摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 2103–2110 2103 Controlling copper(I) halide framework formation using N-donor bridging ligand symmetry: use of 1,3,5-triazine to construct architectures with threefold symmetry Alexander J. Blake,a Neil R. Brooks,a Neil R. Champness,a Paul A. Cooke,a Anne M. Deveson,b Dieter Fenske,b Peter Hubberstey,*a Wan-Sheung Li a and Martin Schröder *a a School of Chemistry, University of Nottingham, Nottingham, UK NG7 2RD. E-mail: m.schroder@nottingham.ac.uk; peter.hubberstey@nottingham.ac.uk b Institut für Anorganische Chemie, Universität Karlsruhe, Engesserstr.Geb-Nr. 30.45, 76128, Karlsruhe, Germany Received 23rd March 1999, Accepted 18th May 1999 The formation of co-ordination polymers between copper(I) halides and 1,3,5-triazine (tri), a potentially tridentate N-donor bridging ligand with threefold symmetry, has been studied. Complexes with both 3 : 1 and 2 : 1 molar ratios are formed by both CuBr and CuI.The compounds [Cu3X3(tri)]• (X = Br or I) are structurally similar, despite crystallising in diVerent space groups. They are composed of (CuX)• columns linked by triazine molecules to generate three-dimensional constructions with non-crystallographically imposed threefold symmetry. The (CuX)• columnar motif can be described as a series of perpendicularly stacked Cu3X3 chairs, alternately rotated by 608 and linked by Cu–X contacts. The tetrahedral co-ordination geometry of the copper centres is completed by a tridentate triazine bridge which links two copper atoms in separate columns. Thus, each (CuX)• column is linked to six adjacent (CuX)• columns.The structure of [Cu2Br2(tri)]• comprises (CuBr)• columns and castellated (CuBr)• chains linked by triazine molecules to generate a construction with crystallographically imposed threefold symmetry. The (CuBr)• columns are similar to but more regular than those found in [Cu3Br3(tri)]•. In this case, however, each column is linked to six adjacent chains.The (CuBr)• castellated chain motif is very unusual. The tetrahedral copper centres are co-ordinated by two adjacent bromide anions and by two triazine molecules each of which links a second chain and a column. Consequently, each chain is linked to four neighbouring chains and two neighbouring columns. Despite a stoichiometry identical to that of [Cu2Br2(tri)]•, [Cu2I2(tri)]• has a completely diVerent structure. The triazine molecules act as bidentate bridging ligands to link (CuI)• layers thereby giving alternating inorganic and organic layers.The tetrahedral co-ordination geometry of the copper centres in the (CuI)• layers, which are eVectively undulating hexagonal nets, is provided by three iodide anions from the layers and by a bridging triazine molecule. Co-ordination polymers are normally constructed by linking transition metal centres through bridging ligands and a rich diversity of architectures has been reported ranging from simple one-dimensional chains to three-dimensional matrices (Scheme 1).1 The anions in these systems act either as co-ordinated, nonbridging, spectator ligands or as non-co-ordinated, void-filling species.Copper(I) halides, however, can themselves form both oligomeric (normally dimers or tetramers) and polymeric frameworks.2–7 Consequently, copper(I) halide co-ordination polymers can be solely metal-halide based (Scheme 2), solely metal–ligand based with oligomeric molecular units linked by bridging ligands (Scheme 3), or a combination of both with metal-halide frameworks linked by bridging ligands (Scheme 4).The influence of ligand geometry and/or symmetry on the metal–ligand co-ordination assembly is self-evident.1 In this paper we report the syntheses and structures of copper(I) halide frameworks linked by N-donor ligands and provide considerable evidence to support the contention that ligand geometry and/or symmetry also influences metal–halogen framework formation.With monodentate N-donor ligands, the copper(I) halide frameworks range from mononuclear species through dinuclear and tetranuclear discrete molecular moieties to polymeric Scheme 1 Co-ordination polymers formed by linking transition metal centres (d) through molecular rods ( ): (a, b) one-dimensional chains, (c, d) two-dimensional sheets and (e) three-dimensional matrices. • • • • • • • • • • • • • • • • • • • • • • •• • • • • • • • • •• ••• • • • • • • • • • • • • • • •• • • • • • • • • •• • • • • • • • • • • • • • •• • •• •• ••• • • • • • • • • • • • (a) (b) • • (c) • (d) (e)2104 J.Chem. Soc., Dalton Trans., 1999, 2103–2110 structures.2–9 The less sterically demanding the ligand, the greater the complexity of the copper(I) halide architectures. The mononuclear species comprise two- and three-co-ordinate molecular units of stoichiometry [CuXL] 8 or [CuXL2] 9 (X = halide, L = N-donor ligand).The dinuclear species are based Scheme 2 Copper(I) halide frameworks: (a, b, c, d) one-dimensional chains. d Copper(I) cations; s, halide anions. • O • • O • O • O O O • O • • O • O • O O • O O • • O O • • O O • • O O • • O O • • O O • • O • • O O • O • O • O • O • O • O • O • O • O • O • O • O • O • (c) (d) (a) (b) Scheme 3 Co-ordination polymers formed by linking copper(I) halide molecular units through molecular rods ( ): (a, b) one-dimensional chains, (c) two-dimensional sheets.d, Copper(I) cations; s, halide anions. • O • O • O O • • O • O • O O • • O • O • O O • • O O• •O O• •O O• •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • •O O • (a) (b) • O O • (c) Scheme 4 Co-ordination polymers (two-dimensional sheets) formed by linking copper(I) halide frameworks through molecular rods. Key as in Scheme 3. • O • • O • O • O O O • • O • O • O O • O • O • • O • O • O O O • O • • O • O • O O • O • O • O O • O • O • O O • O • • O • O • O O O • O • O • O • O O O • O • O O O • O • • • • O • O • O O O • O • O O O • O • • • • O • O • O O O • O • O O O • O • • • • O • O • O O O • O • O O O • O • • • • • • • • • • • • (a) (b) (c) on rhomboid dimers of stoichiometry [Cu2X2L2] 2,8 or [Cu2X2L4] 3,5,9 and the tetranuclear species on cubane 2,10 or stepped cubane 2 tetramers of stoichiometry [Cu4X4L4] and [Cu4X4L6], respectively (X = halide, L = N-donor ligand).All the copper(I) halide polymeric species are based on onedimensional chains; no examples of two-dimensional sheets have been observed.Although the majority of the chains are single-strand split-stair 2–4 or double-strand staircase 3,5,6 frameworks of stoichiometry [CuXL]•, there are a limited number of examples containing linked cubane tetramers2 or steppedcubane7 tetramers of stoichiometry [Cu4X4L3]• and [Cu4X4L4]•, respectively (X = halide, L = N-donor ligand). Bidentate bridging N-donor ligands have immense influence on the metal–halogen framework adopted.They are best classified according to the disposition of their lone pairs, which varies from linear (bridging angle 1808; e.g. pyrazine) through obtuse (bridging angle 1208, e.g. pyrimidine) and acute (bridging angle 608, e.g. pyridazine) to parallel (bridging angle 08; e.g. 1,8-naphthyridine). The copper(I) halide frameworks formed in the presence of parallel bridging N-donor ligands are complex chain, ribbon or layer motifs (Scheme 5).11,12 They are all based on fused Cu3Br3 hexagons, two copper atoms of which are spanned by the bidentate ligand.It is the small bridging angle which dictates the relative positioning of the copper centres and hence the geometry of the copper(I) halide frameworks. With dipyrido(1,2-a:29,39-d)imidazole (dpi),11 CuCl forms [Cu2Cl2(dpi)]• chains, which comprise a linear array of hexagons (Scheme 5a); in contrast, CuBr forms [Cu3Br3(dpi)]• ribbons, which comprise two chains linked by a conventional [CuBr]• staircase chain (Scheme 5b). With acrylonitrile (acn), CuCl forms [Cu2Cl2(acn)]• layers in which the chains are fused to generate a buckled, severely distorted, hexagonal net (Scheme 5c).12 Only a small number of copper(I) halide co-ordination polymers have been obtained with obtuse bridging bidentate N-donor ligands. 2-Cyanoguanidine (cnge), which uses its nitrile and imino nitrogens to give a bridging angle of 1208,13 forms with CuX (X = Cl or Br) a two-dimensional structure in which parallel (CuX2Cu)• rack chains (Scheme 4c) are linked by cnge molecules.Linear bidentate bridging N-donor ligands, so-called molecular rods, favour linear frameworks 14 which in turn generate one-dimensional copper(I) halide frameworks. Thus, pyrazine (pyz) forms two complexes with CuCl both of which adopt Scheme 5 Copper(I) halide [CuX]• co-ordination polymers generated from parallel bridging bidentate ligands: (a) one-dimensional chain, (b) one-dimensional ribbon and (c) two-dimensional layer.O • O • O • O • • O O • O • O • O • O • O • • O O • O • • • O • O • O O • O • O • O O • O • • O O • • O • O • O O • O O • O • O • O • • O • • O • • O • O • O • O • O O • • O O • O • (a) (b) (c)J. Chem. Soc., Dalton Trans., 1999, 2103–2110 2105 two-dimensional sheet structures; in [CuCl(pyz)]• the pyrazine molecules link (CuCl)• split stair chains (Scheme 4a),15 whereas in [Cu2Cl2(pyz)]• they link (CuCl)• staircase chains (Scheme 4b).16 Phenazine (phz) forms 1: 2 complexes [Cu2- X2(phz)]• (X = Cl or Br) structurally similar to [Cu2Cl2- (pyz)]•.17 4-Cyanopyridine (pyridine-4-carbonitrile; pycn), a molecular rod with both pyridinyl and nitrile bases, gives 1 : 1 complexes [CuX(pycn)]• (X = Cl or Br) which adopt a twodimensional sheet structure analogous to that of [CuCl(pyz)]•, with 4-cyanopyridine molecules linking (CuCl)• split-stair chains.18 The remaining copper(I) halide frameworks found in coordination polymers containing linear bidentate bridging N-donor ligands are discrete dinuclear or tetranuclear species.In [Cu2I2(phz)]•, phenazine molecules link Cu2I2 rhomboids to form a one-dimensional chain structure (Scheme 3a).17 A similar one-dimensional chain structure occurs in [Cu4I4- (pycn)5]• in which [Cu4I4(pycn)4] stepped cubane tetramers are linked by a fifth bidentate bridging pycn molecule (Scheme 3b).18 The extended structure of [CuCl(4,49-bipy)]•, comprises an interpenetrated arrangement of mutually perpendicular sheets based on an hexagonal arrangement of Cu2Cl2 rhomboids linked by alternating single and double 4,49-bipyridine (4,49-bipy) bridges (Scheme 3c).19 To support our contention that the choice of copper(I) halide framework is dependent on ligand symmetry, we now report the results of our study of complex formation between the tridentate bridging ligand 1,3,5-triazine (tri) and copper(I) halides.Thus far, only two co-ordination polymers involving triazine have been reported; they are both silver(I) complexes, [Ag(tri)(CF3SO3)]?H2O20 and [Ag6(tri)8][BF4]6?H2O.21 Whereas all of the triazine molecules in the former are triconnected, only six are triconnected in the latter, the other two acting as bidentate ligands. Results and discussion Treatment of copper(I) bromide or iodide with triazine has yielded four crystalline products of stoichiometry [Cu3X3(tri)] (X = Br 1 or I 2) and [Cu2X2(tri)] (X = Br 3 or I 4).Good quality crystals of all four products were obtained by conventional crystallisation techniques. Crystals of 1–3 were obtained by layering solutions of CuX (X = Br or I) in MeCN and solutions of 1,3,5-triazine in CH2Cl2 or Et2O. Crystals of 4 were obtained by vapour phase diVusion of Et2O into a mixture of CuI in MeCN and 1,3,5-triazine in CH2Cl2. The four products were characterised by single crystal X-ray diVraction methods.Although [Cu2X2(tri)]• (X = Br or I) adopt markedly diVerent structures, [Cu3X3(tri)]• (X = Br or I) are similar, although they adopt diVerent space groups (Pbca for [Cu3Br3(tri)]• and Pnma for [Cu3I3(tri)]•). The molecular structures of the complexes are shown in Figs. 1–4; selected interatomic distances and angles are summarised and compared in Table 1. The reactivity of the products increases markedly from the iodides, which were prepared in bulk, to the bromides, which could only be obtained in very limited quantity as crystalline samples.The chlorides were so sensitive that analogous experiments with copper(I) chloride were unsuccessful despite the careful application of Schlenk methods. Crystal and molecular structure of [Cu3X3(tri)]• (X 5 Br or I) The structures of [Cu3X3(tri)]• (X = Br 1 or I 2) are composed of (CuX)• columns linked by triazine molecules to generate three-dimensional constructions with non-crystallographically imposed threefold symmetry (Fig. 1a).The numbering schemes for the asymmetric unit of 1, which contains three copper cations, three bromide anions and a single triazine molecule, and 2, which contains two copper cations and two iodide anions, of which Cu(3) and I(1) lie on a mirror plane, and half a triazine molecule of which C(1) and N(2) lie on a mirror plane, are shown in Fig. 2a and 2b, respectively. The location, in 2, of one of the copper cations on a symmetry element results in a more symmetrical network and diVerent packing of the triazine molecules.The (CuX)• columnar framework motif (Fig. 1b; Scheme 6b) has been observed once before, in the N-methylpyrazinium iodocuprate, [Cu2I3(Mepyz)].22 The motif can be described as a series of perpendicularly stacked Cu3X3 chairs, alternately rotated by 608 and linked by Cu–X contacts [for 1, Cu ? ? ? Br 2.418(1)–2.491(1), average 2.46 Å; for 2, Cu ? ? ? I 2.588(1)– 2.621(1), average 2.61 Å].Each of the three copper atoms in a chair is linked through a tridentate triazine bridge [for 1, Cu ? ? ? N 2.058(4)–2.090(4), average 2.07 Å; for 2, Cu ? ? ?N 2.082(5), 2.083(8) Å] to two copper atoms in separate columns (Fig. 2a). Thus, each column is linked to six adjacent columns (Fig. 1a) to give the three-dimensional construction. Crystal and molecular structure of [Cu2Br2(tri)]• The structure of [Cu2Br2(tri)]• 3 (Fig. 3a) comprises (CuBr)• columns (Fig. 3b; Scheme 6b), similar to those found in 1, and castellated (CuBr)• chains (Fig. 3c; Scheme 6a) linked by triazine molecules to generate a construction with crystallographically imposed threefold symmetry (Fig. 3a). The numbering scheme for the asymmetric unit, which contains two copper cations and two bromide anions, all of which lie on mirror planes, and half a triazine molecule of which C(4) and N(2) lie on a mirror plane, is shown in Fig. 2c. The castellated (CuBr)• chains (Fig. 3c; Scheme 6a) in complex 3 are unusual.Much less common than split-stair chains, castellated chains have been observed previously only in conjunction with the bidentate chelating S-donor ligand, tetrakis(methylsulfanyl)tetrathiafulvalene (tmtttf). In [Cu2Cl2- (tmtttf)]•,23 tmtttf molecules bridge parallel castellated (CuCl)• chains to form two-dimensional sheets. The (CuBr)• column in complex 3 is more regular than that in 1. The Cu ? ? ? Br interatomic distances within the column of 3 [2.423(3), 2.485(2) Å] are similar to those in the column of 1 [2.418(1)–2.491(1) Å] and cover a narrower range than those in the chain of 3 [2.394(3), 2.523(3) Å].Each of the three copper atoms in a chair of the column in complex 3 is co-ordinated by a triazine molecule, Cu ? ? ?N 2.076(13) Å, which uses its other two nitrogen atoms to bridge to two copper atoms, Cu ? ? ? N 2.095(9) Å, in separate chains (Fig. 2b). Owing to the alternating 608 rotation of the chairs a view down the column shows sixfold symmetry as each column is linked to six adjacent chains (Fig. 3a). Each copper atom of a chain is co-ordinated by two triazine molecules, Cu ? ? ?N 2.095(9) Å, each of which links to a second chain and a column. Fig. 1 Molecular structure of [Cu3X3(tri)]• (X = Br or I): (a) view down the columnar axis and (b) view of the one-dimensional columnar framework (Cu, hatched; X, cross-hatched; N, dotted; C, shaded).2106 J. Chem. Soc., Dalton Trans., 1999, 2103–2110 Fig. 2 Co-ordination properties of triazine in [Cu3X3(tri)]• [X = Br (a) or I (b)], [Cu2Br2(tri)]• (c) and [Cu2I2(tri)]• (d) showing the numbering schemes [Cu, hatched; X, cross-hatched; N, dotted; C, shaded; symmetry codes: (a) iii 2x, 20.5 1 y, 0.5 2 z; vii 2x, 0.5 1 y, 0.5 2 z; (b) iii 2x, 2y, 1 2 z; vi 1 1 x, y, z].Consequently, each chain is linked through triazine molecules to four other neighbouring chains and two neighbouring columns. The entire system has threefold symmetry as shown in the schematic diagram depicting the arrangement of chains, tubes and triazine molecules perpendicular to the threefold axis of symmetry (Fig. 3d). A corollary of the non-centrosymmetric character of the complex is the fact that all the triazine molecules point in the same direction (Fig. 3a), giving rise to a polar crystal structure. Crystal and molecular structure of [Cu2I2(tri)]• Despite a stoichiometry identical to that of complex 3, [Cu2I2(tri)]• 4 has a completely diVerent structure (Fig. 4). It comprises (CuI)• sheets (Fig. 4b; Scheme 6c) linked by triazine molecules acting as bidentate bridging ligands (Fig. 2d). The numbering scheme for the asymmetric unit, which contains one copper cation, one iodide anion, and half a triazine molecule of which C(2) and N(2) lie on a mirror plane, is shown in Fig. 2d. The (CuI)• layer (Fig. 4b; Scheme 6c) is eVectively an undulating hexagonal net, each six-membered ring adopting a boat conformation.It is similar to, but much more regular than, the (CuCl)• sheet found in [Cu2Cl2(acn)]•.12 The geometrical parameters in 4 [Cu–I 2.5864(10)–2.6338(9), average 2.616 Å; I–Cu–I 108.39(3)–117.10(4), average 112.58; Cu–I–Cu 105.55(3)– 110.06(3), average 107.88] fall into much more limited ranges than those in [Cu2Cl2(acn)]• [Cu–Cl 2.269(5)–2.720(4), average 2.398 Å; Cl–Cu–Cl 96.52(13)–107.30(13), average 101.88; Cu–Cl–Cu 104.2(2)–127.0(2), average 116.38].12 The fourth position of the tetrahedral copper centre in 4 is occupied by a nitrogen, Cu–N 2.064(5) Å, of a triazine which bridges to a copper atom in a second (CuI)• layer (Fig. 2d) giving alternating copper(I) iodide and triazine layers (Fig. 4a). The alignment of the triazine molecules is such that the unco-ordinated nitrogen atom points directly at the centre of an adjacent triazine molecule (dihedral angle between triazine rings = 82.28; nitrogen–ring centre distance = 2.93 Å; Fig. 4).J. Chem. Soc., Dalton Trans., 1999, 2103–2110 2107 Fig. 3 Molecular structure of [Cu2Br2(tri)]•: (a) view down the columnar axis, (b) view of the one-dimensional columnar framework, (c) view of the one-dimensional castellated chain and (d) schematic representation of the structure showing the threefold symmetry of the extended assembly (Cu, hatched; Br, cross-hatched; N, dotted; C, shaded). This structure may be considered to be a new topological type as replacement of copper by silicon, iodine by nitrogen and triazine by oxygen generates the Si2N2O (Sinoite) structure.24 Copper halide frameworks All three copper halide polymeric frameworks, chains, tubes and sheets found in the four compounds can be traced to the three-dimensional wurtzite structure of copper(I) halides.24 The one-dimensional castellated chain (Scheme 6a) is the simplest, with distorted tetrahedral copper centres linked by two-coordinate halides leaving two co-ordination sites free to bind separate triazine molecules.The novel columnar structural form (Scheme 6b), although previously described as a series of perpendicularly stacked Cu3X3 chairs, alternately rotated by 608 and linked by Cu–X contacts, can also be described as three parallel castellated chains arranged with threefold symmetry, each chain being connected to its two neighbours by Cu–X contacts. The sheets, which have a corrugated appearance (Scheme 6c), are constructed in a similar fashion to the columns, a series of parallel castellated chains being connected by Cu–X contacts.They have essentially the same geometry as the (100) plane of the wurtzite structure.24 For both the columnar and sheet architectures, distorted tetrahedral copper centres are linked by three co-ordinate halides leaving a single co-ordination site free to bind a triazine molecule. 1,3,5-Triazine co-ordination The triazine ligand bridges the copper halide frameworks to generate complex three-dimensional constructions. The dispositions of the copper(I) halide frameworks around the triazine molecules are compared in Fig. 2. Although potentially tridentate, the triazine molecule does not invariably use all three N-donor sites. In 4 it acts as a bidentate ligand by linking two copper iodide sheets (Fig. 2d) to give alternating copper iodide and triazine layers (Fig. 4). The triazine is tridentate in the other three componds. In 1 and 2 each triazine co-ordinates to2108 J.Chem. Soc., Dalton Trans., 1999, 2103–2110 three copper(I) centres in adjacent columns (Fig. 2a and 2b) to generate a system with non-crystallographically imposed threefold symmetry (Fig. 1). The packing of the triazine molecules diVers in the two co-ordination polymers, the location, in 2, of one of the copper centres on a mirror plane resulting in a more symmetrical network. In 3 each triazine molecule bridges a column and two castellated chains (Fig. 2c), to give the threedimensional structure (Fig. 3). Conclusion Co-ordination polymers of copper(I) halides with N-donor ligands dramatically exemplify the influence of ligand geometry and/or symmetry on the type of copper(I) halide framework geometry observed. Thus, monodentate ligands generate either molecular units or one-dimensional chains; parallel bidentate bridges form one-dimensional chains or ribbons and twodimensional layers all of which are based on distorted hexagonal Cu3X3 building blocks; linear bidentate bridges encourage formation of linear arrays in the form of staircase or split-stair chains and the trigonal symmetry of the tridentate triazine molecule promotes the formation of a copper(I) halide framework with threefold symmetry in the form of the columnar constructions in 1, 2 and 3.Experimental Synthesis; general procedures All reagents (Aldrich) were used as received. When using CuBr all procedures were carried out under argon and all solvents were freshly distilled, degassed and dried by literature procedures. 25 Elemental analysis (C, H, N) was performed by the Nottingham University School of Chemistry Microanalytical Service using a Perkin-Elmer 240B instrument. Infrared spectra were obtained (as KBr pressed pellets) using either a Perkin-Elmer 1600 series or a Nicolet Avatar 360 FTIR spectrometer. [Cu3Br3(tri)]• 1. A solution of CuBr (0.050 g; 3.49 × 1024 mol) in dry MeCN (10 cm3) was added to a solution of 1,3,5- triazine (0.0094 g; 1.16 × 1024 mol) in dry MeCN (10 cm3).Dry Et2O was added and the resultant orange precipitate filtered oV, Scheme 6 Novel copper(I) halide frameworks formed in triazine based co-ordination polymers: (a) one-dimensional castellated chains, (b) onedimensional [CuX]• columns and (c) two-dimensional [CuX]• sheets. • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o • o o • o • o • • o • o • o o • o • o • • o • o • o o • o • o • • o • o • o (a) castellated chain (c) sheet (b) column washed with dry MeCN and dry Et2O and dried in vacuo (yield: 0.043 g, 0.84 × 1024 mol, 72%) [Found (calc.for CHBrCuN): C, 7.00 (7.05); H, 0.50 (0.60); N, 8.00 (8.20)%]. IR n& /cm21 (triazine unless stated otherwise): 1556s, 1406s, 1171w, 705m and 699m. Orange block-shaped crystals suitable for X-ray diVraction studies were grown by layering a solution of triazine in Et2O on top of a CuBr solution in MeCN (molar ratio 1 : 3).[Cu3I3(tri)]• 2. A solution of CuI (0.050 g; 2.63 × 1024 mol) in MeCN (10 cm3) was added to a solution of 1,3,5- triazine (0.0071 g; 0.88 × 1024 mol) in Et2O (20 cm3). The mixture was left at room temperature overnight and the resulting red crystalline precipitate filtered oV, washed with CH2Cl2 and dried in vacuo (yield: 0.0125 g, 0.19 × 1024 mol, 22%) [Found (calc. for CHCuIN): C, 5.50 (5.50); H, 0.40 (0.45); N, 6.30 (6.45)%]. IR n& /cm21 (triazine unless stated otherwise): 1554s, 1406s and 1172m. Red block-shaped crystals suitable for X-ray diVraction studies were grown either by slow evaporation of a 2 : 1 CuI : triazine solution in MeCN or CH2Cl2 or by layering a solution of triazine in Et2O on top of a CuI solution in MeCN (molar ratio 1 : 3).[Cu2Br2(tri)]• 3. Orange acicular crystals suitable for X-ray diVraction studies were grown by layering a CuBr solution in pre-dried MeCN on top of a solution of triazine in pre-dried Et2O.All attempts to prepare the product in bulk resulted in the formation of [Cu3Br3(tri)]• (by microanalysis and X-ray powder diVraction analysis). [Cu2I2(tri)]• 4. A solution of CuI (0.020 g; 1.05 × 1024 mol) in MeCN (10 cm3) was added to a solution of 1,3,5-triazine (0.0043 g; 0.53 × 1024 mol) in CH2Cl2 (10 cm3). The mixture was left at room temperature overnight and the resulting yellow precipitate filtered oV, washed with CH2Cl2 and dried in vacuo (yield: 0.0167 g, 0.36 × 1024 mol, 68%) [Found (calc.for Fig. 4 Molecular structure of [Cu2I2(tri)]•: (a) view showing the alternating [CuI]• and triazine layers and (b) view of the two-dimensional [CuI]• layer (Cu, hatched; I, cross-hatched; N, dotted; C, shaded).J. Chem. Soc., Dalton Trans., 1999, 2103–2110 2109 Table 1 Selected bond lengths and angles for [Cu3Br3(tri)]• 1, [Cu3I3(tri)]• 2, [Cu2Br2(tri)]• 3 and [Cu2I2(tri)]• 4 1 a 2 b 3 c 4 d Cu–X/Å Cu(1)–Br(1) Cu(1)–Br(2) Cu(1)–Br(3) Cu(2)–Br(1) Cu(2i)–Br(3) Cu(2ii)–Br(3) Cu(3)–Br(1) Cu(3i)–Br(2) Cu(3ii)–Br(2) 2.4779(10) 2.4204(8) 2.4832(9) 2.4565(8) 2.4898(9) 2.4184(8) 2.4708(8) 2.4305(9) 2.4912(8) Cu(3)–I(1) Cu(4)–I(1) Cu(3)–I(2) Cu(4i)–I(2) Cu(4ii)–I(2) 2.592(2) 2.5875(12) 2.6206(12) 2.616(2) 2.620(2) Cu(3)–Br(1) Cu(3i)–Br(1) Cu(4)–Br(2) Cu(4ii)–Br(2) 2.423(3) 2.485(2) 2.523(3) 2.394(3) Cu(1)–I(1) Cu(1i)–I(1) Cu(1ii)–I(1) 2.6286(14) 2.5864(10) 2.6338(9) Cu–N/Å Cu(1)–N(1) Cu(2)–N(3iii) Cu(3)–N(5iv) 2.090(4) 2.076(4) 2.058(4) Cu(3)–N(2iii) Cu(4)–N(1) 2.083(8) 2.082(5) Cu(3)–N(2) Cu(4)–N(1iii) 2.076(13) 2.095(9) Cu(1)–N(1) 2.064(5) Angles at X/8 Cu(1)–Br(1)–Cu(3) Cu(2)–Br(1)–Cu(1) Cu(3)–Br(1)–Cu(1) Cu(1)–Br(2)–Cu(3i) Cu(1)–Br(2)–Cu(3ii) Cu(3i)–Br(2)–Cu(3ii) Cu(2ii)–Br(3)–Cu(1) Cu(2ii)–Br(3)–Cu(2i) Cu(1)–Br(3)–Cu(2i) 103.71(3) 108.02(3) 100.95(3) 107.71(3) 100.74(3) 109.36(2) 98.26(4) 101.68(2) 107.41(3) Cu(4)–I(1)–Cu(4iv) Cu(4)–I(1)–Cu(3) Cu(4iv)–I(1)–Cu(3) Cu(4i)–I(2)–Cu(4ii) Cu(4i)–I(2)–Cu(3) Cu(4ii)–I(2)–Cu(3) 97.71(5) 107.25(3) 107.25(3) 107.53(3) 103.59(4) 103.25(4) Cu(3i)–Br(1)–Cu(3) Cu(4ii)–Br(2)–Cu(4) 107.34(7) 93.32(7) Cu(1i)–I(1)–Cu(1) Cu(1i)–I(1)–Cu(1ii) Cu(1)–I(1)–Cu(1ii) 105.55(3) 110.06(3) 107.84(3) Angles at Cu/8 N(1)–Cu(1)–Br(2) N(1)–Cu(1)–Br(1) Br(2)–Cu(1)–Br(1) N(1)–Cu(1)–Br(3) Br(2)–Cu(1)–Br(3) Br(1)–Cu(1)–Br(3) N(3iii)–Cu(2)–Br(3v) N(3iii)–Cu(2)–Br(1) Br(3v)–Cu(2)–Br(1) N(3iii)–Cu(2)–Br(3vi) Br(3v)–Cu(2)–Br(3vi) Br(1)–Cu(2)–Br(3vi) N(5iv)–Cu(3)–Br(2vi) N(5iv)–Cu(3)–Br(1) Br(2vi)–Cu(3)–Br(1) N(5iv)–Cu(3)–Br(2v) Br(2vi)–Cu(3)–Br(2v) Br(1)–Cu(3)–Br(2v) 108.05(11) 102.04(12) 117.76(3) 99.22(11) 119.42(3) 107.30(3) 106.26(11) 103.71(12) 119.27(3) 101.92(12) 115.96(3) 107.53(3) 116.65(12) 102.10(11) 111.24(3) 102.68(11) 109.04(3) 114.96(3) N(2iii)–Cu(3)–I(1) N(2iii)–Cu(3)–I(2) I(1)–Cu(3)–I(2) N(2iii)–Cu(3)–I(2iv) I(1)–Cu(3)–I(2iv) I(2)–Cu(3)–I(2iv) N(1)–Cu(4)–I(1) N(1)–Cu(4)–I(2vi) I(1)–Cu(4)–I(2vi) N(1)–Cu(4)–I(2vii) I(1)–Cu(4)–I(2vii) I(2vi)–Cu(4)–I(2vii) 114.2(2) 103.09(11) 110.73(3) 103.09(11) 110.73(3) 114.64(6) 106.5(2) 104.5(2) 114.59(3) 101.3(2) 118.84(4) 109.12(3) N(2)–Cu(3)–Br(1v) N(2)–Cu(3)–Br(1) Br(1v)–Cu(3)–Br(1) Br(1)–Cu(3)–Br(1vi) N(1ii)–Cu(4)–N(1) N(1iii)–Cu(4)–Br(2vii) N(1iii)–Cu(4)–Br(2) Br(2vii)–Cu(4)–Br(2) 117.5(4) 103.5(2) 109.43(7) 113.54(12) 110.3(5) 110.1(2) 100.4(3) 124.59(10) N(1)–Cu(1)–I(1iii) N(1)–Cu(1)–I(1) I(1iii)–Cu(1)–I(1) N(1)–Cu(1)–I(1iv) I(1iii)–Cu(1)–I(1iv) I(1)–Cu(1)–I(1iv) 109.9(2) 106.2(2) 112.07(4) 102.2(2) 117.10(4) 108.39(3) Symmetry operators: a i x, y 1 1, z; ii 2x 1 0.5, y 1 0.5, z; iii 2x, y 2 0.5, 2z 1 0.5; iv 2x, 2y 1 1, 2z 1 1; v 2x 1 0.5, y 2 0.5, z; vi x, y 2 1, z.b i x 2 1, y, z; ii x 2 0.5, y, 2z 1 0.5; iii 2x, 2y, 2z 1 1; iv x, 2y 1 0.5, z; v x, 2y 2 0.5, z; vi x 1 1, y, z; vii x 1 0.5, y, 2z 1 0.5. c i x 2 y, x 2 1, z 2 0.5; ii 2x 1 2, 2y 1 1, z 2 0.5; iii x, x 2 y, z; iv 2x 1 y 1 2, y, z; v y 1 1, 2x 1 y 1 1, z 1 0.5; vi 2x 1 y 1 2, 2x 1 1, z; vii 2x 1 2, 2y 1 1, z 1 0.5.d i 2x 1 0.5, 2y 2 0.5, z 1 0.5; ii x, 2y, z 1 0.5; iii 2x 1 0.5, 2y 2 0.5, z 2 0.5; iv x, 2y, z 2 0.5. Table 2 Crystallographic parameters for [Cu3Br3(tri)]• 1, [Cu3I3(tri)]• 2, [Cu2Br2(tri)]• 3 and [Cu2I2(tri)]• 4 1 2 3 4 Chemical formula Formula weight Crystal system Space group T/K a/Å b/Å c/Å V/Å3 Z m(Mo-Ka)/mm21 Reflections collected Unique reflections, Rint Observed reflections [I > s(I)] R1 [I > 2s(I)] wR2 [all data] C3H3Br3Cu3N3 511.43 Orthorhombic Pbca 220(2) 16.096(3) 6.5180(13) 18.010(4) 1889.5(7) 8 19.322 6952 2074, 0.0959 1776 0.0391 0.0968 C3H3Cu3I3N3 652.40 Orthorhombic Pnma 150(2) 6.913(4) 9.777(5) 15.570(9) 1064.5(10) 4 14.589 3125 1108, 0.0366 961 0.0293 0.0755 C6H6Br4Cu4N6 735.97 Hexagonal P63mc 150(2) 14.247(6) — 6.415(3) 1127.6(9) 3 16.205 1277 410, 0.0961 377 0.0328 0.0819 C3H3Cu2I2N3 461.96 Orthorhombic Cmc21 150(2) 15.380(3) 7.725(2) 6.9130(14) 821.3(3) 4 12.621 2653 751, 0.1324 747 0.0257 0.06422110 J.Chem. Soc., Dalton Trans., 1999, 2103–2110 C3H3Cu2I2N3): C, 7.70 (7.80); H, 0.50 (0.65); N, 8.75 (9.10)%]. IR n& /cm21 (triazine unless stated otherwise): 1573s, 1545s, 1408s, 1169w, 1124w, 711s and 682m. Yellow block-shaped crystals suitable for X-ray diVraction studies were grown by layering a CuI solution in MeCN on top of a solution of triazine in CH2Cl2 (molar ratio 2 : 1).Crystallography Crystal data and summaries of the crystallographic analyses for complexes 1–4 are given in Table 2. DiVraction data were collected on either a Stoë Stadi-4 diVractometer equipped with an Oxford Cryosystems open flow cryostat 26 using w–q scans and graphite monochomated Mo-Ka radiation (for 2, 3 and 4) or a Stoë IPDS image plate diVractometer equipped with an Oxford Cryosystems open flow cryostat (for 1).26 Data were corrected for Lorentz and polarisation eVects.Absorption corrections were applied either numerically (2 , 3 and 4) or using XABS2 (1).27 The structures were solved by direct methods using SHELXS 9728 and full-matrix least squares refinement undertaken using SHELXL 97.29 All hydrogen atoms were placed in geometrically calculated positions and thereafter refined using a riding model with Uiso(H) = 1.2 Ueq(C). All non-hydrogen atoms were refined with anisotropic displacement parameters except for the carbon and nitrogen atoms in 3 which displayed poor displacement parameters and were refined isotropically.For all four structures the largest residual electron density features lie near the copper and halogen centres. The assignment of the absolute structures for 3 and 4 was confirmed by the refinement of Flack enantiopole parameters to values of 20.02(5) and 0.03(5), respectively. CCDC reference number 186/1469. See http://www.rsc.org/suppdata/dt/1999/2103/ for crystallographic files in .cif format.Acknowledgements We thank the EPSRC for the provision of a diVractometer and financial support (to N. R. B., W.-S. L. and P. A. C.). References 1 S. R. Batten and R. Robson, Angew. Chem., Int. Ed., 1998, 37, 1460; O. M. Yaghi and G. Li, Angew. Chem., Int. Ed. Engl., 1995, 34, 207; M. Munakata, L. P. Wu and T. Kuroda-Sowa, Adv. Inorg. Chem., 1998, 46, 17; A. J. Blake, N. R. Champness, P. Hubberstey, W.-S. Li, M. A. Withersby and M. Schröder, Coord. Chem. Rev., 1999, 183, 117; N.R. Champness and M. Schröder, Curr. Opin. Solid State Mater. Chem., 1998, 3, 419. 2 P. C. Healy, C. Pakawatchai, C. L. Raston, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1983, 1905. 3 J. A. Campbell, C. L. Raston and A. H. White, Aust. J. Chem., 1977, 30, 1937. 4 P. C. Healy, J. D. Kildea, B. W. Skelton and A. H. White, Aust. J. Chem., 1989, 42, 115. 5 N. P. Roth, J. L. Maxwell and E. M. Holt, J. Chem. Soc., Dalton Trans., 1986, 2449. 6 K. Nilsson and A. Oskarsson, Acta Chem.Scand., Sect. A, 1985, 39, 663; J. P. Jasinski, N. P. Roth and E. M. Holt, Inorg. Chim. Acta, 1985, 97, 91; M. Massaux, J.-M. Bernaud and M.-T. Le Bihan, Bull. Soc. Chim. Fr., Mineral Crystallogr., 1969, 92, 118; M. Massaux and J.-M. Bernaud, Acta Crystallogr., Sect. B, 1971, 27, 2419; P. Jones, Acta Crystallogr., Sect. C, 1992, 48, 1307; M. Massaux and M.-T. Le Bihan, Acta Crystallogr., Sect. B, 1976, 32, 1586; 2032; I. D. Brown and J. D. Dunitz, Acta Crystallogr., 1960, 13, 28; E.Eitel, D. Oelkrug, W. Hiller and J. Strähle, Z. Naturforsch., Teil B, 1980, 35, 1247; M. Bolte and M. Massaux, Inorg. Chim. Acta, 1981, 52, 191; P. C. Healy, J. D. Kildea, B. W. Skelton and A. H. White, Aust. J. Chem., 1989, 42, 79; 93. 7 M. Massaux, G. Ducreux, R. Chevalier and M.-T. Le Bihan, Acta Crystallogr., Sect. B, 1978, 34, 1863. 8 P. C. Healy, J. D. Kildea and A. H. White, Aust. J. Chem., 1989, 42, 137. 9 P. C. Healy, C. Pakawatchai and A.H. White, J. Chem. Soc., Dalton Trans., 1983, 1935. 10 V. Schramm, Inorg. Chem., 1978, 17, 714; J. C. Dyason, P. C. Healy, L. M. Engelhardt, C. Pakawatchai, V. A. Patrick, C. L. Raston and A. H. White, J. Chem. Soc., Dalton Trans., 1985, 831; M. R. Churchill, G. Davies, M. A. El-Sayed, J. P. Hutchinson and M. W. Rupich, Inorg. Chem., 1982, 21, 995; C. L. Raston and A. H. White, J. Chem. Soc., Dalton Trans., 1976, 2153; V. Schramm and K. F. Fischer, Naturwissenschaften, 1974, 61, 500; V.Schramm, Cryst. Struct. Commun., 1982, 11, 1549; L. M. Engelhardt, P. C. Healy, J. D. Kildea and A. H. White, Aust. J. Chem., 1989, 42, 107. 11 J. Y. Lu, B. R. Cabrera, R.-J. Wang and J. Li, Inorg. Chem., 1998, 37, 4480. 12 M. Massaux, M.-T. Le Bihan and R. Chevalier, Acta Crystallogr., Sect. B, 1977, 33, 2084. 13 M. J. Begley, O. Eisenstein, P. Hubberstey, S. Jackson, C. E. Russell and P. H. Walton, J. Chem. Soc., Dalton Trans., 1994, 1935. 14 A. J. Blake, N. R. Brooks, N. R. Champness, P. A. Cooke, M. Crew, A. Deveson, D. Fenske, L. R. Hanton, P. Hubberstey and M. Schröder, Crystal Engineering, in the press. 15 J. M. Moreno, J. Suarez-Valera, E. Colacio, J. C. Avila-Rosón, M. A. Hidalgo and D. Martin-Ramos, Can. J. Chem., 1995, 73, 1591. 16 S. Kawata, S. Kitagawa, H. Kumagai, S. Iwabuchi and M. Katada, Inorg. Chim. Acta, 1998, 267, 143. 17 M. Munakata, T. Kuroda-Sowa, M. Maekawa, A. Honda and S. Kitagawa, J. Chem. Soc., Dalton Trans., 1994, 2771. 18 A. J. Graham, P. C. Healy, J. D. Kildea and A. H. White, Aust. J. Chem., 1989, 42, 177. 19 O. M. Yaghi and G. Li, Angew. Chem., Int. Ed. Engl., 1995, 34, 207. 20 D. Venkataraman, S. Lee, J. S. Moore, P. Zhang, K. A. Hirsch, G. B. Gardner, A. C. Covey and C. L. Prentice, Chem. Mater., 1996, 8, 2030. 21 M. Bertelli, L. Carlucci, G. Ciani, D. M. Proserpio and A. S. Sironi, J. Mater. Chem., 1997, 7, 1271. 22 D. Adam, B. Hirrschaft and H. Hartl, Z. Naturforsch., Teil B, 1991, 46, 738. 23 M. Munakata, L. P. Wu and T. Kuroda-Sowa, Bull. Chem. Soc. Jpn., 1997, 70, 1727; M. Munakata, T. Kuroda-Sowa, M. Maekawa, A. Hirota and S. Kitagawa, Inorg. Chem., 1995, 34, 2705. 24 A. F. Wells, Structural Inorganic Chemistry, 5th edn., Clarendon Press, Oxford, 1984. 25 D. D. Perrin and W. L. F. Armarego, Purification of Laboratory Chemicals, Pergamon, Oxford, 3rd edn., 1988. 26 J. Cosier and A. M. Glazer, J. Appl. Crystallogr., 1986, 19, 105. 27 S. Parkin, B. Moezzi and H. Hope, J. Appl. Crystallogr., 1995, 28, 53. 28 G. M. Sheldrick, SHELXS 97, Acta Crystallogr., Sect. A, 1990, 46, 467. 29 G. M. Sheldrick, SHELXL 97, University of Göttingen, 1997. Paper 9/02290B

