J. MATER. CHEM., 1994, 4( 8), 1275-1282 Synthesis and Characterization of Ni,Sb,(OEt),, and its Hydrolysis Products Gunnar Westin and Mats Nygren* Department of Inorganic Chemistry, Arrhenius Laboratory, Stockholm University, S-1069I Stockholm, Sweden The hydrolysis pathways of Ni,Sb,(OEt),, in toluene, ethanol and toluene-ethanol mixtures have been studied. Five intermediate compounds have been isolated from solutions having different H20/Ni2Sb,(OEt),, ratios (h values) with h<6. A detailed description of the routes used in connection with the preparation of Ni,Sb,(OEt),, and of the five intermediates, Ni5Sb30p(OEt),5(HOEt),, Ni7Sb404(OEt)18, Ni6Sb404(0Et)16(HOEt),, NiSb303(0Et), and Ni6Sb,,0,8(OEt),8(HOEt)2are given. The compositim of three of the intermediates compounds have been established from single-crystal structural studies.The suggested compositions of the two others, Ni7Sb404(OEt)18 and NiSb 303(OEt)5, are based on elemental analysis. The chemical compositions of the sol and of the gel formed for h>6, NiSb,OS(OEt),, are estimated from chemical considerations. A brief investigation of the hydrolysis pathways of Sb(OEt), in the same solvents is also included. The great interest in bimetallic alkoxides stems from the possibility of utilizing them as catalysts and in sol-gel pro- cessing of advanced If, in the sol-gel processing of the alkoxides, the hydrolysis is controlled carefully one can obtain gels of high homogeneity, and by applying different preparation routes gels possessing various physical properties and morphologies are f~rmed.~,~ Knowledge about the reac- tions that occur in the precursor solution during the hydrolysis process, which finally leads to a gel, is thus of great importance in designing the properties of the product obtained.Such information is scarce, however, especially for bimetallic alkox- ide systems. Studies of monometallic systems such as Ti(OEt), show that many different molecular 0x0-alkoxides are formed as intermediates before the gelling process takes place.9-" The molecular structures of these intermediates are often very complicated. In bimetallic systems it can be expected that the intermediates have metal ratios different from the precursor alkoxide because the metal ions involved have different affin- ities for oxygen and different coordination requirements. Such shifts in stoichiometry have indeed been observed for the first hydrolysis product in some of the bimetallic alkoxide systems studied.', If oxoalkoxides with various metal ratios are formed in the hydrolysis process, these compounds can find use as precursors for further processing.In a previous article we reported the sol-gel processing of M-Sb ethoxides (M=Mn, Fe, Co and Ni) with an M:Sb ratio of 1:2.13 These samples were hydrolysed using the humidity in the atmosphere, but the hydrolysis pathway was not evaluated. In this article, however, we describe the hydroly- sis pathways of Ni,Sb,(OEt),, dissolved in toluene, ethanol and toluene+thanol mixtures.Six intermediates have been isolated and the preparation routes and some characteristic properties of Ni2Sb4(OEt),6 and the intermediates formed are described. For some of these compounds it has been possible to prepare single crystals and accordingly the crystal structures of these have been The preparation aspects of Ni,Sb,(OEt),, and of the intermediates formed are emphas- ized in this article, while a forthcoming article will deal with the structural relationships between compounds formed in the hydrolysis processes." During the course of this work it was found necessary to make a brief investigation of the hydrolysis pathway for Sb(OEt),, and these results are also included here. This study has been conducted within a research pro- gramme aiming at understanding the reactions involved in sol-gel processing of bimetallic alkoxides.Experimental Fourier-transform infrared (FTIR) spectra were obtained using a Mattson Polaris FTIR instrument with deuteriated triglycine sulfate (DTGS) detectors and KBr or polyethylene (PE) beam splitters for the mid- and far-IR studies. respect- ively. The IR spectra of the solids were recorded with KBr or PE tablets, and the dissolved alkoxides were investigated with a KBr cell. The spectra in the far-IR region were scaled to fit the mid-IR spectra and connected with them at 400-450 cm-l. Owing to the strong background absorption from the solvents, significant spectra of the dissolved alkoxides could be obtained only in the region 690-390 cm-' (there is only one relative strong solvent peak in the region around 460 cm-I).Elemental analysis of the Ni :Sb ratio of hydrolysed samples were routinely performed using a scanning electron micro- scope (SEM) equipped with an energy-dispersive specs rometer (EDS)(JEOL 820 and Link AN 10000)with a detect ion limit of ca. 1 wt.