J O U R N A L O F C H E M I S T R Y Materials Gas-phase formation of zinc/cadmium chalcogenide cluster complexes and their solid-state thermal decomposition to form II–VI nanoparticulate material Nigel L. Pickett,a† Steven Lawson,a W. Gregor Thomas,a Frank G. Riddell,a Douglas F. Foster,a David J. Cole-Hamilton* and John R. Fryerb aSchool of Chemistry, The University of St. Andrews, St. Andrews, Fife, UK KY16 9ST.E-mail: djc@st-and.ac.uk bDepartment of Chemistry, Glasgow University, Glasgow, UK G12 8QQ Received 14th August 1998, Accepted 30th September 1998 Gas-phase reactions between R2Zn (R=Me and Et) and tBuSH produce cluster complexes of the type [RZnStBu]n. These clusters, along with [MeZnStBu(py)]2 (py=pyridine), have been characterised by 13C{1H} solid-state NMR. On heating to 100 °C in the solid-state, the complexes [MeZnStBu]5 and [MeZnStBu(py)]2 release dimethylzinc (Me2Zn) to form the zinc bis(thiolate) compound, [Zn(StBu)2]n, with further heating (>200 °C) leading to the formation of ZnS.The ethyl analogue, [EtZnStBu]5, does not lose Et2Zn on heating and thermogravimetric analysis (TGA) suggests a diVerent decomposition pathway, one which mainly involves loss of the organic moieties without the concurrent loss of volatile Zn or S compounds, although ZnS is again the final thermal decomposition product.The decomposition of the involatile pentamers, [MeZnStBu]5 and [EtZnStBu]5, and the dimer, [MeZnStBu(py)]2, proceeds at higher temperature (200–350 °C) to give agglomerates of ME nanoparticulate material, with the individual particles having diameters of 2–20 nm in all cases.The mechanistic pathway by which these clusters decompose appears to be highly dependent upon the R group (Me or Et) present within the cluster. Preliminary results suggest that complexes of the type [RMEtBu]n are also produced from the gas-phase reactions of Me2Zn with tBuSeH and from Me2Cd with tBuSH.others,11 allowed us to propose a mechanism by which particle Introduction nucleation, growth, and suppression of growth by pyridine The use of metal organic vapour phase epitaxy (MOVPE) to can be explained.10c produce single-crystal layers of wide band gap 12–16 (II–VI ) In the absence of a Lewis base (pyridine), particle growth materials (ZnS, ZnSe, CdS and CdSe) is now an established occurs by initial association between R2M and H2E leading technique.The most common method still employs the ther- to the formation of [RMEH]2, which grows rapidly into mally controlled reaction between a group 12 dialkyl com- clusters of the type, HEn(ME)xMRm, consisting of a central pound, R2M (R=Me, Et etc.; M=Zn and Cd) and a group ME core with E–H and M–R fragments on the surface of the 16 hydride (H2S or H2 Se).1 However, a premature reaction particles, which are suspended within the carrier-gas.Surface (prereaction) which takes place between the two precursors in bound M–R fragments react with gas-phase H2E leading to the cold zone of the growth cell, upstream of the heated the elimination of RH and formation of new E–H surface substrate, can cause adverse eVects on the properties of the bound fragments.These E–H sites in turn react with gasgrown epilayers. It has been shown that the introduction of a phase R2M leading to the formation of new M–R sites on the s-donor compound to the reaction system can lead to a continuously growing particles. Termination of particle growth dramatic reduction of the prereaction. Compounds which have may occur by pyridine binding to the reactive surface M–R been investigated as prereaction suppressants include: 1,4- sites, preventing their reaction with H2E by a blocking dioxane,2–4 thioxane,4 triethylamine,5,6 1,3,5-trimethylhexahy- mechanism.10c dro-1,3,5-triazine,6 N,N,N¾,N¾-tetramethylethane-1,2-diamine,7 One step of this growth mechanism involves the reaction of N,N,N¾,N¾-tetramethyldiaminomethane8 and pyridine.9 surface bound E–H fragments with R2M to eliminate RH.While investigating the nature of the prereaction, recent Thus, one other approach that can be used in an eVort to studies of ours10 have shown that the gas-phase reactions that eliminate prereactions in the growth of II–VI semiconductors occur between H2S/H2Se and Me2Cd/Me2Zn result in the is to reduce or eliminate E–H bonds in the group 16 precursors.formation of chalcogenide deposits (ZnS, ZnSe, CdS and A number of research groups have, in their eVorts to control CdSe), with the deposits consisting of poorly formed nanocrys- II–VI epitaxial growth, used a range of dialkyl sulfides/ talline material of the hexagonal phase within the size range selenides and thiols/selenols as the group 16 precursor.The 10–100 nm. The addition of small amounts of pyridine to the most promising alternative group 16 precursors are those reaction system greatly