ISSN:1477-9226    

DOI:10.1039/a902290b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2109–2110 2109 Metal-template synthesis and co-ordination properties of a palladium complex containing a novel and stable imidazole-substituted phosphine C]P bidentate chelate Huifang Lang, Jagadese J. Vittal and Pak-Hing Leung * Department of Chemistry, National University of Singapore, Kent Ridge 119260, Singapore An organopalladium complex promoted Diels–Alder reaction between 1-phenyl-3,4-dimethylphosphole and 1-vinylimidazole gave a novel imidazole-substituted phosphanorbornene bidentate ligand which co-ordinated to the palladium template via the C2 carbon atom of the imidazole group and the bridgehead phosphorus donor atom. Transition-metal complexes containing imidazole and its derivatives play an important role in bioinorganic chemistry.1 These compounds are frequently considered as models in the development of metal-based enzymes and proteins.In terms of co-ordination chemistry, imidazole may be considered as an ambidentate ligand. For instance, it has been well established that the imidazole unit co-ordinates to transition-metal ions, such as palladium(II) and platinum(II), predominantly via one of its nitrogen atoms.In some rare cases, however, the cyclic unit may also co-ordinate as a carbene or an amidine ligand via its C2 carbon atom. This interesting mode of organometallic bonding has been observed in a small number of Cr0, Fe0, RuII and RuIII complexes.2 Apart from their biological interest, the availability of stable C-bound imidazole heavy-metal complexes is important for the study of the trans influence 3 and other related phenomena that are pertinent to the design and development of eYcient support ligands for homogenous catalysis.4 Indeed, it has been reported that palladium complexes containing C-bound imidazole ligands are eYcient catalysts for crosscoupling reactions.5 In general, however, palladium complexes containing monodentate ligands are kinetically labile.We believe that the development of a new class of stable C-bound imidazole complexes may significantly influence the design of catalysts in homogenous catalysis. Here we report the palladium-template synthesis of the first imidazole-substituted tertiary phosphine bidentate ligand in which the C2 carbon of the imidazole unit is involved in metal chelation. In the absence of the transition-metal ion, no Diels–Alder reaction was observed between 1-phenyl-3,4-dimethylphosphole (DMPP) and 1-vinylimidazole.Upon co-ordination to palladium, however, DMPP is activated towards the [4 1 2] cycloaddition reaction. Thus, when [PdCl2(DMPP)2] or the organopalladium complex 1 6 was treated with the dienophile in 1,2-dichloroethane at 84 8C, the exo-cycloadduct 2 was obtained as the sole product (Scheme 1).† When the template complex 1 was used, the cycloaddition reaction was completed in 30 d. Interestingly, during the course of heating, the N3 nitrogen of the imidazole unit underwent a parallel N-alkylation reaction with 1,2-dichloroethane and produced HCl as the side product.7 The HCl thus generated led to further chemoselective cleavage of the orthometallated benzylamine ligand from the palladium template and thus facilitate the formation of the dichloro complex 2 as the final product.6 The * E-Mail: chmlph@nus.edu.sg molecular structure and the co-ordination chemistry of 2 have been determined by X-ray structural analysis (Fig. 1).‡ The study reveals that the reaction of 1 with 1-vinylimidazole in 1,2- dichloroethane has resulted in the removal of the benzylamine chelate and the imidazole substituted exo-phosphanorbornene ligand created co-ordinates to palladium as a bidentate chelate via the bridgehead phosphorus atom and the C2 carbon atom of the imidazole group. The palladium atom is in a slightly distorted square-planar geometry with the bond angles in the ranges 84.8(1)–94.8(2) and 167.0(1)–172.2(2)8. Due to the aromaticity of the imidazole ring, all C]C and C]N bonds [1.350(8)–1.383(6) Å] within the five-membered ring are noticeably shorter than the two attached C]N bonds [1.473(7)– 1.453(7) Å].The Pd(1)]C(11) distance of 1.993(5) Å is similar to the Pd]C bonds observed in other reported complexes containing the orthometallated benzylamine [2.004(11) Å] and naphthylamine units (2.006 Å) which experience a similar trans electronic influence from a chloro ligand.8 The Pd(1)]P(1) distance of 2.207(1) Å is also within the normal range observed for this class of phosphanorbornene complex.The Pd ? ? ? Cl(3) distance is 3.622(2) Å indicating that there is no interaction between the two heavy atoms. Interestingly, when [PdCl2(DMPP)2] was used for the Diels– Alder reaction, a much lower reaction rate was observed. The 31P NMR studies indicated that 80% of [PdCl2(DMPP)2] Scheme 1 Pd N P Cl Me Me Me Ph P Cl Pd Cl Ph Me or [PdCl2(DMPP)2] N N N N 1 2 ClCH2CH2 ClCH2CH2Cl 2 3 1 † Preparation of complex 2.A solution of the organopalladium complex 1 (0.46 g, 0.97 mmol) in 1,2-dichloroethane (40 cm3) was treated with silver perchlorate (0.2 g, 0.97 mmol) in water (1 cm3) for 30 min. The resulting mixture was filtered through a layer of Celite to remove silver chloride and the organic layer was dried over anhydrous MgSO4. The dried solution was treated with 1-vinylimidazole (0.37 g, 3.87 mmol) and the reaction mixture was then stirred at 84 8C for 30 d.The solution was washed with water (50 cm3) and then dried over MgSO4. The solvent was removed under reduced pressure to give a yellow residue. Upon crystallization from acetonitrile–diethyl ether, the dichloro complex 2 was obtained as yellow prisms (0.21 g, 40%), m.p. 256– 258 8C (decomp.). 31P NMR (CDCl3): d 105.3 (s). ‡ Crystal data for 2: C19H22Cl3N2PPd, M = 522.11, orthorhombic, space group Pna21, a = 14.9398(1), b = 16.5294(1), c = 8.6207(1) Å, U = 2128.85(3) Å3, Z = 4, Dc = 1.629 g cm23, T = 293 K, m(Mo-Ka) = 13.29 cm21, F(000) = 1048, R1 = 0.0408, wR2 = 0.0885 for 4214 independent observed reflections [I > 2s(I), 1.84 < 2q < 29.318] and 236 parameters.The Flack parameter was refined to 20.02(4). CCDC reference number 186/1018.2110 J. Chem. Soc., Dalton Trans., 1998, Pages 2109–2110 remains unchanged after the complex was treated with 1-vinylimidazole for 30 d under similar reaction conditions.Nevertheless, a small quantity of 2 was observed in the 31P NMR spectra of the reaction mixture. Clearly, the co-ordinated cyclic diene in Fig. 1 Molecular structure and co-ordination chemistry of complex 2. Selected bond lengths (Å) and angles (8): Pd(1)]P(1) 2.207(1), Pd(1)]C(11) 1.993(5), Pd(1)]Cl(1) 2.391(2), Pd(1)]Cl(2) 2.358(2), P(1)]C(1) 1.859(6), P(1)]C(4) 1.842(5), N(1)]C(6) 1.473(7), N(1)]C(11) 1.375(7), N(1)]C(9) 1.383(6), C(9)]C(10) 1.350(8), C(10)]N(2) 1.381(7), N(2)]C(11) 1.357(7), N(2)]C(12) 1.453(7), C(13)]Cl(3) 1.781(9); P(1)]Pd]Cl(1) 167.0(1), P(1)]Pd]Cl(2) 84.8(1), P(1)]Pd] C(11) 87.8(2), C(11)]Pd]Cl(1) 94.8(2), C(11)]Pd]Cl(2) 172.2(2), Cl(1)] Pd]Cl(2) 92.0(1), C(1)]P(1)]C(4) 81.8(3), C(6)]N(1)]C(11) 124.0(4), N(1)]C(9)]C(10) 106.3, C(9)]C(10)]N(2) 107.7(4), C(10)]N(2)]C(11) 110.7(4), N(2)]C(11)]N(1) 104.4(4) the template complex 1 receives a higher degree of activation as compared with its counterparts in [PdCl2(DMPP)2]. We are currently investigating the optical resolution and the catalytical properties of transition-metal complexes containing the imidazole-substituted phosphine ligand.Acknowledgements We are grateful to the National University of Singapore for support of this research (Grant No. RP972667) and research scholarships (to H. F. L.). References 1 R. J. Sundberg and R. B. Martin, Chem. Rev., 1974, 74, 471; M. Grebl and B. Krebs, Inorg. Chem., 1994, 33, 3877. 2 F. A. Cotton and G. Wilkinson, in Advanced Inorganic Chemistry, John Wiley & Son, New York, 5th edn., 1988, ch. 10. 3 M. F. Tweedle and H. Taube, Inorg. Chem., 1982, 21, 3361. 4 K. H. Weiss, in Transition Metal Carbene Complexes, Verlag Chemie, Weinheim, 1983, p. 227. 5 W. A. Herrmann and C. Kocher, Angew. Chem., Int. Ed. Engl., 1997, 36, 2162. 6 S. Selvaratnam, P. H. Leung, A. J. P. White and D. J. Williams, J. Organomet. Chem., 1997, 542, 61. 7 C. A. Ghilardi, S. Midollini, S. Moneti, A. Orlandini and J. A. Ramirez, J. Chem. Soc., Chem. Commun., 1989, 304. 8 J. Albert, J. Granell and J. Sales, Organometallics, 1995, 14, 1393; S. Y. Siah, P. H. Leung and K. F. Mok, J. Chem. Soc., Chem. Commun., 1995, 1747. Received 20th May 1998; Communication 8/03820A

ISSN:1477-9226    

DOI:10.1039/a803820a

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 2111–2113 2111 Turning dihydrogen gas into a strong acid. Formation and reactions of the very acidic ruthenium dihydrogen complexes trans-[Ru(H2)- (CNH){PPh2(CH2)nPPh2}2][O3SCF3]2 (n 5 2 or 3) † Tina P. Fong,a Alan J. Lough,a Robert H. Morris,*,a Antonio Mezzetti,b Eliana Rocchini c and Pierluigi Rigo *,c a Department of Chemistry, University of Toronto, 80 St. George St., Toronto, Ontario, Canada M5S 3H6 b Laboratorium für Anorganische Chemie, ETH-Zentrum, Universitätstrasse 16, CH-8092 Zürich, Switzerland c Dip.di Scienze e Tecnol. Chimiche, Università di Udine, Via del Cotonificio 108, I-33100 Udine, Italy New, very acidic ruthenium dihydrogen complexes containing the hydrogen isocyanide ligand have been synthesised; when formed under 1 atm H2 they have been shown to spontaneously eliminate trifluoromethylsulfonic acid. Some of us recently reported that the protonation of trans-[FeH(CN)(dppe)2] or trans-[FeH(CNH)(dppe)2]OTf with CF3SO3H (HOTf) ‡ in the appropriate ratio gives trans- [Fe(H2)(CNH)(dppe)2][OTf]2, which is very acidic but surprisingly stable with respect to loss of H2(g).1 Similar reaction pathways are observed for the related ruthenium and osmium complexes (Scheme 1).2 We now find that the very acidic ruthenium analogues trans-[Ru(H2)(CNH)L2][OTf]2 (L = dppe 4a, L = dppp 4b) can be generated by reaction of the new triflate complexes trans-[Ru(OTf)(CNH)L2]OTf (5a, 5b) with dihydrogen gas.These complexes then eliminate HOTf in the absence of excess acid although it is not known whether the proton comes from the H2 or the CNH ligand. This is a signifi- cant new reaction pathway involving dihydrogen complexes: the in situ production of a very strong acid, in this case HOTf, triggered by the reaction of non-acidic H2(g) with a co-ordination complex which is not a strong Brønsted acid. Although very acidic dihydrogen complexes have been reported,1,3–8 there is only one other complex which is prepared from dihydrogen gas.9 This one case involves an unstable iridium dihydrogen complex which can protonate the tetraphenylborate anion in THF.9 There is evidence for the elimination of triflic acid from some iridium hydride complexes but it is not known whether dihydrogen complexes are involved.10–12 The reaction of the complexes trans-[RuH(CN)L2] 12 in CH2Cl2 solution under 1 atm of H2 with an excess of HOTf gives the dihydrogen complexes trans-[Ru(H2)(CNH)L2][OTf]2 4a,§ 4b.¶ They can also be prepared by reaction of complexes trans-[RuH(CNH)L2]OTf 3a, 3b2 with excess HOTf in CH2Cl2 (Scheme 1).The related osmium complexes have also been prepared. 2 The highly acidic ruthenium dihydrogen complexes have so far only been characterized in solution. The 31P-{1H} NMR spectrum of 4a is a sharp singlet at room temperature while that of 4b is a broad singlet. At 183 K the latter complex gives the A2X2 pattern that has been observed for trans-[MXY(dppp)2] † Non-SI unit employed: atm = 101 325 Pa.‡ Abbreviations used: dppe = 1,2-bis(diphenylphosphino)ethane; dppp = 1,3-bis(diphenylphosphino)propane; dtpe = 1,2-bis(ditolylphosphino) ethane; OTf = trifluoromethylsulfonate. species.4 The presence of the NH group in complexes 4 is signalled by a broad resonance in the 1H NMR spectrum in the region at d 9.6 for 4a and 13.7 for 4b. The latter signal is observed only at 183 K; at 293 K the resonance is averaged with Scheme 1 [M] is the fragment [Ru(dppe)2] or [Ru(dppp)2] [M] H C N [M]+ H-H C N [M]+ H C N [M]2+ H-H C N H H 1 2 3 4 H+ H+ H+ H+ § trans-[Ru(h2-H2)(CNH)(dppe)2][OTf]2 4a.Method 1: trans-[RuH- (CN)(dppe)2] (1a, 100 mg, 0.11 mmol) was dissolved in 10 mL of CH2Cl2 producing a clear colourless solution. Excess triflic acid (60 mg, 0.40 mmol) was added to the solution and the resulting light yellow solution was stirred for 1 h.The solvent was removed in vacuo, producing a yellow oil. Method 2: trans-[RuH(CNH)(dppe)2][OTf] (3a, 15 mg, 0.02 mmol) was dissolved in 5 mL of CD2Cl2 and triflic acid (7 mg, 0.05 mmol) was added to the solution. The spectra were recorded immediately. 1H NMR (300 MHz, CD2Cl2): d 12.7 (s, HOTf), 9.6 (br, NH), 7.8–6.8 (m, Ph), 2.9–2.4 (m, 8 H, CH2), 25.9 [br, Ru(h2-H2)]. T1(min): 300 MHz, CD2Cl2, 13.6 ms, 246 K. 31P-{1H} NMR (120.5 MHz, CD2Cl2): d 52.2 (s). trans-[Ru(HD)(CND)(dppe)2][OTf]2, 4a-d2.Method 2 was followed except deuteriated triflic acid (DOTf) was used instead. 1H NMR (300 MHz, CD2Cl2): d 26.0 [t, 1J(HD) = 32.4 Hz, Ru(HD)]. 31P-{1H} NMR (120.5 MHz, CD2Cl2): d 52.2 (s). ¶ trans-[Ru(h2-H2)(CNH)(dppp)2][OTf]2 4b. trans-[RuH(CN)(dppp)2] (20 mg, 21 mmol) was dissolved in 0.5 mL of CD2Cl2 under H2 in an NMR tube and CF3SO3H (6 mL, 68 mmol) was added thereto by means of a syringe. IR (CH2Cl2), cm21: n(CN) 2125 (s). 1H NMR (CD2Cl2, 293 K, 200 MHz): d 7.6–6.9 (m, Ph), 2.4 (br, 8 H, PCH2), 1.9 (br, 2 H, PCH2CH2), 1.6 (br, 2 H, PCH2CH2), 24.2 (br, 2 H, RuH2). 31P-{1H} NMR (CD2Cl2, 293 K, 81 MHz): d 8.9 (br), T = 183 K, d 3.2 (t), 15.6 [t, J(P,P9) = 30.1 Hz]. trans-[Ru(HD)(CND)(dppp)2][OTf]2 4b-d2. Excess DOTf was used in the method above. 1H NMR (200 MHz, CD2Cl2): d 24.2 [t, 1J(HD) = 31.8 Hz, Ru(HD)]. trans-[Ru(h2-H2)(13CNH)- (dppp)2][OTf]2. 31P-{1H} NMR (CD2Cl2): d 8.9 [d, J(13C31P) 13.5 Hz]. 13C-{1H} NMR (CD2Cl2): d 149.9 [q, J(13C31P) 13.6 Hz].2112 J.Chem. Soc., Dalton Trans., 1998, Pages 2111–2113 that of free HOTf because of fast proton exchange. This signal splits into a doublet with 1J(H15N) 108.1 Hz when 4b is prepared with the C15NH ligand. The CNH ligand has also been detected by IR and 13C NMR. The dihydrogen ligand in complexes 4a and 4b gives a broad resonance at d 25.9 and 24.2, respectively, with a characteristically short minimum T1 time of 13.6 ms (at 246 K, 300 MHz) and 5.9 ms (at 223 K, 200 MHz).The corresponding h2-HD complexes are prepared by reacting complexes 1b or 3a with excess DOTf in CD2Cl2. The large 1J(HD) coupling constants of 32.4 Hz for 4a and 31.8 Hz for 4b combined with the T1(min) data indicate that 4a and 4b have rapidly spinning H2 ligands with H]H distances of 0.88 and 0.89 Å, respectively.13 The high acidity of these complexes is illustrated by the chemistry of 4a. When a CD2Cl2 solution of 4a under H2(g) is treated with an excess of the weak base, diethyl ether, complex 2a || forms immediately [equation (1)].The dihydrogen ligand of trans-[Ru(h2-H2)(CNH)(dppe)2]21 1 Et2O 4a trans-[Ru(h2-H2)(CN)(dppe)2]1 1 Et2OH1 (1) 2a 2a is identified by a broad peak at d 25.5 with a minimum T1 of 12.4 ms at 240 K, 300 MHz. The corresponding HD complex has 1J(HD) 32.0 Hz. These two data indicate that the H2 ligand in 2a is fast spinning with an H]H distance of 0.89 Å. Complexes 4b are also deprotonated by diethyl ether to give a mixture of the dihydrogen complex trans-[Ru(h2-H2)(CN)(dppp)2]1 2b and the hydrogen isocyanide complex trans-[Ru(H)(CNH)- (dppp)2]1 3b.2 The dicationic dihydrogen complexes 4 are less stable with respect to loss of H2 than the analogous iron complex.1 Evaporation of solvent leaves yellow oils of complexes 4 and excess acid. These oils slowly lose H2 under Ar to give mainly the complexes trans-[Ru(OTf)(CNH)L2]OTf [equation (2), L = dppe 5a,** L = dppp 5b ††].Complexes 5 can be identified trans-[Ru(h2-H2)(CNH)L2][OTf]2 4a or 4b trans-[Ru(OTf)(CNH)L2]OTf 1 H2 (2) 5a or 5b || trans-[Ru(h2-H2)(CN)(dppe)2][HOTf–OTf] 2a. A yellow oil containing 4a in HOTf was stirred for 30 min in Et2O under 1 atm H2 to form the product. 1H NMR (300 MHz, CD2Cl2): d 13.1 (s, TfOH–OTf), 7.8–6.6 (m, Ph), 2.5–3.0 (m, 8 H, CH2), 25.5 [br, Ru(h2-H2)]; T1(min): 12.4 ms, 240.3 K. 31P-{1H} NMR (120.5 MHz, CD2Cl2): d 54.2 (s). trans-[Ru(h2- HD)(CN)(dppe)2]1.Diethyl ether was added to the yellow oil of 4a-d2 to produce a light yellow precipitate. The solvent was decanted and the product was quickly dried under argon. The product under Ar loses HD and must be isolated and analysed without delay. 1H NMR (300 MHz, CD2Cl2): d 25.5 [t, 1J(HD) = 32.0 Hz, Ru(HD)]. 31P-{1H} NMR (120.5 MHz, CD2Cl2): d 54.1 (s). ** trans-[Ru(OTf)(CNH)(dppe)2]OTf 5a. Diethyl ether was added to the yellow oil of 4a under Ar, producing a light yellow precipitate. The solvent was decanted and the precipitate was washed twice with 5 mL of diethyl ether and dried in vacuo.Yield of crude 5a 60%. Yellow crystals were obtained by slow evaporation of a concentrated solution of the product in CH2Cl2. 1H NMR (300 MHz, CD2Cl2): d 10.5 [t, 1J(HN) = 79 Hz, NH], 7.8–6.6 (m, Ph), 3.0–2.8 (m, 8 H, CH2). 31P-{1H} NMR (120.5 MHz, CD2Cl2): d 48.8 (s) (Found: C, 53.66; H, 4.35; N, 1.32. Calc. for C55H49F6NO6P4RuS2: C, 54.01; H, 4.04; N, 1.14%).†† trans-[Ru(CNH)(OTf)(dppp)2]OTf 5b. trans-[RuH(CN)(dppp)2] (1b, 200 mg, 0.21 mmol) was dissolved in 20 mL of CH2Cl2. Triflic acid (60 ml, 0.68 mmol) was added and the solution was stirred at room temperature for 20 min under argon bubbling. The solvent was removed in vacuo and diethyl ether was added producing a white-pale yellow precipitate. The product was filtered oV, washed with diethyl ether, and dried in vacuo. Recrystallization from CH2Cl2–diethyl ether yielded 0.21 g, 80% (Found: C, 53.86; H, 4.33; N, 1.10.Calc. for C57H53F6- NO6P4RuS2: C, 54.72; H, 4.27; N, 1.12%). IR (Nujol), cm21: n (CN) 2074w. 1H NMR (CD2Cl2, 293 K, 200 MHz): d 7.6–6.7 (m, PC6H5), 2.5 (br, 8 H, PCH2), 2.1 (br, 4 H, PCH2CH2). 31P-{1H} NMR (CD2Cl2, 293 K, 81 MHz): d 1.8 (br), T = 193 K, d 27.3 (t), 0.9 [t, J(P,P9) = 32.7 Hz]. by a characteristic 1HN 1 : 1 : 1 triplet in the 1H NMR spectrum at d 10.5 [1J(NH) 79 Hz] for 5a or by a broad singlet at d 11.0 at 183 K for 5b.Complexes 5 give singlets in the room temperature 31P-{1H} NMR spectra at d 48.8 for 5a and 1.8 for 5b, respectively. A single-crystal X-ray diVraction study of 5a ‡‡ reveals the presence of a co-ordinated triflate and a triflate anion which is hydrogen bonded to an NH group of a slightly bent CNH unit (C]N]H 170.48) (Fig. 1). The CNH ligand has similar dimensions to the one of the complex trans-[FeH- (CNH)(dtpe)2]BF4.15 The Ru]O(1) distance of 2.299(2) Å is long in comparison to the range of Ru]O distances of 2.177(4) to 2.233(2) Å observed in other ruthenium(II)–triflate complexes.16–18 The crowded Ru(dppe)2 site and the high trans influence of the CNH ligand cause a weakening of the Ru]O bond and this allows the weak dihydrogen ligand to co-ordinate in its place (see below).Complex 5a is a weak Brønsted acid. It is not deprotonated by diethyl ether or triphenylphosphine. When complex 5a in CD2Cl2 with excess HOTf is reacted with 1 atm H2, complex 4a is formed in less than 5 min as expected for the reverse of equation (2).Significantly, when complex 5a in CD2Cl2 is placed under 1 atm H2 in the absence of HOTf, the dihydrogen complex trans-[Ru(h2-H2)(CN)- (dppe)2]1 2a is produced along with 1 equivalent of triflic acid, probably present mainly as [TfO–HOTf]27 (Scheme 2). The hydrogen-bonded triflic acid–triflate cluster is identified by 1H NMR spectroscopy as a broad peak at d 13.1. Complex 4a is the likely intermediate in this reaction.However, since it is only Fig. 1 An ORTEP14 diagram of complex 5a. Thermal ellipsoids represent the 50% probability surface. The hydrogen on the nitrogen was located in Fourier electron diVerence map. Selected bond lengths (Å) and angles (8): Ru]O(1) 2.299(2), Ru]C(5) 1.883(3), Ru]P(1), 2.3938(7), Ru]P(2) 2.3851(8), Ru]P(3) 2.4363(8), Ru]P(4) 2.4144(8), C(5)]N(1) 1.149(4), N(1)]H(1N) 0.77, H(1N)]O(3S) 1.86; O(1)]Ru]C(5) 171.3(1), Ru]C(5)]N(1) 177.2(3), C(5)]N(1)]H(1N) 170.4, N(1)]H(1N)]O(3S) 173.4 ‡‡ Crystal data for 5a: C55H49F6NO6P4RuS2, M = 1223.02, monoclinic, space group P21/c (no. 14), a = 9.8064(12), b = 22.121(2), c = 25.213(3) Å, b = 93.210(8)8, U = 5460.6(11) Å3, Dc = 1.488 g cm23, Z = 4, T = 173(2) K, m = 0.552 mm21. For reflections with 2.56 < q < 27.008, R(F) = 0.0365 for 7908 observed reflections [I > 2s(I)] and wR(F2) = 0.0914 for all 10 773 reflections. CCDC number 186/1011. See http:// www.rsc.org/suppdata/dt/1998/2111/ for crystallographic files in .cif format.J.Chem. Soc., Dalton Trans., 1998, Pages 2111–2113 2113 stable in the presence of excess HOTf (see above), it must eliminate triflic acid. The product expected from the heterolytic splitting of dihydrogen would be the monohydride complex trans-[Ru(H)(CNH)(dppe)2]OTf 3a. However as indicated by equation (1), 2a is the thermodynamically stable product. A similar, slower reaction between 5b and H2 produces a mixture of both 2b and 3b.However complex 2b can be quantitatively formed in CH2Cl2 solution by treating 5b with 1 equivalent of NEt3 and then reacting the product with 1 atm H2. Studies of the factors that influence the stability of the tautomers 2 and 3 and the properties of related complexes containing iron and osmium and the diphosphine ligands PEt2CH2CH2PEt2 and PPh2CH2PPh2 are in progress. Acknowledgements This work was supported by an NSERC operating grant and loan of ruthenium salts from Johnson Matthey PLC (to R.H. M.) and by Italian Ministero dell’Università e della Ricerca Scientifica (to P. R.). E. R. gratefully acknowledges the Regione Friuli-Venezia Giulia for a fellowship. Scheme 2 [Ru] is the fragment [Ru(dppe)2] or [Ru(dppp)2] [Ru]+ O C N H [Ru]2+ H-H C N H S O CF3 O [Ru]+ H-H C N [Ru]+ H C N H O3SCF3 – O3SCF3 – CF3SO3 – CF3SO3 – H2 5 + 4 3 2 –HOSO2CF3 References 1 C. E. Forde, S. E. Landau and R. H. Morris, J. Chem. Soc., Dalton Trans., 1997, 1663. 2 T. P. Fong, A. J. Lough, R. H. Morris, T. Stephan, P. Rigo and E. Rocchini, unpublished work. 3 M. S. Chinn, D. M. Heinekey, N. G. Payne and C. D. Sofield, Organometallics, 1989, 8, 1824. 4 E. Rocchini, A. Mezzetti, H. Ruegger, U. Burckhardt, V. Gramlich, A. Del Zotto, P. Martinuzzi and P. Rigo, Inorg. Chem., 1997, 36, 711. 5 M. Schlaf, A. J. Lough, P. A. Maltby and R. H. Morris, Organometallics, 1996, 15, 2270. 6 M. Schlaf, A. J. Lough and R. H. Morris, Organometallics, 1996, 15, 4423. 7 R. M. Bullock, J. S. Song and D. J. Szalda, Organometallics, 1996, 15, 2504. 8 T. A. Luther and D. M. Heinekey, Inorg. Chem., 1998, 37, 127. 9 C. Bianchini, S. Moneti, M. Peruzzini and F. Vizza, Inorg. Chem., 1997, 36, 5818. 10 R. C. Schnabel and D. M. Roddick, Organometallics, 1996, 15, 3550. 11 R. C. Schnabel and D. M. Roddick, Organometallics, 1993, 12, 704. 12 B. P. Cleary and R. Eisenberg, Inorg. Chim. Acta, 1995, 240, 135. 13 R. H. Morris and R. Wittebort, Magn. Res. Chem., 1997, 35, 243. 14 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 15 P. I. Amrhein, S. D. Drouin, C. E. Forde, A. J. Lough and R. H. Morris, Chem. Commun., 1996, 1665. 16 P. W. Blosser, J. C. Gallucci and A. Wojcicki, Inorg. Chem., 1992, 31, 2376. 17 C. Gemel, D. Kalt, K. Mereiter, V. N. Sapunov, R. Schmid and K. Kirchner, Organometallics, 1997, 16, 427. 18 J.-P. Sutter, S. L. James, P. Steenwinkel, T. Karlen, D. M. Grove, N. Veldman, W. J. J. Smeets, A. L. Spek and G. van Koten, Organometallics, 1996, 15, 941. Received 30th April 1998; Communication 8/03256D

