DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 249–253 249 Complexes of lithium tetrahydroaluminate with N,N,N9,N9-tetramethylethane- 1,2-diamine (tmen). Crystal structures of [{Li(tmen)- (AlH4)}2] and [Li(tmen)2][AlH4] and the use of the 6Li-{1H} nuclear Overhauser eVect to study LiAlH4 and LiBH4 in donor solvents Mbolatiana M. Andrianarison, Anthony G. Avent, Miles C. Ellerby, Ian B. Gorrell, Peter B. Hitchcock, J. David Smith * and David R. Stanley School of Chemistry, Physics and Environmental Science, University of Sussex, Falmer, Brighton, UK BN1 9QJ Two crystalline complexes (1 : 1 and 1 : 2) of LiAlH4 with N,N,N9,N9-tetramethylethane-1,2-diamine (tmen) have been isolated. A crystal structure determination of the 1 : 1 complex showed that it formed centrosymmetrical dimers in which [AlH4]2 anions and [Li(tmen)]1 cations are linked by m-hydrogen bridges.Bond lengths and angles within the non-planar eight-membered rings are Al]H 1.55(3), Li]H 1.99(3) Å, H]Li]H 131(1) and H]Al]H 112(1)8 and the exocyclic Al]H distance was 1.53(4) Å.Cryoscopic data indicated that the dimeric structure with Li]H]Al bridges was preserved in benzene solution and measurements of the 6Li-{1H} nuclear Overhauser effect (NOE) showed the presence of Li ? ? ?H]Al interactions. The 1 : 2 adduct [Li(tmen)2][AlH4] 2, which crystallised with an ionic NaCl structure, was converted into 1 by heating under vacuum at 120 8C. The NOE measurements on solutions of 2 in benzene showed the presence of Li ? ? ?H]Al interactions even when a large excess of tmen was present.Aluminium-bound hydrogen was shown to be near to lithium in solutions of LiAlH4 in diethyl ether, tetrahydrofuran, mono- and di-glyme [MeO(CH2)2O(CH2CH2O)nMe, n = 0 or 1]. The NOE measurements on solutions of LiBH4 in Et2O and NMR spectra of partially deuteriated species suggested the formation of contact ion pairs, with Li ? ? ?H4B interactions which were fluxional on the NMR time-scale.Lithium tetrahydroaluminate, LiAlH4, finds widespread use in both inorganic and organic chemistry as a reducing agent and as a hydride or hydroaluminate transfer reagent.1,2 It is usually added to a reaction mixture as a solution in diethyl ether (Et2O) or tetrahydrofuran (thf) and the species in solution have been studied by a variety of techniques, e.g. conductance and ebullioscopic measurements on both Et2O and thf solutions,3,4 vibrational spectroscopy on samples in Et2O, thf and monoglyme,5,6 27Al7–10 and 7Li NMR spectroscopy8,9 on Et2O, thf, mono-, di- and tri-glyme [MeO(CH2)2O(CH2CH2O)nMe (n = 0, 1 or 2)] solutions, and dielectrometry in Et2O.6 Results in Et2O have been interpreted as evidence for ion pairs in dilute solution with multiple-ion formation at higher concentrations. 3–5,8 It has not been established whether the ions within a pair are separated by solvent molecules or whether they are linked by hydrogen bridges as in [(Et2O)2Li(m-H)2AlH2].6 Solvent-separated ion pairs were thought to be the major species in thf 3–5,9 and glyme solvents.5,9,10 In this paper we describe the structural characterisation of N,N,N9,N9-tetramethylethane-1,2-diamine (tmen) complexes of LiAlH4 [{Li(tmen)(AlH4)}2] 1 and [Li(tmen)2][AlH4] 2, and the use of 6Li-{1H} heteronuclear Overhauser effect spectroscopy to study their solutions in benzene.This last technique, which relies on the low quadrupole moment of 6Li and the consequent domination of its relaxation by dipolar interaction with nearby protons, has been used previously in one-dimensional form to study, for example, lithium hydro- [tris(trimethylsilyl)methyl] compounds of boron, aluminium, gallium and indium 11 and in two-dimensional form to study organolithium compounds.12 We also describe 6Li-{1H} nuclear Overhauser effect (NOE) measurements on solutions of LiAlH4 in