% while the accuracy of the EDS measurements was of the order of 6-8% depending on the morphology of the sample. Two of the bimetallic alkoxides were analysed for their Ni, Sb, 0,C and H contents at Analytische Laboratorien, Gummersbach, Germany. The melting behaviour of samples, sealed in glass capillaries, was studied in a solid-block Gallenkamp melting-point apparatus. Anhydrous NiC1, was used as the Ni source, and Sb(OEt), was prepared from SbCl, using the ammonia route" and then distilled in vacuum. Three solvents were used: (i) toluene, which was dried with Na chips; (ii) ethanol, which was distilleg over CaH,; and (iii) acetonitrile, which was dried with 3 A molecular sieves.After they had been dried, the solvents were saturated with dry argon. The KBr and the PE used in the spectroscopic studies were dried at 240 and 110"C, respect-ively, at a pressure of 0.1 Torr. Synthetic Procedure All syntheses and preparations of samples for spectroscopic studies were performed in a glove-box provided uith dry oxygen-free argon. All glassware was kept at 150"C for >3 h before use and was then transferred into the glove-box while it was still hot.Preparation of Ni2Sb4(OEt)16 Typically, Na (0.855 g, 37.2 mmol) was dissolved ilk 40 ml ethanol and Sb(OEt), (9.58 g, 37.3 mmol) was then added. After 16 h NiCl, (2.41 g, 18.6 mmol) was added under stirring, and the stirring was continued for 24 h. The NaCl formed in the reaction was then allowed to precipitate and after ca. 24 h the green solution was separated from the precipitate. A green crystalline mass was formed upon evaporation of this solution. The yield was in the range 85-90%. The material formed was characterized structurally and found to have the composition Ni2Sb,(OEt)16,'4 SEM-EDS studies also revealed a Ni :Sb ratio close to 1 :2. Preparation of Ni,Sb,O,(OEt),,( HOEt), Ni,Sb,(OEt),, (0.5 g, 0.377 mmol) was dissolved in 2.0 ml of a toluene: ethanol (4: 1) solvent and 0.30 ml of a toluene-ethanol solvent containing H,O (1 mol I-' with respect to H20) was added slowly under vigorous stirring, yielding a yellow solution.The H,O :Ni,Sb,(OEt),, ratio (h value) in the final solution was thus 0.8. Square yellow crystals were formed in a yield of 75-80% upon evaporation of this solution. Single-crystal structure analysis has shown that the composi- tion of this compound is Ni,Sb,02(OEt),,( HOEt),.', Ni,Sb,0,(OEt)l,(HOEt)4 also is formed in high yield by spontaneous decomposition of Ni,Sb,(OEt),, in dilute ethanol solutions. Thus, when a solution formed by dissolving Ni,Sb,(OEt),, (0.5 g, 0.377 mmol) in 10 ml ethanol was kept for 2 days at room temperature, well formed crystals of Ni,Sb302(0Et),,(HOEt), were obtained in a yield of 80%.However, if this solution was kept for longer times, Ni,Sb,O,(OEt),,( HOEt), (see below) was formed, which later decomposed into an unidentified oily product. Preparation of Ni7Sb404(OEt)18 Ni,Sb,O,(OEt)l, was prepared by dissolving Ni,Sb,(OEt),, (0.5 g, 0.377 mmol) in 2.0 ml toluene followed by the addition of 0.43 ml of a tolueneeethanol(4: 1) solution (1 mol I-' with respect to H20 content). This corresponds to an h value of 1.14 in the final solution. After a few minutes this solution was evaporated to half of its original volume and acetonitrile was added intermittently under stirring. Yellow-green crystals were formed in a yield of 75-80%.The SEM-EDS studies revealed that the crystals formed had an Ni: Sb ratio of ca. 7 :4. Elemental analysis of the crystals (expressed in wt.%) yielded the following content [data given in parentheses refer to the calculated content according to the formula Ni,Sb,O,(OEt),,]: C, 24.24 (24.39); H, 5.05 (5.12); 0, 19.94 (19.85); Ni, 22.90 (23.17); Sb, 27.65 (27.47)%. The observed and calculated element contents agree within 1.3%. Preparation of Ni6Sb,04(OEt),6( HOEt), Ni,Sb,O,(OEt),,( HOEt), was formed by dissolving Ni,Sb,(OEt),, (0.5 g, 0.377 mmol) in 2.0ml of a toluene-ethanol (1 :2) solvent followed by slow addition of 0.50 ml of a 1 moll-' H20 solution of toluene-ethanol (1 :2). An h value of 1.33 was thereby achieved in the final solution.After 2 h, this solution was intermittently evaporated till a green crystalline mass was formed. The precipitate was washed with ethanol [in order to remove remaining Sb(OEt),]. The single-crystal structure investigation showed that the composition of this phase, formed with a yield of 75'/0, is Ni,Sb,0,(OEt),,(HOEt)4.'6 As described above, Ni6Sb404(OEt)16(HOEt), can also be formed by spontaneous decomposition of dilute solutions of Ni,Sb,(OEt),,. Crystals of Ni,Sb,O,(OEt),,(HOEt), in a yield of 75% were thus formed when a 10ml ethanol solution containing Ni,Sb,(OEt),, (0.5 g) was aged for 4 days at room tempera- ture. If this solution was kept for longer times the J. MATER. CHEM..,1994, VOL. 4 Ni,Sb,04(0Et)16( HOEt), crystals redissolved and evapor- ation of this solution yielded an oily product.Preparation of NiSb,O,(OEt), NiSb,O,(OEt), was formed by dissolving of Ni,Sb,(OEt),, (0.200g, 0.151 mmol) and 0.08 g of Sb(OEt), in 1.0ml of tolueneeethanol (4: 1) whereupon 0.53 ml of ;z solution of toluene-ethanol (4: 1) (1 moll-' with respect to H20) was slowly added under vigorous stirring. These amounts corre- spond to an Ni: Sb ratio of 1:3 and an h value of 3.5 in the final solution. After 2 h the solvent was evaporated and 0.2 ml of the solvent was immediately added. The amount of solvent was then gradually reduced by evaporation