasonably strong intramol- ecular binding.Hence a higher energy will be required to break up the molecule. Measurement of film refractive index by ellipsometry gave an average value of 1.474 for T,>430°C. This value is an exact agreement with the literature value for bulk BaFZ,I7 and indicates a high quality crystalline film. The quality of the films was confirmed by X-ray diffraction of BaF, layers deposited at 610 "C which showed that they were highly oriented in the ( 111) direction (Fig. 4). Channelling analysis with Rutherford back scattering (RBS) for 0.5" steps over +5" with respect to the normal to the surface showed no evidence of either planar or axial channeling, again showing the high degree of orientation of the crystalline films.However, such good quality films were only obtained if growth was carried out under mass transport control. For temperatures <ca. 430 "C there was not only a fall off in the growth rate (cf. Fig. 3) but also in the refractive index of the layers (Fig. 5). Layers grown at these lower temperatures were, in addition, very poorly adherent to the substrate. Moving into a depos- ition region where there is some kinetic control means that not all of the precursor reaching the surface will undergo decomposition, and unreacted, or partially reacted, species could be incorporated into the layer, hence leading to a low value of refractive index and a poor-quality film.This concept is supported by the fact that as the partial pressure of the precursor was lowered, at a temperature corresponding to h v) (111).g 10000-2 8000-Y I>-5 6000-c Q)c .E 4000--([I .g 2000--1A. I01 !I:20 40 60 80 100 2tYdegrees Fig. 4 XRD of BaF, on Si. Total pressure= 10 torr; precursor temperature =96 "C; carrier gas flow =200 sccm; deposition tempera- ture =610 "C. mDm 1.4' 8 Q) -0c.-1.21 ' ' . '. ' " . '. I 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 lo3 WT Fig. 5 Dependence of refractive index of barium-containing films on deposition temperature. Total pressure = 10 Torr; precursor tempera- ture =96 "C; carrier gas flow =200 sccm. mixed transport and kinetic control at a higher precursor partial pressure, the refractive index rose until it became equal to 1.474 again (Fig.6) and a good-quality adherent BaF, film was obtained. This is understandable in terms of the shape of a curve for growth rate as a function of reciprocal temperature, where as reactant partial pressure decreases the transport controlled region extends to lower and lower temperatures.I6 It is interesting to note that Ba(TDFND), does not show this 1.6 r X$ 1.4 -.-m. .-9. c.0 e-2 1.2-1.o 0 50 100 150 200 precursor Ar flow/cm3 (std) min-' Fig. 6 Dependence of refractive index of barium containing films on precursor partial pressure. Total pressure = 10 Torr; precursor tem- perature =96 "C;deposition temperature =388 "C. J.MATER. CHEM., 1994, VOL. 4 r of BaF,. It can also be used in the preparation of supercon- ducting YBCO. In this context it is particularly valuable since its volatility is comparable to that of Y(TMHD), and Cu(TMHD), which are generally used as sources for yttrium and copper. Control of the different gas flows then becomes standard procedure for the three cases. Of course, for YBCO preparation, in situ or postdeposition hydrolysis of the fluoride is required and this is a drawback. However, no non-fluorine- containing barium precursors, to our knowledge, show the same degree of stability, volatility and reproducibility that we 1i have reported here for [Ba( TDFND), tetraglyme]. Therefore, this material is an extremely interesting and promising candi- 01date not just for the MOCVD of BaF, but for high-0 50 100 150 200 250 300 TIK Fig.7 Resistance-temperature plot for an YBCO film prepared with the use of [Ba(TDFND), tetraglyme]. Total pressure= 10 Torr; total carrier gas flow =600 sccm; deposition temperature =660 "C. effect; good-quality BaF, films with a constant value for the refractive index can be obtained for deposition well into the kinetically controlled regi~n.