Et2O, thf, mono- and di-glyme and solutions of LiBH4 in Et2O and compare the conclusions with those derived previously by other methods.Results and Discussion LiAlH4–tmen complexes The reaction between LiAlH4 and 1 equivalent of tmen in thf gave, after recrystallisation from toluene, colourless, airsensitive crystals of the known13 adduct, Li(tmen)AlH4 in good yield. A crystallographic study showed the structure to consist of centrosymmetric dimers 1 (Fig. 1) based on eight-membered rings similar to those observed in the structures of Li[AlH3- {C(SiMe2Ph)3}]?2thf 3,14 Li[AlH3(C6H2Ph3-2,4,6)]?1.5Et2O 4, Li[AlH3(C6H2But 3-2,4,6)]?2thf 5, Li[AlH3{N(SiMe3)2}]?2Et2O 6 15 and LiAlH4?HN(But)CH(But)CH2N(But)H 7, and suggested for that of LiAlH4?HN(But)CH(But)CH]] NBut.16 Discussion of bond lengths and angles (Table 1) is inevitably tentative because of the high standard deviations associated with bonds to hydrogen.However, the average Al]H bond length within the Al2H4Li2 ring of compound 1, 1.55(3) Å, is at the short end of the range found in the related hydroaluminate derivatives 3–714–16 [cf. 1.62(4) in 3, 1.62(4) in 4, 1.57(4) in 5, Fig. 1 Molecular structure of compound 1250 J. Chem. Soc., Dalton Trans., 1998, Pages 249–253 1.62(4) in 6 and 1.56(4) Å in 7]. The average terminal Al]H bond length is 1.53(4) Å so that, as in 3–7, terminal and bridging Al]H bond lengths are similar. The Li]H bond length in 1 [1.99(3) Å] is long [cf. 1.77(4) in 4, 1.78(4) in 6, 1.85(5) in 7, 1.93(4) in 3 and 2.00(5) Å in 5] and the Li]N bond lengths are normal.The ring in 1 is puckered with the two aluminium atoms 0.9 Å above and below, and the two lithium atoms 0.1 Å above and below, the plane defined by the ring hydrogen atoms. Similar puckered rings are found in 3–5 and 7 but the ring in 6 is planar. The endocyclic H]Al]H angle in 1 is 112(2)8. Other H]Al]H angles differ somewhat from this value but the mean [109(2)8] is close to those for 3 [105(1)], 4 [106(2)] and 5 [107(3)8] and not significantly different from the tetrahedral value.The H]Li]H angles [131(1) in 1, 117(1) in 3, 126(2) in 4, 103(2) in 5, 107(2) in 6 and 91(2)8 in 7] show much wider variation. These data suggest that within the dimers the [AlH4]2 ions retain their integrity so that the species 1 is best described as comprising two [AlH4]2 anions and two [Li(tmen)]1 cations. The nonplanarity of the ring in 3 was attributed to the presence of the large alkyl group attached to aluminium 14 but this steric constraint cannot apply to 1.Here the puckering is probably associated with the narrow exocyclic N]Li]N angle imposed by the tmen ligand, which in turn affects the endocyclic H]Li]H angle. The structure of 1 differs from that of [{Li(tmen)- (BH4)}2] in which each [BH4]2 group bridges two lithium centres through three hydrogens (two m, one m3) so that the metal atoms are six- not four-co-ordinate as in 1.17 This difference in structure is probably related to the greater size of aluminium as illustrated by the mean Li ? ? ? B (2.464) and Li ? ? ? Al (2.972 Å) distances.The boron compound did not react with an excess of tmen but the reaction between LiAlH4 and an excess of tmen in thf gave LiAlH4?2tmen 2 in high yield. An X-ray study of the colourless, air-sensitive crystals showed that separate [Li- (tmen)2]1 cations and [AlH4]2 anions (Fig. 2) were packed in a structure of the sodium chloride type (mean Li ? ? ? Al 5.66 Å).Fig. 2 Molecular structure of compound 2 Table 1 Bond lengths (Å) and angles (8) for [{Li(tmen)(AlH4)}2] 1 Al]H(7) Al]H(8) Li]N(2) Li]H(8) H(7)]Al]H(9) H(9)]Al]H(8) H(9)]Al]H(10) N(2)]Li]N(1) N(1)]Li]H(8) N(2)]Li]H(109) 1.55(4) 1.52(3) 2.090(5) 2.04(3) 113(2) 117(2) 111(1) 88.2(2) 114.3(8) 109.3(8) Al]H(9) Al]H(10) Li]N(1) Li]H(109) H(7)]Al]H(8) H(7)]Al]H(10) H(8)]Al]H(10) N(2)]Li]H(8) N(1)]Li]H(109) H(8)]Li]H(109) 