with, as above, small additions of a solvent richer in ethanol to keep the ratio of toluene :ethanol at 4 :1. Green crystals were then formed in a yield of 30-50%.The SEM-EDS studies of these crystals revealed an Ni:Sb ratio of ca. 1:3, and elemental analysis of the compound gave [data given in parentheses refer to calculated content according to the formula NiSb,O,(OEt),]: Ni, 8.28 (8.42); Sb, 52.35 (52.39); 0, 17.29 (18.36); C, 18.12 (17.23); H, 3.68 (3.61)%. These amounts fit the proposed formula to within 6.2%. Preparation of Ni6Sb14018(OEt)18(HOEt), Ni6Sb,4018(0Et)18( HOEt), was formed by slow hydrolysis of Ni,Sb,(OEt),, (0.500g, 0.377 mmol) dissolved in 2.0 ml etha- nol with 1.51 ml aqueous ethanol (1 moll-' with respect to H20) under stirring. This yielded an h value of 4 in the final solution. After 2 h, 3 ml of the solution was evaporated off.Crystallization was performed by intermittent evaporation of the remaining solution over a period of normally <1 week, yielding green crystals. The crystal structure of this com-pound has been determined17 and its composition was found to be Ni6Sb,,018(OEt),,(HOEt),. The green crystals rapidly become opaque in the absence of ethanol. Ni6Sb14018(0Et)18(HOEt)2was formed for 3 <h <5 in pure ethanol in rather low yields with a maximum at h =4 (ca. 30%). Preparation of NiSb,O,(OEt),, Sb,O(OEt), and Sb,O, When Ni,Sb,(OEt),, was dissolved in ethanol-containing solvents green sols were formed rather sharply at h=6. If all the added water is consumed in the hydrolysis, the formed compound ought to have a composition of NiSb,O,(OEt),.Sb,O(OEt), was prepared by hydrolysing a mixture of Sb(OEt), (0.30 g, 1.2 mmol) and 1 ml solvent (toluene-ethanol 4: 1, 1:2 or 0:l) by adding 0.40ml of the corresponding solvent containing H20 (1 moll-' with respect to H,O, yielding h=0.3 in the final solution) drop by drop. After 2 h the mixture was rapidly evaporated to dryness and a white insoluble residue was formed. Assuming that all the added water was consumed in the hydrolysis, the precipitate ought to have the composition Sb,O(OEt),. Sb(OEt), (0.30 g, 1.2 mmol) was dissolved in 1 ml solvent (toluene-ethanol with volume ratios of 4: 1, 1 : 2 and 0: 1) and hydrolysed by addition of 1.78 ml of the corresponding aque- ous solvent (1 moll-' yielding h= 1.5 in the final solution), drop by drop.After 2 h the solvent was evaporated off. The compound formed was identified by its IR spectrum (see below) to be Sb203. Results and Discussion Hydrolysis of Ni,Sb, (OEt),, in Toluene-Ethanol Ni,Sb,(OEt),, was hydrolysed in toluene-thanol mixtures with volume ratios of 1:0, 4: 1, 1 :2 and 0: I. The alkoxide J. MATER. CHEM., 1994, VOL. 4 was dissolved to form 1moll-' solutions with respect to Ni and Sb, and then the same solvent mixture containing water (1 mol I-' with respect to H20) was added. (As water is poorly soluble in toluene the hydrolysis in this case was performed with a toluene-ethanol (4: 1) solution, but after a few minutes most of the ethanol was azeotropically removed by evaporation.) The H20 solutions were added at a rate of ca. 1 equivalent of H20 per Ni2Sb,(OEt)16 per 5 min, drop by drop, with vigorous stirring.After 2 h the solvent was rapidly evaporated off until most of the solvent had been removed. Immediately afterwards, the same solvent was added to form ca. 3 moll-' metal solutions. Crystallization was achieved by slow evaporation of the solvent. When toluene- ethanol solvents are used one might expect a shift of the composition of the solvent during evaporation as the formed gas phase has a different toluene-thanol composition (the azeotrope has a 30:70 composition2') from the original solvent. Even if efforts were made to keep the toluene :ethanol ratio constant during the crystallization process (by intermit- tent addition of a solvent richer in ethanol) some drift away from this ratio might still have occurred.With the precursor Ni2Sb4(0Et)',, the yield of crystals of the various bimetallic ethoxides was high for h<1.5 [h= H,O/Ni,Sb,(OEt),,], while for 1.5 < h -= 5 it was normally < 50% and for h 36 sols were obtained, which formed gels upon evaporation. The hydrolysis experiments were repeated several times and were, with few exceptions, quite reproducible. The present study has focused on the preparation of crystal- line compounds, but non-crystalline materials were also stud- ied. The main tools used to differentiate one intermediate from another were IR studies and SEM-EDS analyses, but we also utilized the fact that the intermediates formed have slightly different colours.The compositions of the compounds formed were determined either from their crystal structures, elemental analysis or from chemical considerations. We will first present some characteristic properties of Ni,Sb,(OEt),, and its hydrolysis products. Next we will discuss the different hydrolysis pathways of Ni,Sb,(OEt),, when dissolved in toluene, ethanol and toluene-ethanol mixtures. Characteristic Properties of Ni,Sb,(OEt),, IR spectra of the solid and dissolved Ni,Sb,(OEt),, [0.17 moll-' toluene-thanol (4: 1) solution] are shown in Fig. 1. Characteristic peaks of