ISSN:1477-9226    

DOI:10.1039/a803256d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 2115 Mobility of gold and silver ions around a macrocyclic polyphosphane. Supramolecular architecture of a digold–calix[4]arene complex Cedric B. Dieleman,a Dominique Matt,*,†,a Ion Neda,*,‡,b Reinhard Schmutzler,*,§,b Holger Thönnessen,b Peter G. Jones b and Anthony Harriman c a Groupe de Chimie Inorganique Moléculaire, UMR 7513 CNRS, 1 rue Blaise Pascal, F-67008 Strasbourg Cedex, France b Institut für Anorganische und Analytische Chemie der Technischen Universität, Postfach 3329, D-38023 Braunschweig, Germany c Ecole Européenne de Chimie, Polymères et Matériaux, 1 rue Blaise Pascal, F-67008 Strasbourg Cedex, France The calixarenes 5,11,17,23-tetra-tert-butyl-25,26,27,28-tetrakis(diphenylphosphinomethoxy)calix[4]arene (L1) and 5,11,17,23-tetra-tert-butyl-25,26,27,28-tetrakis(1,3,5-trimethyl-4,6-dioxo-1,3,5-triaza-2-phosphorinan- 2-yl)calix[4]arene (L2), substituted at the lower rim, reacted with 4 equivalents of [AuCl(tht)] (tht = tetrahydrothiophene) to yield respectively the tetragold complexes [(AuCl)4L1] 1 and [(AuCl)4L2] 2, both adopting a cone conformation.The solid state structure of 1 has been determined by an X-ray diVraction study, which reveals crystallographic twofold symmetry. The reaction of L2 with 2 equivalents of [AuCl(tht)] resulted in the formation of a mixture of two compounds. One, [(AuCl)2L2] 3, could be isolated and was characterised by an X-ray diVraction study. The gold atoms of 3 are tethered to two distal P atoms.Intermolecular Au ? ? ?Au contacts are observed for each gold atom [3.2879(7) Å], resulting in a loose polymeric structure in the solid state. Reaction of L1 with [AuCl(tht)] and TlPF6 resulted in quantitative formation of the dinuclear complex [Au2L1][PF6]2 4 where each gold atom is chelated by two proximal phosphino groups. Fast rotation of the gold atoms around the axis of the calixarene occurs in solution. In this co-operative intramolecular movement each gold atom switches from one pair of P atoms to an adjacent one (DG‡ = 79.4 kJ mol21).The silver analogue 5 was obtained by treating L1 with AgBF4. The calix[4]arene matrix is a macrocyclic building block that possesses a remarkable structural versatility and may undergo controlled multiple functionalisation.1–7 It has become a widely employed tool for engineering of highly sophisticated molecular and supramolecular systems.8–12 Among the most striking developments in this area are sensors for the detection of neutral and anionic species,13 selective complexants for the extraction of valuable metal ions,14–18 novel metallomesogenic materials,19 and molecular capsules for the entrapment of reactive organometallic fragments.20,21 A topic of growing interest deals with the combination of calixarenes and transition metals.22,23 Transition metals have notably been used for shaping calixarenes,24 and generating molecular-sized cages from calixarenes.25 Recent investigations by us and others have shown that tethering of four PIII-containing units to the narrow rim of coneshaped calix[4]arenes provides a complexation domain suitable for the binding of up to four transition metals.26–31 Such preorganised assemblies are expected to possess important properties and, in particular, allow examination of co-operative interactions between metal centres maintained in close proximity.It may also be anticipated that these units are highly appropriate vehicles for studying dynamic processes involving metal fragments bound to a calixarene surface. This latter topic is relevant to the understanding of metal ion transport in multitopic molecular systems.11,32,33 In the present study we report on the synthesis and properties of multimetal species obtained from the previously reported tetrafunctionalised calixarenes L1 and L2.30,34 These ligands, which contain four trivalent phosphorus atoms, adopt a cone- † E-Mail: dmatt@chimie.u-strasbg.fr ‡ E-Mail: i.neda@tu-bs.de § E-Mail: schmutzler@mac1.anchem.nat.tu-bs.de shaped structure in solution. X-Ray diVraction studies of a tetra- and a di-nuclear complex are described. We also report the first examples of dynamic processes involving two metal centres moving on a calixarene-subtended P4 surface.The ability of ligand L1 to form tetranuclear complexes with palladium has recently been outlined in a preliminary communication.29 Results and Discussion Tetranuclear gold complexes In order to investigate its multiple binding properties towards gold centres, L1 was allowed to react with 4 equivalents of [AuCl(tht)] (tht = tetrahydrothiophene). This reaction led to the tetranulcear complex 1 (Scheme 1) in quantitative yield.The FAB mass spectrum of 1 shows an intense peak at m/z 2335, corresponding to the [M 2 Cl] cation. The NMR data (1H, 31P, 13C) are consistent with a C4-symmetrical structure in solution (sharp signals), but from many studies it now appears likely that such a symmetry is only virtual and reflects a fast C2v–C2v R2P PR2 PR2 R2P PPh2 Ph2P Ph2P PPh2 L1 L2 N N N Me Me Me O O R2 = O O But But O But O But O O But But O But O But2116 J.Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 exchange process.35 A single-crystal X-ray determination (Fig. 1 and Table 1) confirmed the presence of four gold atoms anchored to the calixarene ligand and revealed that the highest symmetry element in the solid state is actually a crystallographic C2 axis.The calixarene matrix adopts a distorted cone shape with two facing phenolic units almost parallel (interplane angle 4.28) and the other two being essentially perpendicular (89.48). The distance between the centroids of the two parallel rings is 5.37 Å and that between the other two rings is 7.40 Å. Two opposite P]Au]Cl units are oriented parallel to their appended phenoxy rings, whilst the other two lie roughly orthogonal to the calixarene axis.The particular alignment of the two PAuCl fragments adjacent to the corresponding phenoxy rings (see Fig. 1) is likely to minimise steric interactions between the phosphino groups and the calixarene backbone and is not considered to reflect any bonding interaction between the gold atom and the p system of the appended phenoxy rings [shortest Au ? ? ? C contact 3.593(10) Å; Au(2) ? ? ? C(36)].The intramolecular Au(1) ? ? ?Au(1) distance is 5.277(1) Å and the shortest intermolecular distance is Au(1) ? ? ?Au(1) 5.269(1) Å (operator 1 2 x, 2y, 1 2 z). Other structural parameters are given in Table 1. The related tetranuclear gold complex 2 was obtained using a procedure similar to that outlined in Scheme 1 but starting from L2. Its 13C NMR spectrum shows that the cone shape of the calixarene is maintained upon complexation [d(C6H2CH2) = 31.44].Complex 2 shows a broad peak (d 101.6 vs. 91.8 for the free calixarene) in its 31P NMR spectrum, suggesting that the complex is either exhibiting dynamic behaviour in solution or exists as a mixture of structurally similar isomers. Careful examination of the 1H NMR spectrum reveals that the resonances of the C(O)NMe groups display a complex pattern. A possible explanation for this observation is that rotation of the relatively bulky phosphorus heterocycles around the P]O Scheme 1 S O O But But O But O But PPh2 Ph2P Ph2P PPh2 O O But But O But O But PPh2 Ph2P Ph2P PPh2 Au Au Au Au 4 [AuCl(tht)] Cl Cl Cl Cl tht = L1 CH2Cl2 1 Fig. 1 Molecular structure of complex 1 bonds is restricted, so that the four phosphorinane fragments adopt various orientations with respect to the calixarene axis. Note that rotamers have neither been detected in 1 nor in the related 25,26,27,28-tetra(diphenylphosphinooxy)calix[4]arene tetragold complex recently reported by Puddephatt and coworkers. 28 Other characteristic data for 2 are given in the Experimental section. Non-exhaustive auration of L1 and L2: formation of neutral complexes Since compound L1 contains four discrete binding sites we surmised that it would also allow formation of complexes with a lower gold content.As an approach to this subject, NMR spectroscopy was used to monitor the addition to L1 of 1, 2 and 3 equivalents of [AuCl(tht)]. For each experiment the room temperature 1H NMR spectrum, recorded 0.5 h after the reaction was completed, virtually corresponded to that of a new single species distinct from L1 and 1 and having a highly symmetrical structure (one But signal, a single AB quartet for the C6H2CH2 groups, one peak for the m-H; see Experimental section).These spectra were identical to those obtained upon reaction of the tetranuclear complex 1 with 3, 2 or 1 equivalent of L1, respectively. These findings unambiguously indicate that (i) partially aurated compounds must be formed during the addition experiments, and (ii) the resulting species undergo dynamic processes.The latter may be inter- or intra-molecular, or both. Similar observations arose from investigations of the reaction mixtures of L2 with various amounts of [AuCl(tht)]. Despite variable temperature studies on each of the above-mentioned reaction mixtures, the exact number of species formed in each of the gold addition (or displacement) experiments could not be determined.Neither was it possible to determine whether these reaction mixtures correspond to equilibria between species of different nuclearity. To date most studies demonstrating the lability of gold fragments have been carried out on cluster compounds.36 Very recently scrambling of Au(PPh3)1 units between two phosphorus( III) centres of a unique diphosphine was demonstrated by Assmann and Schmidbaur.37 In this case, intermolecular exchange occurs in competition to a bimolecular process.Thus, R2P PR2 P P N N N Me O Me Me O N N N Me Me Me O O Au Au Au Au Cl Cl Cl Cl 2 N N N Me Me Me O O R2 = O O But But O But O But Table 1 Selected bond lengths (Å) and angles (8) for complex 1 Au(1)]P(1) Au(2)]P(2) P(1)]C(51) P(1)]C(22) P(2)]C(71) P(1)]Au(1)]Cl(1) C(51)]P(1)]Au(1) C(22)]P(1)]Au(1) C(71)]P(2)]Au(2) 2.226(3) 2.223(3) 1.805(11) 1.845(10) 1.817(12) 175.35(11) 116.1(4) 115.0(3) 111.9(4) Au(1)]Cl(1) Au(2)]Cl(2) P(1)]C(61) P(2)]C(81) P(2)]C(42) P(2)]Au(2)]Cl(2) C(61)]P(1)]Au(1) C(81)]P(2)]Au(2) C(42)]P(2)]Au(2) 2.292(3) 2.271(3) 1.828(11) 1.797(11) 1.849(10) 176.28(14) 108.1(3) 114.7(3) 115.7(3)J.Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 2117 it is conceivable that, when a partially aurated complex is formed with L1 or L2 in one of the reactions described above, the AuCl units jump between molecules, so as to form species of diVering nuclearity. On the other hand, partially aurated species may also undergo intramolecular fluxional processes in which the gold atoms migrate between phosphorus sites.Addition of 0.5 equivalent of free L1 to a 1 : 1 Au:L1 reaction mixture caused the appearance of signals of free L1 besides those of the original dynamic species (see above), both in the 1H and 31P NMR spectra. This result shows that among the compounds formed in the 1 : 1 mixture there is at least one species in which intramolecular gold scrambling is favoured over intermolecular exchange.The probability of intermolecular gold transfer in an L1 complex was deduced from the [AuCl(tht)]- addition experiments described above, starting from the tetranuclear gold complex 1. The conclusive proof for the formation of partially aurated species was provided on crystallising a dichloromethane– toluene solution containing L2 with 2 equivalents of [AuCl(tht)]. Careful analysis of the single crystals formed revealed that at least two types of gold-containing species were present.The molecular structure of one of them, namely the digold complex 3, was established by a single crystal X-ray diffraction study (Fig. 2). For the other isolated crystals the presence of gold atoms was established, but disorder problems prevented a precise analysis (see Experimental section). Compound 3 crystallised as a solvate with toluene and adventitious water (1 : 1.5 : 1). As expected, the calix[4]arene framework is present in the cone conformation (Fig. 2), with interplanar angles between the aromatic rings of the macrocycle and the calixarene reference Fig. 2 Molecular structure of complex 3 R2P PR2 P P N N N Me Me Me O O N N N Me O Me Me O N N N Me Me Me O O Au Au Cl Cl R2 = 3 O O But But O But O But plane (average plane defined by the four bridging methylene carbon atoms) of 84 [C(31) ring], 53 [C(42) ring], 86 [C(53) ring] and 538 [C(64) ring]. The gold-bearing aromatic rings are essentially parallel (interplanar angle 28) whereas the other two subtend an angle of 748.The distances between the centres of the two pairs of opposite phenolic rings are 5.47 and 7.60 Å. The 3-NMe vectors of all the 1,3,5-trimethyl-4,6-dioxo-1,3,5- triaza-2-phosphorinan-2-yl groups point away from the calix- [4]arene cavity, thus minimising the steric interactions with the neighbouring heterocycles. These heterocycles adopt orientations with respect to the corresponding phenolic rings that are best described by the following torsion angles: N(2) ? ? ? P(1)]O(9) ? ? ? C(34) 5, N(5) ? ? ? P(2)]O(10) ? ? ? C(45) 224, N(8) ? ? ? P(3)]O(11) ? ? ? C(56) 23 and N(11) ? ? ? P(4)]O(12) ? ? ? C(67) 718.Thus, heterocycles P(1) and P(3) (linked to the Au]Cl fragments) tend to ‘overlap’ with the corresponding phenolic rings, resulting in an ‘eclipsed’ arrangement, whereas heterocycles P(2) and P(4) are staggered with respect to the corresponding phenolic rings (Fig. 2). The partial superposition of the heterocycles P(1) and P(3) and the two corresponding phenol rings ensures minimum steric interaction between the PAuCl fragments and the uncomplexed heterocycles. p-Stacking between the two rings can be ruled out.Interestingly, the P(2) and P(4) doublets point roughly in the same direction, so that the two P]Au]Cl fragments become formally inequivalent. Note that the absence of a symmetry element in the solid-state structure of 3 is reminiscent of the NMR peculiarities (several signals for the NMe groups) described above for complex 2.The six-membered heterocycles display a typical 30 half-boat conformation in which the phosphorus atoms lie 0.56–0.60 Å outside the plane formed by the remaining atoms. The P]Au]Cl groups are almost linear [P(1)]Au(1)]Cl(1) 174.40(8) and P(3)]Au(2)]Cl(2) 170.12(8)8]. Intermolecular Au ? ? ?Au contacts of 3.2879(7) Å are observed, so that a loose polymeric structure emerges in the solid state (Fig. 3).Such gold–gold (aurophilic) interactions are well known.38–41 The intramolecular Au(1) ? ? ?Au(2) distance is 7.800(1) Å. See Table 2 for further structural details. Cationic digold and disilver complexes From molecular models 42 it is apparent that the phosphorus– phosphorus separation between neighbouring phosphino Fig. 3 View showing the supramolecular structure of the dinuclear gold complex 3 [intermolecular Au ? ? ?Au distance 3.2879(7) Å] Table 2 Selected bond lengths (Å) and angles (8) for complex 3 Au(1)]P(1) Au(1) ? ? ?Au(2I) Au(2)]Cl(2) P(1)]Au(1)]Cl(1) Cl(1)]Au(1)]Au(2I) P(3)]Au(2)]Au(2II) O(9)]P(1)]Au(1) N(1)]P(1)]Au(1) N(9)]P(3)]Au(2) 2.209(2) 3.2879(7) 2.272(2) 174.40(8) 79.37(6) 110.69(6) 105.2(2) 116.3(2) 118.1(3) Au(1)]Cl(1) Au(2)]P(3) P(1)]Au(1)]Au(2I) P(3)]Au(2)]Cl(2) Cl(2)]Au(2)]Au(1II) N(3)]P(1)]Au(1) O(11)]P(3)]Au(2) N(7)]P(3)]Au(2) 2.270(2) 2.206(2) 106.04(6) 170.12(8) 78.51(6) 118.9(2) 104.9(2) 117.9(3) Symmetry transformation used to generate equivalent atoms: I x, 1 2 y, 2��� 1 z; II x, 1 2 y, ��� 1 z.2118 J.Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 groups in L1 is suitable for linear P]M]P co-ordination. Such ligating behaviour was indeed found in complex 4, which was obtained quantitatively according to the reaction sequence outlined in Scheme 2. Note that 4 was also formed as the sole product when L1 was treated with an excess of [Au(tht)- (MeCN)]1. Complex 4 is stable upon standing in solution.It is worth mentioning that chloride abstraction from complex 1 (with 4 equivalents of AgBF4) followed by addition of 1 equivalent of L1 aVorded the BF4 analogue of 4 as the sole product. These findings illustrate the strong tendency ave as a strong double chelator towards Au1, rather than as a bridging ligand. The FAB mass spectrum of complex 4 displays an intense peak at 1980, with the profile exactly as expected for the [(Au2L1)PF6] monocationic species.Note that there was no indication for formation of oligomeric species. The 1H NMR spectrum of 4 recorded at 20 8C displays two AB quartets (intensity 1 : 1) for the C6H2CH2 groups and one sharp peak for the But groups, whereas the 31P NMR spectrum shows a single peak. These observations are fully consistent with the C2vsymmetrical structure as drawn. A remarkable feature of one of the two C6H2CHAHB quartets is the large separation between the corresponding A and B parts (Dd 2.5 vs. 0.78 for the other one). We tentatively assign this AB system to the methylene groups that form a 12-membered metallomacrocycle with the gold atoms. In these methylene groups one CH bond probably comes very close to the metal atom and it is likely that this creates high anisotropy around the CH2 group. On heating a solution of 4 in C2D2Cl4 the C6H2CH2 signals first broaden, then coalesce to a single non-resolved AB system. The original spectrum reappeared upon cooling the solution to room temperature (Fig. 4). These findings can be interpreted in terms of a fast equilibration of the type shown in Scheme 3 (homomerisation), in which the gold ions move between the diVerent pairs of neighbouring phosphorus centres. To rule out intermolecular gold transfer, we repeated the variable temperature study at a 10-fold lower concentration and verified that this has no influence on the coalescence temperature. The observed intramolecular dynamics, in which the two metal atoms move simultaneously (both clockwise or counterclockwise), is reminiscent of those found in some recently Scheme 2 P P P P Au Au PPh2 Ph2P Ph2P PPh2 a. 2 [AuCl(tht)] b. 2 TlPF6 P P P P 2 PF6 L1 Ag Ag 5 2 BF4 4 O O But But O But O But O O But But O But O But O O But But O But O But CH2Cl2–MeCN reported Li2 complexes.43 The experimentally determined thermodynamic data reveal that there is no important entropic contribution to this process (DS‡ = 20.008 kJ K21 mol21, DH‡ = 76.020 kJ mol21). It appears therefore very likely that the migration of the gold ions from one site to the other does not involve a significant structural change.The related silver complex 5, which was prepared by treating L1 with 2 equivalents of AgBF4, displayed similar dynamic behaviour. In the temperature range 273–330 K the 31P NMR spectra showed a single signal with the expected J(AgP) coupling constants. As for 4, the 1H NMR spectrum of 5 displays two AB systems for the C6H2CH2C6H2 hydrogen atoms.Again one AB separation is rather large (Dd 2.23) compared to the other one (0.85). The dynamics observed in complexes 4 and 5 involves two metal ions and four phosphorus centres. The exact nature of the Fig. 4 Variable temperature 1H NMR spectra (500 MHz, C2D2Cl4) for complex 4 Scheme 3 Rotation of the gold atoms around the C2 axis of the calixarene in complex 4 P P P P P P P P 2+ 2+ * 4 * O O But But O But O But O O But But O But O ButJ.Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 2119 driving force which allows metal migration is not known, but it is likely that the M]P bond cleavages that must occur during this process are favoured by the strain existing within the two 12-membered metallomacrocycles. Dissociation of phosphine ligands in gold complexes are not unusual and have been observed in a number of mononuclear complexes of the type [Au(phosphine)2]1.44 In summary, we have shown that cone-shaped calix[4]arenes substituted at the lower rim by four phosphorus(III) centres behave as a small binding surface for four AuCl units. In the dinuclear complexes 4 and 5 a unique motion of two metal ions along the P4 surface could be established.Further investigations are needed to assess which factors will allow the introduction of similar dynamics in the solid state. Experimental Unless otherwise stated, materials were obtained from commercial suppliers and used without further purification.Solvents were dried over suitable reagents and freshly distilled under dry nitrogen before use. All reactions were carried out using modified Schlenk techniques under a dry nitrogen atmosphere. Routine 1H 13C-{1H} and 31P-{1H} NMR spectra were recorded on a Bruker AC-200 spectrometer at 200.1, 50.3 and 81.00 MHz, respectively. The 1H chemical shifts are reported relative to residual protiated solvents (CDCl3, d 7.26; CD2Cl2, d 5.32), the 13C chemical shifts relative to deuteriated solvents (CDCl3, d 77.00; CD2Cl2, d 53.8) and the 31P NMR data relative to external H3PO4. Mass spectra were recorded on a ZAB HF VG or a KRATOS MS 50 RF analytical spectrometer using m-nitrobenzyl alcohol as a matrix.Elemental analyses were performed by the Service de Microanalyse, Centre de Recherche Chimie, Strasbourg, and by the Analytisches Laboratorium des Instituts für Anorganische und Analytische Chemie der Technischen Universität, Braunschweig.Melting points were determined with a Büchi capillary melting-point apparatus and are uncorrected. Compounds L1,34 L2 30 and [AuCl(tht)] 45 were synthesized according to published procedures. Preparations [(AuCl)4L1] 1. To a stirred solution of compound L1 (0.300 g, 0.21 mmol) in CH2Cl2 (10 cm3) was added a solution of [AuCl(tht)] (0.268 g, 0.84 mmol) in CH2Cl2 (10 cm3). After 0.5 h the solution was filtered over a bed of Celite, concentrated to ca. 5 cm3, and addition of pentane precipitated complex 1 as a colourless powder which was dried in vacuo.Yield 0.438 g, 88%; m.p. 158–163 8C (decomp.). 1H NMR (200 MHz, 293 K, CDCl3): d 7.82–7.25 (m, 40 H, PPh2), 6.52 (s, 8 H, m-H), 5.30 (s br, 8 H, OCH2PPh2), 4.16 and 2.79 (AB quartet, J = 13 Hz, 4 H each, C6H2CH2) and 1.01 (s, 36 H, tert-butyl). 13C-{1H} NMR (50 MHz, 293 K, CDCl3): d 151.75, 146.04, 134.61, 131.79, 127.93 and 125.87 (quaternary aryl C), 134.36, 134.10, 132.08, 129.21, 129.04 and 125.87 (aryl CH), 71.75 (d, JPC = 42 Hz, OCH2PPh2), 33.85 [s, C(CH3)3], 32.01 (s, C6H2CH2) and 31.35 [s, C(CH3)3]. 31P-{1H} NMR (81 MHz, 293 K, CDCl3): d 22.7 (s). FAB mass spectrum: m/z 2335 (M1 2 Cl, 100%) (Found: C, 48.76; H, 4.52. Calc. for C96H100Au4Cl4O4P4: C, 48.62; H, 4.25%). [(AuCl)4L2] 2. To a stirred solution of compound L2 (0.255 g, 0.19 mmol) in CH2Cl2 (20 cm3) was added a solution of [AuCl(tht)] (0.244 g, 0.76 mmol) in CH2Cl2 (10 cm3). After 0.5 h a precipitate was formed and the solvent was evaporated to dryness, yielding a colourless residue.Analytically pure complex 2 was obtained by recrystallisation from toluene–CH2Cl2 (1 : 1). Yield 0.318 g, 55%; m.p. 140 8C (decomp.). 1H NMR (200 MHz, 293 K, CD2Cl2): d 6.82 (s br, 8 H, m-H), 3.91 and 3.25 (AB quartet, J = 13 Hz, 8 H, C6H2CH2), 3.60 (m, 24 H, PNMe), 2.78 {m, 12 H, CH3N[C(O)]2} and 1.07 (br s, 36 H, tert-butyl). 13C-{1H} NMR (50 MHz, 293 K, CD2Cl2): d 150.79–127.90 (aryl C), 37.60 [d of complex m, PNMe, 3J(PMe) ª 15 Hz], 37.50 (m, NMe), 34.66 [s, C(CH3)3], 31.44 [C(CH3)3 and C6H2CH2] and 30.88 [s, NCH3(CO)2]. 31P-{1H} NMR (81 MHz, 293 K, CD2Cl2): d 101.6 (s).FAB mass spectrum: m/z 2235.3 (M1 2 Cl, 100%) (Found: C, 32.44; H, 3.82. Calc. for C64H88Au4Cl4N12O12P4?2CH2Cl2: C, 32.47; H, 3.79%). Partially aurated species derived from L1 and L2 (the signals arising from tht have been omitted). L1 : [AuCl(tht)] = 1 : 1. 1H NMR (200 MHz, 293 K, CDCl3): d 7.37–7.12 (40 H, PPh2), 6.71 (br s, 8 H, m-H), 5.17 (br s, 8 H, OCH2PPh2), 4.24 and 3.02 (AB quartet, J = 12 Hz, 4 H each, C6H2CH2) and 1.05 (s, 36 H, tert-butyl). 31P-{1H} NMR (81 MHz, 293 K, CDCl3): d 16.6 (br s). L1 : [AuCl(tht)] = 1 : 2. 1H NMR (200 MHz, 293 K, CDCl3): d 7.90–7.28 (40 H, PPh2), 6.61 (br s, 8 H, m-H), 4.97 (br s, 8 H, OCH2PPh2), 4.53 and 2.92 (AB quartet, J = 12 Hz, 4 H each, C6H2CH2) and 1.04 (s, 36 H, tert-butyl). 31P-{1H} NMR (81 MHz, 293 K, CDCl3): d 31.4 (br s). L1 : [AuCl(tht)] = 1 : 3. 1H NMR (200 MHz, 293 K, CDCl3): d 7.66–7.19 (40 H, PPh2), 6.58 (br s, 8 H, m-H), 4.96 (br s, 8 H, OCH2PPh2), 4.46 and 2.90 (AB quartet, J = 12 Hz, 4 H each, C6H2CH2) and 1.01 (s, 36 H, tert-butyl). 31P-{1H} NMR (81 MHz, 293 K, CDCl3): d 34.5 (br s) and 25.2 (br s). L2 : [AuCl(tht)] = 1 : 1. 1H NMR (200 MHz, 293 K, CD2Cl2): d 6.76 (br s, 8 H, m-H), 3.38 and 3.19 (AB quartet, J = 13 Hz, 4 H each, C6H2CH2), 3.44 (br d, 24 H, PNMe), 2.64 {br s, 12 H, [C(O)]2NMe} and 1.03 (s, 36 H, tert-butyl). 31P-{1H} NMR (81 MHz, 293 K, CD2Cl2): d 96.8 (br s) and 91.0 (br s). 31P-{1H} NMR (81 MHz, 203 K, CD2Cl2): d 107.2 (br s) and 89.2 (s). L2 : [AuCl(tht)] = 1 : 2. 1H NMR (200 MHz, 298 K, CD2Cl2): d 6.81 (br s, 8 H, m-H), 3.89 and 3.23 (AB quartet, J = 13 Hz, 4 H each, C6H2CH2), 3.53 (br d, 24 H, PNMe), 2.70 {br s, 12 H, [C(O)]2NMe} and 1.04 (br s, 36 H, tert-butyl). 31P-{1H} NMR (81 MHz, 293 K, CD2Cl2): d 99.1 (br s) and 96.0 (br s). L2 : [AuCl(tht)] = 1 : 3. 1H NMR (200 MHz, 298 K, CD2Cl2): d 6.5 (br s, m-H), 3.90 and 3.33 (AB quartet, J = 13 Hz, 4 H each, C6H2CH2), 3.70 (br s, 24 H, PNMe), 2.75 {br s, 12 H, [C(O)]2NMe} and 1.00 (br s, 36 H, tert-butyl). 31P-{1H} NMR (81 MHz, 293 K, CD2Cl2): d 106.2 (br s). [Au2L1][PF6]2 4. To a solution of compound L1 (0.151 g, 0.10 mmol) in CH2Cl2 (10 cm3) was added a solution of [AuCl(tht)] (0.067 g, 0.21 mmol) in CH2Cl2 (10 cm3). After 0.5 h the solution was added to a suspension of thallium hexafluorophosphate (0.073 g, 0.21 mmol) in acetonitrile (2 cm3).After stirring for 5 min the white precipitate was filtered through a bed of Celite, and the filtered solution concentrated to ca. 5 cm3. Addition of pentane yielded complex 4 as a colourless precipitate. Yield 0.209 g, 94%; mp 120–123 8C (decomp.). 1H NMR (200 MHz, 293 K, CDCl3): d 7.79–7.25 (40 H, PPh2), 6.82 and 6.33 (AB quartet, J = 2, 4 H each, m-H), 5.23 and 2.73 (AB quartet, J = 13, 2 H each, C6H2CH2), 5.14 and 4.88 (AB quartet, J = 13, 4 H each, OCH2PPh2, JPA = JPB ª 0), 4.48 and 3.70 (AB quartet, J = 13 Hz, 2 H each, C6H2CH2) and 0.98 (s, 36 H, tert-butyl). 13C-{1H} NMR (50 MHz, 293 K, CDCl3): d 152.65–125.14 (aryl C), 73.00 (vt, JPC13JPC = 46 Hz, OCH2PPh2), 33.93 [s, C(CH3)3], 32.00 and 31.70 (2s, C6H2CH2) and 31.31 [s, C(CH3)3]. 31P-{1H} NMR (81 MHz, 293 K, CDCl3): d 40.6 (br s). FAB mass spectrum: m/z 1980.4 ([M 2 PF6]1, 20%) (Found: C, 54.30; H, 4.88. Calc. for C96H100Au2F12O4P6: C, 54.25; H, 4.74%).[Ag2L1][BF4]2 5. To a stirred solution of silver(I) tetrafluoroborate (0.027 g, 0.14 mmol) in tetrahydrofuran (2 cm3) was added a solution of compound L1 (0.100 g, 0.07 mmol) in tetrahydrofuran (10 cm3). After 5 min the solvent was evaporated to dryness, yielding a colourless residue. The latter was dissolved in CH2Cl2 (5 cm3) and the resulting solution filtered over a bed of Celite. Addition of hexane precipitated complex 5 as2120 J. Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 a colourless powder.Yield 0.102 g, 80%; mp 141–143 8C (decomp.). 1H NMR (200 MHz, 293 K, CDCl3): d 7.70–7.26 (br m, 40 H, PPh2), 6.76 and 6.30 (2 br s, 4 H each, m-H), 5.13 and 4.56 (AB quartet, J = 11, 4 H each, OCH2PPh2), 4.40 and 2.17 (AB quartet, J = 12, 2 H each, C6H2CH2), 4.22 and 3.37 (AB quartet, J = 12 Hz, 2 H each, C6H2CH2) and 0.98 (s, 36 H, tertbutyl). 13C-{1H} NMR (50 MHz, 293 K, CDCl3): d 152.63– 124.66 (aryl C), 73.57 (vt, JPC 1 3JPC = 30 Hz, OCH2PPh2), 33.66 [s, C(CH3)3], 31.64 and 29.95 (2s, C6H2CH2C6H2) and 31.09 [s, C(CH3)3]. 31P-{1H} NMR (81 MHz, 293 K, CDCl3): d 4.6 [2d, J(107Ag]P) = 509 Hz, J(109Ag]P) = 585 Hz, PPh2]. FAB mass spectrum: m/z 1743 ([M 2 BF4]1, 8) (Found: C, 63.16; H, 5.69. Calc. for C96H100Ag2B2F8O4P4: C, 62.97; H, 5.50%). Variable temperature NMR experiments The 1H, 13C-{1H} and 31P-{1H} NMR spectra for the variable temperature studies were recorded on a Bruker ARX-500 spectrometer (1H, 500.1 MHz) for L1 complexes and an AC-200 spectrometer (1H, 200.1; 31P, 81.0 MHz) for L2 complexes.All temperatures were corrected to the ethylene glycol spin temperature, and are estimated to be reliable at ±5 K. The spectra were recorded every 10 K for each experiment to determine the coalescence temperature Tc/K. The free energy barrier for the observed process (DG‡ = 79.4 kJ mol21) is derived from the Eyring equation: DG‡ = 1023RTc[22.96 1 (ln Tc/Dn)] kJ mol21.46 The calculations for the determination of DH‡ and DS‡ are based on the two exchange processes PCH2�ÆPCH92 and m-H �Æ m-H9.Crystallography Crystallographic data for complexes 1 and 3 are given in Table 3. Colourless crystals of 1 suitable for diVraction were obtained by slow diVusion of hexane into a chlorobenzene solution of the pure complex. Single crystals of 3 were obtained by addition of toluene to a dichloromethane solution (1 : 1, v/v) containing a 2 : 1 mixture of [AuCl(tht)] and L2.Crystals were mounted on glass fibres in inert oil and transferred to the cold gas stream of the diVractometer (Siemens P4 with LT-2 low temperature attachment). The orientation matrices for 1 and 3 were refined from setting angles of 64 (81) reflections in the 2q range 5–258 (monochromated Mo-Ka radiation). Absorption corrections were based on y scans. Both structures were solved by direct methods and refined aniostropically on F2 (program SHELXL 9347). Hydrogen atoms were included using Table 3 Crystallographic data for complexes 1 and 3 * Formula M Crystal colour, habit Crystal dimensions/ mm a/Å b/Å c/Å b/8 U/Å3 Z Dc/g cm23 m/mm21 No.reflections measured No. independent reflections Rint R [F > 4s(F)] 1 C96H100Au4Cl4O4P4 2371.31 Colourless prism 0.50 × 0.40 × 0.25 29.370(4) 22.809(3) 18.807(3) 114.79(2) 11 438(3) 4 1.377 5.30 12 264 9932 0.048 0.050 3?1.5C6H5CH3?H2O C74.5H104Au2Cl2N12O13P4 1964.41 Colourless tablet 0.70 × 0.30 × 0.15 37.230(6) 27.211(4) 21.269(4) 126.671(10) 17 721(5) 8 1.473 3.57 23 002 15 517 0.091 0.049 * Details in common: monoclinic, space group C2/c; 173 K; (D/s)max < 0.001; R = S Fo| 2 |Fc /S|Fo|.a riding model or rigid methyl groups. Weighting schemes of the form w21 = [s2(Fo 2) 1 (aP)2 1 bP] were employed, with P = (Fo 2 1 2Fc 2)/3. Special features of refinement. In compound 1 various regions of poorly resolved electron density indicate the presence of disordered solvent, but no appropriate model could be refined. In the crystal data, values for Mr, Dc, etc.do not include this extra solvent. In 3 the tert-butyl group C(61)–C(63) is disordered over two positions and one toluene is disordered over an inversion centre. From the experiment leading to 3 {1 equivalent [AuCl- (tht)] 1 2 equivalents L2} further crystals were collected. An X-ray determination of ligand L2 with 1.5 ‘AuCl’ was conducted: we found one AuCl site to be fully occupied and the neighbouring one half-occupied; the latter represents alternative (overlapping) sites in neighbouring molecules. The calixarene system displayed a cone conformation.One solvent molecule and the half occupied Au atom also overlapped; in view of these disorder problems, the full data of this compound will not be presented. Crystal data: monoclinic, space group P21/n, Z = 4, a = 17.783(4), b = 19.528(4), c = 28.508(6), b = 107.573(12)8. CCDC reference number 186/986.See http://www.rsc.org/suppdata/dt/1998/2115/ for crystallographic files in .cif format. Acknowledgements Financial support from the Royal Society of Chemistry (Journal Grant to D. M.), from the Deutscher Akademischer Austauschdienst (DAAD, to C. B. D.), and from the Fonds der Chemischen Industrie (to R. S.) is gratefully acknowledged. We are indebted to Johnson Matthey for a generous loan of precious metals, and to DEGUSSA AG for a gift of HAuCl4. References 1 C. D. Gutsche, Monographs in Supramolecular Chemistry, Vol.I: Calixarenes, ed. J. F. Stoddart, The Royal Society of Chemistry, Cambridge, 1989. 2 A. McKervey and V. Böhmer, Chem. Br., 1992, 724. 3 S. Shinkai, Tetrahedron, 1993, 49, 8933. 4 E. van Dienst, W. I. I. Bakker, J. F. J. Engbersen, W. Verboom and D. N. Reinhoudt, Pure. Appl. Chem., 1993, 65, 387. 5 D. V. Khasnis, J. M. Burton, J. D. McNeil, C. J. Santini, H. Zhang and M. Lattman, Inorg. Chem., 1994, 33, 2657.. Loeber, C. Wieser and D.Matt, in Stereoselective reactions of metal-activated molecules, eds. H. Werner and J. Sundermeyer, Vieweg, Wiesbaden, 1995, pp. 191–194. 7 V. Böhmer, Angew. Chem., 1995, 107, 785; Angew. Chem., Int. Ed. Engl., 1995, 34, 713. 8 T. M. Swager and B. Xu, J. Inclusion Phenom., 1994, 19, 389. 9 A. Zanotti-Gerosa, E. Solari, L. Giannini, C. Floriani and A. Chiesi-Villa, Chem. Commun., 1996, 119. 10 H. Budig, S. Diele, R. Paschke, D. Ströhl and C. Tschierske, J. Chem. Soc., Perkin Trans. 2, 1996, 1901. 11 A. Ikeda and S. Shinkai, Chem. Rev., 1997, 97, 1713. 12 P. J. Stang, D. H. Cao, K. Chen, G. M. Gray, D. C. Muddiman and R. D. Smith, J. Am. Chem. Soc., 1997, 119, 5163. 13 P. D. Beer, Chem. Commun., 1996, 689. 14 J.-C. G. Bünzli and J. M. Harrowfield, in Calixarenes, a versatile class of macrocyclic compounds, eds. J. Vicens and V. Böhmer, Kluwer, Dordrecht, 1990, p. 211. 15 R. Ungaro, A. Casnati, F. Ugozzoli, A. Pochini, J.-F. Dozol, C. Hill and H. Rouquette, Angew.Chem., 1994, 106, 1551; Angew. Chem., Int. Ed. Engl., 1994, 33, 1506. 16 J. F. Malone, D. J. Marrs, M. A. McKervey, P. O’Hagan, N. Thompson, A. Walker, F. Arnaud-Neu, O. Mauprivez, M.-J. Schwing-Weill, J.-F. Dozol, H. Rouquette and N. Simon, J. Chem. Soc., Chem. Commun., 1995, 2151. 17 M. R. Yaftian, M. Burgard, A. El Bachiri, D. Matt, C. Wieser and C. Dieleman, J. Inclusion Phenom., 1997, 29, 137. 18 M. R. Yaftian, M. Burgard, D. Matt, C. B. Dieleman and F. Rastegar, Solvent Extr.Ion Exch., 1997, 15, 975. 19 B. Xu and T. M. Swager, J. Am. Chem. Soc., 1995, 117, 5011.J. Chem. Soc., Dalton Trans., 1998, Pages 2115–2121 2121 20 R. Cameron, S. J. Loeb and G. P. A. Yap, Inorg. Chem., 1997, 36, 5498. 21 C. Wieser, D. Matt, J. Fischer and A. Harriman, J. Chem. Soc., Dalton Trans., 1997, 2391. 22 D. M. Roundhill, Prog. Inorg. Chem., 1995, 43, 533. 23 C. Wieser, C. B. Dieleman and D. Matt, Coord. Chem. Rev., 1997, 165, 93. 24 P. Faidherbe, C.Wieser, D. Matt, A. Harriman, A. De Cian and J. Fischer, Eur. J. Inorg. Chem., 1998, 4, 451. 25 P. Jacopozzi and E. Dalcanale, Angew. Chem., 1997, 109, 665; Angew. Chem., Int. Ed. Engl., 1997, 36, 613. 26 C. Floriani, D. Jacoby, A. Chiesi-Villa and C. Guastini, Angew. Chem., 1989, 101, 1430; Angew. Chem., Int. Ed. Engl., 1989, 28, 1376. 27 W. Xu, J. P. Rourke, J. J. Vittal and R. J. Puddephatt, J. Chem. Soc., Chem. Commun., 1993, 145. 28 W. Xu, R. J. Puddephatt, L. Manojlovic-Muir, K. W. Muir and C. S. Frampton, J. Inclusion Phenom., 1994, 19, 277. 29 C. Dieleman, C. Loeber, D. Matt, A. De Cian and J. Fischer, J. Chem. Soc., Dalton Trans., 1995, 3097. 30 I. Neda, H.-J. Plinta, R. Sonnenburg, A. Fischer, P. G. Jones and R. Schmutzler, Chem. Ber., 1995, 128, 267. 31 M. Stolmàr, C. Floriani, A. Chiesi-Villa and C. Rizzoli, Inorg. Chem., 1997, 36, 1694. 32 A. M. Rouhi, Chem. Eng. News, 1997, 34. 33 R. A. Pufahl, C. P. Singer, K. L. Peariso, S.-J. Lin, P. J. Schmidt, C. J. Fahrni, V. Cizewski Culotta, J. E. Penner-Hahn and T. V. O’Halloran, Science, 1997, 278, 853. 34 C. B. Dieleman, D. Matt and P. G. Jones, J. Organomet. Chem., 1997, 545–546, 461. 35 M. Conner and S. L. Regen, J. Am. Chem. Soc., 1991, 113, 9670. 36 P. Braunstein and J. Rose, Gold Bull., 1985, 18, 17. 37 B. Assmann and H. Schmidbaur, Chem. Ber., 1997, 130, 217. 38 P. G. Jones, Gold Bull., 1981, 14, 102. 39 H. Schmidbaur, Interdiscip. Sci. Rev., 1992, 17, 213. 40 S. S. Panthaneni and G. R. Desiraju, J. Chem. Soc., Dalton Trans., 1993, 319. 41 P. M. Van Calcar, M. M. Olmstead and A. L. Balch, Inorg. Chem., 1997, 36, 5231. 42 HGS Biochemistry Molecular Models, MARUZEN Co. Ltd., Tokyo. 43 M. Veith, M. Zimmer, K. Fries, J. Böhnlein-Maus and V. Huch, Angew. Chem., 1996, 108, 1647; Angew. Chem., Int. Ed. Engl., 1996, 35, 1529. 44 M. J. Mays and P. A. Vergnano, J. Chem. Soc., Dalton Trans., 1979, 1112. 45 R. Uson, A. Laguna and M. Laguna, Inorg. Synth., 1989, 26, 85. 46 H. Günther, NMR-Spektroskopie, 3rd edn., Thieme, Stuttgart, 1992, p. 310. 47 G. M. Sheldrick, SHELXL 93, Program for Crystal Structure Refinement, University of Göttingen, 1993. Received 25th February 1998; Paper 8/01586D