improves the crystallinity exhibited which are bound to tert-butyl groups i.e. tBu2S12 and tBu2Se,13 by the particles, while addition of larger quantities of pyrialong with tBuSH14,15 and tBuSeH.16 dine retains the improved crystal quality whilst also lead- Although successful II–VI growth has been achieved with ing to a decrease in particle size.The ability of pyridine to tBuSH in combination with Me2Zn, Et2Zn and Me2Cd,14,15 it influence the particle size decreases in the order ZnS> has been suggested that the prereaction is only totally elimin- CdS>CdSe>ZnSe. These results, along with the work of ated when Et3N is added to the growth system.17 Herein, we report that in the gas-phase tBuEH and R2M (R=Et, Me; M=Zn, Cd; E=S, Se) react to form zinc/cadmium chalcogen- †Current address: School of Chemistry and Biochemistry, and School ide cluster complexes of the type [RMEtBu]n.The formation, of Materials Science and Engineering and Molecular Design Institute, Georgia Institute of Technology, Atlanta, Georgia 30332–0400, USA.characterisation and solid-state thermal decomposition of these J. Mater. Chem., 1998, 8, 2769–2776 2769clusters is described and discussed, with the goal of advancing fractional distillation at normal pressure through a high eYciency (30 theoretical plates) SpaltrohrA (Fisher Scientific the understanding of the epitaxial growth process using both conventional and single-source precursors.18 UK) distillation column.Bp 79–80 °C. Yield after purification 48 g (70%). 13C{1H} NMR (C6D6), d 36.06 (s, SeC(CH3)3) and 38.77 (s, SeC). 1H NMR (C6D6), d 0.10 (s, 1H, SeH) and Experimental 1.40 [s, 9H, SeC(CH3)3]. General (C2D5)2Zn. A Grignard solution of CD3CD2MgBr in Microanalytical data were obtained at the University of St. 100 cm3 of Et2O was prepared from 23 g of CD3CD2Br Andrews. Analysis of samples by powder X-ray diVraction (0.20 mol) and 6 g of Mg turnings (0.25 mol). This Grignard (PXRD) were carried out on a Sto�e STADI/P diVractometer solution was added dropwise to a rapidly stirred solution of using Cu-Ka radiation with data collected in the transmission 13.0 g of ZnCl2 (0.095 mol) in 100 cm3 of Et2O at a rate mode.Transmission electron micrographs (TEM) were suYcient to maintain a steady reflux (ca. 30 min). The resulting obtained at the University of St. Andrews on a Phillips EM suspension was externally heated to reflux for 1 h after com- 301 microscope at 80 keV and at Glasgow University on a plete addition of the Grignard reagent.All volatiles were JEOL 1200 EX operated at 120 keV (point resolution of collected by trap-trap distillation in vacuo into a -196 °C cold 0.3 nm) or on a ABT 002B operated at 200 keV (point trap and the (C2D5)2Zn subsequently purified by fractional resolution of 0.18 nm). Particle sizes were determined by direct distillation. Bp 118 °C. Yield 11.4 g, 90%. 13C{1H} NMR measurement of individual particles from the transmission (C6D6), d 5.74 (qnt, JD–C=18.5 Hz, ZnCD2CD3), 9.24 (spt, electron micrographs. 13C{1H} solid-state NMR spectra were JD–C=19.0 Hz, ZnCD2CD3). 2H NMR (C6H6–C6D6), d 0.27 obtained on a Bru�ker MSL 500 spectrometer using CPMAS (br s, 2 2H, ZCD3), 1.20 (br s, 3 2H, ZnCD2CD3). accumulation techniques with chemical shifts referenced to the CH2 resonance (d 38.56) of an external adamantane sample.[MeZnStBu]5. This compound was prepared according to Solution NMR data were recorded on a Bru�ker Associates the literature procedure,20 from the low temperature (-78 °C) AM300 spectrometer operating in the Fourier transform mode reaction between Me2Zn and tBuSH in light petroleum. with (for 13C) noise proton decoupling.The 13C{1H} and 1H Yielding, after work-up, a white powder of the product. NMR spectra were run in deuteriated solvents for the lock Found: C, 35.29; H, 7.65; C25H60S5Zn5 requires C, 35.41; H, signal, as given for each spectrum, with chemical shifts in ppm 7.13%. Solid-state 13C{1H} NMR, d-8.20 (s, ZnCH3),-6.76 to high frequency of tetramethylsilane (TMS) as the internal (s, ZnCH3), -5.78 (s, 2ZnCH3), -0.84 (s, ZnCH3), 35.39 [s, reference.The 2H NMR spectra were run in non-deuteriated SC(CH3)3], 36.09 [s, SC(CH3)3], 36.43 [s, SC(CH3)3], 36.50 [s, solvent, as given for each spectrum, with a few drops of the SC(CH3)3], 36.80 [s, SC(CH3)3], 46.64 (s, SC), 50.83 (s, SC), deuteriated analogue to act as both the lock signal and internal 51.59 (s, SC), 52.45 (s, SC), 52.74 (s, SC).reference—the 2H NMR signals were then back-referenced to TMS by using the known shift of the solvent relative to TMS. [MeZnStBu(py)]2. This compound was prepared according Simultaneous thermal analyses [STA, thermogravimetric to the literature procedure,20 from the low temperature analysis—diVerential temperature analysis (TGA–DTA)] were (-78 °C) reaction between