the solid compound in the C-0 and M-0 regions (1200-50cm-') are: 1096, 1053, 1036, 879, 604, 553, 497, 493, 457, 425, 374, 292, 271 and 135cm-'.The spectrum of the dissolved compound has an M-0 band maximum at 515 cm-'. The spectra of concen-trated toluene-ethanol solutions also contain a broad band with a maximum at 515 cm-', but in addition they contain weak shoulders at positions where the solid sample exhibits strong and sharp peaks (see also below). Note also that the spectra of a 0.17 moll-' ethanol solution and a 0.17 mol I-' toluene-ethanol (4 : 1)solution are identical. When the compound was heated in the melting point apparatus a change in colour from green to purple was observed at 63-65"C, and the material melted to form a viscous liquid at 96-98 "C. The purple compound formed by heat treatmeni at 70°C for 10min reverted to the original green modification of Ni,Sb,(OEt),, in a few hours.Furthermore, a green solution was formed when a toluene-thanol solvent was added to the purple compound. Subsequent evaporation of thc solvent yielded crystals of Ni,Sb,(OEt),,. Ni,Sb,(OEt),, is very soluble in mixtures of ethanol and toluene C0.4 moll-' both in solutions of pure ethanol and of toluene-thanol(4 : 1)I, in both cases yielding green solutions. In less concentrated solutions Ni,Sb,02(OEt),,( HOEt), and later Ni6Sb404(0Et)16(HOEt)4 are formed in high yields within a few days (see also above). The chemical reactions leading to the formation of these compounds are unknown, but it is known that alkoxides can decompose to yield oxoalkoxides and ethers.2'-22 Ni,Sb,(OEt),, is stable for months in more concentrated solutions and in the solid state.Dissolving Ni2Sb,(OEt),, in toluene initially yielded a solution with a very strong purple colour, but after ca. 1h the solution became faintly green-brown and a purple precipi- tate was formed. The purple precipitate reverted to the green modification of Ni2Sb4(OEt)16 within 1 week when it was stored. Addition of toluene-ethanol solvent to thc purple precipitate resulted in partial dissolution, yielding a green solution which, as above, upon evaporation yielded Ni2Sb4(0Et),,. Examining the purple precipitate in the SEM revealed it to have an Ni: Sb ratio of ca. 1: 2. The IR spectrum of the purple precipitate is shown in Fig. 2. The spectrum is composed of a rather broird band containing a number of peaks, some of which can be ascribed to Ni2Sb4(OEt),, (see Fig.1). The IR spectrum of the purple compound formed by heat treatment is similar, but the contribution from Ni,Sb,(OEt),, is more pronounced. This is in agreement with the latter compound reverting to Ni,Sb,(OEt),, much faster than the former (see above). The two spectra are, however, sufficiently similar to suggest that the purple precipitate and the compound formed by heat treatment are the same. The similarity in the chemical behav- iour of the two samples also supports this conclusion. The IR spectra of Ni2Sb,(OEt),, dissolved in toluene- ethanol solutions (4: 1 and 0: 1) and that of solid Ni,Sb,(OEt),, differ greatly, suggesting that the molecular -% 5 es:13 d --%, 4 Fig.1 IR spectra of Ni,Sb,(OEt),,: (a) solid and (b) dissolved in Fig. 2 IR spectrum of the purple compound formed by precipitation toluene-ethanol (4 : 1) of Ni,Sb,(OEt),, in toluene structure of Ni,Sb,(OEt),, is not the same in solution and in the solid state. However, the IR spectra of concentrated ethanol solutions contain weak shoulders which coincide with the strongest peaks occurring in the spectrum of the solid compound. SnSb2(OR),23 and PbSb,(OR)824 have been reported to occur as monomers in refluxing parent alcohol or in benzene. Ni alkoxides are known to have either tetrahedral or octa- hedral coordination around the Ni ions, depending on the bulk and electron-donating properties of the ligands.The Ni ions in Ni,Sb,(OEt),, are octahedrally c00rdinated.l~ Ni(OR),," NiNbz(OR),2 and NiTa2(OR)1226 with ethoxy ligands exhibit octahedrally coordinated Ni ions, while isopro- poxy ligands yield tetrahedral coordination. However, Ni-isopropoxy compounds dissolved in electron-donating sol- vents like pyridine and THF yield alkoxides with octahedrally coordinated Ni ions. On the other hand, the Ni ions in NiAl,(OEt), are tetrahedrally coordinated in the solid state but octahedrally coordinated in ethanol solution, while the corresponding isopropoxy compound contains tetrahedrally coordinated Ni ions both in the solid state and in isopropyl alcohol solutions.27 These observations seem to suggest that the Ni ions are always tetrahedrally coordinated by isopro- poxy ligands but can be either octahedrally or tetrahedrally coordinated by ethoxy ligands.In an excess of ethanol the Ni ions seem to prefer octahedral coordination, but at some concentration level a small shift in the concentration of ethanol may cause a change in coordination from octahedral to tetrahedral or vice versa. Note also that the colour of compounds with Ni in an octahedral environment is typically green.25-28 Based on these considerations, it seems plausible that: (i) the Ni,Sb,(OEt),, occurs mainly as monomeric units in solution, but in