ISSN:1477-9226    

DOI:10.1039/a801586d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 2119–2126 2119 Sodium hydrotris(methimazolyl)borate, a novel soft, tridentate ligand: preparation, structure and comparisons with sodium hydrotris(pyrazolyl)borate † John Reglinski,* Mark Garner, Iain D. Cassidy, Paul A. Slavin, Mark D. Spicer and David R. Armstrong Department of Pure & Applied Chemistry, University of Strathclyde, Glasgow, UK G1 1XL. E-mail: j.reglinski@strath.ac.uk Received 3rd March 1999, Accepted 29th April 1999 The hydrotris(methimazolyl)borate anion (Tm), a soft analogue of the hydrotris(pyrazolyl)borate anion (Tp), has been synthesized. This novel ligand system has been designed to maintain the tripodal geometry around the boron while allowing the replacement of the three nitrogen donor atoms by three sulfur (thione) donor atoms, thus providing a complementary soft, tridentate, face capping ligand system.The two ions, Tm and Tp were compared by X-ray analysis and ab initio calculations in an attempt to explore the eVects of exchanging the hard donor atoms for soft donor atoms in this type of ligand. The compound NaTm is essentially salt like with discrete anions and hydrated sodium cations.The structure of NaTp crystallised under identical conditions is observed to be an infinite ribbon containing monodentate, bridging and pendant pyrazolyl units. The co-ordination sphere of the sodium cation in NaTp is completed by two water molecules. Ab initio calculations at the Hartree–Fock level using a 6-31G* basis set on these anions and their sodium complexes suggested that while both ions are in general similar in nature, there are subtle diVerences which will influence their chemistry.Ab initio calculations were also used to provide a rational analysis of the formation of the two sodium salts obtained and on the analogous copper complexes further to clarify the hard and soft nature of the two ligand systems. Introduction Transition metal complexes are a rich source of a diverse range of catalysts, reagents and “smart” materials.Although the reactivity of these compounds is based on the metal, and it is the choice of metal which provides the basic reactive profile, the exact properties exhibited by a metal arise from the symbiotic relationship between the metal and its ligands. At a gross level, modulation of the reactivity of a metal complex by a ligand involves a modification of solubility and charge. At a more subtle level, ligands control the number and position of vacant co-ordination sites, the size or shape of the substrate admitted to a metal centre and the stability of electron rich or electron poor metal centres.Considering the number of factors involved, it is unsurprising that the design of new metal complexes is approached in a systematic manner using homologous series of ligands in an attempt to engineer a graded change in chemistry.1 However, the utility of a ligand is not just determined by its properties and those of its metal complexes but also by the ease with which it can be prepared.The more successful ligand systems (e.g. phosphines) tend to be those which can be constructed relatively simply through one, or two step, high yield reactions.1 On occasions a ligand system is developed which is capable of dramatically altering the behaviour of a metal. This was the case with the hydrotris(pyrazolyl)borate anion (Tp),2,3 which, for instance, once complexed with copper generated a unique carbonyl complex.4,5 Consistent with the above, pyrazolylborates are relatively simple to prepare and have been found to produce complexes with a vast array of metal ions spanning the wider disciplines of classical co-ordination chemistry, organometallic chemistry and inorganic biochemistry.3,6–8 As a ligand system, Tp has found great favour amongst synthetic chemists † Methimazolyl = 3-methyl-2-thioimidazolinyl.as a facial tridentate six electron donor sometimes compared with and used as an alternative to the isoelectronic cyclopentadienyl anion (Cp).3 While it is possible to alter the steric requirements of the three donor nitrogens in Tp by placing substituent groups (e.g. Me, tBu or Ph) 3,9 on the pyrazole ring adjacent to the donor nitrogens, it is more diYcult to alter the electron donor properties of the ligand. Thus, the pyrazolylborate ligand system (cf. phosphines) is somewhat restricted should a markedly softer ligand be required.Soft tridentate sulfur containing macrocycles such as trithiacyclononane have been prepared 10 which oVer a six electron bonding set comparable with that found for Cp and Tp. However, since thioether macrocycles do not naturally carry an overall charge, they are not readily able to oVer a controlled, graded alteration of the chemistry of a metal in conjunction with Cp and Tp. Clearly it would be of some value to extend the scope of the existing structural motif simply by changing the donor set.This problem has been addressed by Riordan and co-workers 11–14 whose elegant chemistry produced soft tripodal borate ligands based on thioethers, eqn. (1). Subsequent complexation of this tridentate sulfur based system with molybdenum carbonyl produced a similar structural motif to Tp.11 The existence of these species confirms that soft ligands based on the tetrahedral borate anion can be prepared and furthermore that they can generate the structural types observed with the more popular Tp and2120 J.Chem. Soc., Dalton Trans., 1999, 2119–2126 Cp ligand systems. However, these soft aliphatic species lack the extended conjugated system which is present in the pyrazolyl moieties thus mitigating against a wider distribution of the electron density within the ligand. Furthermore, due to the orientation of the electron pairs around the sulfur it is diYcult to generate a protected pocket around the metal centre via the incorporation of bulky groups adjacent to the donor atom.We envisaged that any expansion of the Tp ligand system would be best achieved by using the traditional high yield reaction of Trofimenko2 but employing an alternative organic ring fused to the boron. Our previous work15 had highlighted that in species such as methimazole the acidic hydrogen lay on the nitrogen rather than on the sulfur and that consistent with other workers 16,17 these species are best described as amine thiones, eqn.(2). This suggested to us that species such as methimazole could undergo an elimination of hydrogen on reaction with the BH4 2 anion in a melt to generate a soft analogue of Tp, eqn. (3).18 We have recently demonstrated this to be the case 18 and hence report in full the synthesis and structure of our soft tripodal ligand, the hydrotris(methimazolyl)borate anion, Tm. As much of the utility of this anion is likely to be driven by its chemical analogy with Tp, we have also synthesized and crystallised the parent Tp system under identical conditions.Furthermore, both anions have been subjected to ab initio calculations to try and provide an insight into the subtle chemical diVerences which may be expected as a result of progressively replacing methimazole for pyrazole in these systems. Experimental All chemicals were commercially obtained and used without further purification.All NMR spectra were recorded on a Bruker AMX 400 spectrometer operating at 400.1 MHz for 1H and 100 MHz for 13C. Preparation Sodium hydrotris(pyrazolyl)borate (NaTp). The preparation of sodium hydrotris(pyrazolyl)borate follows that of Trofi- menko.2 Briefly, sodium tetrahydroborate (1.5 g, 0.040 mol) and pyrazole (10 g, 0.15 mol) were placed in a 50 ml round bottom flask and the temperature gently raised to 160 8C. The melt gently evolved hydrogen gas, which was collected and its volume measured. The reaction was stopped when 3 mole equivalents (ª0.12 mol, ª2.7 l) of hydrogen had been collected signifying that the dominant product would be the tris- (pyrazolyl)borate.The mixture was allowed to cool and washed with hexane. The resulting white powder was recrystallised from hexane–toluene at 24 8C. Crystals for structure determination were obtained directly from the liquors, prior to the retrieval and drying (dehydration) of the bulk material.Spectroscopic properties were consistent with literature values 2 (Found: C, 45.50; H, 4.15; N, 35.08. Calc. for C9H10BN6Na: C, 45.80; H, 4.27; N, 35.61%). Sodium hydrotris(methimazolyl)borate (NaTm). 3-Methylimidazoline- 2-thione (methimazole, 13.2 g, 0.115 mol) and sodium tetrahydroborate (1.1 g, 0.029 mol) were mixed together in a 50 ml round bottom flask, which was fitted with an air jacket condenser. The vessel was placed in an oil-bath and the temperature raised slowly to 160 8C.The mixture melted at approximately 136–140 8C (mp methimazole = 144 8C) whereupon the vigorous evolution of hydrogen gas began. This was collected as above and the reaction allowed to proceed until 3 mole equivalents (ª0.090 mol, 2.0 l) of hydrogen gas had evolved. Excessive heating (>180 8C) led to the mixture turning deep pink/purple indicative of undesirable products forming. Once the reaction was complete the mixture was allowed to cool. The resulting solid was washed with hexane to remove excess of methimazole and Soxhlet extracted into chloroform.The resulting white powder was filtered oV and dried, yielding anhydrous NaTm (65%) (Found: C, 38.30; H, 4.22; N, 21.78; S, 25.61. C12H16BN6NaS3 requires C, 38.51; H, 4.31; N, 22.45; S, 25.70%). dH (400.1 MHz; solvent (CD3)2SO) 3.38 (s, 3 H, Me), 6.40 (d, 1 H, CH) and 6.79 (d, 1 H, CH, J 2.2 Hz). dC (100.1 MHz; solvent (CD3)2SO) 33.6 (CH3), 116.5 (CH), 120.8 (CH) and 163.4 (Cquat).n& /cm21 (Nujol mull): 2478 (B–H). The compound thus obtained is suYciently pure for further use. However, it can be recrystallised from methylene chloride at 24 8C in the presence of moist air to yield a hydrated form (Found: C, 31.31; H, 5.33; N, 18.34; S, 21.21. C12H25BN6- NaO4.5S3 requires C, 31.65; H, 5.53; N, 18.45; S, 21.12%). Thallium(I) hydrotris(methimazolyl)borate (TlTm). Anhydrous NaTm (0.50 g, 1.3 mmol) in 75 ml of acetone was added to TlNO3 (0.35 g, 1.3 mmol) suspended in 25 ml of acetone and the mixture was refluxed for 4 h.The solid formed was filtered oV, washed with water to remove NaNO3 and any residual TlNO3 filtered and dried (0.61g, 81%) (Found: C, 26.56; H, 2.90; N, 15.25. C12H16BN6S3Tl requires C, 25.94; H, 2.90; N, 15.12%). n& /cm21 (Nujol mull): 2475 (B–H). dH (400.1 MHz; solvent (CD3)2SO, 313 K) 3.98 (s, 9 H, CH3), 6.44 (d, J = 3, 3 H, CH) and 6.81 (d, J = 3, 3 H, CH). Crystal structure determinations Crystals of NaTp?1H2O and NaTm?4.5H2O were obtained by slow evaporation of methylene chloride solutions in the presence of moist air.In both cases colourless crystals were obtained. They were mounted on glass fibres and all measurements performed at room temperature using graphite monochromated Mo-Ka radiation. Accurate cell dimensions were obtained from 25 accurately centred reflections. Details of the data collection and refinement are given in Table 1. The structure was solved by direct methods 19 and expanded using Fourier techniques.20 The non-hydrogen atoms were refined anisotropically 21 and in both cases the refinement converged satisfactorily. The structure of NaTm was somewhat problematic since the sodium ions and their associated water molecules are badly disordered.It has not been possible to obtain an entirely satisfactory model for this part of the structure and as a consequence the R factors are a little higher than one would wish. Nevertheless, the Tm anion is well defined and the separation of anion and cation is unambiguous. Selected bond lengths and angles are shown in Tables 2 and 3.CCDC reference number 186/1451. See http://www.rsc.org/suppdata/dt/1999/2119/ for crystallographic files in .cif format.J. Chem. Soc., Dalton Trans., 1999, 2119–2126 2121 Ab initio calculations Ab initio calculations at the Hartree–Fock level using the 6-31G* basis set 22–24 were carried out on the Tm and Tp anions and their corresponding sodium derivatives using the GAUSSIAN 94 computational package.25 No symmetry constrictions were applied during the geometry optimisation procedures and frequency calculations were subsequently performed to verify that the optimised structures corresponded to a local minimum.Selected calculated parameters are shown in Table 4. In order to make a simple qualitative assessment of the relative hardness/softness of the Tm and Tp anions, calculations were carried out on the respective copper(I) complexes (cf.sodium) using the DZ (14,11,5)/[8,6,2] copper basis sets of Ahlrichs and co-workers.26 Results and discussion The synthesis of the poly(pyrazolyl)borate anions is driven by the elimination of dihydrogen from the reaction of the tetrahydroborate anion with the acidic proton commonly found in pyrazoles. The extension of this synthetic route to the preparation of softer species became possible once it was realised that many 1,2 imine thiols such as methimazole are better formulated as their thione tautomer.15–17 The acidic hydrogen on the amine group should allow species such as methimazole to react in an analogous manner to pyrazole, eqns.(2) and (3), yielding species in which soft sulfur donors will be available for metal co-ordination. However, there will be a small modification to the ligand architecture around the donor atoms. In particular, in the hydrotris(methimazolyl)borate anion the boron is separated from the donor atoms by three bonds in comparison to two bonds in Tp.Thus, when the Tm anion chelates to metals such as zinc,18 three eight-membered rings form rather than the three six-membered rings found with the Tp anion.9 From the location of the acidic hydrogen the binding of methimazole to boron is expected to be via a B–N linkage. However, it is possible that thiol–thione tautomerisation, eqn. (2),27,28 while not dominant at room temperature, could occur in the melt prior to coupling.Alternatively a 1–3 shift might occur after the initial coupling reaction. Consequently, concern existed that under the prevailing reaction conditions the boron might migrate from the nitrogen to the sulfur. Spectroscopic evidence for the desired B–N coupling in the Tm anion relied heavily on the chemical shift of the thione carbon (d 163.4) in the NMR spectrum and the identification of the C]] S stretch (ª730 cm21) in the infrared spectrum neither of which could be considered as conclusive.Definitive evidence for the desired arrangement in the form of a crystal structure was sought. Suitable crystals proved elusive until it was observed that crystallisation occurred more readily if moist air was allowed to contact the liquors during the process. Single crystals were thus obtained. Crystal structures The structure of our new system consists of discrete Tm anions and disordered one-dimensional chains of hydrated sodium cations, thus rationalising the need for water during the crystallisation process.The anion (Fig. 1) has approximate C3 symmetry, with each methimazolyl group twisted about the B–N bond to minimise the steric eVects of the methyl groups. This results in a “propeller-like” conformation of the rings. The soft nature of the ligand is manifest in the total lack of interaction of the donor atoms with the sodium cation. It is also notable that the ligand is not prearranged for complexation.The rotation about the B–N bond results in an “inverted” con- figuration, with the three sulfur atoms on the same side as the B–H bond. A viable comparison of Tm with Tp required that the Tp anion be crystallised in an analogous manner to the Tm anion, i.e. in the presence of moist air. The structure of the compound thus obtained reveals infinite one-dimensional chains (Fig. 2) in which five-co-ordinate sodium ions (Fig. 3) are bridged by a pyrazolyl nitrogen and a water molecule.The co-ordination sphere of the sodium ion is completed below the plane by an interaction with a second pyrazolyl nitrogen, while the vacant sixth co-ordination site (above the plane) is protected by a pyrazolyl ring. It is tempting to invoke a weak h5-p interaction although the Na–C3N2 ring centroid distance, at 3.01 Å, is somewhat longer than in genuine documented examples such as Na(h5-C5H5)?TMEDA (2.65 Å) 29 and [Na(h5-C5H5)2]2 (2.336(3) Å).30 The conformation of the Tp anion is markedly diVerent to the Tm anion.It co-ordinates as a didentate ligand to one sodium, with one of the co-ordinated pyrazolyl groups bridging to the second sodium centre. The third, pendant, pyrazolyl group is weakly hydrogen bonded (dN(6)–H(12) = 1.8 Å) to the bridging water ligands (Fig. 3). A number of structures of substituted poly(pyrazolyl)borate ligands have been previously determined. Unlike our structure, alkali metal salts of perfluoroalkyl substituted Tp ligands form mono- or di-nuclear structures with significant M–F interactions,31–33 while hydrotris(3,4,5-trimethylpyrazolyl)borate forms a mononuclear complex with three trimethylpyrazoles completing the potassium co-ordination sphere.34 Potassium salts of the related compounds hydrotris(1,2,4-triazolyl)borate 35 and Fig. 1 The structure of the Tm anion showing the atom numbering scheme. The thermal ellipsoids are drawn at the 40% level. Table 1 Experimental details of the crystal structure determination of NaTp?H2O and NaTm?4.5H2O NaTp?H2O NaTm?4.5H2O Molecular formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 Z m(Mo-Ka)/cm21 T/K Reflections measured Unique reflections Observed reflections RR 9 C9H12BN6NaO 254.03 Orthorhombic P212121 (no. 19) 8.508(2) 20.730(3) 7.062(1) 1245.5 4 1.23 293 1692 1692 1212 (I > 1.00s(I)) 0.046 0.043 C12H25BN6NaO4.5S3 455.37 Triclinic P1� (no. 2) 9.962(2) 14.790(2) 8.217(2) 83.55(10) 78.08(2) 72.65(1) 1129.0(4) 2 3.78 293 5227 4935 (Rint = 0.044) 2299 (I > 3.00s(I)) 0.067 0.0872122 J.Chem. Soc., Dalton Trans., 1999, 2119–2126 Table 2 Selected bond lengths (Å) and angles (8) for Na[HB(MeC3H2N2S)3]?4.5H2O X-Ray Ab initio a X-Ray Ab initio a S(1)–C(1) S(2)–C(5) S(3)–C(9) N(1)–C(1) N(1)–C(2) N(1)–B(1) N(2)–C(1) N(2)–C(3) N(2)–C(4) N(3)–C(5) N(3)–C(6) N(3)–B(1) N(4)–C(5) C(1)–N(1)–C(2) C(1)–N(1)–B(1) C(2)–N(1)–B(1) C(1)–N(2)–C(3) C(1)–N(2)–C(4) C(3)–N(2)–C(4) S(1)–C(1)–N(1) S(1)–C(1)–N(2) N(1)–C(1)–N(2) N(1)–C(2)–C(3) N(2)–C(3)–C(2) N(1)–B(1)–N(3) N(1)–B(1)–H(1) N(3)–B(1)–H(1) C(5)–N(3)–C(6) C(5)–N(3)–B(1) C(6)–N(3)–B(1) C(5)–N(4)–C(7) C(5)–N(4)–C(8) C(7)–N(4)–C(8) 1.706(9) 1.699(8) 1.695(8) 1.354(9) 1.37(1) 1.55(1) 1.36(1) 1.36(1) 1.48(1) 1.37(1) 1.393(9) 1.54(1) 1.341(9) 108.2(7) 123.7(7) 128.1(7) 108.8(7) 124.3(9) 126.8(9) 126.9(6) 125.7(6) 107.4(7) 107.9(8) 107.6(8) 108.3(6) 112(3) 106(3) 106.2(6) 126.3(7) 127.5(7) 109.9(7) 125.1(8) 125.0(7) 1.704 1.345 1.382 1.579 1.351 1.385 1.442 1.579 108.2 126.4 125.1 109.4 124.9 125.8 129.3 123.6 107.1 108.8 106.4 107.6 111.3 N(4)–C(7) N(4)–C(8) N(5)–C(9) N(5)–C(10) N(5)–B(1) N(6)–C(9) N(6)–C(11) N(6)–C(12) C(2)–C(3) C(6)–C(7) C(10)–C(11) B(1)–H(1) C(9)–N(5)–C(10) C(9)–N(5)–B(1) C(10)–N(5)–B(1) C(9)–N(6)–C(11) C(9)–N(6)–C(12) C(11)–N(6)–C(12) S(2)–C(5)–N(3) S(2)–C(5)–N(4) N(3)–C(5)–N(4) N(3)–C(6)–C(7) N(4)–C(7)–C(6) S(3)–C(9)–N(5) S(3)–C(9)–N(6) N(5)–C(9)–N(6) N(5)–C(10)–C(11) N(6)–C(11)–C(10) N(1)–B(1)–N(5) N(3)–B(1)–N(5) N(5)–B(1)–H(1) 1.37(1) 1.45(1) 1.351(9) 1.38(1) 1.55(1) 1.36(1) 1.35(1) 1.46(1) 1.35(1) 1.33(1) 1.34(1) 0.97(6) 108.8(6) 124.8(6) 126.3(6) 109.4(7) 123.7(7) 126.9(7) 127.0(6) 125.4(6) 107.6(7) 109.6(7) 106.7(7) 127.2(6) 126.5(6) 106.3(7) 107.6(7) 107.9(7) 108.1(6) 108.7(6) 113(3) 1.331 1.180 a The calculated bond lengths and angles are identical for the three methimazolyl rings in NaTm and the three pyrazolyl rings in NaTp. Thus for clarity ab initio data are provided for a single representative ring system for each anion only.tetrakis(pyrazolyl)borate 36 both form polymeric structures more reminiscent of our structure, in which two water molecules bridge between each potassium ion and the heterocyclic nitrogens only bond terminally. In the case of M[B(pz)4] an h5 interaction is also observed, with the M1–ring centroid distances being 3.097(6) (M = K) and 3.257(4) Å (M = Na). Ab initio calculations The scope of the co-ordination chemistry of Tm is vast and can be estimated from the volume of available information on Tp.In order further to understand the analogy between these two systems and to probe their electronic structures and their relative complexing abilities both anions (Tm/Tp) and their sodium and copper(I) complexes were subjected to geometry optimisation procedures by ab initio calculations. Initially, however, it was thought instructive to begin the ab initio studies of this novel system by analysing the stability of the two possible conformers of Tm; namely that including a B–N linkage (I) and that possible via a B–S linkage (II).Consistent with the experimental results, the structure using a B–N linkage was found to be the more stable by 40.4 kcal mol21. Having established that calculations can identify the most stable conformer of the Tm anion, a viable comparison of the two tripodal ligands, as free anions, using ab initio calculations could be made.Their optimised geometries are shown in Fig. 4. The most noticeable features are that both ions have approximately C3 symmetry and that the rings have twisted about the B–N bonds with the result that the donor atoms are no longer prearranged for complexation. Instead the molecules have “inverted” such that the donors lie in a plane on the same side of the molecule as the B–H bond. It is particularly satisfying that the calculations are able to predict this conformational change.The angles of twist from the plane of the B–H axis are 43 (Tm) and 498 (Tp). Also of interest are the calculated parameters for the five-membered methimazole and pyrazole rings. The rings are planar and the bond lengths lie in the range 1.31–1.41 Å, consistent with extensive delocalisation. The calculated C–S distance is 1.704 Å and may be compared with the values of 1.686 Å for the parent thione structure 3-methylimidazoline- 2-thione and 1.766 Å for the corresponding thiol tautomer calculated in a similar manner.15 The charge distribution (as obtained from a Mulliken population analysis) shows that each ring in the Tm anion carries a charge of 20.62 electron compared with 20.57 electron in the Tp anion.This is compensated by changes in charge on the B–H unit, B being more positive and H less negative in the Tm anion. Thus in the Tm anion we have a ligand behaving as an anionic thione with the charge delocalised across the whole ligand.Overall the agreement of the calculated and observed data (Table 4) is good and deviations can on the whole be ascribed to the eVects of coordination. The calculated Tm structure closely parallels that determined by X-ray crystallography (in which anion–cation separation is observed). A direct comparison of the calculated structure of the free Tp anion with the crystal structure is not strictly meaningfulJ. Chem. Soc., Dalton Trans., 1999, 2119–2126 2123 since in the crystal the Tp is co-ordinated to the sodium ions (see above).Indeed, so far as we can ascertain, there are no examples of non-co-ordinated Tp in the literature. However, the structures of the anhydrous sodium salts of Tm and Tp were calculated (Fig. 4). In both cases the anions act as tridentate ligands, in contrast to the crystal structures. We ascribe this observation to the lack of competing solvent molecules, which forces ion pairing and thus co-ordination to the metal as the most stable arrangemen The two structures diVer somewhat.Most noticeable is that the pyrazole rings in NaTp lie parallel to the Na–B axis, giving approximately C3v symmetry, while in NaTm the methimazole rings lie at an angle of 408 to the Na–B axis, resulting in a lower, C3, symmetry (similar to the geometry observed in Zn(Tm)Br18). Despite the twist of the rings, the Na ? ? ? B distance in NaTm (3.72 Å) is much larger than that in NaTp (3.13 Å) and reflects the increase in size (from six to eight) of the chelate rings formed on complexation.The sodium ion in NaTp carries a greater positive charge (10.76) than that in NaTm (10.59) and since the charge on the B–H unit is identical in both species the greater negative charge is localised on the pyrazole rings. Since the eVect of solvent is clearly significant we then calculated the eVect of addition of one molecule of ammonia to the Fig. 2 The structure of the polymeric ribbon of NaTp?H2O. complex (Fig. 4, Table 4). In both cases the overall enthalpy of formation was more negative than for the unsolvated species. The stabilisation is greater for NaTp(NH3) than for NaTm- (NH3) (215.3 vs. 213.9 kcal mol21) which is in line with the greater charge carried by the sodium in NaTp. The Na–N and Na–S bond lengths increased by 0.039 and 0.055 Å respectively and the Na ? ? ? B distances also increased by 0.071 (NaTp) and 0.141 Å (NaTm). It is clear that solvent addition has a markedly greater eVect on the Tm complex, pulling the metal out of the ligand cavity.It can be envisaged that the addition of further molecules of solvent will further weaken the co-ordination of Tp and Tm to sodium. The ultimate structures observed in the solid state would seem to represent the “ideal” situation for each ligand, the generally less stable interaction of Tm and sodium leading to a pentahydrate and complete decomplexation, while the more tightly bound NaTp complex only takes up two water molecules, retaining some of the Na–N linkages.The impetus for the preparation of Tm was given by the desire to generate a soft tripodal ligand analogous to Tp and so the relative hardness of the ligands was also probed by our ab initio calculations. Co-ordination to the hard anion, Na1, has already been shown to favour Tp, with calculated complexation energies of 2153.4 kcal mol21 for NaTp and 2146.9 kcal mol21 for NaTm, in line with the expected hardness of the ligands, i.e.Tp is harder than Tm. Further to confirm this hypothesis required replacement of Na1 with a softer cation. In this case the Cu1 ion was chosen. The structures of CuTp and CuTm were optimised ‡ (Table 4) and the resulting complexation energies (calculated using the DZ basis set for Cu) were 2169.7 and 2174.0 kcal mol21 for CuTp and CuTm respectively, confirming that Tm is indeed the softer ligand. It is interesting that these calculations are in line with experimental evidence in predicting that CuTm is more stable than CuTp; CuTm is air stable and is resistant to complexing CO, while CuTp is prone to aerobic oxidation and readily forms adducts with p-acid ligands.18 This pattern of reactivity is also consistent with the observations of Riordan and co-workers 11–14 on other soft tripodal ligands.Fig. 3 The local structure of NaTp?H2O showing the atom numbering scheme and the thermal ellipsoids at the 40% level.‡ The more appropriate copper basis set DZ (14,11,5)/ [8,6,2] of Ahlrichs and co-workers 26 was used for these calculations.2124 J. Chem. Soc., Dalton Trans., 1999, 2119–2126 Table 3 Bond lengths (Å) and angles (8) for Na[HB(C3H3N2)3]?H2O X-Ray Ab initio a (for Tp2) X-Ray Ab initio b (for Tp2) Na(1)–O(1) Na(1)–O(1) Na(1)–N(2) Na(1)–N(4) Na(1)–N(4) N(1)–N(2) N(1)–C(1) N(1)–B(1) N(2)–C(3) N(3)–N(4) N(3)–C(6) N(3)–B(1) O(1)–Na(1)–O(1) O(1)–Na(1)–N(2) O(1)–Na(1)–N(4) O(1)–Na(1)–N(4) O(1)–Na(1)–N(2) O(1)–Na(1)–N(4) O(1)–Na(1)–N(4) N(2)–Na(1)–N(4) N(2)–Na(1)–N(4) N(4)–Na(1)–N(4) Na(1)–O(1)–Na(1) N(2)–N(1)–C(1) N(2)–N(1)–B(1) C(1)–N(1)–B(1) Na(1)–N(2)–N(1) Na(1)–N(2)–C(3) N(1)–N(2)–C(3) N(4)–N(3)–C(6) N(4)–N(3)–B(1) C(6)–N(3)–B(1) Na(1)–N(4)–Na(1) Na(1)–N(4)–N(3) 2.392(4) 2.389(4) 2.476(4) 2.427(4) 2.749(4) 1.365(5) 1.337(5) 1.543(5) 1.338(6) 1.366(4) 1.339(5) 1.549(5) 174.73(10) 84.0(1) 91.3(1) 80.7(1) 91.0(1) 87.8(1) 99.5(1) 97.1(1) 74.6(1) 168.9(1) 95.46(10) 110.1(3) 123.5(4) 126.0(4) 121.9(2) 123.8(3) 104.6(4) 110.1(3) 121.0(3) 128.5(3) 85.97(10) 127.6(2) 1.336 1.337 1.561 1.306 1.561 110.6 120.3 129.1 111.4 N(4)–C(4) N(5)–N(6) N(5)–C(7) N(5)–B(1) N(6)–C(9) C(1)–C(2) C(2)–C(3) C(4)–C(5) C(5)–C(6) C(7)–C(8) C(8)–C(9) H(1)–B(1) Na(1)–N(4)–C(4) Na(1)–N(4)–N(3) Na(1)–N(4)–C(4) N(3)–N(4)–C(4) N(6)–N(5)–C(7) N(6)–N(5)–B(1) C(7)–N(5)–B(1) N(5)–N(6)–C(9) N(1)–C(1)–C(2) C(1)–C(2)–C(3) N(2)–C(3)–C(2) N(4)–C(4)–C(5) C(4)–C(5)–C(6) N(3)–C(6)–C(5) N(5)–C(7)–C(8) C(7)–C(8)–C(9) N(6)–C(9)–C(8) N(1)–B(1)–N(3) N(1)–B(1)–N(5) N(3)–B(1)–N(5) H(1)–B(1)–N(1) 1.339(5) 1.362(5) 1.343(5) 1.532(6) 1.324(5) 1.369(7) 1.374(7) 1.384(6) 1.367(6) 1.361(8) 1.367(7) 1.08(3) 127.5(3) 90.9(2) 94.2(3) 104.9(3) 109.1(4) 124.1(3) 126.8(4) 105.5(4) 108.9(4) 104.2(4) 112.2(4) 111.8(4) 104.1(4) 109.1(4) 109.0(5) 104.3(4) 112.1(5) 109.8(3) 112.6(4) 108.6(4) 1.371 1.406 1.196 108.5 103.1 111.4 108.2 110.7 a The calculated bond lengths and angles are identical for the three methimazolyl rings in NaTm and the three pyrazolyl rings in NaTp.Thus for clarity ab initio data are provided for a single representative ring system for each anion only. Fig. 4 The calculated structures of the Tp and Tm anions, the complexes NaTm and NaTp and their ammonia adducts, NaTm?NH3 and NaTp?NH3.J. Chem. Soc., Dalton Trans., 1999, 2119–2126 2125 Table 4 Calculated data for Tm, Tp and their metal complexes Tm (X-ray) Tm Tp NaTm NaT(pm2) NaT(p2m) NaTp NaTm?NH3 NaTp?NH3 CuTma CuTpa Parameter d(B–H)/Å d(B–N)/Å d(C]] S)/Å d(M–S)/Å d(M–N)/Å d(M ? ? ? B)/Å H–B–N/8 N–B–N/8 S–M–S/8 N–M–N/8 N–M–S/8 0.97 1.55 (av.) 1.70 (av.) — 1.180 1.579 1.704 111.3 107.6 1.196 1.561 110.7 108.2 1.204 1.563 1.726 2.725 3.722 105.0 113.5 103.4 1.203 1.536 1.575 (av.) 1.725 (av.) 2.710 (av.) 2.353 3.439 106.0 (av.) 110.2 (av.) 115.5 110.4 90.2 101.4 1.201 1.546 (av.) 1.591 1.725 2.690 2.335 (av.) 3.281 106.2 (av.) 112.5 87.6 92.6 105.4 1.199 1.556 2.333 3.126 107.4 111.4 88.6 1.205 1.562 1.724 2.780 3.863 105.0 113.6 100.1 1.201 1.554 2.372 3.197 107.4 111.5 86.5 1.202 1.565 1.727 2.454 3.389 104.8 113.7 111.7 1.199 1.557 2.155 2.908 108.2 110.8 95.1 Complexation energy/kcal mol21 Charge on M Charge on B Charge on rings 10.89 20.62 10.83 20.57 2146.9 10.59 10.89 20.45 2147.4 10.61 10.88 20.45 (av.) 2149.0 10.67 10.87 20.48 (av.) 2153.4 10.76 10.86 20.50 2160.8 10.62 10.88 20.47 2168.7 10.73 10.87 20.51 2174.0 10.81 10.83 20.50 2169.7 10.90 10.81 20.53 a The copper basis set DZ (14,11,5)/[8,6,2] of Ahlrichs and co-workers 26 was used for these calculations.2126 J.Chem. Soc., Dalton Trans., 1999, 2119–2126 Lastly, we note that Tp and Tm form the extremes of a potential series of ligands with S3, S2N, SN2 and N3 donor sets. The S2N donor with two methimazoles (m) and one pyrazole (p), T(pm2), has recently been prepared by Parkin and co-workers 37 by an alternative synthetic methodology, while the bis(pyrazole) monomethimazole species, T(p2m), remains unknown at this time.However, we have examined the sodium salts of these intermediate ligands by ab initio calculations (Table 4). The resulting energies of complexation show a clearly graded change in behaviour on going from Tp through to Tm. We believe this indicates that this series of ligands will modulate the behaviour of metal centres in a controlled fashion by presenting a series of well defined, closely related donor sets with diVering electron donor properties.We have found, in line with the work of Riordan,38 Parkin 39 and Janiak 40 and their co-workers on related systems, that for the preparation of metal complexes the thallium(I) salt of Tm is more convenient than the sodium salt, particularly when using metal halide precursors. The thallium halide formed during the reaction is insoluble, leaving the clean metal complex in solution.We have repeated the preparation of the previously reported zinc bromide complex, Zn(Tm)Br18 and find that separation is more straightforward and that the yield is improved. The preparation of the thallium salt is outlined in the Experimental section. Acknowledgements We thank Dr. A. R. Kennedy for collection of the X-ray data. References 1 C. A. Tolman, Chem. Rev., 1977, 77, 313. 2 S. Trofimenko, Inorg. Synth., 1970, 12, 99. 3 S. Trofimenko, Chem. Rev., 1993, 93, 943. 4 M. I. Bruce and A. P. P. Ostazewski, J. Chem. Soc., Chem. Commun., 1972, 1124. 5 M. I. Bruce and A. P. P. Ostazewski, J. Chem Soc., Dalton Trans., 1973, 2433. 6 N. Kitajima and W. B. Tolman, Prog. Inorg. Chem., 1995, 43, 419. 7 D. L. Reger, Coord. Chem. Rev., 1996, 147, 571. 8 M. Etienne, Coord. Chem. Rev., 1996, 156, 201. 9 K. Yoon and G. Parkin, J. Am. Chem. Soc., 1991, 113, 8414. 10 S. R. Cooper, Acc. Chem. Res., 1988, 21, 141. 11 P. Ge, B. S. Haggerty, A. L.Rheingold and C. G. Riordan, J. Am. Chem. Soc., 1994, 116, 8406. 12 C. Ohrenberg, P. Ge, P. J. Schebler, C. G. Riordan, G. P. A. Yap and A. L. Rheingold, Inorg. Chem., 1996, 35, 749. 13 C. Ohrenberg, M. M. Saleem, C. G. Riordan, G. P. A. Yap, A. Verma and A. L. Rheingold, Chem. Commun., 1996, 1081. 14 P. J. Schebler, C. G. Riordan, L. Liable-Sands and A. L. Rheingold, Inorg. Chim. Acta, 1998, 270, 543. 15 M. Garner, D. R. Armstrong, J. Reglinski, W. E. Smith, R. Wilson and J.H. McKillop, Bio-org. Med. Chem. Lett., 1994, 4, 1357. 16 J. Elguero, C. Marzin, A. R. Katritzky and P. Linda, Adv. Heterocycl. Chem., 1976, 1, Suppl. p. 400. 17 R. S. Balestrero, D. M. Forkey and J. G. Russell, Magn. Reson. Chem., 1986, 24, 651. 18 M. Garner, J. Reglinski, I. Cassidy, M. D. Spicer and A. R. Kennedy, Chem. Commun., 1996, 1975. 19 SIR 92, A. Altomare, M. C. Burla, M. Camalli, M. Cascarano, C. Giacovazzo, A. Guagliardi and G. Polidori, J. Appl. Crystallogr., 1994, 27, 435. 20 DIRDIF 92, P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman, S. Garcia-Granda, J. M. M. Smits and C. Smykalla, DIRDIF Program System, Technical Report of the Crystallography Laboratory, University of Nijmegen, 1992. 21 TEXSAN, crystal structure analysis package, Molecular Structure Corporation, Woodlands TX, 1985 and 1992. 22 W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys., 1972, 56, 2257. 23 P. C. Harihan and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213. 24 J. D. Dill and J. A. Pople, J. Chem. Phys., 1975, 62, 2921. 25 GAUSSIAN 94, M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. Gill, W. B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. A. Keith, G. A. Peterssoa, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakkara, M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defress, J. Baker, J. P. Stewart, M. Head-Gordon, C. Gonzalez and J. A. Pople, Gaussian Inc., Pittsburg, PA, 1995. 26 A. Schäfer, H. Horn and J. Ahlrichs, J. Chem. Phys., 1992, 97, 2571. 27 J. Kister, G. Assef, G. Mille and J. Metzger, Can. J. Chem., 1979, 57, 813. 28 J. Kister, G. Assef, G. Mille and J. Metzger, Can. J. Chem., 1979, 57, 822. 29 T. Aoyagi, H. M. M. Shearer, K. Wade and G. Whitehead, J. Organomet. Chem., 1979, 175, 21. 30 S. Harder, M. H. Prosenc and U. Rief, Organometallics, 1996, 15, 118. 31 H. V. R. Dias, H.-L. Lu, R. E. RatcliVe and S. G. Bott, Inorg. Chem., 1995, 34, 1975. 32 H. V. R. Dias and H.-J. Kim. Organometallics, 1996, 15, 5374. 33 H. V. R. Dias, W. Jin, H.-J. Kin and H.-L. Lu, Inorg. Chem., 1996, 35, 2317. 34 G. G. Lobbia, P. Cecchi, R. Spagna, M. Colapietro, A. PiVeri and C. Pettinari, J. Organomet. Chem., 1995, 485, 45. 35 C. Janiak, Chem. Ber., 1994, 127, 1379. 36 C. Lopez, R. M. Claramunt, D. Sanz, C. F. Foces, F. H. Cano, R. Faure, E. Cayon and J. Elguero, Inorg. Chim. Acta, 1990, 176, 195. 37 C. Kimblin, T. Hascall and G. Parkin, Inorg. Chem., 1997, 36, 5680. 38 P. J. Schebler, C. G. Riordan, I. A. Guzei and A. L. Rheingold, Inorg. Chem., 1998, 37, 4754. 39 C. M. Dowling, D. Leslie, M. H. Chisholm and G. Parkin, Main Group Chem., 1995, 1, 29. 40 C. Janiak, S. Temizdemir and T. G. Scharmann, Z. Anorg. Allg. Chem., 1998, 624, 755. Paper 9/01703H