Me2Zn, tBuSH and pyridine in performed on a ‘TA Instruments SDT 2960 Simultaneous light petroleum.Yielding, after work-up, a white powder of DTA-TGA’ instrument. the product. Found: C, 47.84; H, 6.83; N, 4.99; C20H34S2N2Zn2 Dry oxygen free nitrogen, helium and argon (BOC), purified requires C, 48.30; H, 6.89; N, 5.63%. Solid-state 13C{1H} by passing through two consecutive columns (2.5 ×80 cm) NMR, d -7.29 (s, 2ZnCH3), 36.77 [s, SC(CH3)3], 37.92 [s, packed with Cr2+ on silica, were used as the carrier gases and SC(CH3)3], 43.42 (s, SC), 43.91 (s, SC), 124.14 (s, pyC), as the inert atmospheres under which all preparations and 124.91 (s, pyC), 138.69 (s, pyC), 149.64 (br s, 2pyC).manipulations were carried out. Greaseless joints and taps were employed and manipulations were carried out using Gas-phase preparation of Zn and Cd chalcogenide cluster standard Schlenk line and catheter tubing techniques.complexes Dimethylcadmium, dimethylzinc and diethylzinc were prepared and purified as described previously.19 ZnCl2, Mg Prereaction experiments were conducted at just above turnings, tBuMgCl (2.0 mol dm-3 in Et2O), C2D5OH, pyridine atmospheric pressure (101.350 Pa) using an experimental setand tBuSH were purchased from Aldrich and the pyridine and up as schematically illustrated in Fig. 1. Carrier-gas containing tBuSH distilled from CaH2 prior to use. Amorphous Se powder specific gas-phase concentrations of reactants [Me2Cd, Me2Zn, (mesh size <325) was purchased from Johnson Matthey.Et2Zn and (C2D5)2Zn together with tBuSeH or tBuSH] were C2D5OH was transformed into C2D5Br by a standard method allowed to meet at the same point along a horizontal quartz on treatment with HBr. Light petroleum (bp 40–60 °C) and tube. Total gas flow through the apparatus was in the range diethyl ether were dried by distillation from sodium 300–450 cm3 min-1 and the gas-phase concentration of indidiphenylketyl and degassed prior to use.vidual reactants in the range 1×10-3–1×10-2 mol dm-3. EZuent gases containing unused reactants and any volatile Preparation of precursors products were allowed to pass through neutralising solutions followed by an industrial scrubber before being released into tBuSeH. A standard solution of tBuMgCl in Et2O (250 cm3, a fumehood.For specific experiments, eZuent gas containing 2.0 mol dm-3, 0.5 mol) was further diluted by the addition of any volatile products was diverted via a gas sampling port to 750 cm3 Et2O. Selenium powder (39.5 g, 0.50 mol ) was added a gas chromatograph (HP 5890)—mass spectrometer (HP in small batches over 2 h to the rapidly stirred Grignard 5972 series mass selective detector) for analysis (GC-MS).solution to form a voluminous white precipitate of tBuSeMgCl. The suspension was stirred overnight before being cooled in an ice bath and hydrolysed by the cautious addition of aqueous [EtZnStBu]5. Using the experimental set-up as described above, the gas-phase mixing of stoichiometric quantities of HCl (600 cm3, 1moldm-3, 0.6 mol). The clear pale-yellow organic layer was separated and the lower colourless aqueous Et2Zn and tBuSH produced a fine white powder along the entire length of the quartz tube.Attempts to obtain a satisfac- layer extracted with 3×100 cm3 of Et2O. The combined organic extracts were dried over MgSO4 followed by CaH2. tory elemental analysis were hampered by the sample constantly losing weight when exposed to air owing to hydrolysis.Finally, the tBuSeH, a clear, colourless and slightly light sensitive liquid, was separated from Et2O and purified by Solid-state 13C{1H} NMR, d 5.85 (s, ZnCH2CH3), 6.30 2770 J. Mater. Chem., 1998, 8, 2769–2776Fig. 1 Schematic diagram of the experimental set-up used both in the preparation and solid-state thermolysis of metal chalcogenides cluster complexes.(s, ZnCH2CH3), 7.19 (s, ZnCH2CH3), 7.69 (s, ZnCH2CH3), before thermolysis, 1.53 g; weight of residue, 0.32 g; 79.1% weight loss. [EtZnStBu]5: weight before thermolysis, 3.15 g; 8.63 (s, ZnCH2CH3), 14.13 (s, ZnCH2CH3), 14.61 (br s, 2ZnCH2CH3), 14.77 (br s, 2ZnCH2CH3), 35.94 [s, SC(CH3)3], weight of residue, 1.45 g; 54.0% weight loss. 36.21 [s, SC(CH3)3], 36.43 [s, SC(CH3)3], 36.75 [s, SC(CH3)3], 36.92 [s, SC(CH3)3], 45.47 (s, SC), 50.18 (s, SC), 50.46 (s, Formation of [Zn(StBu)2]n from [MeZnStBu(py)]2 SC), 50.93 (s, SC), 51.34 (s, SC).Heating [MeZnStBu(py)]2 to 100 °C in vacuo (ca. 1 h) liberated a colourless liquid, which was collected in a cold trap. This [MeZnSetBu]n. The gas-phase mixing of stoichiometric left a dirty white solid. The colourless liquid distillate was quantities of Me2Zn and tBuSeH produced a white deposit.shown to be [Me2Zn(py)2]: 1H NMR, d -0.70 (s, ZnCH3), Attempts to obtain a satisfactory elemental analysis were again 7.32 (m, pyH), 7.72 (m, pyH), 8.58 (m, pyH). 