concentrated solutions (see above) minor amounts of dimeric units might also be present; (ii) the Ni ions are octahedrally coordinated in the solution. An octahedral environment around the Ni ions in the monomer is achieved by the addition of two ethanol units. This implies that the composition of the monomeric unit in the solution ought to be NiSb,(OEt),(HOEt),.Considering the proposed composition of the monomer and recalling that Ni ions can be tetrahedrally coordinated and that the purple colour is typical of Ni in tetrahedral c~ordination,~~-~~it is tempting to suggest that the purple compound, which is formed in the absence of ethanol, contains tetrahedrally coordinated Ni ions. Characteristic Properties of Ni5Sb302(OEt)15(HOEt)4 The IR spectra of Ni5Sb302(0Et)15(HOEt), as a solid and dissolved in a toluene-ethanol (4: 1) solvent (0.07 moll-') are shown in Fig. 3. Characteristic peaks of the solid com- pound are: 1104, 1060, 894, 594, 534, 490, 414, 389 and 261 cm-'.A broad band ascribed to OH stretching is observed around 2500 cm-', with a maximum at 2545 cm-'. The spectrum obtained for the solution exhibited M-0 peaks at 590, 532,488,420 and 390 cm-' and the OH stretching band has its maximum at 2540cm-'. The similarities of the IR spectra of the solid and dissolved samples, as well as the observation that both the solid and the-solution are yellow indicate that the molecular structures in the solid staie and in solution are the same. The colour might be ascribed to the occurrence of five- and six-coordinated Ni ions in the str~cture.'~ When heated, Ni5Sb30,(OEt),,( HOEt), became dark and decomposed around 100-105 "C, yielding a grey-green solid product and liquid Sb(OEt),.Ni,Sb,02(OEt),,(HOEt), is soluble in toluene-ethanol J. MATER. CHEM., 1994, VOL. 4 '1200 . 1000 800 600 400 200 wavenurnber/cm-' Fig. 3 IR spectra of Ni,Sb,O,(OEt),,(HOEt),: (a) solid and (b)dis-solved in toluene-thanol(4 :1) (4: 1, 0.07 mol l-'), yielding an orange solution, but is very sparingly soluble in ethanol, giving a very faintly yellow solutions. The compound is stable for months in the solid state. Characteristic Properties of Ni,Sb404(OEt)18 The IR spectra of solid Ni,Sb,O,(OEt),, and of the substance dissolved in toluene and toluene-thanol (4 :1) solutions are shown in Fig.4. Characteristic peaks in the spectrum of the solid compound are: 1110, 1096, 1062, 895, 636, 515,458 and 259cm-'.In this case no OH band was observed. Characteristic peaks in the 690-390 cm-' region of the spectra of the dissolved samples in toluene solution were found at 635, 598 and 516 cm-' and at 633, 596 and 507 cm-' for the toluene-ethanol solution. Although the spectra of the solid and the dissolved compound exhibit some differences, the similar overall appearance of the spectra suggests that the molecular structure of the solid and dissolved samples are similar. No distinct melting point could be determined, but the compound became black at 195-200 "C. Ni7Sb,0,(OEt),8 is very soluble both in pure toluene and in toluene-thanol (4:l) and 0.2moll-' solutions can be prepared in both cases. Ni7Sb40,(0Et)18 is very sparingly soluble in acetonitrile.Although the crystal structure of this compound has not yet been determined, its yellow-green colour suggests that some of the Ni ions might be five-coordinated. The structure of a related compound, Mn,Sb,O,(OEt),,( HOEt)2,29 has been determined by single-crystal X-ray diffraction. The structure analysis revealed that this compound contains both five- and six-coordinated Mn ions. It has the same M :Sb :0:OEt ratio as Ni,Sb,O,(OEt),,, but contains two additional ethanol ligands. Furthermore, the IR spectra of the two solid com- 1200 1000 800 600 400 200 wavenumberlcm-' Fig. 4 IR spectra of Ni,Sb,O,(OEt),,: (a)solid,(b)dissolved in toluene and (c) dissolved in toluene-ethanol (4 : 1) J. MATER.CHEM., 1994, VOL. 4 pounds are similar. These observations suggest that the struc- tures of the two compounds ought to have similar features. Characteristic Properties of Ni,Sb,O,(OEt),,( HOEt), The IR spectra of Ni,Sb,O,(OEt),,(HOEt), in the solid state and dissolved in a toluene-ethanol(4: 1,O.l moll-l) are given in Fig. 5. Characteristic IR peaks of the solid compound are: 1108,1060,893,646,634,610,579,522,471,278and 261 cm-'. A broad OH stretching band around 2500cm-' with a maximum at 2520cm-' was observed. The spectrum of the dissolved sample exhibits M-0 peaks at 646, 609, 576, 541, 505 and 460cm-' and a broad OH stretching band with a maximum at 2570 cm-l. The main features of the M-0 bands of the solid and the dissolved compound are similar, but there are some differences in the region 400-550 cm-'.Thus the peak at 521 cm-l in the spectrum of the solid seems to be shifted to higher wavenumbers (541cm-') for the dissolved compound, and the peak at 471 cm-' seems to be split in two; one with a maximum at 505 cm-' and another at 460 cm-'. As mentioned above, the solvent exhibits an IR peak at the latter position. However, studying the IR spectra of solutions containing different amounts of solvent confirmed that the dissolved compound has a peak at 460cm-'. The differences