ISSN:1477-9226    

DOI:10.1039/a901703h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 2123 Spectroscopic and structural studies on 1 : 2 adducts of silver(I) salts with tricyclohexylarsine Graham A. Bowmaker,a EVendy,b,c Brian W. Skelton b and Allan H. White b a Department of Chemistry, University of Auckland, Private Bag 92019, Auckland, New Zealand b Department of Chemistry, The University of Western Australia, Nedlands, W.A. 6907, Australia c Jurusan Pendidikan Kimia, FPMIPA IKIP, Malang, Jalan Surabaya 6, Malang, 65145, Indonesia A series of mononuclear complexes [Ag{As(C6H11)3}2X] (X = Cl, Br, I, CN, NCO, OClO3, O2NO or O2CCF3) have been synthesized and characterised by room temperature single crystal X-ray determinations and lowfrequency vibrational spectroscopy.For X = Cl, Br, I or OClO3 the crystal structures are isomorphous with a number of their previously recorded P(C6H11)3 counterparts, crystallising in the familiar monoclinic C2/c array with the Ag]X bond lying on a crystallographic 2 axis (the perchlorate is disordered) which also relates the pair of As(C6H11)3 ligands.The silver environment in all is planar three-co-ordinate XAgAs2. For X = CN, NCO or O2NO the derivative lower-symmetry triclinic P1� array, also common among the P(C6H11)3 analogues is found. The tri- fluoroacetate, resembling the nitrate in that the anion behaves as a small ‘bite’ bidentate ligand, [Ag{As(C6H11)3}2- (O2CCF3)], is monoclinic space group P21/n.The thiocyanate by contrast, utilising the ambidentate capacity of the SCN ligand, remarkably, is a linear polymer, ? ? ? {As(C6H11)3}2Ag(SCN)Ag(SCN) ? ? ?, triclinic, space group P1, in which the silver atom environment is four-co-ordinate, NSAgAs2. Far-IR spectra of the halide and pseudohalide complexes in the series exhibit single bands due to n(AgX) vibrational modes at 230, 165, 139, 262 and 311 cm21 for X = Cl, Br, I, NCO and CN respectively. These are all higher than the values for the corresponding P(C6H11)3 complexes, suggesting that the Ag]X bonding is stronger in the As(C6H11)3 complexes, this being the result of weakened Ag]L bonding in going from L = P(C6H11)3 to As(C6H11)3.In recent years single crystal X-ray determinations in concert with low frequency vibrational spectroscopic studies have defined isomeric forms and their bonding characteristics for a wide spectrum of adducts of coinage metal(I) salts, MX, with unidentate Group V bases. Whereas for donor atoms E = N a variety of stereochemical and electronic characteristics may be employed, linear RCN, trigonal planar pyridine derivatives, tertiary NR3, etc., extensive essays with heavier E = P, As, Sb tend to be largely restricted to tertiary ER3, adducts MX:ER3 (1 : n) for R = Ph being commonly defined for n = 4 as the ionic form [M(EPh3)4]1X2 for X = ClO4 or NO3,1 for n = 3 as mononuclear [M(EPh3)3X] (diverse X),2 for n = 2 as mononuclear [M(EPh3)2X] [small M, E (= Cu, P) only] or binuclear [(Ph3E)2M(m-X)2- M(EPh3)2],3 and for n = 1 as ‘step’ or ‘cubane’ tetramers [{(Ph3E)MX}4] (E = P or As only),4 all except [M(EPh3)2X] having tetrahedral metal atom environments.Nevertheless, studies with alternative R substituents of diVerent stereochemical or electronic characteristics have shown that it is possible to prepare extensive series of complexes of a particular structural type with a wide range of X, as shown recently in the arrays of adducts of silver(I) salts with tricyclohexylphosphine, nmax being 2 in mononuclear [Ag{P(C6H11)3}2X],5 while n = 1 forms may be frequently binuclear rather than tetranuclear.6 In that spirit we now extend our studies of these arrays to encompass adducts of silver(I) salts with tricyclohexylarsine, recording in the present report results obtained for the AgX:As(C6H11)3 (1 : 2) system, and, in the following,7 the 1 : 1 system.We report here syntheses, and structural and spectroscopic characterisation, of AgX :As(C6H11)3 (1 : 2) for X = Cl, Br, I, CN, NCO, NO3, ClO4 or O2CCF3 as mononuclear [Ag{As(C6H11)3}2X] arrays; by contrast, in keeping with the ambidentate nature of the SCN ligand, AgSCN:As(C6H11)3 (1 : 2) is found to be an infinite one-dimensional polymer with four-co-ordinate silver(I).Experimental Syntheses All compounds were obtained by dissolution of the appropriate silver(I) salt (1 mmol) with tricyclohexylarsine (2 mmol) with warming in acetonitrile (5 cm3), colourless crystals depositing on cooling and standing.X = Cl: m.p. > 109 8C (decomp.), no satisfactory analysis data obtained. X = Br: m.p. >125 8C (decomp.) (Found: C, 51.7; H, 7.8. Calc. for C36H66AgAs2Br: C, 51.69; H, 7.95%). X = I: m.p. >131 8C (decomp.) (Found: C, 49.1; H, 7.5. Calc. for C36H66AgAs2I: C, 48.94; H, 7.53%). X = CN: m.p. 148–150 8C (Found: C, 56.8; H, 8.7; N, 1.6. Calc. for C37H66AgAs2N: C, 56.78; H, 8.50; N, 1.7%).X = NCO: m.p. >158 8C (decomp.) (Found: C, 55.4; H, 8.4. Calc. for C37H66AgAs2NO: C, 55.64; H, 8.33%). X = NO3: m.p. >143 8C (decomp.) (Found: C, 52.7; H, 8.0. Calc. for C36H66AgAs2NO3: C, 52.82; H, 8.13%). X = ClO4: m.p. >189 8C (decomp.), no satisfactory analysis obtained. X = O2CCF3: m.p. 146–148 8C (Found: C, 52.2; H, 7.9. Calc. for C38H66AgAs2F3O2: C, 52.48; H, 7.65%). X = SCN: m.p. >171 8C (decomp.) (Found: C, 54.7; H, 7.9. Calc. for C37H66AgAs2NS: C, 54.55; H, 8.17%).Structure determinations Unique room-temperature diVractometer data sets were measured (2q–q scan mode, 2qmax as specified: monochromatic Mo-Ka radiation, l = 0.71073 Å, T ª 295 K) yielding N independent reflections, No with I > 3s(I) being considered ‘observed’ and used in the full matrix least squares refinements after gaussian absorption correction. Anisotropic thermal parameters were refined for the non-hydrogen atoms, (x, y, z, Uiso)H being included at estimated values.Conventional residuals R,R9 on |F| are quoted at convergence, statistical reflection weights being derivative of s2(I) = s2(Idiff) 1 0.0004s4(Idiff). Neutral2124 J. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 atom complex scattering factors were employed, computation using the XTAL 3.2 program system.8 Abnormal feature/idiosyncrasies/ variations in procedure are noted individually below (‘variata’). In all figures, 20% thermal envelopes are shown for the non-hydrogen atoms, hydrogen atoms (where shown) having arbitrary radii of 0.1 Å.Ring carbon atoms are labelled lmn (l = ligand number, m = ring number, n = atom number). A general comment is called for in respect of the treatment of cyanate and cyanide moieties in this and accompanying/ following papers. In all cases assignment of the atoms concerned was attempted on the basis of crystallographic and chemical considerations during refinement by consideration of the behaviour of residuals, thermal parameters, geometries, etc.In a few cases a ‘definitive’ assignment emerged, but in many cases the result was ambiguous, and that presented should be treated with appropriate circumspection. Thus, all cyanates are treated as N-bonded; in e.g. the example presented in the present paper apparent bond length distributions and the behaviour of residuals during refinement support this assignment, but the attachment of the ligand at the nitrogen is non-linear, and ‘thermal motion’ at the peripheral oxygen is abnormally high.A similar observation may be made in respect of cyanides: the weight of evidence supports the probability that these are C-bound in terminally attached species, or in isolated or bridging [NCAgCN]2 arrays, but are scrambled in situations where they behave as a bridging group between two similar atoms. Individual cases are described in detail, particularly for those few where departure from the above norm is suggested.Crystal/refinement data for [AgX{As(C6H11)3}2] º C36H66Ag- As2X. (a) Monoclinic, space group C 2/c (C6 2h, no. 15), Z = 4. (i) X 5 Cl. M = 792.1, a = 17.016(3), b = 9.209(3), c = 24.821(5) Å, b = 108.07(2)8, U = 3698 Å3, Dc = 1.423 g cm23, F(000) = 1648, mMo = 24.2 cm21, specimen 0.23 × 0.40 × 0.11 mm, A* min,max = 16, 1.72, 2qmax = 608, N = 5375, No = 3466, R = 0.050, R9 = 0.062. (ii) X 5 Br. M = 836.6, a = 16.997(7), b = 9.239(2), c = 24.826(8) Å, b = 108.20(3)8, U = 3704 Å3, Dc = 1.500 g cm23, F(000) = 1720, mMo = 34.2 cm21, specimen 0.19 × 0.34 × 0.17 mm, A* min,max = 1.61, 1.82 (analytical correction), 2qmax = 558, N = 4056, No = 2642, R = 0.035, R9 = 0.035.(iii) X 5 I. M = 883.6, a = 17.094(2), b = 9.344(6), c = 24.99(1) Å, b = 108.67(2)8, U = 3781 Å3, Dc = 1.552 g cm23, F(000) = 1792, mMo = 31.1 cm21, specimen 0.58 × 0.44 × 0.10 mm, A* min,max = 1.36, 3.37 (analytical correction), 2qmax = 608, N = 5490, No = 3543, R = 0.039, R9 = 0.040.(iv) X 5 ClO4. M = 856.1, a = 17.03(1), b = 9.568(4), c = 25.185(8) Å, b = 109.57(4)8, U = 3866 Å3, Dc = 1.471 g cm23, F(000) = 1776, mMo = 23.2 cm21, specimen 0.15 × 0.21 × 0.09 mm, A* min,max = 1.18, 1.31 (analytical correction), 2qmax = 508, N = 3471, No = 1898, R = 0.057, R9 = 0.058. Variata. For the above compounds the cell and coordinate setting previously defined5 for certain P(C6H11)3 counterparts was adopted. In the perchlorate the chlorine atom is located on the 2 axis, while the unco-ordinated oxygen atoms are necessarily disordered about the twofold axis in the C2/c model; the coordinated oxygen atom was also modelled as disordered, lying somewhat oV the Ag ? ? ? Cl axis.(b) Triclinic, space group P1� (Ci 1, no. 2), Z = 2. (i) X 5 CN. M = 782.7, a = 9.205(3), b = 9.759(4), c = 23.673(7) Å, a = 95.13(3), b = 97.38(2), g = 117.29(3)8, U = 1847 Å3, Dc = 1.407 g cm23, F(000) = 816, mMo = 23.5 cm21, specimen 0.22 × 0.10 × 0.15 mm, A* min,max = 1.23, 1.41, 2qmax = 508, N = 6527, No = 2861, R = 0.056, R9 = 0.051.(ii) X 5 NCO. M = 798.7, a = 9.327(7), b = 9.816(2), c = 23.692(4) Å, a = 95.54(2), b = 96.52(3), g = 117.41(4)8, U = 1886 Å3, Dc = 1.406 g cm23, F(000) = 832, mMo = 23.0 cm21, specimen 0.46 × 0.20 × 0.06 mm, A* min,max = 1.14, 1.57, 2qmax = 508, N = 6447, No = 4083, R = 0.060, R9 = 0.068. (iii) X 5 NO3. M = 818.7, a = 9.199(2), b = 9.813(3), c = 23.844(5) Å, a = 95.08(2), b = 96.13(2), g = 115.16(2)8, U = 1915 Å3, Dc = 1.419 g cm23, F(000) 852, mMo = 22.7 cm21, specimen 0.42 × 0.55 × 0.32 mm, A* min,max = 1.27, 1.83, 2qmax = 508, N = 6717, No = 5285, R = 0.043, R9 = 0.050. Variata.For the above compounds the cell and coordinate setting previously defined for certain P(C6H11)3 counterparts 5 was adopted. In the cyanate and nitrate various components of various cyclohexyl rings were modelled as disordered over pairs of sites, occupancies set at 0.5 after trial refinement.In the cyanate the thermal parameter refinement of the peripheral component was ill behaved and the isotropic form was adopted, possibly encompassing disorder or isomerism of the potentially ambidentate ligand. Although modelled as C-bound, a C/N, disordered model for the cyanide is indistinguishable on the basis of refinement, a possibility supported by rather aberrant thermal parameters. (c) X = O2CCF3. M = 869.7, monoclinic, space group P21/n [C5 2h, no. 14 (variant)], a = 10.085(5), b = 24.36(1), c = 16.462(8) Å, b = 93.91(4)8, U = 4034 Å3, Dc (Z= 4) = 1.432 g cm23, F(000) = 1800, mMo = 21.7 cm21, specimen 0.29 × 0.34 × 0.47 mm, A* min,max = 1.72, 2.00, 2qmax = 508, N = 6838, No = 4579, R = 0.046, R9 = 0.052. Variata. One of the ligand rings was modelled as disordered over two sets of sites, disordered atom occupancy set at 0.5 after trial refinement. (d ) X = SCN. M = 814.8, triclinic, space group P1 (C1 1, no. 1), a = 13.767(5), b = 13.589(3), c = 11.952(6) Å, a = 100.23(3), b = 90.98(4), g = 118.47(2)8, U = 1920 Å3, Dc (Z = 2) = 1.408 g cm23, F(000) = 848, mMo = 23.1 cm21, specimen 0.09 × 0.10 × 0.46 mm, A* min,max = 1.17, 1.32, 2qmax = 508, N = 6749, No = 5040, R = 0.042, R9 = 0.041 (preferred hand).CCDC reference number 186/983. See http://www.rsc.org/suppdata/dt/1998/2123/ for crystallographic files in .cif format. Spectroscopy Far-infrared spectra were recorded at 4 cm21 resolution at room temperature as Polythene discs on a Digilab FTS-60 Fourier transform infrared spectrometer employing an FTS-60V vacuum optical bench with a 6.25 mm mylar film beam splitter, a mercury lamp source and a pyroelectric triglycine sulfate detector.Discussion Crystal structures Depending on the nature of the anionic moiety X, species of the type [M(ER3)n]X or [{(R3E)nMX}m] may be isolated for M = univalent coinage metal, E = P, As or Sb, n rising to a maximum of 4 in ionic compounds or 3 in those where X is covalently bound, the ligand substituent being variously alkyl or aryl.With R = cyclohexyl, structurally characterised examples are more limited and characterised by diminished n. For simple molecular or ionic adducts, the maximum n recorded in structurally characterised examples is 2, and those for E = P, i.e. P(C6H11)3, as ligand, only. For M = Cu or Ag, ‘simple’ X, X interacts with one metal alone in MX :P(C6H11)3 (1 : 2) complexes, all being mononuclear, regardless of whether X is a (pseudo-)halide or oxyanion, [M{P(C6H11)3}2X] arrays being structurally defined for (a) Cu, X = I,9 N3,10 OClO3,11 FBF3 12 (X unidentate) or O2NO13 [X (small bite) bidentate], (b) Ag, X = Cl, Br, I, CN, SCN, NCO,5 OClO3 5,14 (unidentate) or O2NO5,14 (small bite bidentate).By contrast, for M = Au (which we shall not consider further), all AuX:P(C6H11)3 (1 : 2) adducts take the form of ionic [Au{P(C6H11)3}2]X (X = Cl,15 SCN16 or PF6 17), with a linear two-co-ordinate metal atom in the cation.The results of the present room-temperature single crystal X-ray studies are consistent in stoichiometry, connectivity and stereochemistry with the complexes being of 1 : 2 AgX:As(C6-J. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 2125 H11)3 formulation [the first coinage metal As(C6H11)3 complexes so characterised]. The structures of a selection of the compounds studied are shown in Figs. 1 and 2. For X = Cl, Br, I, CN, NCO, OClO3 or O2NO, all anionic moieties except the nitrate (which is a small bite bidentate) behave as unidentate ligands, so that the metal is (pseudo-)planar three-co-ordinate As2MX.Like their P(C6H11)3 counterparts,5 these complexes form a limited number of isomorphous structural systems, arising from a monoclinic C2/c cell, Z = 4, in which the M]X bond is disposed on a crystallographic 2 axis which relates the two E(C6H11)3 ligands, only one of which is crystallographically independent, or a derivative triclinic P1� array, Z = 2, in which one independent [M{E(C6H11)3}2X] moiety comprises the asymmetric unit of the structure.In the P(C6H11)3 array the C2/ Fig. 1 Molecular projections of (a) the perchlorate, (b) the nitrate, (c) the trifluoroacetate, normal to the As2Ag plane c form was found for the X = Cl, Br, I, CN or SCN adducts (the anionic moiety in the last being disordered); for the present As(C6H11)3 array that form is found for the X = Cl, Br or I adducts, inclusive also of the O-bonded perchlorate complex.In the P(C6H11)3 array the position of the latter was somewhat ambiguous, an earlier determination having described it as in the triclinic P1� for, but further studies suggested the possibility of the achievement of the higher C2/c symmetry in admixture or as an alternative to it, the diVerence between the two forms possibly being very slight (dependent on crystallisation conditions?). Here, with As(C6H11)3 as ligand, we find the array unproblematically C2/c, although, as modelled in that space group, the perchlorate is necessarily disordered about the twofold axis; in detail, the silver atom lies on that axis, as does the chlorine, the latter albeit with a thermal parameter slightly higher than e.g., the anionic ligands, but not so high as to suggest gross disorder.By contrast, the co-ordinated oxygen atom is resolvable into a pair of symmetry-related oV-axis components 1.31(4) Å apart; the associated geometry should be treated appropriately circumspectly.The ligand components match those of the P(C6H11)3 analogues, hich coordination geometries about the metal are compared in Table 1. The alternative form recorded for the remaining P(C6H11)3 complexes of ref. 5 is a triclinic P1� array derivative of the C2/c form, found for X = NCO, O2NO (and OClO3), the two P(C6H11)3 ligands now having diVerent conformations with disorder resolvable in ring 21n.This form is also found among the As(C6H11)3 complexes, again for the X = NCO or O2NO arrays, but also for X = CN, the CN/P(C6H11)3 complex having been found to be C2/c. The disorder characteristics among the ligand rings are diverse, aVecting rings 12n and 21n in varying degree [ring 21n having been modelled as disordered in the previous P(C6H11)3/nitrate complex], manifested either as exalted thermal motion and/or resolvable disorder in either or both of these rings. In the cyanide and the cyanate (particularly the latter) the behaviour of the anionic moiety is not conducive to definitive assignment of co-ordination of one particular end of that ligand; the ‘bent’ co-ordination of the cyanate in particular is at variance with the NC assignment supported by the bond length. The Ag]N]C angle [133(1)8] in the present complex is significantly less than the value [151.2(4)8] in the P(C6H11)3 analogue, also modelled as N-bound.5 Given that in both P(C6H11)3 and As(C6H11)3 adducts the nitrate can co-ordinate as O,O9-bidentate, albeit of small ‘bite’, it may be that the coordination site has a more appropriately spacious ambience (vis-à-vis the C2/c form), permitting high ‘thermal motion’ and/or alternative co-ordination modes more readily. The geometries of these complexes are also compared with those of their P(C6H11)3 counterparts in Table 1.The trifluoroacetate complex, also mononuclear, crystallizes in a form diVerent to the above, being monoclinic, P21/n, Z = 4, with one [Ag{As(C6H11)3}2(O2CCF3)] molecule, devoid of crystallographic symmetry, comprising the asymmetric unit of the structure.It resembles the nitrate in that the anion behaves as a bidentate ligand of small ‘bite’, raising the co-ordination number of the silver to four. The ligand conformation, exhibiting disorder in one of the substituent rings, more nearly resembles that of the triclinic P1� examples than the monoclinic C2/c (which have molecular 2 symmetry). Given the seriously diVerent base strengths of O2CCF3 vs.O2NO, it might be expected that this would be reflected in the parameters of the metal atom environment; changes are relatively minor, perhaps the opposite of expectations which would suggest that O2CCF3 being a stronger base than O2NO should co-ordinate more strongly. The Ag]O distances in the O2CCF3 are appreciably larger, with Ag]As shorter and As]Ag]As enlarged.A comparison of the relationship between the d(Ag]E) bond lengths and the E]Ag]E bond angles for [Ag{E(C6H11)3}2X] (E = P or As) is shown in Fig. 3. This shows that the longer Ag]E bonds are generally accompanied by smaller E]Ag]E angles, but the2126 J. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 Table 1 Molecular core geometries for mononuclear [Ag{E(C6H11)3}2X] complexes. Distances in Å, angles in 8. Values for E = As are given below those for E = P. Values for the triclinic P1� phase (rather than the monoclinic C2/c phase) are characterized by the presence of two distance/angle entries X Cl Br I CN OClO3 NCO O2NO O2CCF3 Ag]E 2.4712(8) 2.5628(7) 2.4734(9) 2.5559(7) 2.478(1) 2.5587(8) 2.4803(9) 2.558(2), 2.568(2) 2.429(1), 2.4323(9) 2.517(1) 2.457(1), 2.454(1) 2.541(2), 2.540(2) 2.451(2), 2.440(2) 2.5243(8), 2.5185(8) 2.432(2), 2.437(2) 2.506(1), 2.511(1) Ag]X 2.489(1) 2.481(2) 2.618(1) 2.556(1) 2.778(1) 2.715(2) 2.153(6) 2.14(2) 2.720(7), 3.014(9) 2.61(2) 2.205(6) 2.38(1) 2.47(1), 2.79(1) 2.452(8), 2.553(6) 2.531(6), 2.666(6) 2.515(7), 2.585(7) E]Ag]E 128.29(3) 122.21(4) 129.42(3) 123.39(3) 130.75(4) 125.48(4) 124.86(4) 120.86(8) 147.34(3) 152.40(7) 133.65(4) 127.31(7) 139.14(9) 134.39(3) 141.19(6) 137.41(4) E]Ag]X 115.85(2) 118.90(2) 115.29(2) 118.31(2) 114.62(3) 117.26(2) 117.57(2) 114.7(6), 124.2(6) 98.5(1), 113.5(1) 102.8(5), 103.9(5) 111.6(1), 114.5(1) 115.5(2), 116.7(2) 110.5(2), 109.0(2) 106.5(2)–116.6(2) 103.5(1)–109.2(1) 105.5(2)–113.2(2) Data for the E = P array are taken from ref. 5, inclusive of data for the perchlorate from ref. 14. For the present As(C6H11)3 arrays, for the cyanide, ‘C]N’ is 0.90(3) Å and ‘Ag]C]N’ 175(2)8; for the cyanate ‘C]N,O’ are 0.75(2), 1.34(3) Å, and Ag]N]C, N]C]O 133(1), 172(2)8 [cf. 1.06(1), 1.25(2) Å; 151.2(4), 174.5(6)8 for the E = P analogue]; 5 for the nitrate Ag]O]N are 101.6(7), 95.2(4)8, and for the perchlorate Ag]O]Cl are 137(2)8. E = P data for the O2CCF3 adduct are as yet unpublished; the E = P structure is isomorphous, albeit devoid of disorder.It should be noted that the present monoclinic C2/c array, characteristic of silver(I) complexes, and with b ª 1088, is diVerent to that found in some copper(I) analogues where b ª 968, the cyclohexyl ring dispositions diVering in the two ‘polymorphs’; see ref. 18. Fig. 2 A single strand of the polymer of the thiocyanate correlation shows a considerable degree of scatter, particularly in the case of the E = As complexes.It has previously been noted that the P]Ag]P angle in some three- and four-coordinate AgX/diphosphine complexes is sensitive to the nature of the Ag]X bonding,19 and a correlation between increasing d(Hg]P) and decreasing q(P]Hg]P) has been observed for some four-co-ordinate triphenylphosphine mercury(II) complexes [HgX2(PPh3)2].20 This correlation has been attributed to an eVect of the anionic ligand X on both of these structural parameters, as a result of competition for electron density between the M]P and the M]X bonds.5 The straight lines in Fig. 3 are approximate correlations for [Ag{E(C6H11)3}2X] with unidentate anionic ligands X, and these show that d(Ag]E) is approximately equally sensitive to changes in X for both E = P and As, but that q(E]Ag]E) is more sensitive to such changes for E = As. This is consistent with the view that As(C6H11)3 is a weaker electron donor than P(C6H11)3, so that the eVects of changes in the Ag]X bonding on the As]Ag]As angle are significantly greater. The points for [Ag{As(C6H11)3}2X] (X = NO3 or O2CCF3) in Fig. 3 are displaced to smaller d(Ag]E) and q(E]Ag]E) relative to the line corresponding to the correlation for the unidentate ligands X = ClO4, CN, Cl, Br or I. The mostJ. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 2127 Table 2 Silver environments in AgSCN:As(C6H11)3 (1 : 2). Distances in Å, angles in 8. The two values in each entry are for silver (associated ligands) with n = 1 or 2 respectively. Primed atoms belong to the adjoining moiety Ag(n)]As(n1) Ag(n)]As(n2) Ag(n)]S(n) Ag(n)]N(n9) S(n)]C(n) N(n)]C(n) Ag(1) ? ? ? Ag(2,29) 2.573(1), 2.574(3) 2.580(2), 2.581(2) 2.602(5), 2.602(5) 2.33(1), 2.36(2) 1.64(2), 1.62(2) 1.12(2), 1.21(2) 6.433(3), 6.435(3) As(n1)]Ag(n)]As(n2) As(n1)]Ag(n)]S(n) As(n1)]Ag(n)]N(n) As(n2)]Ag(n)]S(n) As(n2)]Ag(n)]N(n9) S(n)]Ag(n)]N(n9) Ag(n)]S(n)]C(n) S(n)]C(n)]N(n) C(n)]N(n)]Ag(n9) 121.84(7), 122.11(9) 112.1(1), 111.8(1) 103.3(3), 107.6(4) 111.8(1), 112.6(1) 109.5(4), 102.0(3) 94.2(5), 96.8(4) 98.7(7), 99.6(6) 178(1), 174(2) 144(2), 139(1) likely reason for this is that NO3 and O2CCF3 act as bidentate ligands in these complexes.This is supported by the observation that the corresponding points for [Ag{P(C6H11)3}2X] (X = NO3 or O2CCF3) are much closer to the correlation line for unidentate X; NO3 acts as a unidentate ligand in this case while O2CCF3, although formally bidentate in both cases, shows a greater tendency towards unidentate co-ordination (i.e.a greater diVerence between the two Ag]O bond lengths) in the X = P case. As discussed previously for the [Ag{P(C6H11)3}2X] complexes, the E]Ag]E angle shows an overall decrease as the donor strength of X increases, but the behaviour of the halide complexes is anomalous in this respect, as this angle shows a small increase along the series X = Cl, Br, I, althoe donor strength of X2 is expected to increase along this series.The exact reason for this behaviour is not known at present, but we note that the same trend is observed in the present [Ag- {As(C6H11)3}2X] series. The present work has also defined the nature of the AgSCN:As(C6H11)3 (1 :2) complex. Whereas the P(C6H11)3 counterpart is relatively unexceptional, being mononuclear and of the monoclinic C2/c family,5 here we find a one-dimensional polymer, ? ? ? {As(C6H11)3}2Ag(SCN){As(C6H11)3}2Ag- (SCN) ? ? ?, with four-co-ordinate As2AgSN silver (Fig. 2); comparison is invited with the structure of AgCN:As(C6H11)3 (1 : 1) in the following paper in the context of potential or actual ambidentate co-ordination by these ligands.7 The complex crystallises in the rare triclinic space group P1 with two formula units comprising the asymmetric unit. The structural parameters that define the silver atom environments are given in Table 2. Whereas the distances within the two silver environments are very similar, there are significant and substantial differences in their angular geometries.The structure overall, however, is of higher pseudo-symmetry, the quasi-screw axis relating the two independent components of the asymmetric unit being clearly evident in Fig. 2. It is of interest that the As]Ag]As angles in the present complexes are not greatly Fig. 3 The d(Ag]E) vs. q(E]Ag]E) correlation for [Ag{E(C6H11)3}2X] [E = P (d) or As (j); data for E = P from ref. 5]. The straight lines are drawn in as ad hoc unidentate data diVerent to the counterpart angles in the mononuclear arrays, suggesting that the possibility of polymer or dimer formation in those arrays may not be energetically very diVerent from that in the monomer. Infrared spectroscopy The far-IR spectra of [Ag{As(C6H11)3}2X] (X = Cl, Br, I, NCO or CN) are shown in Figs. 4 and 5. Strong bands due to As(C6H11)3 ligands are evident down to about 280 cm21; ligand bands also occur in the region below this, but these are generally quite weak.Strong bands, the wavenumbers of which are dependent on the nature of X, are also observed in the spectra. Thus, for X = Cl, Br or I, strong, halogen-sensitive bands are observed at 230, 165 and 139 cm21 respectively (Fig. 4), and these are assigned as the n(AgX) modes for these complexes. The wavenumbers of these modes correlate well with those of other AgX complexes. This can be verified by comparison with data for the corresponding P(C6H11)3 complexes,5 as shown in Table 3, and for other AgX complexes with PPh3 and AsPh3 ligands.20 For a range of complexes of the latter type it has been shown that the dependence of n(AgX) on r(AgX) can be represented by the empirical formula (1) where b = 22 340, 22 490, n/cm21 = b(r/Å)2m (1) 41 300 and m = 5.09, 5.18, 5.68 for X = Cl, Br, I respectively.21 These correlations are shown as the continuous lines in Fig. 6, and the data for [Ag{E(C6H11)3}2X] (E = P or As; X = Cl, Br or Fig. 4 Far-IR spectra of [Ag{As(C6H11)3}2X]: X = Cl (a), Br (b) or I (c). The bands assigned to the n(AgX) modes are labelled with their wavenumbers2128 J. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 I) are represented as points on this graph. From this it is evident that the data for the E(C6H11)3 compounds lie very close to the correlation lines that were established for the EPh3 complexes, nearly all of which have structures that are diVerent from those of the E(C6H11)3 complexes.As is evident from the data in Fig. 6, the As(C6H11)3 com- Fig. 5 Far-IR spectra of [Ag{As(C6H11)3}2X]: X = NCO (a) or CN (b). The bands assigned to the n(AgX) modes are labelled with their wavenumbers Fig. 6 Plots of n(AgX) wavenumber against bond length r(AgX) from equation (1). Data points are for [Ag{E(C6H11)3}2X] [X = Cl (d), Br (j) or I (m); data for E = P from ref. 5 are the points with greater r(AgX) and smaller n(AgX) for each X] Table 3 Wavenumbers (cm21) of the n(AgX) modes in the IR spectra of [Ag{E(C6H11)3}2X] X Cl Br I NCO CN E = P* 207 149 121 241 288 E = As 230 165 139 262 311 * Ref. 5 pounds show shorter r(AgX) and greater n(AgX) than for the corresponding P(C6H11)3 complexes. A similar observation has been made recently in connection with the data for [Ag(EPh3)X] (E = P or As; X = Cl, Br or I), and has been attributed to weakened Ag]L bonding in going from L = PPh3 to AsPh3.22 Possible reasons for this have been given in terms of contributions from steric and electronic eVects.22 The data for [Ag{E(C6H11)3}2X] are consistent with those for [Ag(EPh3)3X], but do not seem to oVer any further means of distinguishing between the two possible causes of the observed trends.The assignment of the n(AgX) modes for [Ag{As(C6H11)3}2X] (X = NCO or CN) is complicated by the presence of ligand bands in the same region. However, a comparison of the spectra of these two compounds (Fig. 5) suggests n(AgX) 262 and 311 cm21 for X = NCO and CN respectively.Comparison of these results with those for the corresponding P(C6H11)3 complexes (Table 3) provides strong support for these assignments. In particular, there is an increase in n(AgX) of about 20 cm21 from the P(C6H11)3 to the corresponding As(C6H11)3 complex, as is also found for the X = Cl, Br or I compounds. The increase in n(AgX) for the X = NCO complex is somewhat puzzling, however, as the X-ray data suggest an increase in the Ag]X bond length in this case (Table 1), suggesting that the Ag]X bond is weaker in the As(C6H11)3 complex.This is not consistent with the IR result given above, and is contrary to the trend observed in all of the other complexes in Table 1. In this connection, it is relevant that the Ag]As/As]Ag]As data for this complex, as presented in Fig. 3, are also somewhat anomalous, the point for this compound lying well to the left of the correlation line for the other [Ag{As(C6H11)3}2X] compounds involving unidentate X.The reason for these anomalies is not known at present, although we note that the crystal structure determination for [Ag{As(C6H11)3}2(NCO)] presented some diYculties (see Experimental section). It would appear that a structure in which Ag]P is greater and Ag]X is smaller than the values given in Table 1 would be more consistent with the spectroscopic data for this compound, and the structural data for the compounds in Table 1.Acknowledgements We acknowledge support of this work by grants from the Australian Research Grants Scheme and the University of Auckland Research Committee. We thank Ms. Catherine Hobbis for recording the far-IR spectra. References 1 G. A. Bowmaker, EVendy, R. D. Hart, J. D. Kildea, E. N. de Silva, B. W. Skelton and A. H. White, Aust. J. Chem., 1997, 50, 539 and refs. therein. 2 EVendy, J. D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 587 and refs.therein. 3 G. A. Bowmaker, EVendy, E. N. de Silva and A. H. White, Aust. J. Chem., 1997, 50, 641 and refs. therein. 4 G. A. Bowmaker, EVendy, R. D. Hart, J. D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 651. 5 G. A. Bowmaker, EVendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2449. 6 G. A. Bowmaker, EVendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2459. 7 G. A. Bowmaker, EVendy, P. C. Junk and A. H. White, following paper. 8 S. R. Hall, H. D. Flack and J. M. Stewart (Editors), The Xtal 3.2 Reference Manual, Universities of Western Australia, Geneva and Maryland, 1992. 9 G. L. Soloveichik, O. Eisenstein, J. T. Poulton, W. E. Streib, J. C. HuVmann and K. G. Caulton, Inorg. Chem., 1992, 31, 3306. 10 J. Green, E. Sinn and S. Woodward, Inorg. Chim. Acta, 1995, 230, 231. 11 R. J. Restivo, A. Costin, G. Ferguson and A. J. Carty, Can. J. Chem., 1975, 53, 1949. 12 J. Green, E. Sinn, S. Woodward and R. Butcher, Polyhedron, 1993, 12, 991.J. Chem. Soc., Dalton Trans., 1998, Pages 2123–2129 2129 13 W. A. Anderson, A. J. Carty, G. J. Palenik and G. Schreiber, Can. J. Chem., 1971, 49, 761. 14 M. Camalli and F. Caruso, Inorg. Chim. Acta, 1988, 144, 205. 15 J. A. Muir, M. M. Muir, L. B. Pulgar, P. G. Jones and G. M. Sheldrick, Acta Crystallogr., Sect. C, 1985, 41, 1174. 16 J. A. Muir, M. M. Muir and E. Lorca, Acta Crystallogr., Sect. B, 1980, 36, 931. 17 M. K. Cooper, G. R. Dennis, K. Henrick and M. McPartlin, Inorg. Chim. Acta, 1980, 45, L151. 18 P. C. Healy, R. D. Hart, B. W. Skelton, A. H. White and G. A. Bowmaker, in preparation. 19 M. Barrow, H. B. Bürgi, M. Camalli, F. Caruso, E. Fischer, L. M. Venanzi and L. Zambonelli, Inorg. Chem., 1983, 22, 2356. 20 H. B. Bürgi, R. W. Kunz and P. S. Pregosin, Inorg. Chem., 1980, 19, 3707. 21 G. A. Bowmaker, EVendy, J. D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 577. 22 G. A. Bowmaker, R. D. Hart, E. N. de Silva, B. W. Skelton and A. H. White, Aust. J. Chem., 1997, 50, 553. Received 29th January 1998; Paper 8/00790J