1.50(3) 1.59(3) 2.095(5) 1.94(3) 97(2) 106(2) 112(1) 100.7(9) 104.6(8) 131(1) Symmetry transformation: 9 2x, 2y, 2z.The tetrahydroaluminate anion has been structurally characterised previously as the [NEt4]118 and [AlH2L]119 salts (L = N,N,N9,N0,N0-pentamethyldiethylenetriamine or 1,4,8,11- tetramethyl-1,4,8,11-tetraazacyclotetradecane). The mean Al]H distance [1.52(7) Å] in 2 may be compared with values of 1.61(5) Å in [NEt4][AlH4] and 1.55(3) Å in LiAlH4.20 Compound 2 was converted into 1 when it was heated to 120 8C under vacuum.The dimeric structure of compound 1 in benzene was con- firmed by cryoscopic measurements. Fig. 3 shows the 1H NMR spectrum of a ca. 0.1 M solution in C6D6; the hydride resonance is exceedingly broad and even in spectra obtained from concentrated solutions only just visible above the baseline. The width of the 27Al-{1H} peak, 390 Hz, is somewhat broader than those found for LiAlH4 in donor solvents (ca. 180 Hz in diethyl ether and as low as 14 Hz in diglyme).8 The aluminium is thus in an environment similar to that in [AlH4]2, distorted enough from tetrahedral for 27Al relaxation to broaden the 1H signal but not enough to cause it to collapse to a single line.The 6Li-{1H} difference spectra show enhancements of ca. 50% on selective irradiation in the hydride region of the spectrum and ca. 38% in the NMe region but not elsewhere. They indicate that the lithium nuclei are close to hydrogen atoms of AlH4 and NMe fragments.The NMR and cryoscopic data taken together are in accord with the presence in solution of species which are cyclic as in solid 1. It is not possible to tell whether there is rapid exchange on the NMR time-scale of ionic fragments between rings. The NMR data for compound 2 were similar to those for 1 and the NOE was strong in both the Al]H and N]Me regions. The ions present in the crystals of 2 cannot therefore be fully separated in solution since if this were the case the NOE in the hydride region would disappear, leaving only a weak enhancement near the NMe signal.Similar results were obtained from solutions of 2 in Et2O. Since the concentration of solventseparated ion pairs would be expected to be increased by the addition of base, samples containing a 10-fold excess of tmen were examined. Unfortunately the Al]H and tmen signals now overlapped so that separate enhancements could not be observed but an NOE could still be observed over the whole hydride region.Hence although the solid that separated from solution was the ionic compound 2, significant Li ? ? ?H]Al interactions persisted in the supernatant solution. The ionic compound 2 must therefore either dissociate into 1 and free tmen or give other species in which [AlH4]2 competes successfully with tmen for a place in the co-ordination sphere of lithium. Problems associated with the study of the coordination of tmen in solution have been reviewed: it seems that bonds between lithium and tmen can be broken in a wide range of solvent systems.21 Fig. 3 Proton NMR spectrum of a 0.1 M solution of compound 1 in C6D6, and 6Li-{1H} difference spectra with signals placed at the positions of the selective irradiation in the 1H spectrumJ. Chem. Soc., Dalton Trans., 1998, Pages 249–253 251 Proton and 6Li-{1H} NOE spectra of LiAlH4 in ethers Diethyl ether. The relationship between solute concentration and 27Al NMR signal widths for solutions of LiAlH4 in Et2O has been studied previously.The peak width, which is ca. 1 kHz at a concentration of 1.0 M, decreases in more dilute solutions so that Al]H coupling may be observed.8 Conversely, signals from hydrogen atoms bound to aluminium are sharp at high concentrations but become broad upon dilution so that chemical shifts are difficult to determine. Data for a range of concentrations, wider than that reported