between the spectra of the solid and the dissolved compound may be explained by some rearrangement of ethoxy groups in the molecule or by dissociation of the molecule. The latter reaction mechanism has been observed for Pb,Nb,O,(OEt),,, which has a structure resembling that of Ni6Sb404(OEt)16(HOEt),, containing dimers of Nb(OEt)5.30If a similar dissociation mechanism occurs in this case, Sb(OEt), ought to be formed.However, the IR spectrum does not reveal the presence of Sb(OEt),. On the other hand, Sb(OEt), is not easily detected in the presence of Ni,Sb,O,(OEt),,( HOEt), since the characteristic bands of Sb(OEt), overlap with the M-0 band of Ni6Sb,04(0Et),,(HOEt)4. However, the general features of the IR spectra of solid and of dissolved Ni6Sb404(OEt)16(HOEt), are similar enough to indicate that some structural features of the molecule ought to be present both in the solid state and in solution. It was not possible to determine a distinct melting point.A change of colour from green to yellow occurred in the range 50-80 "C, and the sample became dark around 150-170 "Cin connection with a loss of Sb(OEt),. Ni,Sb,O,(OEt),,( HOEt), is very soluble in toluene-etha- no1 (4:1, 0.36moll-l) but only sparingly soluble in ethanol. The compound is stable for months in the solid state in the presence of ethanol. In ethanol-free environments, however, the crystals turn yellow and decompose within a few minutes. Ni,Sb,0,(OEt)l,(HOEt)4 is formed in high yields when r prepared as described above. In solvents containing <20 vol% ethanol Ni,Sb,O,(OEt),, is formed instead. Attempts to prepare Ni7Sb4O4(0Et),, from Ni,Sb,O,(OEt),,( HOEt), and vice zlersa by dissolving the former in toluene or the latter in toluene-ethanol I 1 :2) and then crystallizing the compounds were unsuccessful and non- crystalline materials were formed, Characteristic Properties of an Ni-Sb Ethoxide having an Ni :Sb Ratio of ca.1 :3 Hydrolysis of Ni,Sb,(OEt),, in toluene-ethanol (4 1) in the range 1.5<h <3 yielded pale green crystals after more than 1 week; according to their IR spectra they were different from all other compounds identified in this system. The maximum yield (20-30Y0)was obtained at h =2.5. The SEM-EDS studies revealed that the crystals had an Ni:Sb ratio of ca. 1 :3. This implies that the remaining solution ought to be more Ni-rich than the starting one. In this connection note that crystallization from a solution containing Ni and Sb in a ratio of 1 :3 did not improve the yield very much.An IR spectrum of this solid compound is given in Fig. 6 and contains characteristic peaks at 1098, 1053, 891, 690, 650, 627,570,550,501,355and 275 cm-'. No OH stretching band was observed. This compound changed colour to yellow and became opaque at 55-60°C, and at higher temperatures a slow decomposition was observed. The compound is very soluble in toluene-ethanol (4: l), yielding a green solution. Characteristic Properties of NiSb,O, (OEt), The IR spectra of solid NiSb,O,(OEt), and of the dissolved compound in toluene+thanol (4:1) are shown in Fig. 7. Characteristic peaks of the solid sample are: 1098,1053, 891, c 1200 1000 800 600 400 200 wavenum ber/cm-' Fig.6 IR spectrum of an Ni-Sb ethoxide with an Ni: Sb ratio of 1:3 formed in toluene-ethanol (4 : 1) at 1.5<h<3 r 1200 1000 800 600 400 260 ' 1200 1000 800 600 400 200 wavenum ber/cm-' waven umber/cm-' Fig. 5 IR spectra of Ni6Sb,04(OEt)16(HOEt)4:(a)solid and (b) dis-Fig. 7 IR spectra of NiSb,O,(OEt),: (a) solid and (b) dissolved in solved in toluene-ethanol (4:1) toluene-ethanol(4 :1) J. MATER. CHEM.. 1994, VOL. 4 706, 653, 613, 529, 500, 367 and 270 cm-'. No OH stretching band was observed. The M-0 peaks of the dissolved sample were found at 654, 613 and 528 cm-'. The spectra of the dissolved and solid samples are very similar, suggesting that the molecular structure of the solid persists in solution.The compound melts at 136-150°C and becomes black at 195-200 T.NiSb,O,(OEt), is very soluble in toluene-ethanol (4:l),yielding a green solution. The crystallization rate and the yield are low, which might be explained by the very high solubility of this compound in toluene-ethanol. Thus the crystallization does not start until the mother liquor has become highly viscous. When the starting solution contained Ni and Sb in a ratio of 1:2, NiSb,O,(OEt), is formed throughout the hydrolysis region 3 <h <5, but the crystallization rates and the yields are lower compared with the findings above. It thus seems that an excess of Ni ions in the solution inhibit the crystallization process, possibly due to formation of more Ni-rich Ni-Sb ethoxides.Characteristic Properties of Ni,Sb,,018 (OEt),,( HOEt), An IR spectrum of the solid modification of Ni,Sb14018(0Et)18(HOEt)2is given in Fig. 8, with character- istic peaks at 1098, 1054, 892, 708, 670, 649, 588, 541, 477, 388,357,297,252 and 151 cm-'. A weak, broad OH stretching band with a maximum around 3000-2500cm-' was also observed in the spectrum. The compound melts around 140 "C to form a green liquid. The Ni :Sb ratio in Ni6Sb14018(OEt)18( HOEt), is some- what lower than that in the starting solution. The solution thus