13C{1H} NMR, hampered by the sample constantly losing weight when exposed d -12.34 (s, ZnCH3), 124.07 (m, pyC), 136.69 (m, pyC), to air owing to hydrolysis.The solid-state 13C{1H} NMR 149.44 (m, pyC). The white solid was identified as impure spectrum consisted of broad peaks which merged into one [Zn(StBu)2]n: Found: C, 37.55; H, 7.47; N, 0.24; C8H18S2Zn another, but which had a similar appearance to those of requires C, 39.41; H, 7.44%. Solid-state 13C{1H} NMR, d 36.96 [MeZnStBu]5. 1H NMR (CD2Cl2), d -0.40 (s, ZnCH3), 1.70 [s, SC(CH3)3], 49.20 (s, SC), 50.08 (s, SC).[s, SC(CH3)3]. [MeCdStBu]n. The gas-phase mixing of stoichiometric Results and discussion quantities of Me2Cd and tBuSH produced a white deposit. The solid-state 13C{1H} NMR again gave broad peaks which Synthesis and characterisation of Zn and Cd chalcogenide cluster complexes merged into one another, but which had a similar appearance to those of [MeZnStBu]5.Solid-state 13C{1H} NMR, d -9.53 Using the system described in the experimental section, the (br, CdCH3), -5.81 (br, CdCH3), 37.44–39.31 [br, gas-phase mixing at room temperature of stoichiometric SC(CH3)3], 48.96–50.29 (br, SC). amounts of Et2Zn and tBuSH (He carrier gas) resulted in the formation of a fine white powder deposit. GC-MS analysis of Solid-state thermal decomposition of Zn chalcogenide cluster the eZuent carrier-gas confirmed that ethane is the only complexes gaseous by-product of this reaction.When Et2Zn was replaced by the deuteriated analogue, (C2D5)2Zn, the ethane produced A similar experimental set-up to that used above in the preparation and collection of prereaction deposits was was pure D5H, formed by protonation of (C2D5)2Zn by the acidic S–H of tBuSH.Upon exposure to air, the white powder employed (Fig. 1), with known amounts of metal chalcogenide cluster complexes contained inside a ceramic crucible placed gave oV a strong smell of thiol, presumably due to hydrolysis rather than from any remaining tBuSH. Without further inside the quartz tube and at the centre of the furnace.A carrier gas (Ar), at just above atmospheric pressure (101.350 purification, solid-state 13C{1H} NMR analysis of the powder gave a spectrum [Fig. 2(a)] which exhibits five diVerent sets of Pa), was passed along the quartz tube at a flow rate of 200 cm3 min-1 while the furnace was heated to a maximum ethyl resonances along with five diVerent tBu resonances. These results confirmed the presence of at least five diVerent temperature of 600 °C.As in the case of the preparation of prereaction deposits, eZuent gas containing unused reactants environments for both Zn and S within the deposit and the probable identity of the complex as the thiolate pentamer, and any volatile products was allowed to pass through neutralising solutions followed by an industrial scrubber before [EtZnStBu]5.A non-quaternary 13C{1H} suppression spectrum (which greatly reduces the intensity of CH and CH2 being released into a fumehood. For specific experiments, eZuent gas containing any volatile products was diverted via resonances) [Fig. 2(b)] shows the methyl signals of the ethyl groups, ZnCH2CH3, to be down field of the methylene, a gas sampling port to the GC-MS.[MeZnStBu]5: weight before thermolysis, 0.52 g; weight of ZnCH2CH3, signals due to the shielding eVect of the zinc. Attempts to establish the structure of the complex by X-ray residue, 0.16 g; 69.2% weight loss. [MeZnStBu(py)]2: weight J. Mater. Chem., 1998, 8, 2769–2776 2771E M E M R E M E M R E M tBu R tBu R tBu R tBu tBu Fig. 3 The basic structure of [RMEtBu]5 (M=Zn, Cd; E=S, Se; R=Me, Et).of the tert-butyl groups to be significantly shifted to higher magnetic field strength relative to the shifts of the other four. This unique resonance probably arises from the tert-butyl group in the ‘handle’ section of the molecule and the shift to higher field is probably due to the sulfur being bonded to two rather than three adjacent zinc atoms (resulting in a higher shielding eVect).Similarly, in [MeZnStBu]5, one of the ZnCH3 carbon resonances is clearly distinct from the others, with a shift this time to lower field, and again, this is probably the resonance of the ZnCH3 group in the ‘handle’ section of the molecule, which is deshielded due to the zinc atom being bonded to only two adjacent sulfur atoms rather than three as for all the other zinc atoms.White powders were also obtained from the gas-phase reactions of Me2Cd with tBuSH and from Me2Zn with tBuSeH. Although both compounds aVord poorly resolved solid-state 13C{1H} NMR spectra, the over-all shapes of the spectra are again similar to that of [MeZnStBu]5, suggesting that cluster complexes of the type [MeMEtBu]n are formed. Although the value of n cannot be established from the spectra alone, it has been