previously, are given in Table 2. They show that in the more concentrated solutions the environment around aluminium is highly distorted from tetrahedral, the electric field gradient is large, and quadrupole relaxation is so fast that coupling to hydrogen is completely suppressed. At lower concentrations the aluminium environment is less distorted from tetrahedral and the 1H signals are broadened by Al]H coupling.Fig. 4(b) shows 6Li-{1H} NMR spectra of LiAlH4 in Et2O obtained with selective irradiation at various frequencies in the proton spectrum [Fig. 4(a)]. The largest NOE (180%) resulted from irradiation of the hydrogen atoms attached to aluminium and there was only a small effect near the methylene and the methyl protons of the diethyl ether. This implies that the hydrogens on aluminium spend a signifi- cant proportion of their time close to lithium. In previous work,8 the 7Li NMR spectra were recorded for solutions of LiAlH4 in Et2O in the concentration range 3.75–0.0075 M. The linewidths (0.55–2.8 Hz) were concentration dependent but as they changed only negligibly (typically <0.1 Hz) on broad-band decoupling it was deduced that persistent Li]H]Al bridges were absent.8 It is however notoriously difficult to observe Li]H coupling.15 Our results suggest that although Li ? ? ?H]Al interactions may not be stable over periods long enough for Li]H coupling constants to be measured they are sufficient to generate a significant NOE.The effect on ebullioscopic and conductimetric results has been described previously.3,4 Tetrahydrofuran.The 1H NMR spectra of a 0.5 M solution of LiAlH4 in thf [Fig. 4(c)] showed only solvent signals at d 1.7 Fig. 4 (a) The 1H NMR spectrum of a 1.0 M solution of LiAlH4 in diethyl ether, (b) 6Li-{1H} spectra placed at positions of selective irradiation in the proton spectrum, (c) the 1H spectrum of a 0.5 M solution of LiAlH4 in thf, and (d ) 6Li-{1H} spectra placed at positions of selective irradiation in the proton spectrum Table 2 Peak widths of 1H NMR resonances in concentrated solutions of LiAlH4 in diethyl ether Concentration/M 1.92 2.36 4.87 5.16 5.85 Dn2� 1 /Hz 20 9.9 9.1 8.2 6.3 and 3.7; the hydrogen atoms attached to aluminium could not be directly observed, suggesting that the signals were broadened as a result of coupling to nearby aluminium nuclei in nearly tetrahedral environments.The 6Li-{1H} spectra are shown in Fig. 4(d ). The largest NOE resulted from irradiation of the downfield a-methylene multiplet of thf.This showed that, in contrast to solutions of LiAlH4 in Et2O, the a-methylene protons of the solvent were close to the lithium. However, the NOE spectra also showed a large effect upon irradiation between the solvent signals in the region where Al]H resonances have been located.22 The results provide direct evidence for the proximity of lithium and hydroaluminate ions and a method for the determination of values of dH for hydrogens attached to aluminium in a system in which direct observation is difficult or heteronuclear decoupling of aluminium is not possible.Although earlier 7Li NMR studies9 indicated that Li]H scalar coupling was absent, the NOE results show that Li ? ? ?H]Al bridges are formed at least transiently in 0.5 M solution. When 1 mol equivalent of HCl was condensed into the thf solution at 278 8C the NOE near the a-protons of the solvent remained, and that over the hydride region was suppressed, showing that Li ? ? ?H]Al interactions were absent. The species in solution were probably solvent-separated ion pairs [Li(thf)4][AlClnH4 2 n] or chloride-bridged species, e.g.