~ The presence of the tetraglyme would appear to affect the adsorption properties of the precursor. This is being investigated further. Because of the varying nature of the layer properties at lower temperature for deposition from [Ba( TDFND),- tetraglyme] it was not possible to examine meaningfully kinetic parameters, such as order of reaction.At deposition temperatures >430 "C the variation of deposition rate with partial pressure of the precursor was determined from a log-log plot of growth rate against carry-over rate. The slope of such a plot at 610°C was ca. 0.9, which is consistent with a mass-transport controlled regime. At the same temperature the dependence of growth rate on the partial pressure of oxygen was determined simply by varying the oxygen flow, but keeping the total gas flow and pressure constant. The slope of the plot was ca. 0.1, which means that the order of reaction with respect to oxygen is effectively zero. This is not too surprising since the films being grown were BaF, and not the oxide.No significant participation of the oxygen in the growth process would therefore be expected. The presence of oxygen may be desirable, however, to help to keep the carbon content of the films low. Fig. 7 shows a resistance-temperature plot for an YBCO film prepared, as described above, using [Ba( TDFND),- tetraglyme] as the barium precursor. Semiconducting resis- tivity is seen at high temperatures and the onset of supercon- ductivity is at ca. 80 K with zero resistance being attained at ca. 26 K. These values could undoubtedly be improved upon by optimising the depositions and annealing conditions,12 and the potential value of the barium precursor is apparent. Conclusions We have shown that [Ba(TDFND), tetraglyme] is a stable, highly volatile precursor suitable for reproducible MOCVD temperature superconducting YBCO as well.We acknowledge the support of the Commission of the European Communities under the BRITE/EI JRAM Programme, Contract No. BREU/0438. We also wish to thank Professor D. J. Cole-Hamilton for helpful discussions, Mr. R. P. McGinty and his colleagues for performing the thermal analyses, and Dr. H. Kheyrandish for the RBS results. References 1 M. L. Hitchman, D. D. Gilliland, D. J. Cole-Hamilton and S. C. Thompson, Znst. Phys. Conf. Ser., 1990, 111,305. 2 E. Fitzer, H. Oetzmann, F. Schmaderer and G. Wahl, in Proc. Eighth Eur. Conf. CVD, ed. M. L. Hitchman and N. J. Archer, Les Editions de Physique, Paris, 1991, p.C2-713. 3 A. P. Purdy, A. D. Berry, R. T. Holm, M. Fatemi and D. K. Gaskill, Inorg. Chem., 1989,28,2799. 4 C. O-Gonzalez, H. Schachner, H. Tippmann and F. J. Trojer, Physicu C, 1988,153-155, 1042. 5 S. C. Thompson, D. J. Cole-Hamilton, D. D. Gilliland and M. L. Hitchman, Adu. Muter. Opt. Electron, 1992,1,81. 6 D. D. Gilliland, M. L. Hitchman, S. C. Thompson and D. J. Cole-Hamilton, J. Phys. ZZZ France, 1992,2,1381. 7 M. L. Hitchman, S. H. Shamlian, D. D. Gilliland, D. I. Cole-Hamilton and S. C. Thompson, results to be published. 8 K. Timmer, C. I. M. A. Spee, A. Mackor and H. A. Meinema, Eur. Put. Appl., 1991,405 634, A2. 9 H. A. Meinema, K. Timmer, E. A. van der Zouwen-Assink, C. I. M. A. Spee, P. van der Sluis and A. L. Spek, XXVllIth Znt. Con5 Coord. Chem., Gera, DDR, 1990,Abstr. 6-82. 10 J. A, T. Norman and G. P. Pez, J. Chem. Soc., Chem. Commun., 1991,971. 11 S. C. Thompson, D. J. Cole-Hamilton, S. L. Cook and D. Barr, Eur. Put. Appl., February 1993,92307390. 12 P. S. Kirlin, R. Binder, R. Gardiner and D. W. Brown, SPIE Processing of Films for High T, Superconducting Electronics, 1989, 1187, 115. 13 S. H. Kim, C. H. Cho, K. S. No and J. S. Chun, J. Murer. Res., 1991,704. 14 M. A. V. Ribeiro da Silva, M. D. M. C. Ribeiro da Silva, A. P. S. M. C. Carvalho, M. J. Akello and G. Pilcher, .1. Chem. Thermodynum., 1984, 16, 137. 15 D. D. Gilliland, PhD Thesis, University of Strathclyde, 1993. 16 M. L. Hitchman, Progr. Cryst. Growth, 1981,4,249. 17 CRC Handbook of Chemistry und Physics, ed. D. R. Lide, CRC Press, Boca Raton, F1. 1993 pp. 4-42. Paper 3/04895K; Received 12th August, 1993