proposed from NMR data that n=4 for [MeCdStBu]n when prepared from solution.21,22 As discussed in the Introduction, for MOVPE growth systems employing R2M and H2E (M=Zn, Cd; R=Me, Et; E=S, Se), pyridine can suppress the growth of particulate material by binding to the surface metal atom sites of particles growing within the gas phase.By a similar process, O’Brien and coworkers have reported that the reaction between Me2Zn, tBuSH and pyridine in solution (benzene) aVords the dinuclear Fig. 2 Solid-state 13C{1H} NMR spectra of (a) [EtZnStBu]5, (b) non- pyridine adduct complex, [MeZnStBu(py)]2, rather than the quaternary suppression spectrum of [EtZnStBu]5 and (c) [MeZnStBu]5. pentamer, [MeZnStBu]5, which forms in the absence of pyridine.11 These results confirm that for all MOVPE growth systems crystallography were thwarted due to our inability to grow X- which involve the gas-phase mixing of thiols/selenols with ray quality crystals. However, the solid-state 13C{1H} NMR dialkylzinc/cadmium compounds, a prereaction (detected or data of the ethyl complex suggested the complex to be isostruc- otherwise) will occur in the cold zone of the reactor.This will tural with the methyl pentamer, [MeZnStBu]5, first synthesised result in the depletion (or complete elimination) of the original by Coates and Ridley,20 and more recently structurally charac- precursors from the gas phase before the gas stream carrying terised by O’Brien and coworkers.11 Thus, in an attempt to the precursors reaches the heated substrate.Moreover, confirm the likely structure of the ethyl complex [EtZnStBu]5, Lovergine et al. have reported17 that the addition of triethyl- [MeZnStBu]5 was prepared by the literature method20 and amine to the growth system, when forming ZnS epilayers from characterised by solid-state 13C{1H} NMR spectroscopy. tBuSH and Me2Zn, although avoiding the formation of solid The solid-state 13C{1H} NMR spectrum of [MeZnStBu]5 prereaction deposits, still leads to non-homogeneous epitaxial does indeed correlate well with that of the ethyl complex, growth owing to some form of depletion of the two precursors essentially exhibiting five diVerent tBu resonances and four from the gas phase (other than through the epitaxial growth methyl resonances in a 1525151 ratio [Fig. 2(c)]. Interestingly, process itself ). This observation may be explained by the [MeZnStBu]5 is fluxional in solution (C6D6) with the 1H NMR formation of [MeZnStBu(Et3N)]2 dimers which then either spectrum exhibiting single resonances for all five ZnCH3 precipitate onto the substrate, or initiate gas-phase particulate groups and for all five tBu groups while the 13C{1H} solution growth, with the particulate matter then being precipitated spectrum of [MeZnStBu]5 also exhibits only one ZnCH3 onto the substrate.It is also possible that such dimers are not resonance and one tert-butyl resonance.11 This fluxionality is formed and that Et3N behaves like pyridine in terminating quenched in the solid-state, at least up to 25 °C.Thus, the particulate growth at an early enough stage that prereaction NMR data strongly support the view that in the solid-state deposits are not visible.10 [EtZnStBu]5 and [MeZnStBu]5 are isostructural, as schematically shown in Fig. 3. The structure consists of a cubic Thermogravimetric analysis and solid-state thermal arrangement of zinc and sulfur atoms with one edge of the decomposition of zinc chalcogenide cluster complexes cube broken open by the addition of the extra Zn–S unit, resembling a supermarket trolley.The solid-state 13C{1H} Because of the current interest in the use of metal chalcogenide complexes as ‘single-source’ precursors, both in MOVPE NMR data of both the methyl and ethyl complexes (Fig. 2) show the resonance for one of the quaternary carbon atoms growth and nanoparticle material preparation,18 as well as the 2772 J.Mater. Chem., 1998, 8, 2769–2776Table 1 Thermal analytical and weight loss data for zinc sulfide cluster complexes Step 1a Step 2a Residue (%)b Weight loss (%)b Weight loss (%)b Complex Tonset/ °C Obs. Calc.c Tonset/ °C Obs. Calc.d TGA Thermolysis Calc.e [MeZnStBu]5 110 24 28 230 45 43 31 31 29 [MeZnStBu(py)]5 70 46 51 230 29 29 26 21 20 [EtZnStBu]5 48 46 27f aFrom TGA (thermogravimetric analysis).bWeight losses and residual weights all relative to initial weight of complex. cLoss of half the Zn content as Me2Zn. dLoss of half the S content as tBu2S. eLoss of organics+half the ZnS content. fLoss of organics only, gives a calculated residue of 53%.role that zinc/cadmium chalcogenide cluster complexes might Me2Zn or [Me2Zn(py)2], and a second sharp endotherm with an onset temperature of ca. 230 °C corresponding to the loss play in MOVPE growth whilst employing separate group 12 and 16 precursors, it is important