(thf)3LiClAlH3 in which electron-deficient Li]H]Al bridges have been replaced by stronger electron-precise Li]Cl]Al links. Mono- and di-glyme. The 1H NMR spectra of a 0.5 M solution of LiAlH4 in monoglyme showed a weak sextet [d 2.90, 1J(Al]H) 172.8 Hz], resulting from coupling of hydrogen to aluminium nuclei (I = 5 2 – ), partly obscured by intense solvent signals between d 3 and 4.In the 6Li-{1H} spectra the largest NOE arose from irradiation of the methylene protons of the solvent, but there was a slightly smaller effect from the hydrogens bound to aluminium with an approximate correlation with maxima at points expected if the signal were a sextet. As with solutions in thf, the NOE results showed that both the methylene protons of the solvent and the hydrogen atoms attached to aluminium were close to the lithium cations.The solutions in diglyme were similar [d 3.10, 1J(Al]H) = 173.5; lit.,10 d 3.08, 1J(Al]H) 173 Hz]. Hence the [AlH4]2 anion is able to compete successfully with the glyme for a place in the lithium coordination sphere even when the aluminium environment is sufficiently symmetrical to allow the sextet in t 1H spectrum to be observed. Hydrogen–deuterium exchange between LiBH4 and LiBD4 in Et2O Solutions of LiBH4 in Et2O have been investigated by ebullioscopic and conductance measurements,3 and by vibrational 6,23 and NMR24 spectroscopy.It has been suggested that the species in solution are either contact ion pairs and multiple ions or hydride-bridged molecular species. Multinuclear NMR and NOE measurements allow a distinction between these two possibilities to be made. The 1H spectra (see Fig. 5) from a 0.5 M solution of LiBH4 in Et2O show, besides solvent peaks, a sharp quartet due to hydrogens coupled to 11B [1J(11B]H) = 81.25 Hz] and a less intense septet from hydrogens coupled to 10B nuclei [1J(10B]H) = 26.6 Hz].The NOE spectra show that the largest effect results from irradiation of the components of the hydridic quartet and septet, and that a smaller effect is produced upon irradiation of the downfield resonance of the solvent. These results suggest (a) that LiBH4 in Et2O gives species in which the lithium and hydrogen atoms are close and (b) that the [BH4]2 anion is not significantly distorted from tetrahedral.Further evidence comes from the isotope shifts in partly deuteriated LiBH4. The 1H NMR spectrum of a 1 : 1 mixture of LiBH4 and LiBD4 in Et2O (Fig. 6) showed four major features252 J. Chem. Soc., Dalton Trans., 1998, Pages 249–253 which could be assigned to the statistical mixture of isotopomers [LiBH4 2 nDn]2 (n = 0–3) observed earlier.24 If there were specific interactions via (m-H)2 or (m-H)3 bridges between lithium and boron there would be a non-uniform equilibrium isotope shift as hydrogen was replaced by deuterium and this non-uniformity would be expected to increase when the sample was cooled.If however the interactions were non-specific and largely electrostatic, only small secondary isotope shifts would be expected. Fig. 6 shows that the second alternative is observed. The resonance centres were shifted only 0.01 ppm to lower frequency for each increase in the value of n and although resolution was lost at 230 8C the spectrum had a very similar envelope to that at 25 8C showing that the effect of changing temperature was small.There was also an isotope effect on the coupling constant 1J(11B]H), viz. LiBH4 81.3, LiBH3D 81.1, LiBH2D2 80.9 and LiBHD3 80.6 Hz, corresponding to a monatonic decrease in the value of 1J as n increased. The isotope effects on the coupling constant 2J(H]D) (1.56 Hz) were too small to be measured. Similar results have been obtained for NaBH4 2 nDn and KBH4 2 nDn in ether solvents.25 An attempt to make a similar assessment of the specificity of the hydride bridges in the LiAlH4–LiAlD4 system was inconclusive. The 1H-{27Al} NMR spectra of samples of LiAlH4 and LiAlH4–LiAlD4 (1 : 1) in Et2O gave singlets with Dn2� 1 7 Hz so that with reasonable estimates of the isotope shifts it was clear that the peaks were