ISSN:1477-9226    

DOI:10.1039/a800790j

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 2127–2132 2127 Crystalline TCNQ and TCNE adducts of the diborane(4) compounds B2(1,2-E2C6H4)2 (E 5 O or S) Todd B. Marder,a Nicholas C. Norman,*b A. Guy Orpen,*b Michael J. Quayle †b and Craig R. Rice b a The University of Durham, Department of Chemistry, Science Laboratories, South Road, Durham, UK DH1 3LE b The University of Bristol, School of Chemistry, Cantock’s Close, Bristol, UK BS8 1TS Received 11th February 1999, Accepted 12th May 1999 Crystal structures of 1 : 1 adducts of diborane(4) compounds and the electron acceptors TCNQ and TCNE, namely B2(1,2-E2C6H4)2?TCNQ (E = O or S) and B2(1,2-E2C6H4)2?TCNE (E = O or S), have been found to show predominantly two-dimensional heteromolecular packing motifs with a variety of interlayer packings.Charge transfer complexes of electron rich alkenes, such as tetrathiafulvalene [2-(1,3-dithiol-2-ylidene)-1,3-dithiole] (TTF) and tetramethyltetraselenafulvalene (TMTSF), and electron acceptors, such as TCNQ 1 (7,7,8,8-tetracyano-p-quinodimethane) and TCNE 2 (tetracyanoethene), constitute an important and much studied class of crystalline organic material with important properties including metallic conductivity and superconductivity.1 Another electron rich alkene to have been studied in this regard, although one which aVords materials with significantly less satisfactory electronic properties, 1 is dibenzotetrathiafulvalene 3, together with its selenium and tellurium analogues 4 and 5.The structure of the adduct 3?1 has been established by X-ray crystallography.2 In view of the close structural relationship between compound 3 and the diborane(4) compound B2(1,2-S2C6H4)2 6 (they diVer by only two electrons), and with the related oxo-derivative B2(1,2- O2C6H4)2 7,3 we sought to investigate whether 6 and 7 would form similar crystalline adducts with TCNQ and TCNE and to explore the nature of the intermolecular interactions. Results and discussion A solution of equimolar quantities of compounds 6 and 1 in CH2Cl2 showed no colour change (6 is colourless, 1 is pale yellow) and 1H and 11B NMR analysis provided no evidence for any adduct formation in solution.4,‡ However, cooling to 230 8C aVorded dark red crystals of the 1 : 1 adduct 6?1a.Red crystals of the 1 : 1 adducts 6?2, 7?1 and 7?2 were prepared similarly. In addition, by employing 1,2-dichloroethane as a solvent, a second polymorphic form of 6?1a (6?1b) was also crystallised.Full experimental details and analytical data are provided in the Experimental section. E B E E B E CN CN NC NC CN CN NC NC E C E E C E 3, E = S; 4, E = Se; 5, E = Te 6, E = S; 7, E = O 1 2 †Present address: Department of Chemical Engineering, UMIST, PO Box 88, Manchester, UK M60 1QD. ‡ The proton and boron solution chemical shifts of compounds 6 and 7 shift markedly upfield when the boron centres are complexed by nitrogen (and phosphorus) donor ligands as discussed in ref. 4. All five crystal structures are notably similar, having the triclinic space group P1� with Z = 1. The component molecules each lie over an inversion centre and the unit cell volumes are 576.3, 582.0, 516.8, 470.5 and 419.2 Å3 for 6?1a, 6?1b, 7?1, 6?2 and 7?2, respectively. The crystal structure of 3?1 which has been previously characterised 2 also crystallises in the space group P1� with Z = 1 and a unit cell volume of 574.9 Å3.Selected crystallographic details for all structures are given in Table 1. All six crystal structures (i.e. the five described here and 3?1) show a predominant two-dimensional packing in which the layers consist of the component species in a 1 : 1 ratio. In the third dimension, the crystal structures contain an ABCABC packing arrangement in which the layers are stacked parallel above each other along the crystallographic c axis direction. Despite the overall similarity of the structures, however, the two-dimensional arrangements of the component molecules do diVer between each structure, utilising a range of intermolecular interactions to pack eYciently.The two-dimensional crystal structures for 6?1a, 3?1, 7?1, 6?2 and 7?2 are shown in Figs. 1–5 respectively. A discussion for each is given first followed by a description of the three-dimensional packing. Simplified representations of the two-dimensional structures are shown in Fig. 6. Common to all six crystal structures are parallel homomolecular ribbons in the crystallographic a direction linked by heteromolecular interactions.Intermolecular interactions were recorded with cut-oV limits S ? ? ? S 3.80, S ? ? ? H 3.25 and N? ? ? H 2.95 Å. As shown in Fig. 7 three angles have been employed to describe the orientations of the molecules within the layers. The angle q describes misalignment of the molecular axes relative to each other, i.e. the direction of propagation in the heteromolecular ribbon.This has been calculated as the angle between the molecular axes of each species [through the B–B bond for 6 and 7, through the C]] C(CN2) unit for 1 and through the C]] C bond for 2]. The angles w and f (f9 is used for 2) describe the angle of propagation of the homomolecular ribbons. These are calculated as the angle between the molecular axis of one molecule and the line relating an atom in one molecule to the corresponding atom in the next molecule in the homomolecular ribbon, i.e.the a axis. Table 2 records these angles for each crystal structure. The two-dimensional structures of adducts 6?1a (Fig. 1) and 6?1b (not shown) are essentially the same. The homomolecular ribbons of 6 are apparently driven primarily by S ? ? ? S and S ? ? ? H contacts, there being two S ? ? ? S interactions and eight S ? ? ? H contacts per molecule including a three-centred interaction between S(2), S(1a) and H(6a). The symmetry independ-2128 J.Chem. Soc., Dalton Trans., 1999, 2127–2132 Table 1 Selected crystallographic details for the complexes 6?1a, 6?1b, 7?1, 6?2, 7?2 and 3?1 6?1a 6?1b 7?1 6?2 7?2 3?1a Empirical formula Formula weight Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 T/K Z m/mm21 Total reflections Independent reflections Rint R1 [I > 2s(I)] (data) C24H12B2N4S4 506.24 Triclinic P1� 7.803(3) 7.855(4) 9.867(4) 73.85(5) 85.71(3) 83.18(3) 576.3(4) 173(2) 1 0.434 4904 1963 0.0488 0.0470 (1324) C24H12B2N4S4 506.24 Triclinic P1� 7.862(2) 8.386(2) 9.745(2) 71.97(3) 72.71(2) 88.28(2) 582.0(2) 173(2) 1 0.430 6120 2637 0.0337 0.0365 (1867) C24H12B2N4O4 442.00 Triclinic P1� 7.1996(11) 8.560(3) 9.414(3) 67.00(2) 83.16(2) 75.47(2) 516.8(3) 173(2) 1 0.098 4293 1749 0.0427 0.0547 (1164) C18H8B2N4S4 430.14 Triclinic P1� 6.861(3) 7.172(3) 10.862(3) 80.80(2) 79.61(3) 64.00(2) 470.5(3) 173(2) 1 0.517 4889 2123 0.0398 0.0470 (1523) C18H8B2N4O4 365.90 Triclinic P1� 5.962(3) 6.024(4) 12.686(6) 97.88(2) 97.66(3) 108.91(3) 419.2(4) 173(2) 1 0.103 2677 1847 0.0416 0.0576 (1508) C26H12N4S4 508.65 Triclinic P1� 9.215(3) 10.644(4) 7.734(2) 113.32(3) 122.28(2) 67.64(3) 574.9(3) Room temperature 1 a Details taken from ref. 2. ent intermolecular interactions for all crystal structures are listed in Table 3 and are categorised according to the ribbon type they form. The angle of propagation w in the homomolecular direction of 6 in 6?1a is 54.98 [54.18 in 6?1b] and for 1 in 6?1a f is 59.58 [58.48 in 6?1b].These ribbons of 1 have eight C]] ] N? ? ? H contacts per molecule. The two-dimensional layers in 3?1 (Fig. 2) are similar to those in 6?1a and 6?1b, with homomolecular ribbons propagating at angles of w = 55.58 and f = 57.88. These ribbons also contain the same intermolecular distances, with a similar length for the chalcogen-containing molecular ribbons and slightly longer (by ca. 0.2 Å) lengths for the ribbons of 1. The structures of 6?1a, 6?1b and 3?1 show near linear heteromolecular ribbons [the molecular axes of 6 and 1 are canted at angles q of only 7.3, 4.3 and 9.28 in 6?1a, 6?1b and 3?1 respectively] which contain six C]] ] N? ? ? H (<2.95 Å) contacts per molecule of 1.Thus, it appears that the homomolecular interactions are at least numerically dominant in these crystaindicated by the greater number of interactions formed in the homomolecular directions. The layers are slightly ruZed with the angle between the mean Fig. 1 A view of the two-dimensional packing arrangement of molecules of 6 (red) and 1 (green) in adduct 6?1a showing the atom numbering scheme. Atoms are drawn as spheres of arbitrary radius. Table 2 Crystal packing parameters Adduct q/8 w/8 f/8 Interlayer separation/Å 6?1a 6?1b 3?1 7?1 6?2 7?2 7.3 4.3 9.2 52.3 8.9 56.9 54.9 54.1 55.5 56.6 77.2 54.4 59.5 58.4 57.8 271.1 265.6 a 68.6 a 3.59 3.43 3.44 3.44 3.60 3.39 a Values given are for f9 (see Fig. 7). planes through 6 and 1 being 11.9, 10.9 and 3.68 in 6?1a, 6?1b and 3?1 respectively. In adduct 7?1 (Fig. 3) the arrangement of molecules of 7 is similar to those of 6 in 6?1, with an angle of propagation w of 56.68 [cf. 54.98 in 6?1a]. Thus there are two O ? ? ? O and eight O? ? ? H–C contacts per molecule and the three-centred H ? ? ?E interaction is also present. Molecules of 1 propagate at similar magnitudes of f (271.18, cf. 59.58 in 6?1a), but rotated in the opposite sense.As a result there are only four short C]] ] N? ? ?H contacts per molecule rather than eight. The lateral displacement of molecules of 1 in the direction orthogonal to the homomolecular axis precludes the interaction N(2) ? ? ? H(11A) (present in 6?1a). The main diVerence in the two crystal structures lies in the angle of propagation along the heteromolecular chains, which are less linear in 7?1 (q = 52.38) [cf. 7.38 in 6?1a]. Thus, the heteromolecular ribbons are better termed zigzag, although as in 6?1a and 6?1b they extend along the crystallographic b axis and contain six C]] ] N? ? ? H–C contacts per Fig. 2 A view of the two-dimensional packing arrangement of molecules of 3 (red) and 1 (green) in adduct 3?1. Details as in Fig. 1. Fig. 3 A view of the two-dimensional packing arrangement of molecules of 7 (red) and 1 (green) in adduct 7?1. Details as in Fig. 1.J. Chem. Soc., Dalton Trans., 1999, 2127–2132 2129 Table 3 Selected intermolecular distances for adducts 6?1a, 6?1b, 3?1, 7?1, 6?2 and 7?2 6?1a 6?1b 3?1 Aa AA BB CCC AAA B CCCC S1 ? ? ? S1i S2 ? ? ? H6ii S1 ? ? ? H6ii N1 ? ? ? H11iii N1 ? ? ? H12iii N1 ? ? ? H4 N2 ? ? ? H5 N2 ? ? ? H3iv 7?1 O1 ? ? ? H3i O2 ? ? ? H3ii O2 ? ? ? O2ii N2 ? ? ? H12iii N1 ? ? ? H4iv N2 ? ? ? H5v N1 ? ? ? H6v 3.582(2) 3.177(4) 3.245(4) 2.653(4) 2.860(5) 2.906(5) 2.708(5) 2.679(5) 2.995(3) 3.047(4) 3.805(3) 2.572(3) 2.737(4) 2.684(4) 2.833(4) S1 ? ? ? S1i S2 ? ? ? H6ii S1 ? ? ? H6ii N1 ? ? ? H11iii N1 ? ? ? H12iii N1 ? ? ? H4 N2 ? ? ? H5 N2 ? ? ? H3iv 6?2 S1 ? ? ? S1i b S1 ? ? ? S2ii S2 ? ? ? H6iii CN? ? ?CNc N8 ? ? ? H4iv N8 ? ? ? H5v N9 ? ? ? H3vi N9 ? ? ? H5vi 3.613(1) 3.236(2) 3.193(2) 2.682(3) 2.731(3) 2.926(3) 2.751(3) 2.625(3) 3.578(2) 4.064(2) 3.366(3) 3.231 2.624(4) 2.831(4) 2.922(4) 2.658(4) S2 ? ? ? S2i S1 ? ? ? H3ii S2 ? ? ? H3i N1 ? ? ? H8iii N1 ? ? ? H9iii N1 ? ? ? H5iv N2 ? ? ? H4iv N2 ? ? ? H6v 7?2 O1 ? ? ? H3i O2 ? ? ? H3ii O1 ? ? ? O1i CN? ? ?CNc N8 ? ? ? H5iii N9 ? ? ? H4 N9 ? ? ? H6iv 3.716 2.904 3.277 2.913 2.874 2.997 2.592 2.770 2.685(3) 2.702(3) 3.491(4) 3.239 2.603(3) 2.787(3) 2.763(3) a Letter types refer to homomolecular interactions (A) [for 3, 6 and 7] and (B) [for 1 and 2] or heteromolecular interactions (C).Symmetry codes: 6?1a, i 1 2 x, 2 2 y, 2z, ii 21 2 x, y, z, iii 1 2 x, 21 2 y, 1 2 z, iv 1 1 x, y, z; 6?1b, i 2x, 2y, 2 2 z, ii 1 1 x, y, z, iii 1 2 x, 21 2 y, 1 2 z, iv 21 1 x, y, z; 3?1, i x, y, 21 1 z, ii 2x, 2y, 21 1 z, iii 1 2 x, 2y, 21 2 z, iv 1 2 x, 1 2 y, 2z, v 1 2 x, 1 2 y, 1 2 z; 7?1, i 21 2 x, y, z, ii 2x, 2 2 y, 1 2 z, iii 1 1 x, y, z, iv 2x, 2y, 1 2 z, v 1 2 x, 2y, 1 2 z; 6?2, i 2 2 x, 2y, 2z, ii 21 1 x, y, z, iii 21 1 x, 1 1 y, z, iv 1 2 x, 21 2 y, 1 2 z, v 2 2 x, 21 2 y, 1 2 z, vi x, 1 1 y, z, vii x, 21 1 y, z; 7?2, i 2x, 1 2 y, 1 2 z, ii 1 1 x, 1 1 y, z, iii 2 2 x, 2y, 2 2 z, iv 21 2 x, 21 1 y, z.b Interlayer contact. c Calculated as the distance between the midpoint of each C]] ] N group.molecule. The layers remain planar, however, as in 6?1a with the angle between the two mean planes of 7 and 1 being only 6.88. Homomolecular ribbons are also present in adduct 6?2 (Fig. 4) and the two-dimensional layer is similar to that observed in 6?1a, although the angle of propagation for molecules of 6, w, is somewhat more inclined (77.28, cf. 54.98 in 6?1a). This more canted structure for the ribbon results from the translation of molecules of 6 in the crystallographic b direction and gives rise to six S ? ? ? S and four S ? ? ? H–C interactions per molecule, as opposed to two S ? ? ? S and eight S ? ? ? H–C contacts per molecule in 6?1a.The three-centred interaction observed in the crystal structures containing 1 is absent. Whilst the relative strengths of these interactions are not known with certainty (all are near the sum of the van der Waals radii in length), it is noted, however, that the same number is present in each case.In 6?2 molecules of 2 propagate at an angle of f = 265.68 resulting in antiparallel C]] ]N? ? ?C]] ] N interactions, which are notably absent in the crystal structures of 6?1 and 7?1. In 6?2 q is 8.98 leading to a near linear heteromolecular ribbon, containing eight C]] ] N? ? ? H–C contacts per molecule of 2, more than are present in 6?1 and 7?1. In adduct 7?2 (Fig. 5) the homomolecular ribbons of 7 contain the same interactions as in 7?1, and the propagation angle w is 54.48 (cf. 56.68 in 7?1), although the shortest of the Fig. 4 A view of the two-dimensional packing arrangement of molecules of 6 (red) and 2 (green) in adduct 6?2. Details as in Fig. 1. interatomic distances are shorter (by ca. 0.3 Å, see Table 3). Molecules of 2 in 7?2 are arranged in a diVerent orientation from that in 6?2, rotated by ca. 1358 (f = 268.6 and 65.68 for 7?2 and 6?2 respectively). The ribbons of 2 show C]] ] N? ? ?C]] ] N interactions as is the case in 6?2.However, as in 7?1, the heteromolecular ribbons are not linear, the misalignment angle q being 56.98, cf. 52.38 in 7?1. The heteromolecular contacts Fig. 5 A view of the two-dimensional packing arrangement of molecules of 7 (red) and 2 (green) in adduct 7?2. Details as in Fig. 1. Fig. 6 Simplified representations of the two-dimensional structures 6?1a, 6?1b, 3?1, 7?1, 6?2 and 7?2. Molecules of 3, 6 and 7 in red, molecules of 1 and 2 in green.(6.1a), (6.1b) and (3.1) (7.1) (7.2) (6.2)2130 J. Chem. Soc., Dalton Trans., 1999, 2127–2132 include three short and one long C]] ] N? ? ? H–C interaction per molecule of 2. The layer packings in the structures of adducts 6?1a, 6?1b, 3?1, 7?1, 6?2 and 7?2 show considerable variation (see Figs. 8– 13), but all have an ABCABC type packing arrangement. As noted above, the two-dimensional structures in 6?1a, 3?1 and 6?1b are essentially identical. It is in the third dimension that these crystal structures diVer, accompanied by changes in the unit-cell volume [V = 576.3(4) and 582.0(2) Å3 for 6?1a and 6?1b, respectively].In 6?1a (Fig. 8) a degree of arene ring stacking is evident with the extremes of the arene groups of 6 being eclipsed in adjacent layers. Arene ring overlap for molecules of 1 seems unimportant, with antiparallel C]] ] N orientations prominent. In polymorph 6?1b (Fig. 9) molecules of 1 again have little arene ring overlap between layers. The most notable Fig. 7 Parameters w, q, f, f9 used to define molecular packing in the structures of adducts 6?1a, 6?1b, 3?1, 7?1, 6?2 and 7?2. B E E B E E B E E B E E NC NC CN CN NC NC CN CN w q j B E E B E E CN CN CN CN B E E B E E CN CN CN CN w j; j' = 90° - j q Fig. 8 A view of three of the two-dimensional layers in adduct 6?1a showing the three-dimensional structure (black above magenta above cyan). Fig. 9 A view of three of the two-dimensional layers in adduct 6?1b.Details as in Fig. 8. diVerence between the two polymorphs is that in 6?1b there is only slight overlap of the five-membered rings of 6 and none involving their six-membered rings. Since the layers in the polymorphs are similar, 6?1b can be described as being derived from 6?1a by the slipping of sheets. There is a striking diVerence, however, in the comparison of the three-dimensional structures of both polymorphs of 6?1 with that of 3?1. It is apparent that the heteromolecular ribbons in 3?1 (Fig. 10) align almost directly over each other, these stacks containing a multitude of heteromolecular p–p interactions with molecules of 1 sandwiched between the aromatic groups of molecules of 3 in adjacent layers. In the structures of adducts 7?1, 6?2 and 7?2 some degree of p stacking is apparent, but not as much as in 3?1. In 7?1 (Fig. 11) there are some heteromolecular p stacks formed through the partial eclipsing of 7 with 1 in adjacent layers. There is no significant overlap of molecules of 6.Heteromolecular stacking is more evident in 6?2 (Fig. 12) where molecules of 2 are sandwiched between molecules of 7. Finally, homomolecular interlayer packing is prominent in 7?2 (Fig. 13) with the aromatic rings of molecules of 7 overlapping. Having established that compounds 6 and 7 formed cocrystals with 1 and 2, as is the case for 3, it was of interest to determine the extent of any charge transfer. In 3?1 and other related crystal structures the degree of charge transfer, z, from the donor to the acceptor molecules, has been assessed by spectroscopic methods (IR, Raman and UV-vis) and by variations in molecular geometries.5 In this latter method the charge transferred from the HOMO of the donor molecule (i.e. 3 or Fig. 10 A view of three of the two-dimensional layers in adduct 3?1. Details as in Fig. 8. Fig. 11 A view of three of the two-dimensional layers in adduct 7?1. Details as in Fig. 8.J. Chem. Soc., Dalton Trans., 1999, 2127–2132 2131 analogue) to the LUMO of the acceptor molecule (i.e. 1 or analogue) can be estimated by contrasting the relevant bond lengths in the crystal structures of the pure acceptor (and donor) with those in the co-crystallised materials. Such contrasts, which are supported by experiment, recognise that on population of the LUMO of TCNQ (1) the structure becomes more benzenoid and less quinoid in character. Similar principles hold for structures of TCNE (2) and its adducts.The degree of charge transfer can be evaluated in terms of the ratios of bond distances. However, as shown in Tables 4 and 5, changes in the molecular structures of 1 and 2 from the native molecule to that in the crystalline adducts reported here are negligible; selected bond lengths for the molecular stuctures of 6 and 7 in the parent structures and in the adducts are noted in Fig. 12 A view of three of the two-dimensional layers in adduct 6?2. Details as in Fig. 8. Fig. 13 A view of three of the two-dimensional layers in adduct 7?2. Details as in Fig. 8. Table 4 Selected bond lengths (Å) for crystals containing compound 1 1 a 6?1a 6?1b 3?1 7?1 C]] C (ring) C–C (ring) C]] C C–C C]] ] N 1.346(5) 1.448(4) 1.374(5) 1.441(5) 1.140(5) 1.339(5) 1.448(5) b 1.383(5) 1.437(5) b 1.158(4) b 1.343(2) 1.442(3) b 1.375(3) 1.436(3) b 1.143(2) b 1.346(6) 1.442(6) 1.384(4) 1.438(4) b 1.138(5) b 1.337(3) 1.450(3) b 1.364(3) 1.441(3) b 1.150(3) b a See ref. 6. b Average of two values. Table 5 Bond lengths (Å) for crystals containing compound 2 2 a 2 b 6?2 7?2 C]] C C–C C]] ] N 1.344(3) 1.437(2) 1.135(2) 1.348(2) 1.434(2) c 1.146(2) c 1.356(6) 1.438(4) c 1.143(4) c 1.356(5) 1.439(3) c 1.139(3) c a See ref. 7; cubic form. b Ref. 8; monoclinic form. c Average of two values. Tables 6 and 7 and also show negligible changes. Thus, despite the red colour associated with the co-crystals of 6/7 with 1/2 described here, there is no structural evidence for any appreciable charge transfer although this is not unexpected due to the relatively poor electron donating ability of benzenoid derivatives such as 3 as noted in the Introduction; preliminary electrochemical studies on 6 and 7 9 together with PES, EHMO and UV-vis results 3a also confirm that they are poor electron donors.§ Nevertheless, the fact that the crystalline adducts described here do form indicates that boron analogues of TTF and TMTSF are worthy synthetic targets that are likely to aVord crystalline adducts with 1/2 where the degree of charge transfer (and hence conductivity) may be much greater.As a final aspect of this study, lattice energy calculations 10a,11 have been performed to evaluate and compare the relative stability of each crystal structure, particularly for the polymorphic pair 6?1a and 6?1b, and to assess the importance of the intermolecular interactions. The procedure followed the methods used by Buttar et al.12 The results are presented in Table 8 and suggest that the packing in polymorph 6?1a is 7.12 kcal mol21 more stable than that in 6?1b (as might be expected given its higher density and lower unit cell volume). Lattice energy calculations using the experimental structure have been asserted to reproduce experimental values successfully.10 The experimental lattice energy results for the polymorphs 6?1a and 6?1b appear to be consistent with current theories on diVerences in lattice energies between polymorphs being less than ca. 10% of the total lattice energy.10b,12 Since the two-dimensional layers of 6?1a, 6?1b and 3?1 are essentially indentical, one might expect that the relative diVerences in lattice energies be associated with layer stacking and particularly to arise from favourable p–p interactions. If this were the case, 3?1 would be the most stable structure type since, as noted above, the ribbons align above each other in a more organised fashion than in 6?1a and 6?1b. However, the lattice energy of 3?1 is less than that of either 6?1a or 6?1b suggesting that arene stacking is of little Table 6 Bond lengths (Å) for crystals containing compound 6 6 a 6?1a 6?1b 6?2 B–B B–S S–C 1.675(5) 1.792(2) 1.756(2) 1.681(8) 1.791(4) b 1.754(4) b 1.676(4) 1.791(3) b 1.758(2) b 1.682(6) 1.789(3) b 1.755(3) b a See ref. 3(b). b Average of two values. Table 7 Bond lengths (Å) for crystals containing compound 7 7 a 7?1 7?2 B–B B–O O–C 1.678(5) 1.388(2) 1.387(2) 1.677(6) 1.388(3) b 1.396(3) b 1.684(5) 1.389(3) b 1.384(3) b a See ref. 3(b). b Average of two values. Table 8 Lattice energies (kcal mol21) for experimental structures Adduct Lattice energy 6?1a 6?1b 3?1 7?1 6?2 7?2 273.92 266.80 263.96 270.43 263.31 250.95 § Measurement of the electrical conductivity of the crystalline adducts reported here has been hampered by the small size of the crystals.2132 J. Chem. Soc., Dalton Trans., 1999, 2127–2132 importance here (at least insofar as the force fields are able to reproduce such interactions).The apparent thermodynamic preference for the formation of co-crystals in all these cases is striking and presumably favoured by formation of heteromolecular contacts. There may, however, be (kinetic) crystallisation eVects at work, since the co-crystals are presumably less soluble than the individual component species. In view of the dominance of the two-dimensional layers and the results of the lattice energy calculations, it becomes apparent that interlayer p stacking is not important in the formation of these co-crystals. Indeed, as evidenced by the existence of polymorphs 6?1a and 6?1b the layers can slip at very little energy penalty.Attempted structure optimisations led to significant layer slippings, without aVecting the sequence of lattice energies, and with very small eVects on the intralayer structures. This may be taken as support for the empirical observations above, that it is the intralayer structure that is the robust motif in these structures.Experimental General procedures All reactions were performed using standard Schlenk techniques under an atmosphere of dry, oxygen-free dinitrogen. All solvents were distilled from appropriate drying agents immediately prior to use (sodium for Et2O and hexanes and CaH2 or 3 Å molecular sieves for chlorocarbons). Microanalytical data were obtained at the University of Bristol. Compounds 6 and 7 were prepared by the literature methods,3a 1 and 2 were procured commercially.Preparations In a typical preparation, a pale yellow solution of compound 2 (0.027 g; 0.21 mmol) in CH2Cl2 (2 cm3) was added to a colourless solution of 6 (0.063 g; 0.21 mmol) in CH2Cl2 (2 cm3) resulting in no noticeable colour change. This reaction solution was then cooled to 2308 C and maintained at this temperature for 24 h. After this time a crop of small dark red crystals had formed.The remaining solution was then removed by syringe and the resulting crystals of 6?2 washed with Et2O (1 cm3) and hexane (2 × 2 cm3) and dried under vacuum (0.051 g, 56%). One of these was used for X-ray crystallography (C9H4BN2S2 requires C, 50.3; H, 1.9; N, 13.0. Found: C, 50.4; H, 1.8; N, 13.3%). Crystals of 6?2 with the same unit cell dimensions were also obtained from chlorobenzene and 1,2-dichloroethane. All other compounds were prepared similarly. Crystals of adduct 6?1a were obtained from CH2Cl2 solution (60%) (C12H6- BN2S2 requires C, 56.9; H, 2.4; N, 11.1.Found: C, 58.7; H, 2.5; N, 13.0%), of 6?1b from 1,2-dichloroethane solution (54%) (Found: C, 57.8; H, 2.7; N, 12.5%), of 7?1 from CH2Cl2 solution (53%) (C12H6BN2O2 requires C, 65.2; H, 2.7; N, 12.7. Found: C, 64.1; H, 2.9; N, 9.6%) and of 7?2 from CH2Cl2 solution (39%) (C9H4BN2O2 requires C, 59.1; H, 2.2; N, 15.3. Found: C, 57.0; H, 2.4; N, 14.3%). Crystals of 7?2 with the same unit cell as those grown from CH2Cl2 were also obtained from 1,2-dichloroethane. X-Ray crystallography Many of the details of the structure analyses are listed in Table 1.X-Ray diVraction measurements on single crystals coated in a hydrocarbon oil mounted on a glass fibre under argon were made with graphite-monochromated Mo-Ka X-radiation (l – = 0.71073 Å) using a Siemens SMART area diVractometer. CCDC reference number 186/1460. See http://www.rsc.org/suppdata/dt/1999/2127/ for crystallographic files in .cif format.Lattice energy calculations Atomic point charges for each component molecule were assigned individually using GAUSSIAN13 at the 6-31G level. Lattice energy calculations were performed on the experimental geometry using the Crystal Packer module in Cerius 2 14 and with unrestricted geometry optimisation with the Dreiding 15 force field. Acknowledgements T. B. M. thanks Natural Sciences and Engineering Research Council of Canada (NSERC), N. C.N. thanks the EPSRC, Laporte plc and The Royal Society and A. G. O. thanks EPSRC for research support and for a studentship (to M. J. Q.). T. B. M. and N. C. N. also thank NSERC and The Royal Society for supporting this collaboration via a series of Bilateral Exchange Awards. References 1 M. R. Bryce and L. C. Murphy, Nature (London), 1984, 309, 119; - J. M. Williams, M. A. Beno, H. H. Wang, P. C. W. Leung, T. J. Emge, U. Geiser and K. D. Carlson, Acc. Chem. Res., 1985, 18, 261. 2 T. J. Emge, F. Mitchell Wiygul, J. S. Chappell, A. N. Bloch, J. P. Ferraris, D. O. Cowan and T. J. Kistenmacher, Mol. Cryst. Liq. Cryst., 1982, 87, 137. 3 (a) F. J. Lawlor, N. C. Norman, N. L. Pickett, E. G. Robins, P. Nguyen, G. Lesley, T. B. Marder, J. A. Ashmore and J. C. Green, Inorg. Chem., 1998, 37, 5282; (b) W. Clegg, M. R. J. Elsegood, F. J. Lawlor, N. C. Norman, N. L. Pickett, E. G. Robins, A. J. Scott, P. Nguyen, N. J. Taylor and T. B. Marder, Inorg. Chem., 1998, 37, 5289; (c) G. Lesley, T. B. Marder, N. C. Norman and C. R. Rice, Main Group Chem. News, 1997, 5, 4. 4 W. Clegg, C. Dai, F. J. Lawlor, T. B. Marder, P. Nguyen, N. C. Norman, N. L. Pickett, W. P. Power and A. J. Scott, J. Chem. Soc., Dalton Trans., 1997, 839. 5 M. Decoster, F. Lonan, J. E. Guerchais, Y. Le Mist, J. Sala Pala, J. C. JeVery, E. Faulgues, A. Leblon and P. Molinine, Polyhedron, 1995, 14, 1741. 6 R. E. Long, R. A. Sparks and K. N. Trueblood, Acta Crystallogr., 1965, 18, 932. 7 R. E. Little, D. Pautler and P. Coppens, Acta Crystallogr., Sect. B, 1971, 27, 1493. 8 U. Druck and H. Guth, Z. Kristallogr., 1982, 161, 103. 9 J. M. Williams, personal communication; G. R. Whittell, unpublished work. 10 (a) R. Docherty and W. Jones, Organic Molecular Solids, Properties and Applications, ed. W. Jones, CRC Press, New York, 1997; (b) A. Gavezotti and G. Filippini, J. Am Chem. Soc., 1995, 117, 12299; - (c) M. H. Charlton, R. Docherty and M. G. Hutchings, J. Chem. Soc., Perkin Trans. 2, 1995, 2203. 11 A. Gavezotti, J. Am Chem. Soc., 1991, 113, 4622. 12 D. Buttar, M. H. Charlton, R. Docherty and J. Starbuck, J. Chem. Soc., Perkin Trans. 2, 1998, 763. 13 GaussView v. 1.01, Gaussian Inc., Semichem, Pittsburgh, 1997. 14 Cerius 2 v. 3.5, Molecular Simulations Inc., Cambridge, 1997. 15 S. L. Mayo, B. D. Olafson and W. A. Goddard (III), J. Phys. Chem., 1990, 94, 8897. Paper 9/01169B