to understand the thermal of ‘tBu2S’. The residue consists of ZnS with a mass corresponding to one half of the total ZnS mass present in the original properties of complexes of the type: [EtZnStBu]5, [MeZnStBu]5 and [MeZnStBu(py)]2.Simultaneous thermal analyses sample. Macroscopic thermolysis also gave residual weights consistent with this sequence of reactions (Table 1). (STA), thermogravimetric analysis–diVerential temperature analysis (TGA–DTA) carried out on [MeZnStBu]5 and GC-MS analysis of the gas-phase products from the macroscopic thermolysis of [MeZnStBu]5, under Ar flow, [MeZnStBu(py)]2, indicates that these complexes demonstrate similar thermal behaviour [Table 1 and Fig. 4(a) and (b)]. confirmed that Me2Zn is released but that tBu2S is not the only S containing volatile decomposition product, 2-methyl- There is a broad endotherm at ca. 100 °C with a weight loss corresponding to the loss of all the zinc bound Me groups, as propene and tBuSH are also produced.Thus, at 100 °C, a trace amount of methane was seen while at 150 °C, traces of methane and 2-methylpropene along with large quantities of Me2Zn were detected. At 200 °C, large quantities of 2-methylpropene and tBuSH were detected along with traces of tBu2S. For the complex [MeZnStBu(py)]2, at 100 °C, a large quantity of pyridine was detected along with trace amounts of methane and toluene (resulting from the preparative method), while at 150 °C, pyridine, methane and 2-methylpropene (Me2Zn was not detected, see below) were observed, and above 200 °C, 2- methylpropene along with trace amounts of tBu2S and methane were seen.These results for [MeZnStBu(py)]2 do not demonstrate the production of Me2Zn.We propose that this is due to the presence of pyridine which condenses on the walls of the stainless steel tubing used to connect the thermolysis cell to the gas-sampling valve of the GC-MS. This pyridine will complex with any Me2Zn released to the gas-phase, forming the solid adduct complex, [Me2Zn(py)2]. At room temperature, dissociation and revapourisation of [Me2Zn(py)2] is at such a slow rate that gas-phase concentrations of Me2Zn are too low to detect.To confirm that [MeZnStBu(py)]2 decomposes by a similar mechanistic pathway to that of [MeZnStBu]5, [MeZnStBu(py)]2 was heated in the solid state under vacuum to 100 °C. The volatile products were collected in a cold trap and at room temperature gave a viscous liquid, which was shown by 1H NMR to contain Me2Zn and pyridine.Elemental analysis on the remaining solid, without washing or further purification, suggested it to be the complex [Zn(StBu)2]n. This was confirmed by solid-state 13C{1H} NMR as shown in Fig. 5(a). The 13C{1H} NMR of the original complex, [MeZnStBu(py)]2, is shown in Fig. 5(b) for comparison. The complex, [Zn(StBu)2)]n, shows two peaks of almost equal intensity due to the SCtBu carbons showing that both syn and anti orientations of the tBu groups occur with equal probability.Many complexes of this general formula, M(ER)2, including [Zn(StBu)2)]n,23f have previously been prepared and used as potential ‘single-source’ precursors for II–VI materials.23 It has been proposed by Steigerwald and Sprinkle that the use of Me2Cd and R2Te under MOVPE growth conditions results in the in situ formation Cd(TeR)2 which then decomposes to give CdTe.23h This could also be occurring in the growth of ZnS from Me2Zn and tBuSH.Although a detailed mechanistic pathway cannot be fully Fig. 4 Simultaneous thermal analyses (STA, thermogravimetric established it appears, from these results, that the overall analysis—diVerential temperature analysis (TGA–DTA)) of (a) [MeZnStBu]5, (b) [MeZnStBu(py)]2 and (c) [EtZnStBu]5. pathway for the thermal decomposition of [MeZnStBu]5 is as J.Mater. Chem., 1998, 8, 2769–2776 2773Fig. 6 X-Ray powder diVraction pattern of ZnS formed from the thermolysis of (a) [EtZnStBu]5, (b) [MeZnStBu]5 and (c) [MeZnStBu(py)]2. Standard X-ray powder diVraction patterns of (d) hexagonal (wurtzite) and (e) cubic (sphalerite) phases of ZnS are also shown.The peaks at 2h 21 and 24° arise from the Vaseline support. Fig. 5 Solid-state 13C{1H} NMR spectra of (a) the residue remaining after heating [MeZnStBu(py)]2 to 100 °C and (b) [MeZnStBu(py)]2 (the peaks in the range d 55–85 and d 190–220 are spinning side abstraction from the tert-butyl group to form 2-methylpropene bands of the main pyridine resonances at d 120–155).and [EtZnSH] with reductive elimination from the latter aVording the observed ethane and ZnS. Compounds similar to [EtZnSH] are known to give alkane and ZnS upon thermal shown in Scheme 1. This involves a rearrangement of the decomposition. Why this process should occur for [EtZnStBu]5 methyl radicals to liberate the volatile compound Me2Zn, rather than for [MeZnStBu]5 