too broad for signals from individual isotopomers like those in the LiBH4–LiBD4 system to be observed.The asymmetry of the broad signals should however have been sufficient to show the presence of mixed species LiAlHnD4 2 n.We were surprised therefore to find that the 1H signal in the LiAlH4–LiAlD4 mixture was symmetrical indicating that under our conditions (in Et2O at 50 8C for 17 h) H–D exchange had not taken place. Hydrogen–deuterium exchange has been observed previously in strong donor solvents such as thf,26 diglyme9,26 or MeCN.26 Fig. 5 (a) The 1H NMR spectrum of a 0.5 M solution of LiBH4 in Et2O, (b) 6Li-{H} difference spectra with irradiation at various positions in the proton spectrum.The peak marked S is attributed to silicone grease Fig. 6 Proton NMR spectra of a 1 : 1 mixture of LiBH4 and LiBD4 in Et2O at (a) 243 and (b) 288 K Conclusion The solid obtained from a toluene solution containing LiAlH4 and an equivalent of tmen is the complex 1 in which Li and Al are joined by m-H bridges. The solid that separates from a toluene solution of LiAlH4 containing an excess of tmen is [Li(tmen)2][AlH4] 2.There are no specific bonds between the anion and cation in the solid 2 but 6Li-{1H} NOE measurements indicate that [AlH4]2 ions in solution are brought near to Li. Similar Li ? ? ?H]Al contacts are detected in solutions of LiAlH4 in ether solvents in the absence of amine donors. The Li ? ? ?H]Al interactions are stronger in diethyl ether or thf than in mono- or di-glyme and apparently determine the structure of the solvent-free solid which crystallises from diethyl ether.20 The NOE measurements in themselves are not sufficient to distinguish between hydride bridges between specific lithium and aluminiun atoms in well defined species such as 1 and less discriminate fluxional interactions within ion pairs.For LiBH4 in Et2O the interactions are of the second kind. For LiAlH4 in Et2O the nature of the interaction is still unclear. The NOE measurements appear to show that whereas the symmetry of the [AlH4]2 ions is significantly perturbed by the formation of hydrogen bridges, that of the [BH4]2 ion remains essentially tetrahedral.Although this observed difference between boron and aluminium may result from genuine differences in the nature of the hydride-bridged species, it may simply reflect the different sensitivities of the boron and aluminium nuclei to the symmetry of their environments. More work is required to resolve this problem. Interactions analogous to those described in this paper have been postulated in solutions of the complex [Li(hmpa)4][BF4] [hmpa = hexamethylphosphoramide, P(NMe2)3O] in aromatic solvents.27 Experimental All materials were manipulated by standard Schlenk techniques using a conventional vacuum manifold and argon as a blanket gas.Solvents were dried, by heating under reflux with LiAlH4 for ethers and with sodium for toluene, and subsequently distilled. The compound LiAlH4 (Aldrich) was used as received and tmen (Aldrich) was dried over CaH2 and then distilled.Solutions of LiAlH4 were made by heating 1 g under reflux in the relevant ether (20 cm3), filtering through Celite, determining the concentration by decomposing aliquots in 1 M sulfuric or hydrochloric acid, measuring the dihydrogen evolved with a Sprengel pump,28 and diluting to give the concentrations required. Melting points were measured on samples in sealed capillaries and IR spectra were recorded as Nujol mulls on a Perkin-Elmer 1720 FT spectrometer.For the H–D exchange reaction between LiBH4 and LiBD4 in Et2O a solution of LiBD4 (0.32 g, 12.5 mmol) and LiBH4 (0.27 g, 12.5 mmol) in Et2O (50 cm3) was heated under reflux for 30 min in an atmosphere of argon, allowed to cool to room temperature, and filtered through a medium-porosity glass frit. All NMR spectra were