ought to contain minor amounts of another alkoxide, richer in Ni than Ni,Sb,,O,,(OEt),,( HOEt),. However, we have not been able to obtain crystals of any composition other than Ni,Sb,,O,,(OEt),,( HOEt), in the region 3 <h<5.Characteristic Properties of NiSb,O,(OEt), The IR spectra of the gels of nominal composition NiSb,O,(OEt), produced in 0 :1,1: 2 and 4 :1toluene-ethanol solutions were similar and exhibited very broad M-0 bands (the IR spectrum of the gel produced in a 4: 1 solvent is shown in Fig. 9). Characteristic IR peaks of the gels were found at 1096, 1050, 888 and 656 cm-'. No OH stretching band was observed, neither were any peaks stemming from toluene. The absence of OH groups indicates that the peaks at 1096, 1050 and 888 can be attributed to ethoxy groups bonded to Ni and/or Sb ions. Hydrolysis of Sb(OEt), in Toluene-Ethanol Sb(OEt), solutions (1moll-') in toluene+thanol with volume ratios ranging from 0: 1 to 4: 1 were hydrolysed by #-'I 1 8 C (d-2 za (d I wavenurnber/crn-' Fig.8 IR spectrum of solid Ni,Sb,,O,,(OEt),,( HOEt), 8 c a 42 0u) n (d 1200 1000 800 600 400 200 wavenurnberkrn-' Fig.9 IR spectrum of a gel formed by hydrolysis of Ni,Sb,(OEt),, dissolved in toluene-ethanol (4 : 1) at h =6 slow addition of a 1moll-' solution of H,O in toluene- ethanol. The hydrolysis of Sb(OEt), followed the same pathway in all toluene-ethanol mixtures studied. Insoluble hydrolysis products were formed for all h values. Two products were identified, one at h=0.3 and one at h= 1.5. The first product coexists with Sb(OEt), (Fig. 10) up to h =0.3 and for 0.3 <h <1.5 two hydrolysis products coexist.For h> 1.5 only Sb203 was formed, and this hydrolysis seems to be quantitative. Characteristic Properties of Sb,O(OEt), and Sb,O, The IR spectrum of Sb,O(OEt), (Fig. 11)shows characteristic peaks at 1098, 1045, 892, 630, 555, 307, 196 and 159cm-'. No OH stretching band was found. Corresponding studies for O<h<0.33 showed that the hydrolysis product in this h range is composed of two phases, namely liquid Sb(OEt), and the solid phase formed at h=0.33. The IR spectrum of the compound formed at h = 1.5 is identical with that of senarmontite (Sb203), which is built up from Sb406 units.,' Corresponding studies for 0.33<h <1.5 r 1200 1000 800 600 400 200 wavenurn ber/cm-' Fig. 10 IR spectrum of liquid Sb(OEt), 1200 1000 800 600 400 200 waven umbe r/cm-' Fig.11 IR spectrum of Sb,O(OEt), J. MATER. CHEM., 1994, VOL. 4 r 1200 1000 800 600 400 200 waven umbedcm-' Fig. 12 IR spectrum of Sb20, revealed that the precipitates consisted of mixtures of senarmontite and Sb,O(OEt),. Hydrolysis Pathway of Ni2Sb4(OEt),,in Toluene These studies were limited to h<1.5. As mentioned above, a purple precipitate starts to form within 1h when Ni,Sb,(OEt),, is dissolved in toluene, implying that the hydrolysis studies had to be performed immediately after the purple solution was formed. Unreacted Ni,Sb,(OEt),, was found in the range h< 1.14 and in larger amounts at lower h values. In addition, Ni,Sb,O,(OEt),, and liquid Sb(OEt), were found for 0.5 <h< 1.5.When a stoichiometric amount of water was added (h= 1.14), Ni7Sb404(OEt)18 and Sb(OEt), were the only species present, which suggests that the following reaction occurs: 7Ni,Sb,(OEt),, +8H20+2Ni,Sb,04(0Et),, + 20Sb(OEt), +16EtOH Hydrolysis Pathway of Ni2Sb4(OEt)16in Toluene-Ethanol (4:1) Ni,Sb,O,(OEt),,( HOEt), was found along with Ni,Sb,(OEt),6 and liquid Sb(OEt), for 0.1<h<0.8, while only Ni,Sb,02(OEt),,(HOEt), and liquid Sb(OEt), were formed at h =0.8. This observation suggests that the following reaction occurs at h =0.8: 5Ni,Sb,(OEt),, +4H,O -+2Ni,Sb,O,(OEt),, (HOEt), + 14Sb(OEt), Between h=0.8 and 1.33, one obtains a mixture of Ni,Sb,O,(OEt),,(HOEt), Ni,Sb,0,(OEt),,(HOEt)4 and Sb(OEt),, while for h=1.33 the solution contains only Ni6Sb40,(0Et),,( HOEt), and Sb(OEt),.The reaction that occurs at h= 1.33 can thus be written: 3Ni,Sb,(OEt),6 -I-4H2O +Ni$b404(OEt)16( HOEt), + 8Sb(OEt),+4EtOH Ni,Sb40,(0Et),,(HOEt), was found for h values up to ca. 1.5. Between h= 1.5 and h= 3 the crystallization yield was low (20-30%) with a maximum around h =2.5. The composi- tion of the crystals formed in this interval is not known, but SEM-EDS analysis indicates that they have an Ni: Sb ratio of 1:3. The comparatively large Sb content of these crystals suggests that one or more Ni-rich compounds ought to occur within this h range. The IR studies of non-crystalline materials formed within the range indicate the presence of another compound in relatively large quantities. So far we have not been able to isolate and characterize it.In the range 3 <h <5, NiSb,O,(OEt), could be crystallized with maximum yield at 1281 h=3.5. At h36, a sol is formed which yields a gel upon evaporation of the solvent. The hydrolysis pathway of Ni2Sb,(0Et),, in toluene-etha- no1 (2: 1) is identical to that found for toluene-ethanol (4: 1) solutions, but in the former case no crystals are formed for 1.5<h<3. In conclusion, our findings so far indicate that the hydrolysis pathway for Ni,Sb,(OEt),, in toluene-ethanol solutions involves the following species: Ni,Sb,(OEt),,+Ni,Sb302(OEt)15(HOEt), +Sb(UEt), -+ Ni,Sb,O, (OE t ),, (HOE t ), +Sb (OEt ), +NiS b303 (OE t)5 + NiSb,O,(OEt), (sol) Hydrolysis Pathway of Ni,sb,(OEt),, in Ethanol The first part of the hydrolysis pathway of Ni,Sb,(OEt),, in ethanol resembles that found for the toluene-thanoh solvents.Thus, the hydrolysis of Ni,Sb,(OEt),, in ethanol in the range 0 <h<1.5 yielded Ni,Sb,O,(OEt),,( HOEt), at lower h values, with a maximum yield at h=0.8, and at higher water contents Ni6Sb40,(0Et),,(HOEt), in a maximum yield at h=1.33. In the range 1.5<h<3 no compounds could be isolated or identified. In the range h<3 the Ni-containing alkoxides are very sparingly soluble. For 3.0 <h <5 Ni6Sb140,,(OEt),,( HOEt), is formed, with maximum yield at h=4. At even higher h values, ha6, a sol of the nominal composition NiSb,O,(OEt), is formed. The hydrolysis pathway for Ni,Sb,(OEt),, in ethanol involves the following species: Ni,Sb,(OEt),,+Ni,Sb,02(OEt)15(HOEt), +Sb(OEt),+ Ni,Sb,O,(OEt),,( HOEt), +Sb(OEt),+ Ni,Sb,,O,, (OEt),, (HOEt), -+NiSb,O, (OEt), (sol I Conclusions The hydrolysis pathways of Ni,Sb,(OEt),, in toluene.ethanol and toluene-thanol mixtures have been surveyed. Six inter- mediate compounds were identified. It was found that Ni-rich Ni-Sb oxoethoxides were formed in almost quantitative yields together with Sb(OEt), for small additions of H,O. This indicates that Ni2+ ions are more susceptible to hydrolysis than Sb3+ ions. When more water was added (h> 1.5) free Sb(OEt), was no longer observed, neither were ariy of its hydrolysis products. Instead, various Ni-Sb oxoethoxides were formed.It was also found that the hydrolysis pathways differ somewhat depending on the solvent used, but sols were formed for ha6 in all solvents used. Note also that in most cases the molecular structures of the Ni-Sb oxoethoxides in the solid and dissolved state appeared to be similar. The formation of different Ni-Sb oxoethoxides upon hydrolysis makes new compounds with different Ni :Sb ratios accessible for further processing. Many of these corn pounds, especially those formed at low h values, are obtained in high yield. Single-crystal structural studies of some of the intermediates formed have already been accomplished, while corresponding studies for the others are in progress. The structural features of the Ni-Sb oxoethoxides, obtained in conjunction with the hydrolysis pathways described above, will be discussed in a forthcoming article.', The stimulating and fruitful collaboration with Prof.R. Norrestam and Dr. U. Bemm is gratefully acknowledged. 1282 J. MATER. CHEM.. 1994, VOL. 4 This study was financially supported by the Swedish National Research Science Council. 15 16 U. Bemm, R. Norrestam, M. Nygren and G. Westin, Inorg. Chem., 1992,31,2050. U. Bemm, R. Norrestam, M. Nygren and G. Westin, Inorg. Chem., 1993,32,1597. References 17 18 19 U. Bemm, R. Norrestam, M. Nygren and G. Westin, manuscript in preparation. U. Bemm and R. Norrestam, manuscript in preparation. Haslam, U.S. Pat. 2 839 554, 1958; Chem. Ahstr., 1959,53, 1144e. 1 2 3 4 5 6 7 8 9 10 11 12 K.G. Caulton and L. G. Hubert-Pfalzgraf, Chem. Rev., 1990, 90,969. M. Guglielmi and G. Carturan, J. Non Cryst. Solids, 1988,100, 16. L. G. Hubert-Pfalzgraf, New J. Chem., 1987, 11,663. C. Sanchez and J. Livage, New J. Chem., 1990,14,513. G. Yi and M. Sayer, Ceram. Bull., 1991,70,1173. D. C. Bradley, Chem. Rev., 1989,89,1317. S. Livage, M. Henry and C. Sanchez, Prog. Solid State Chem., 1988, 18,249. C. J. Brinker and G. W. Scherer, Sol-Gel Science, Academic Press, New York, 1990. V. W. Day, T. A. Eberspacher, W. G. Klemperer, C. W. Park and F. S. Rosenberg, J. Am. Chem. SOC., 1991,113,8190. Y. N. Chen, W. G. Klemperer and C. W. Park, Better Ceramics through Chemistry V, Proc. Material Research SOC. Symp. 1992, in the print. R. Schmid, A. Mosset and J. Galy, J. Chem. Soc., Dalton Trans., 1991,1999. C. D. Chandler, C. Roger and M. J. Hampden-Smith, Chem. Rev., 1993,93,1205. 20 21 22 23 24 25 26 27 28 29 30 Handbook of Chemistry and Physics, ed. R. C. Weast, CRC Press, Boca Raton, 67th edn., 1986-1987, p. D-15. A. I. Yanovsky, M. I. Yanovskaya, V. K. Limar. V. G. Kessler, N. Ya. Turova and Y. T. Struchkov, J. Chem. SOC., Chem. Commun., 1991,1605. N. Ya Turova, V. G. Kessler and S. I. Kucheiko, Polyhedron, 1991, 10, 2617. T. Athar, R. Bohra and R. C. Mehrotra, Main Group Met. Chem., 1987, 10, 399. T. Athar, R. Bohra and R. C. Mehrotra, Synth. React. Inorg. Met.- Org. Chem., 1989,19,195. B. P. Baranwal and R. C. Mehrotra, Aust. J. Chem., 1980,33, 37. R. Jain, A. K. Rai and R. C. Mehrotra, Z. Nutuforsch., Teil B, 1985,40,1371. R. C. Mehrotra and J. Singh, Can. J. Chew., 1984.62,1003. G. Garg, R. K. Dubey, A. Singh and R. C. Mehrotra, Polyhedron, 1991,10,1733. U. Bemm, R. Norrestam, M. Nygren and G. Westin, manuscript in preparation. R. Papiernik, L. G. Hubert-Pfalzgraf, J-C. Daran and Y. Jeannin, J. Chem. SOC., Chem. Commun., 1990,695. 13 G. Westin and M. Nygren, J. Muter. Sci., 1992,27, 1617. 31 C. Svensson, Acta Crystallogr., Sect. B, 1975,31,2016. 14 U. Bemm, R. Norrestam, M. Nygren and G. Westin, Acta Crystallogr., Sect. C, submitted. Paper 4/01289E; Received 3rd March, 1994