is not clear but demonstrates leaving the zinc-bis(thiolate), [Zn(StBu)2]n, as a residue.how a small diVerence within the precursor can radically alter Further heating leads to the loss of tert-butyl groups, mainly the pathway by which decomposition takes place. as 2-methylpropene and tBuSH along with far smaller amounts of tBu2S.H2S might be expected from the decomposition of Analysis of the solid thermolysis residues tBuSH, but we have shown that the extent of decomposition The macroscopic thermolysis experiments described above of tBuSH is very low below 300 °C.24 For [MeZnStBu(py)]2, produced grey or white residues, which were scraped from the a similar process occurs initially to give [Me2Zn(py)2] and pyrolysis tube and handled in air before being analysed by [Zn(StBu)2]n.powder X-ray diVraction (PXRD) and transmission electron The STA analysis of [EtZnStBu]5 is rather diVerent from microscopy (TEM). PXRD analyses of all the residues show those of the other compounds, since there is no evidence for broad peaks (Fig. 6) and although on first inspection the formation of [Zn(StBu)2]n (no sharp exotherm near 230 °C) diVraction patterns appear to resemble the PXRD of cubic and the decomposition occurs in three stages [Fig. 4(c)]. The (sphalerite) ZnS, previous studies of ours10,25 along with the residual mass obtained from both STA and macroscopic calculations of others26 suggest the residues are more likely to thermolysis (ca. 47%) is between those expected for the simple consist mainly of nanoparticles (responsible for the line broad- loss of the organic moieties (53.1%) and for the type of ening) of distorted hexagonal (wurtzite) phase ZnS (having a mechanism described above for [MeZnStBu]5 (26.6%). In the high number of stacking faults and leading to peak dependent thermolysis of [EtZnStBu]5, below 100 °C, trace amounts of line broadening).However, the observation of a peak at 2h= ethane were detected, at 150 °C, ethane along with larger 33° does suggest that some cubic phase is also present. The amounts of 2-methylpropene were seen, while above 200 °C, diVraction pattern of the residue from [MeZnStBu]5 [Fig. 6(b)] large amounts of 2-methylpropene along with trace amounts is somewhat broader than the others suggesting that the of tBu2S were detected.Although trace amounts of compounds average particle size is smaller. containing sulfur were detected, the mechanistic pathway of TEM images for the residue obtained from [EtZnStBu]5 thermal decomposition involves no major loss of zinc or sulfur. [Fig. 7(a)] show the ZnS to be in the form of circular agglomer- Thus, the decomposition pathway for [EtZnStBu]5 is rather ates 0.5–2 mm in diameter made up of individual ZnS nanopart- diVerent from that for [MeZnStBu]5 in that Et2Zn is not icles in the size range of 5-10 nm, whilst those from observed as a product, but rather ethane and 2-methylpropene [MeZnStBu]5 appear as laths [Fig. 7(b)], which probably arise are the major products. One possible decomposition route as a function of sample preparation.Individual crystals which explains the volatile product formation involves b-H within the aggregates are in the size range 5–20 nm. Although high resolution TEM images of the residue formed from [MeZnStBu]5 [Fig. 7(c)] reveal many of the nanocrystals to be twinned (or to be composed of a number of subcrystals), there are also a number of perfect single crystals within the residue.The particles produced on thermolysis by all three complexes are somewhat diVerent from those formed during gas-phase reactions between R2M and H2S (M=Zn, Cd) where in [MeZnStBu]5 2.5 [Zn(StBu)2] n + 2.5 Me2Zn T > 250 °C 2.5 ZnS + 2.5 tBuSH + 2.5 T = 100 °C general, and especially upon the addition of pyridine, more Scheme 1 Schematic representation of the processes occurring in the solid-state thermolysis of [MeZnStBu]5.homogeneous nanoparticles are formed.10 The particles 2774 J. Mater. Chem., 1998, 8, 2769–2776conditions used, the decomposition of the involatile chalcogenide cluster complexes proceeds at higher temperatures (>200 °C) to give agglomerates of nanometer sized particles of metal chalcogenides.However, the pathway by which decomposition occurs seems to be highly dependent upon the alkyl group within the chalcogenide cluster complex. Acknowledgements We thank the EPSRC for financial support (N.L.P., via a ROPA award, S.L. and D.F.F.) and Ciba Specialty Chemicals for a studentship (W.G.T.). References 1 (a) S. Fujiita, Y. Matsuda and A. Sasaki, J.Cryst. Growth, 1984, 68, 231; (b) A. Yoshikawa, S. Muto, S. Yamaga and H. Kasai, J. Cryst. Growth, 1988, 86, 279; (c) S. Yamaga, A. Yoshikawa and H. Kasai, J. Cryst. Growth, 1988, 86, 252. 2 P. J. Wright, B. Cockayne, A. J. Williams, A. C. Jones and E. D. Orrell, J. Cryst. Growth, 1987, 84, 552. 