recorded on samples in sealed glass tubes or tubes with rotationally symmetrical poly(tetrafluoroethylene) valves.The 1H NMR spectra were recorded at 90, 250, 360 or 500 MHz using respectively Perkin-Elmer R32, Bruker AC 250, WM 360 or AMX 500 instruments. The 7Li NMR spectra were recorded on a Bruker WP 80 FT spectrometer at 31.14 MHz, a Bruker AC 250 instrument at 73.6 MHz, or a Bruker WM 360 spectrometer at 139.9 MHz; chemical shifts are relative to external aqueous LiNO3. The 6Li NMR spectra were recorded on a Bruker WM 360 spectrometer at 52.99 MHz or an AMX 500 spectrometer at 73.59 MHz.The 6Li NOE difference spectra were obtained by selectively irradiating a resonance in the 1H spectrum for 14 s before obtaining a 6Li spectrum using a 308 pulse and no decoupling during the 8.4 s acquisition period. The procedure was repeated with selective irradiation of an empty region of the 1H spectrum and the two 6Li spectra wereJ. Chem. Soc., Dalton Trans., 1998, Pages 249¡V253 253subtracted to give the difference spectrum. The value of theNOE enhancement cited was obtained by absolute integrationof the two 6Li spectra; the theoretical maximum enhancement is340%.The power level for the selective irradiation was the sameas that used in the more familiar 1H¡V1H NOE experiment. The27Al NMR spectra were recorded on a Bruker WM 360 spectrometerat 93.8 MHz; chemical shifts are relati to external[Al(H2O)6]31.PreparationsLiAlH4?tmen 1. A mixture of tmen (0.8 cm3, 5.3 mmol) andLiAlH4 (0.2 g, 5.27 mmol) was stirred in thf (25 cm3) for 24 h at20 8C.Filtration through Celite followed by removal of solventand crystallisation from toluene gave colourless crystals ofcompound 1 (0.57 g, 70%), m.p. 179 8C (Found: C, 46.7; H, 12.8;N, 17.7%; M = 310. C12H40Al2Li2N4 requires C, 46.7; H, 13.1; N,18.2%; M = 308). n& max/cm21 (Al]H) 1660s (br) and 1730 (sh);dH(C6D6) 1.73 (4 H, CH2) and 2.07 (12 H, CH3); dC(C6D6)46.0 (CH3) and 56.5 (CH2); dLi(C6D6) 0.47; dAl(C6D6) 103[qnt, J(Al]H) 172 Hz, Dn2 1(27Al-{1H}) 391 Hz].LiAlH4?2tmen 2.A mixture of tmen (2.0 cm3, 13.3 mmol) andLiAlH4 (0.2 g, 5.27 mmol) was stirred at 60 8C in thf (25 cm3)for 6 h. After cooling to 20 8C and filtration through Celite thesolvent was removed to give a white solid. Crystallisation fromtoluene gave colourless crystals of compound 2 (1.28 g, 90%)(Found: C, 52.5; H, 13.1; Al, 10.0; Li, 2.7; N, 20.8. C12H36-AlLiN4 requires C, 53.3; H, 13.4; Al, 10.0; Li, 2.6; N, 20.7%).n& max/cm21 (Al]H) 1651s (br); dH(C6D6) 2.0 (8 H, CH2) and 2.07(24 H, CH3); dC(C6D6) 46.0 (CH3) and 57.3 (CH2); dLi(C6D6)0.45; dAl(C6D6) 103 [qnt, J(Al]H) 171 Hz, Dn2 1(27Al-{1H}) = 365Hz].CrystallographyCrystal data.For 1, C12H40Al2Li2N4, M = 308.3, monoclinic,space group P21/n (no. 14), a = 7.839(2), b = 15.802(5), c =9.132(6) , b = 98.17(4)8, U = 1119.7(9) 3, Z = 2, Dc = 0.91Mg m23, F(000) = 344. Colourless, air-sensitive block 0.4 ¡Ñ0.3 ¡Ñ 0.3 mm, m(Mo-Ka) = 0.13 mm21, T = 173 K.Data collection and processing: CAD4 diffractometer, q¡V2qscan, Mo-Ka radiation, l = 0.710 73 , 2 < q < 258, 1967 uniquereflections giving 1400 with I > 2s(I), no absorption or decaycorrection.Structure analysis and refinement: direct methods(SHELXS 86),29 full-matrix least-squares refinement on all F2using SHELXL 93,30 all non-H atoms anisotropic, hydride Hatoms located on a difference map and freely refined with isotropicthermal parameters, ligand H atoms included in ridingmode with Uiso(H) = 1.2 Ueq(C) or 1.5 Ueq(C) for methylgroups.R1 = S( Fo| 2 |Fc )/S|Fo| = 0.049 [for I > 2s(I)], wR2 =[Sw(Fo2 2 Fc2)2/Sw(Fo2)2] = 0.138, S = 1.03 (for all data).For 2, C12H36AlLiN4, M = 270.4, orthorhombic, space groupPnma, a = 18.146(2), b = 11.616(5), c = 9.438(3) , U =1989.4(11) 3, Z = 4, Dc = 0.90 Mg m23, F(000) = 608, colourlessair-sensitive needle 1.0 ¡Ñ 0.4 ¡Ñ 0.15 mm (in a