ISSN:1477-9226    

DOI:10.1039/a901169b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 2131 Spectroscopic and structural studies on 1 : 1 adducts of silver(I) salts with tricyclohexylarsine Graham A. Bowmaker,a EVendy,b,c Peter C. Junkb and Allan H. White b a Department of Chemistry, University of Auckland, Private Bag 92019, Auckland, New Zealand b Department of Chemistry, The University of Western Australia, Nedlands, W.A. 6907, Australia c Jurusan Pendidikan Kimia, FPMIPA IKIP, Malang, Jalan Surabaya 6, Malang, 65145, Indonesia Adducts of a number of silver(I) salts, AgX, with tricyclohexylarsine, of 1 : 1 AgX:As(C6H11)3 stoichiometry have been synthesized for X = Cl, Br, I, NO3, NCO or CN and subjected to room temperature single crystal X-ray determinations and studies of their low frequency vibrational spectra.The chloride is binuclear [{(C6H11)3As}Ag(m-Cl)2Ag{As(C6H11)3}], isomorphous with its previously recorded P(C6H11)3, X = Cl, Br counterparts, the silver environment being quasi-planar, trigonal AsAg(m-Cl)2; the nitrate is isomorphous, with a central planar Ag(m-O)2Ag array, a pair of oxygen atoms, one from each of a pair of symmetry-related nitrate groups, bridging the two silver atoms.The present bromide, unlike its P(C6H11)3 counterpart, is not a dimer, but, like the iodide, a ‘cubane’ tetramer; the iodide is isomorphous with its P(C6H11)3 analogue, a crystallographic 2 axis passing through a pair of opposed faces of the tetramer, but the bromide is of a new type, rhombohedral space group R3� .Remarkably, the cyanate is also of the cubane form, the first recorded (other than an organometallic) incorporating a first-row atom, obtained unsolvated and bis(pyridine) solvated. The cyanide is a linear polymer . . . {(C6H11)3As}2AgNCAgCN{As(C6H11)3}2AgNC . . . (‘a’ phase, from 2-methylpyridine); a second ‘b’ phase was obtained from 2,4,6-trimethylpyridine, of similar form, while from pyridine, a solvate of 3:4:2 AgCN:As(C6H11)3 :py stoichiometry [{Ag[As(C6H11)3]2(py)}2(CN)][Ag(CN)2] was obtained.The structure of the E = Px = CN unsolvated 1 : 1 analogue was also determined, also being a linear polymeric array like its E = As counterpart. The far-IR spectra of the halide complexes exhibit bands due to n(AgX) vibrational modes at 229, 148 cm21 (X = Cl), 167, 151, 125, 109 cm21 (X = Br) and at 111, 86 cm21 (X = I). These spectra were interpreted in terms of idealised C2h Ag2Cl2 and Td Ag4X4 structures of the silver halide cores.In the preceding paper 1 we have recorded the results of synthetic, structural (room-temperature, single crystal X-ray) and low-frequency vibrational spectroscopic studies on adducts of silver(I) salts, AgX, with tricyclohexylarsine, of 1 : 2 stoichiometry, accessing and defining an extensive novel, mononuclear array of As2AgX environments, paralleling those found similarly with tricyclohexylphosphine.2 With P(C6H11)3 a family of 1 : 1 adducts has also been defined, the chloride and bromide being binuclear and the iodide tetranuclear,3 contrasting with the linear two-co-ordinate array found in AuCl:P(C6H11)3 (1 : 1),4 paralleled in many more AuX:P(C6H11)3 (1 : 1) species with more complex X.In the present report we describe the results of a counterpart study of the parallel 1 : 1 array, specifically for X = Cl, Br, I, NO3 or NCO; while, unsurprisingly, the halides and nitrate are Ag(m-X)2Ag or Ag4X4 binuclear or tetranuclear ‘cubane’ arrays, the cyanate, remarkably, is also a cubane tetramer, obtained in two forms, as is the cyanide, a linear polymer.We have also obtained the 1 : 1 P(C6H11)3:CN analogue of the arsenic complex described above; as with many of the cyanide arrays described here and in other studies,5 the joint objectives of obtaining nicely crystalline material suitable for X-ray work and substantial pure homogeneous bulk sample for elemental analysis and spectroscopy become increasingly diYcult of simultaneous achievement, some complexes being loosely solvated while others appear to provide crystals in an ill defined matrix of, presumably, various polymer forms, which the cyanides are prone to form, and where characterisation depends heavily on the X-ray work.Caveats concerning the latter are described at length in ref. 1. Experimental Syntheses All compounds were obtained by the dissolution of the appropriate silver(I) salt (1 mmol) with tricyclohexylarsine (1 mmol) in pyridine (5 cm3) [exceptions: the nitrate (methanol) and the cyanide (2-methyl- or 2,4,6-trimethyl-pyridine)] with warming, followed by cooling and standing, whereupon colourless crystals deposited.X = Cl: m.p. 208–211 8C (Found: C, 46.3; H, 7.10. Calc. for C36H66Ag2As2Cl2: C, 46.23; H, 7.11%). X = Br: m.p. >203 8C (Found: C, 42.4; H, 6.3. Calc. for C72H132Ag4As4Br4: C, 42.21; H, 6.49%). X = I: m.p. >199 8C (decomp.) (Found: C, 38.8; H, 6.1. Calc.for C72H132Ag4As4I4: C, 38.67; H, 5.95%). X = NO3: m.p. > 112 8C (Found: C, 43.6; H, 6.9; N, 3.0. Calc. for C36H66Ag2As2N2O6 :C, 43.74; H, 6.73; N, 2.83%). X = NCO: m.p. >211 8C (decomp,), no satisfactory analysis obtained. X = CN (Found: C, 50.1; H, 7.2; N, 3.1. Calc. for C38H66Ag2As2N2: C, 49.8; H, 7.26; N, 3.06%). The 1 : 1 AgCN:P(C6H11)3 adduct was obtained similarly from pyridine, m.p. 136–138 8C (Found: C, 54.7; H, 7.7; N, 4.5. Calc. for C19H33AgNP: C, 55.08; H, 8.03; N, 3.38%).Structure determinations The general procedure is given in the preceding paper;1 specific details are as follows. (a) Triclinic, space group P1� (Ci 1, no. 2), Z = 1 dimer. (i) X 5 Cl. C36H66Ag2As2Cl2, M = 935.4, a = 8.827(2), b = 9.451(2), c = 13.338(2) Å, a = 99.41(3), b = 93.66(2), g = 114.96(2)8, U = 984.1 Å3, Dc = 1.578 g cm23, F(000) = 476, mMo = 28.2 cm21, specimen 0.13 × 0.30 × 0.20 mm; A*min,max = 1.53, 1.90, 2qmax = 558, N = 4519, No = 3121, R = 0.036, R9 = 0.039.(ii) X 5 NO3. C36H66Ag2As2N2O6, M = 988.5, a = 9.244(3), b = 9.628(3), c = 13.343(2) Å, a = 96.77(2), b = 90.72(2), g = 117.32(3)8, U = 1044.8 Å3, Dc = 1.571 g cm23, F(000) = 504, mMo = 25.5 cm21, specimen 0.42 × 0.26 × 0.18 mm, A*min,max = 1.51, 1.89, 2qmax = 508, N = 3669, No = 2809, R = 0.040, R9 = 0.047. Variata. The cell and coordinate setting follows that previ-2132 J. Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 ously defined for the isomorphous chloride P(C6H11)3 counterpart. 3 For the chloride, (x, y, z, Uiso)H were refined.In the nitrate high ‘thermal motion’ on two of the rings suggests unresolved disorder. (b) X 5 Br. C72H132Ag4As4Br4, M = 2048.7, rhombohedral, space group R3� (C2 3i, no. 148), a = 22.482(5), c = 28.35(2) Å, U = 12 413 Å3 (hexagonal setting), Dc (Z = 6 tetramers) = 1.645 g cm23, F(000) = 6144, mMo = 44.9 cm21, specimen 0.12 × 0.15 × 0.21 mm, A*min,max = 1.75, 3.73 (analytical correction), 2qmax = 508, N = 4857, No = 1471, R = 0.057, R9 = 0.049.Variata. The three rings of ligand 1 were modelled as disordered over two sites, occupancy 0.5, with isotropic thermal parameter forms. (c) X 5 I. C72H132Ag4As4I4, M = 2236.7, monoclinic, space group C2/c (C6 2h, no. 15), a = 15.660(7), b = 29.284(9), c = 19.312(6) Å, b = 108.67(3)8, U = 8390 Å3, Dc (Z= 4 tetramers) = 1.770 g cm23, F(000) = 4384, mMo = 39.9 cm21, specimen 0.28 × 0.56 × 0.38 mm, A*min,max = 2.60, 3.45 (analytical correction), 2qmax = 508, N = 7378, No = 3736, R = 0.046, R9 = 0.045.Variata. The compound is isomorphous with its P(C6H11)3 counterpart 3 and was refined in that cell and coordinate setting. ‘High thermal motion’ appears a foil for widespread disorder, resolved in ring 22. (d) X 5 NCO. This compound was obtained, unsolvated from MeCN, and as a bis(pyridine) solvate from pyridine, the structures of both having been determined. The latter, being the more precise, is recorded in detail here, further details for the unsolvated form being deposited. (i) Unsolvated.C76H132Ag4As4N4O4, M = 1897.1, rhombohedral, space group R3 (C3 4, no. 146), a = 24.181(4), c = 12.615(2) Å, U = 6379 Å3 (hexagonal setting), Dc (Z = 3 tetramers) = 1.479 g cm23, F(000)4, mMo = 24.9 cm21, specimen 0.32 × 0.12 × 0.41 mm, A*min,max = 1.35, 1.74, 2qmax = 558; N = 3254, No = 1352; R = 0.058, R9 = 0.058 (preferred hand). (ii) Bis(pyridine) solvate.C76H132Ag4As4N4O4?2C5H5N, M = 2055.3, monoclinic, space group C2/c (C6 2h, no. 15), a = 27.536(9), b = 14.667(7), c = 24.008(9) Å, b = 106.51(3)8, U = 9297 Å3, Dc (Z= 4 tetramers) = 1.468 g cm23, F(000) = 4208, mMo = 22.9 cm21, specimen 0.44 × 0.07 × 0.32 mm, A*min,max = 1,18, 2.31, 2qmax = 508, N = 7841, No = 2506, R = 0.050, R9 = 0.042. Variata. For the unsolvated specimen, data although extensive were weak, and, in the context of a non-centrosymmetric space group and disorder in the axial ligand substituents, as well as the axial cyanate which inclines to the axis from its point of attachment, would support meaningful anisotropic thermal parameter refinement for Ag, As only, with some further constraints in ligand and anion geometries.High ‘thermal motion’ was also evident in certain of the rings of the solvated form, modelled as such in the context of the inability meaningfully to deconvolute any disordered components. For both forms, the assignment of anion atoms was problematic, favouring N coordination, but with some dubiousness occasioned by occurrences such as a need to refine the central carbon of anion 2 in the solvated form with an isotropic thermal parameter and in the unsolvated form to constrain components of the anion geometry.(e) X 5 CN. C38H66Ag2As2N2, M = 916.6. a Phase. Monoclinic, space group P21/n (C 5 2h, no. 14, variant), a = 10.140(4), b = 25.542(9), c = 15.520(4) Å, b = 90.63(3)8, U = 4020 Å3, Dc (Z = 4) = 1.514 g cm23, F(000) = 1872, mMo = 26.3 cm21, specimen 0.27 × 0.08 × 0.05 mm, A*min,max = 1.12, 1.26, 2qmax = 488, N = 6309, No = 2858; R = 0.058, R9 = 0.058.Variata. Assignment of the cyanide C, N components, made on the basis of refinement behaviour and consistent with the usual NCAgCN component norm, should, none the less, be taken with a grain of salt in consequence of a rather weak body of data, consistent with small crystal size. b Phase. Monoclinic, space group P21/c (C5 2h, no. 14), a = 9.815(3), b = 17.829(7), c = 24.216(8) Å, b = 103.63(3)8, U = 4118 Å3, Dc (Z = 4) = 1.478 g cm23, F(000) = 1872, mMo = 25.7 cm21, specimen 0.50 × 0.28 × 0.13 mm, A*min,max = 1.34, 1.79, 2qmax = 508, N = 7248, No = 3704, R = 0.053, R9 = 0.051. Variata. This material was obtained by crystallisation from 2,4,6-trimethylpyrdine. A similar caveat to the above applies in respect of cyanide C, N assignment. High thermal motion on rings 21, 23 is probably a foil for unresolved disorder.( f ) [{Ag[As(C6H11)3]2(py)}2(CN)][Ag(CN)2]. Crystallization of the cyanide from pyridine yielded small crystals of an adduct of 3 : 4 : 2 AgCN:As(C6H11)3 :py stoichiometry (Found: C, 55.2; H, 7.6; N, 3.5. C85H142Ag3As4N5 requires C, 54.97; H, 7.71; N, 3.77%), also the subject of a rather imprecise albeit useful determination: C85H142Ag3As4N5, M = 1857.4, monoclinic, space group P21 (C2 2, no. 4), a = 14.887(6), b = 14.429(3), c = 20.636(7) Å, b = 97.79(3)8, U = 4392 Å3, Dc (Z = 2) = 1.404 g cm23, specimen 0.21 × 0.33 × 0.12 mm, A*min,max 1.24, 1.64, 2qmax = 508, N = 8040, No = 2839, R = 0.075, R9 = 0.070 (preferred hand).Variata. The cyanide C, N assignment caveat applies, atoms in the cation being modelled as composites and in the anion as C-bound. Data were weak and limited in scope, supporting, in the context of a non-centrosymmetric space group, meaningful anisotropic thermal parameter refinement for Ag, As only.( g) X 5 CN, E 5 P. This is similar in unit cell to the b form of the arsine derivative (e) above but is triclinic: C38H66Ag2- N2P2, M = 828.7, triclinic, space group P1� (Ci 1, no. 2), a = 23.063(9), b = 17.557(6), c = 10.17(1) Å, a = 90.02(5), b = 94.88(6), g = 101.06(3)8, U = 4027 Å3, Dc (Z = 4) = 1.367 g cm23, F(000) = 1728, mMo = 9.8 cm21, specimen 0.30 × 0.14 × 0.52 mm, A*min,max = 1.14, 1.33, 2qmax = 508, N = 12 235, No = 6647; R = 0.069, R9 = 0.073. Variata.Cyanide groups within the NCAgCN array were modelled as C-bound, albeit undefinitively. CCDC reference number 186/984. See http://www.rsc.org/suppdata/dt/1998/2131/ for crystallographic files in .cif format. Spectroscopy Far-infrared spectra were recorded at 4 cm21 resolution at room temperature as Polythene discs on a Digilab FTS-60 Fourier transform infrared spectrometer employing an FTS- 60V vacuum optical bench with a 6.25 mm mylar film beam splitter, a mercury lamp source and a pyroelectric triglycine sulfate detector. Discussion Crystal structures Crystallization of a variety of silver(I) salts, AgX (X = Cl, Br, I or NO3), with tricyclohexylarsine in 1 : 1 stoichiometric ratio from acetonitrile results in the production of a number of colourless crystalline complexes; the results of roomtemperature single crystal X-ray determinations, coupled with elemental analysis, are consistent with the stoichiometry and connectivity implied in the AgX :As(C6H11)3 (1 : 1) formulation. The structures of a selection of the compounds studied are shown in Figs. 1–4. The chloride is isomorphous with its P(C6H11)3 counterpart, [and the AgBr :P(C6H11)3 (1 : 1) adduct] crystallizing in a triclinic P1� cell which contains one centrosymmetric dimer of the form [{(C6H11)3E}Ag(m-X)2Ag{E(C6- H11)3}], one half of which comprises the asymmetric unit of the structure. Geometries of the systems are compared in Table 1, together with that of a counterpart nitrate adduct, also iso-J.Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 2133 morphous, and isolated with the present array; there appear to be no other structurally characterised oxyanion counterparts hitherto described for any AgX:P(C6H11)3 (1 : 1) system. By contrast, AgNO3 :EPh3 (1 : 1) (E = P,6 As 7 or Sb 8) have all been defined as one-dimensional polymers, rather than the present dimeric type. The form of the present nitrate is of interest, the Fig. 1 Projection of the binuclear nitrate, normal to the Ag(m-O)2Ag plane; 20% thermal envelopes are shown for the non-hydrogen atoms, hydrogen atoms having arbitrary radii of 0.1 Å Table 1 Dimer core geometries in [{[(C6H11)3E]AgX}2].Distances in Å, angles in 8; primed atoms are centrosymmetrically related E/X Ag]X Ag]X9 Ag]E Ag ? ? ? Ag9 X? ? ?X Ag]X]Ag9 X]Ag]X9 E]Ag]X E]Ag]X9 P/Cl 2.640(3) 2.454(2) 2.354(2) 3.488(2) 3.718(3) 86.33(6) 93.67(7) 118.85(6) 147.48(6) As/Cl 2.643(2) 2.456(1) 2.4519(9) 3.3809(9) 3.822(2) 82.97(4) 97.03(5) 115.76(4) 147.21(4) As/ONO2 2.428(6) 2.269(5) 2.4198(8) 3.807(1) 2.757(7) 108.2(2) 71.8(2) 130.0(1) 158.3(1) In the nitrate Ag ? ? ? O(29) is 2.792(5) Å; N]O(1,2,3) are 1.245(6), 1.197(8), 1.225(9) Å, with angles opposed being 121.3(5), 121.8(6), 116.9(5)8.anion bridging by way of a single oxygen atom, rather than various alternatives. The silver–silver distance is much longer than in the chloride, but the two Ag ? ? ? O distances may be regarded as unsymmetrical here also, as also are the As]Ag]O angles.The nitrate itself is also unsymmetrically disposed vis-àvis the two silver atoms, so that O(2), O(3) ? ? ? Ag, Ag9 distances are unsymmetrical, with O(2) displaying a semichelate interaction. The nitrate plane is tilted relative to the Ag(m-O)2Ag plane, the dihedral angle being 44.3(3)8. The iodide is also isomorphous with its P(C6H11)3 counterpart, being a tetramer of the ‘cubane’ type, lying disposed about a crystallographic 2 axis in monoclinic space group C2/c (the axis passing through the midpoints of an opposed pair of Ag2I2 faces); geometries are compared in Table 2.The bromide, unlike its P(C6H11)3 analogue which is a dimer, is also a tetramer of the cubane form, but not isomorphous with the iodide, crystallising in rhombohedral space group R3� , a crystallographic 3 axis passing through bromide, silver and arsenic sequence disposed along a cube body diagonal; the core geometry is given in Table 3.The crystal packing is of some interest here, the molecules being disposed in normal to the 3 axis, the axial ligands projecting and plugging into cavities in the adjacent layers (Fig. 3). The cyanate, also a tetramer, obtained in both unsolvated and bis(pyridine) solvated forms is remarkable; although cubane types are well known among coinage metal(I) salt : Group V unidentate base adducts of 1 : 1 stoichiometry, their occurrence is hitherto restricted to halides X = Cl, Br or I, with the ‘step’ form as a less frequent alternative.The present is the first recorded occurrence of a pseudohalogen example, the anion participating by co-ordination of one of its termini within the cube, the remaining linearly disposed atoms projected outwards along the putative 3 axis. Which end of the anion coordinates is not definitively established by the X-ray work here, crystals of both forms being deficient in respect of the precision achievable within the determination; for the present, in lieu of a better experiment, we accept indication on the basis of the refinement that it is the nitrogen atom, as might be expected on the basis of previous studies.2,9–11 The present complex is the first example of an L4M4X4 ‘cubane’ structure in which X = pseudohalide.Other examples of such structures in which the anion component X is a first row atom donor ligand are known, including organometallic compounds (e.g.