3 M. J. Almond, M. P. Beer, M. G. B. Drew and D. A. Rice, J. Organomet. Chem., 1991, 421, 129. 4 B. Cockayne, P. J.Wright, A. J. Armstrong, A. C. Jones and E. D. Orrell, J. Cryst. Growth, 1988, 91, 57. 5 P. J. Wright, P. J. Parbrook, B. Cockayne, A. C. Jones, E. D. Orrell, K. P. O’Donnell and B. Henderson, J. Cryst. Growth, 1989, 94, 441. 6 P. J.Wright, B. Cockayne, P. J. Parbrook, A. C. Jones, P. O’Brien and J. R.Walsh, J. Cryst. Growth, 1990, 104, 601. 7 M. J. Almond, M.P. Beer, K. Hagen, D. A. Rice and P. J.Wright, J. Mater. Chem., 1991, 1, 1065. 8 O. Briot, M. DiBlasio, T. Cloitre, N. Briot, P. Bigenwald, B. Gil, M. Averous, R. L. Aulombard, L. M. Smith, S. A. Rushworth and A. C. Jones, Mater. Res. Soc. Symp. Proc., 1994, 340, 515. 9 P. J. Wright, B. Cockayne, P. J. Parbrook, P. E. Oliver and A. C. Jones, J. Cryst. Growth, 1991, 108, 525. 10 (a) N. L. Pickett, D. F. Foster and D. J. Cole-Hamilton, J. Mater. Chem., 1996, 6, 507; (b) N. L. Pickett, D. F. Foster and D. J. Cole- Hamilton, J. Cryst. Growth, 1997, 170, 476; (c) N. L. Pickett, F. G. Riddell, D. F. Foster, D. J. Cole-Hamilton and J. R. Fryer, J. Mater. Chem., 1997, 7, 1855. 11 M. A. Malik, M. Motevalli, J. R. Walsh, P. O’Brien and A. C. Jones, J. Mater. Chem., 1995, 5, 731. 12 T. Obinata, K. Uesugi, G. Sato, I. Suemune, H. Machida and N. Shimoyama, Jpn. J. Appl. Phys., 1995, 34, 4143. 13 (a) W. Kuhn, A. Naumov, H. Stanzl, S. Bauer, K. Wolf, H. P. Wagner, W. Gebhardt, U. W. Pohl, A. Krost, W. Richter, U. Dumichen and K. H. Thiele, J. Cryst. Growth, 1992, 123, 605. 14 (a) D. N. Armitage, H. M. Yates, J. O. Williams, D. J. Cole- Hamilton and I.L. J. Patterson, Adv. Mater. Opt. Electron., 1992, 1, 43; (b) D. F. Foster, I. L. J. Patterson, L. D. James, D. J. Cole- Hamilton, D. N. Armitage, H. M. Yates, A. C. Wright and J. O. Williams, Adv. Mater. Opt. Electron., 1994, 3, 163. 15 K. Nishimura, K. Sakai, Y. Nagao and T. Ezaki, J. Cryst. Growth, 1992, 117, 119. 16 K. Nishimura, Y. Nagao and K. Sakai, J. Cryst. Growth, 1993, Fig. 7 TEM images of (a) ZnS from the thermolysis of [EtZnStBu]5, 134, 293. (b) ZnS from the thermolysis of [MeZnStBu]5 and (c) as (b) but at 17 N. Lovergine, M. Longo, C. Gerardi, D. Manno, A. M. Mancini higher magnification. and L. Vasanelli, J. Cryst. Growth, 1995, 156, 45. 18 (a) P. O’Brien and R. Nomura, J. Mater. Chem., 1995, 5, 1761; (b) P. O’Brien, Precursors for Electronic Materials, in Inorganic produced from these chalcogenide cluster complexes are similar Materials, ed.D. W. Bruce and D. O’Hare, John Wiley, London, to, but larger than, those reported by Steigerwald and 1992, p. 499; (c) M. Bochmann, Chem. Vap. Deposition, 1996, 2, 85. coworkers from the solid-state thermolysis of Zn(SPh)2 and 19 D. F. Foster and D. J. Cole-Hamilton, Inorg. Synth., 1997, 31, 29. [Zn(SPh)2(depe)] [depe=1,2-bis(diethylphosphino)ethane].23i 20 G. E. Coates and D. Ridley, J. Chem. Soc., 1965, 1870. 21 G. E. Coates and A. Lauder, J. Chem. Soc. A, 1966, 264. 22 J. D. Kennedy and W. McFarlane, J. Chem. Soc., Perkin Trans. 2, Conclusion 1977, 1187. Gas-phase mixing of thiols/selenols (REH) with dialkylzinc/ 23 (a) M. Bochmann, K. J. Webb, M. Harman and M. B. Hursthouse, Angew. Chem., Int. Ed. Engl., 1990, 29, 638; cadmium compounds (R2M) results in the precipitation of (b)M. Bochmann, K. J.Webb, J. E. Hails and D.Wolverson, Eur. zinc/cadmium chalcogenide cluster complexes of general J. Solid State Inorg. Chem., 1992, 29, 155; (c) G. Krau� ter and formula [tBuEMR]n. This demonstrates the importance of W. S. Rees, Jr., J. Mater. Chem., 1995, 5, 1265; (d)M. Bochmann, avoiding the low temperature premixing of such precursor A. P. Coleman and A. K. Powell, Polyhedron, 1992, 11, 507; combinations in the growth of II–VI materials by MOVPE, (e) J. G. Brennan, T. Siegrist, P. J. Carroll, S. M. Stuczynski, as this can lead to gas-phase depletion of precursors and non- L. E. Brus and M. L. Steigerwald, J. Am. Chem. Soc., 1989, 111, 4141; ( f )W. S. Rees, Jr. and G. Krau�ter, J. Mater. Res., 1996, 11, uniformity in the epitaxial growth. Under the reaction J. Mater. Chem., 1998, 8, 2769–2776 27753005; (g) R. D. Schluter, G. Krau� ter and W. S. Rees, Jr., J. Cluster 25 S. W. Haggata, X. Li, D. J. Cole-Hamiltion and J. R. Fryer, J. Mater. Chem., 1996, 6, 1771. Sci., 1997, 8, 123; (h)M. L. Steigerwald and C. R. Sprinkle, J. Am. Chem. Soc., 1987, 109, 7200; (i) J. G. Brennan, T. Siegrist, 26 M. G. Bawendi, A. R. Kortan, M. L. Steigerwald and L. E. Brus, J. Chem. Phys., 1989, 91, 7282. P. J. Carroll, S. M. Stuczynski, P. Reynders, L. E. Brus and M. L. Steigerwald, Chem.Mater., 1990, 2, 403. 24 N. L. Pickett, D. F. Foster, D. Ellis and D. J. Cole-Hamilton, J. Mater. Chem., to be submitted. Paper 8/06421K 2776 J. Mater. Chem., 1998, 8, 2769&ndas