capillary),m(Mo-Ka) = 0.10 mm21, T = 293 K, 1833 reflections, for2 < q < 258, no absorption or decay correction.Refinement on F2 using SHELXL 93, non-H atoms anisotropic,hydride H atoms located on a difference map and positionsrefined, ligand H atoms in riding mode, R1 = 0.072 [for502 reflections with I > 2s(I) and 98 parameters], wR2 =0.367, S = 1.02 (for all data). Both the anion and cation lie on acrystallographic mirror plane, with consequent averaged positionsfor the disordered central CH2CH2 groups of the tmenligands.CCDC reference number 186/790.AcknowledgementsThe authors thank the EPSRC for financial support and ICI plcfor a CASE studentship for M.C. E.References1 See, for example, A. R. Barron and G. Wilkinson, Polyhedron, 1986,5, 1897; B. M. Bulychev, Polyhedron, 1990, 9, 387 and refs. therein.2 Comprehensive Organic Synthesis, eds. B. M. Trost and I. Fleming,Pergamon, Oxford, 1991, vol. 8.3 E. C. Ashby, F. R. Dobbs and H. P. Hopkins, J. Am. Chem. Soc.,1975, 97, 3158.4 E. C. Ashby, F. R. Dobbs and H. P. Hopkins, J. Am. Chem. Soc.,1973, 95, 2823.5 A.E. Shirk and D. F. Shriver, J. Am. Chem. Soc.,1973, 95, 5904.6 S. V. Sigalova, I. P. Romm, E. N. Gur¡¦yanova, R. R. Shifrina,E. S. Shcherbakova, L. N. Margolin, L. P. Ivanov and A. I.Gorbunov, J. Gen. Chem. USSR., 1989, 59, 301.7 S. Her£¾mnek, O. Kr£¾iz, J. Ples£¾ek and T. Hanslk, Chem. Ind.(London), 1975, 42.8 H. Nth, Z. Naturforsch., Teil B, 1980, 35, 119.9 H. Nth, R. Rurlnder and P. Wolfgardt, Z. Naturforsch., Teil B,1981, 36, 31.10 V. P. Tarasov, V. I.Privalov, A. A. Gorbik and S. I. Bakum, Sov. J.Coord. Chem. (Engl. Transl.), 1985, 11, 935.11 A. G. Avent, C. Eaborn, M. N. A. El-Kheli, M. E. Molla, J. D. Smithand A. C. Sullivan, J. Am. Chem. Soc., 1986, 108, 3854.12 See, for example, W. Bauer, in Lithium Chemistry A Theoretical andExperimental Overview, eds. A.-M. Sapse and P. von R. Schleyer,Wiley, New York, 1995, ch. 5; H. Balzer and S. Berger, Chem. Ber.,1992, 125, 733 and refs. therein.13 J. A. Dilts and E. C. Ashby, Inorg. Chem., 1970, 9, 855.14 C. Eaborn, I. B. Gorrell, P. B. Hitchcock, J. D. Smith andK. Tavakkoli, Organometallics, 1994, 13, 4143.15 A. Heine and D. Stalke, Angew. Chem., Int. Ed. Engl., 1992, 31, 854;R. J. Wehmschulte, J. J. Ellison, K. Ruhlandt-Senge and P. P. Power,Inorg. Chem., 1994, 33, 6300.16 M. G. Gardiner, S. M. Lawrence and C. L. Raston, Inorg. Chem.,1995, 34, 4652; 1996, 35, 1349.17 D. R. Armstrong, W. Clegg, H. M. Colquhoun, J. A. Daniels,R. E. Mulvey, I. R. Stephenson and K. Wade, J. Chem. Soc., Chem.Commun., 1987, 630.18 K. Semenenko, A. L. Dorosinskii and . B. Lobkovskii, J. Struct.Chem. (Engl. Transl.), 1973, 14, 700.19 J. L. Atwood, K. D. Robinson, C. Jones and C. L. Raston, J. Chem.Soc., Chem. Commun., 1991, 1697.20 N. Sklar and B. Post, Inorg. Chem., 1967, 6, 66921 D. B. Collum, Acc. Chem. Res., 1992, 25, 448.22 D. A. Horne, J. Am. Chem. Soc., 1980, 102, 6011.23 A. E. Shirk and D. F. Shriver, J. Am. Chem. Soc., 1973, 95, 5901.24 B. D. James, B. E. Smith and R. H. Newman, J. Chem. Soc., Chem.Commun., 1974, 294; B. E. Smith, B. D. James and R. M. Peachey,Inorg. Chem., 1977, 16, 2057.25 I. A. Oxton, A. G. McInnes and J. A. Walter, Can. J. Chem., 1979,57, 503.26 V. P. Tasarov, S. I. Bakum, V. I. Privalov and Yu. A. Buslaev, Dokl.Akad. Nauk SSSR, 1982, 266, 1423; V. P. Tasarov, S. I. Bakum,V. I. Privalov, Yu. A. Buslaev and A. M. Kuznetsov, Koord. Khim.,1983, 9, 822.27 D. Barr, K. B. Hutton, J. H. Morris, R. E. Mulvey, D. Reed andR. Snaith, J. Chem. Soc., Chem. Commun., 1986, 127.28 A. Weissberger (Editor), Techniques of Organic Chemistry, 2nd edn.,Interscience, London, 1957, vol. 3, Part II, p. 365.29 G. M. Sheldrick, SHELXS 86, University of Gttingen, 1986.30 G. M. Sheldrick, SHELXL 93, University of Gttingen, 1993.Received 15th August 1997; Paper 7/06003C