[Cu4{P(C6- H4Me-p)3}4(m3-h1-C]] ] CPh)4]),12 and further examples may prove to be achievable (there being no recorded fluoride examples) with species such as azide. The ‘cube’ in the present form is the Table 2 Tetramer core geometries in cubane [{[(C6H11)3E]AgX}4]. Distances in Å, angles in 8; primed atoms are related by the internal 2 axis. The two values in each entry are for E = P, As respectively for X = I, followed by X = NCO Ag(1)]X(1) Ag(1)]X(2) Ag(1)]X(19) Ag(1)]E(1) Ag(1) ? ? ? Ag(19) Ag(1) ? ? ? Ag(29) Ag(1) ? ? ? Ag(2) Ag(2) ? ? ? Ag(29) Ag(1) ? ? ? X(29) X(1)]Ag(1)]X(2) X(1)]Ag(1)]X(19) X(2)]Ag(1)]X(19) E(1)]Ag(1)]X(1) E(1)]Ag(1)]X(2) E(1)]Ag(1)]X(19) Ag(1)]X(1)]Ag(2) Ag(1)]X(1)]Ag(19) Ag(2)]X(1)]Ag(19) 2.889(2), 2.897(2); 2.46(1) 2.902(2), 2.900(2); 2.36(1) 2.975(2), 2.933(1); 2.50(1) 2.470(4), 2.560(2); 2.461(2) 3.555(2), 3.341(2); 3.554(2) 3.749(2), 3.551(2); 3.510(2) 3.356(2), 3.203(2); 3.360(2) 3.818(2), 3.589(2); 3.465(2) 4.879(2), 4.820(2); 4.22(1) 112.38(5), 114.82(4); 91.1(4) 103.41(6), 107.43(5); 88.4(4) 91.95(4), 94.69(3); 87.8(5) 110.4(1), 107.25(5); 124.8(3) 124.4(1), 121.38(6); 133.3(3) 110.3(1), 109.83(5); 118.6(3) 71.60(5), 67.88(4); 87.6(4) 74.63(5), 69.94(4); 91.6(4) 80.12(4), 75.91(3); 91.8(4) Ag(2)]X(1) Ag(2)]X(2) Ag(2)]X(29) Ag(2)]E(2) X(1) ? ? ? X(19) X(1) ? ? ? X(29) X(1) ? ? ? X(2) X(2) ? ? ? X(29) Ag(2) ? ? ? X(19) X(1)]Ag(2)]X(2) X(1)]Ag(2)]X(29) X(2)]Ag(2)]X(29) E(2)]Ag(2)]X(1) E(2)]Ag(2)]X(2) E(2)]Ag(2)]X(29) Ag(1)]X(2)]Ag(2) Ag(1)]X(2)]Ag(29) Ag(2)]X(2)]Ag(29) 2.848(2), 2.839(1); 2.39(1) 3.095(2), 3.050(2); 2.41(1) 2.861(2), 2.851(2); 2.48(2) 2.435(2), 2.536(1); 2.467(2) 4.602(2), 4.699(2); 3.46(2) 4.226(1), 4.290(2); 3.36(2) 4.811(2), 4.884(2); 3.44(2) 4.512(2), 4.601(2); 3.46(2) 5.040(2), 4.964(2); 4.20(1) 108.02(5), 112.00(4); 91.4(4) 95.50(4), 97.85(4); 87.3(5) 98.45(4), 102.40(4); 89.9(5) 122.07(9), 117.91(5); 127.4(3) 100.9(1), 100.12(5); 124.0(3) 128.7(1), 125.77(5); 125.5(3) 67.97(4), 65.08(3); 89.6(4) 81.7(4), 76.25(3); 93.0(5) 79.62(5), 74.82(4); 90.1(5) In the cyanate N]C, C]O for anions 1 and 2 are 0.81(2), 1.39(2) and 1.06(2), 1.24(2) Å; N]C]O are 176(2) and 177(2)8.Ag(1,19,2)]N]C are 123(1), 125(2), 128(2) (anion 1) and Ag(2,29,1)]N]C are 125(1), 125(1), 124(1)8 (anion 2).2134 J. Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 Fig. 2 (a), (b) The bromide and iodide ‘cubane’ structures, projected down their 3 and 2 axes respectively.(c) The cyanate [bis(pyridine) solvate form], projected (i) down and (ii) normal to the 2 axis most nearly ‘cubic’ example, angles at the silver and nitrogen atoms within the array lying closer to 908 than in any other systems hitherto recorded. The cyanide, by contrast, is a polymer, found so far, in two phases, a and b, depending on the crystallisation solvent, and of similar type, taking the form of a 1 : 2 complex with the complex ion [Ag(CN)2]2 as the counter ion which then bridges successive Ag{As(C6H11)3}2 units in the manner of the 1 : 2 thiocyanate complex described in the previous paper:1 ? ? ? {(C6- H11)3As}2Ag(NCAgCN){(C6H11)3}2Ag ? ? ? [Fig. 4(a)], assignment of the C, N components of the anion being tentatively made on the basis of refinement behaviour in the context of weak data and consistent with C-bound cyanide within the anionic (NCAgCN) component. One ? ? ? {(C6H11)3As}2Ag- (NCAgCN) ? ? ? unit comprises the asymmetric unit of the polymer which is generated by the unit a translation in the a form and the 21 screw of the monoclinic space group P21/c in the b form.Selected structural parameters are listed in Table 4. The geometry of the linear NCAgCN anionic component is unremarkable, being comparable to those of the numerous examples of it as a discrete entity. About the ‘cationic’ silver, Ag(1), in an As2AgN2 environment, we find that the As]Ag]As angle is considerably enlarged here [144.20(8), 137.20(6)8, cf.the thiocyanate 1 121.84(7), 122.11(9)8], as is N]Ag]N, cf. N]Ag]S 108.3(5), 104.0(4) vs. 94.2(5), 96.8(4)8. The Ag]As distances are only slightly shorter here 2.539(2), 2.568(2), 2.576(1), 2.545(2) Å, cf. 2.573(1)–2.581(2) Å, while Ag]N are intermediate [2.43(1), 2.51(2); 2.41(1), 2.38(1) Å] cf. 2.33(1), 2.36(2) (Ag]N) and 2.602(5), 2.602(5) (Ag]S) of the thiocyanate. An interesting feature of both phases is the non-linear co-ordination of one of the anionic cyanides to the cationic silver [Ag]N]C 134(1), 145(1)8]. The E = P analogue, whose characterisation rests on the single crystal X-ray study, is similar in nature, crystallising in a derivative triclinic cell, wherein a pair of independent pseudo-symmetrically related strands of polymer lying parallel to a each contribute one repeat unit to the asymmetric unit ofJ.Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 2135 Fig. 3 The bromide structure projected normal to (a) and down (b) (one sheet only) the 3 axis, showing the disposition of the tetramers in sheets, with axial ligands interlocking adjacent sheets the structure, the strands being similar in aspect to those of their E = As counterparts.The geometry diVers significantly however (Table 4); the non-linear co-ordination of one end of the NCAgCN unit to the ‘cationic’ silver atom, noted above in the E = As analogue, is now exacerbated [Ag]N(29)]C(29) 119(1), 121(1)8] with a marked disparity of ca. 0.5 Å now evident in the Ag]N(1,29) distances, i.e. N(29) is now incipiently dissociating and the P2AgN(1) angle sums now (ca. 3568) approaching planarity closely. Crystallisation of the E = As reaction mixture from pyridine, rather than 2-methyl- or 2,4,6-trimethyl-pyridine, results in solvation of the cationic silver atoms, displacing some of the co-ordinated anions, to yield an interesting adduct of 3:4:2 AgCN:As(C6H11)3 :py stoichiometry.The general nature of the2136 J. Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 Fig. 4 (a) A single strand of the cyanide polymer in the a (i) and b form (ii). (b) The [{Ag[As(C6H11)3]2(py)}2(CN)]1 cation of the 3:4:2 AgCN:As(C6H11)3 :py adduct Table 3 ‘Cubane’ core geometry in [{[(C6H11)3As]AgBr}4]. Distances in Å, angles in 8; singly and doubly primed atoms are generated by 1 2 x, x 2 y, z and 1 2 x 1 y, 1 2 x, z respectively Ag(1)]As(1) Ag(1)]Br(2) Ag(1) ? ? ? Ag(2) Ag(2) ? ? ? Ag(29) Ag(1) ? ? ? Br(1) As(1)]Ag(1)]Br(2) Br(2)]Ag(1)]Br(29) Ag(2)]Br(1)]Ag(20) 2.497(5) 2.726(3) 3.448(4) 3.553(4) 4.599(6) 114.43(8) 104.1(1) 80.6(1) Ag(2)]As(2) Ag(2)]Br(1) Ag(2)]Br(2) Ag(2)]Br(20) Br(1) ? ? ? Br(2) Br(2) ? ? ? Br(29) Ag(2) ? ? ? Br(29) As(2)]Ag(2)]Br(1) As(2)]Ag(2)]Br(2) As(2)]Ag(2)]Br(20) Br(1)]Ag(2)]Br(2) Br(1)]Ag(2)]Br(20) Br(2)]Ag(2)]Br(20) Ag(1)]Br(2)]Ag(2) Ag(1)]Br(2)]Ag(29) Ag(2)]Br(2)]Ag(29) 2.494(4) 2.747(4) 3.035(3) 2.616(5) 4.268(5) 4.300(5) 4.813(5) 115.2(1) 95.0(1) 135.5(1) 95.01(8) 105.4(1) 98.8(1) 73.31(9) 80.4(1) 77.52(9) complex is definitively, albeit imprecisely, established by elemental analysis coupled with a single crystal X-ray study as [{Ag[As(C6H11)3]2(py)}2(CN)][Ag(CN)2] [Fig. 4(b)], the [Ag- (CN)2]2 counter ion now being discrete, the two cationic silver atoms bridged by a simple linearly co-ordinated cyanide, modelled with C,N scrambled, and with the fourth coordination site occupied by pyridines.The imprecise geometries are unexceptional, but it is of interest that the pair of py groups within the cation are aligned so as to be eclipsed relative to the Ag ? ? ? Ag line. Infrared spectroscopy The far-IR spectra of [{[(C6H11)3As]AgCl}2] and [{[(C6H11)3- As]AgX}4] (X = Br or I) are shown in Fig. 5. Strong bands due to the As(C6H11)3 ligand are evident down to about 280 cm21; ligand bands also occur in the region below this, but these are generally quite weak.Strong bands, the wavenumbers of which are dependent on the nature of X, are also observed in these spectra. For the chloride the two very strong bands at 229 and 148 cm21 are assigned to the n(AgCl) modes of the Ag2Cl2 unit that is present in this complex (see above). The situation is similar to that found previously for the [{[P(C6H11)3]AgX}2] and [{(Ph3P)2AgX}2] (X = Cl or Br) complexes, which contain similar silver halide core units.3,13 For an Ag2X2 unit of D2h symmetry (x parallel to the Ag ? ? ? Ag diagonal, y parallel to the X? ? ? X diagonal) two IR-active modes of B2u and B3u symmetry are predicted, involving displacement of X and Ag along the positive and negative x directions respectively (B3u) or a similar vibration in the y direction (B2u). For a perfectly square Ag2X2 unit these two modes would have the same frequency, and this situation has been observed in the chloroform solvates [{(Ph3P)2AgX}2]?2CHCl3 (X = Cl or Br).13 A distortion in which two bonds on opposite sides of the square are shortened and the other two are lengthened results in a lowering of the symmetry from D2h to C2h, and the two IR-active n(AgX) modes both have Bu symmetry.The forms of these modes are such that one mainly involves stretching of the two short Ag]X bonds, the other stretching of the two long Ag]X bonds, so that the two modes should give rise to bands at significantly diVerent wavenumbers. This situation has previously been observed in unsolvated [{(Ph3P)2AgCl}2] 13 and in [{[(C6H11)3P]AgCl}2],3 Fig. 5 Far-IR spectra of (a) [{[As(C6H11)3]AgCl}2], (b) [{[As(C6H11)3]- AgBr}4], (c) [{[As(C6H11)3]AgI}4]. The bands assigned to the n(AgX) modes are labelled with their wavenumbersJ. Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 2137 Table 4 Selected geometries: cyanide adducts. Distances in Å, angles in 8 (a) AgCN:As(C6H11)3 (1 : 1). The two values in each entry are for a,b forms respectively; primed atoms are generated by x 2 1, y, z and 2 2 x, y 2 ��� , ��� 2 z respectively (i) The ‘cationic’ silver atoms [Ag(1)] Ag]As(1) Ag]As(2) As(1)]Ag]As(2) As(1)]Ag]N(1) As(1)]Ag]N(29) As(2)]Ag]N(1) 2.539(2), 2.576(1) 2.568(2), 2.545(2) 144.20(8), 137.20(6) 107.6(4), 99.8(3) 97.1(4), 99.5(3) 96.4(4), 107.2(2) Ag]N(1) Ag]N(29) As(2)]Ag]N(29) N(1)]Ag]N(29) Ag]N(1)]C(1) Ag]N(29)]C(29) 2.43(1), 2.41(1) 2.51(2), 2.38(1) 100.3(4), 105.3(3) 108.3(5), 104.0(4) 134(1), 145(1) 163(2), 168(1) (ii) The ‘anionic’ silver atoms [Ag(2)] Ag]C(1) Ag]C(2) C(1)]Ag]C(2) 2.08(2), 2.04(1) 2.02(2), 2.04(1) 177.5(7), 174.4(6) C(1)]N(1) C(2)]N(2) Ag]C(1)]N(1) Ag]C(2)]N(2) 1.11(2), 1.12(2) 1.08(2), 1.12(2) 171(2), 175(1) 178(2), 177(1) (b) AgCN:P(C6H11)3 (1 : 1); primed atoms are generated by the unit a translation, the two values in each entry being for strands 1,2 respectively (i) The ‘cationic’ silver atoms [Ag(1)] Ag]P(1) Ag]P(2) P(1)]Ag]P(2) P(1)]Ag]N(1) P(1)]Ag]N(29) P(2)]Ag]N(1) 2.420(4), 2.429(4) 2.414(3), 2.413(3) 147.9(1), 149.4(1) 99.6(3), 98.4(3) 97.0(3), 94.8(2) 108.3(3), 108.7(3) Ag]N(1) Ag]N(29) P(2)]Ag]N(29) N(1)]Ag]N(29) Ag]N(1)]C(1) Ag]N(29)]C(29) 2.41(1), 2.44(1) 2.97(1), 2.92(1) 93.5(3), 94.3(3) 99.8(5), 99.9(4) 168(1), 171(1) 119(1), 121(1) (ii) The ‘anionic’ silver atoms [Ag(2)] Ag]C(1) Ag]C(2) C(1)]Ag]C(2) 2.04(1), 2.06(1) 2.10(1), 2.09(1) 174.5(6), 176.0(6) C(1)]N(1) C(2)]N(2) Ag]C(1)]N(1) Ag]C(2)]N(2) 1.17(2), 1.12(2) 1.16(2), 1.18(2) 167(1), 169(1) 177(1), 176(1) (c) AgCN:As(C6H11)3:py (3:4:2) (i) The cation: the two values for each entry are for moieties 1,2 Ag]As(1) Ag]As(2) Ag]N(1) As(1)]Ag]As(2) As(1)]Ag]N(1) As(1)]Ag]C/N As(2)]Ag]N(1) 2.54(5), 2.57(2) 2.55(2), 2.55(3) 2.49(3), 2.58(4) 128(1), 124.1(6) 107(1), 104.1(9) 108(1), 112(2) 106(1), 106(2) Ag]C/N C]N As(2)]Ag]C/N N(1)]Ag]C/N Ag]C/N]N/C 2.34(5), 2.40(5) 0.84(6) 108(2), 112.6(9) 95(2), 93(1) 170(4), 166(3) (ii) The anion Ag]C C(1)]Ag]C(2) 1.94(7), 2.04(6) 177(2) C]N Ag]C]N 1.21(8), 1.06(7) 165(5), 172(5) and is the reason for the large separation between the two n(AgX) bands observed for [{[(C6H11)3As]AgCl}2] (Fig. 5).It has previously been shown that the wavenumbers of the n(MX) modes of Group 11 metal halide complexes are sensitive to the strength of the M]X bonds. Thus, the n(MX) wavenumbers have been empirically correlated with the M]X bond length r via relationship (1) where b and m are constants.14,15 Somewhat n/cm21 = b(r/Å)2m (1) surprisingly, this relationship appears to be valid for vibrations involving both terminal and bridging halogen atoms X.3,15,16 The most likely reason for this is that the M]X]M angles in the bridged complexes are close to 908, so that the vibrations of the M]X bonds involved in the bridge are essentially uncoupled, and so are independent of each other.The angles in the Ag2Cl2 unit in [{[(C6H11)3As]AgCl}2] are close to 908 (Table 1), so the frequencies of the n(AgCl) modes should fit the previously established correlation (1) for AgCl complexes.Fig. 6 shows that this is indeed the case. The spectra of the cubane tetramers [{[(C6H11)3As]AgX}4] (X = Br or I) show n(AgX) bands below 200 cm21. For an Ag4X4 unit of Td symmetry two IR-active n(AgX) modes of T2 symmetry are predicted.17–19 Thus, the two bands at 111 and 86 cm21 for the X = I complex (Fig. 5) can be assigned to these modes. The positions of these bands are quite similar to those of the corresponding P(C6H11)3 complex (113, 83 cm21).3 The situation for the X = Br complex is less straightforward; four partially resolved bands are observed at 167, 151, 125 and 109 cm21 (Fig. 5).A possible reason for this is the significant distortion of the Ag4Br4 core from Td symmetry that is observed in this complex. The Ag]Br bond lengths range from about 2.6 to 3.0 Å (Table 3), compared with a range 2.8 to 3.0 Å for the Ag]I bond lengths in the iodide (Table 2).The recently established correlation between the n(AgX)2138 J. Chem. Soc., Dalton Trans., 1998, Pages 2131–2138 wavenumbers and the Ag]X bond lengths 15 suggests that the n(AgBr) bands of the bromide tetramer should lie in the range 100 to 160 cm21, which agrees well with the observed spectrum. The reduction in the symmetry of the Ag4Br4 core from Td to C3v would result in the two T2 modes being split into four components. This is the most likely explanation for the relatively complex pattern of bands observed in the far-IR spectrum of the bromide.Acknowledgements We acknowledge support of this work by grants from the Australian Research Grants Scheme and the University of Auckland Research Committee. We thank Ms. Catherine Hobbis for recording the far-IR spectra. Fig. 6 Plot of wavenumber of the n(AgCl) band against Ag]Cl bond length. The data (d) and the best fit curve using equation (1) are from ref. 15.The data (j) are for [{[(C6H11)3As]AgCl}2] References 1 G. A. Bowmaker, EVendy, B. W. Skelton and A. H. White, preceding paper. 2 G. A. Bowmaker, EVendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2449. 3 G. A. Bowmaker, EVendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2459. 4 J. A. Muir, M. B. Pulgar, P. G. Jones and G. M. Sheldrick, Acta Crystallogr., Sect. C, 1985, 1174. 5 G. A. Bowmaker, EVendy, J. C. Reid, C. E. F. Rickard, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., following paper. 6 R. A. Stein and C. Knobler, Inorg. Chem., 1977, 16, 242. 7 M. Nardelli, C. Pelizzi, G. Pelizzi and P. Tarasconi, J. Chem. Soc., Dalton Trans., 1985, 321. 8 EVendy, J. D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 671. 9 K. F. Chew, W. Derbyshire and N. Logan, Chem. Commun., 1970, 1708; R. N. Dash and D. V. Ramana Rao, Z. Anorg. Allg. Chem., 1972, 393, 309. 10 O. H. Ellestad, P. Klaeboe, E. E. Tucker and J. Songstad, Acta. Chem. Scand., 1972, 26, 3579; D. Britton and J. D. Dunitz, Acta Crystallogr., 1965, 18, 424. 11 W. Beck and W. P. Fehlhammer, in MTP Int. Rev. Sci., 1972, 2, 253; K. Vrieze and G. van Koten, in Comprehensive Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon, Oxford, 1987, vol. 2, p. 189. 12 V. W.-W. Yam, W.-K. Lee and K.-K. Cheung, J. Chem. Soc., Dalton Trans., 1996, 2335. 13 G. A. Bowmaker, EVendy, J. V. Hanna, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1993, 1387. 14 G. A. Bowmaker, P. C. Healy, J. D. Kildea and A. H. White, Spectrochim. Acta, Part A, 1988, 44, 1219. 15 G. A. Bowmaker, EVendy, J. D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 577. 16 G. A. Bowmaker, R. D. Hart, B. E. Jones, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1995, 3063. 17 B. K. Teo and D. M. Barnes, Inorg. Nucl. Chem. Lett., 1976, 12, 681. 18 G. A. Bowmaker, R. J. Knappstein and S. F. Tham, Aust. J. Chem., 1978, 31, 2137. 19 G. A. Bowmaker and P. C. Healy, Spectrochim. Acta, Part A, 1988, 44, 115. Received 29th January 1998; Paper 8/00793D

ISSN:1477-9226    

DOI:10.1039/a800793d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 2133–2139 2133 The phase behaviour of 1-alkyl-3-methylimidazolium tetrafluoroborates; ionic liquids and ionic liquid crystals John D. Holbrey and Kenneth R. Seddon The QUESTOR Centre, The Queen’s University of Belfast, Belfast, UK BT9 5AG. E-mail: j.holbrey@qub.ac.uk; k.seddon@qub.ac.uk; http://www.ch.qub.ac.uk/staV/personal/krs Received 9th April 1999, Accepted 10th May 1999 Air- and water-stable 1-alkyl-3-methylimidazolium tetrafluoroborate salts with the general formula [Cn-mim][BF4] (n = 0–18) have been prepared by metathesis from the corresponding chloride or bromide salts.The salts have been characterised by 1H NMR and IR spectroscopy, microanalysis, polarising optical microscopy and diVerential scanning calorimetry. Those with short alkyl chains (n = 2–10) are isotropic ionic liquids at room temperature and exhibit a wide liquid range, whereas the longer chain analogues are low melting mesomorphic crystalline solids which display an enantiotropic smectic A mesophase.The thermal range of the mesophase increases with increasing chain length and in the case of the longest chain salt prepared, [C18-mim][BF4], the mesophase range is ca. 150 8C. The investigation of ambient temperature ionic liquids has largely focused on systems containing 1,3-dialkylimidazolium and N-alkylpyridinium organic cations and tetrachloroaluminate( III) anions.1–4 These systems possess all the intrinsically useful characteristics of ambient temperature ionic liquids (negligible vapour pressure, wide liquidus, thermal stability, high ionic conductivity and a large electrochemical window), but they are both highly reactive to certain materials and are susceptible to moisture.Following the relatively recent reports of air-stable room temperature 1-ethyl-3-methylimidazolium based ionic liquids containing tetrafluoroborate 5 and toluenesulfonate 6 anions, a wide range of new ionic liquids have been developed incorporating many diVerent anions 5,7,8 which demonstrate the potential specifically to tailor the properties (i.e.moisture stability, viscosity, miscibility with other cosolvents) of the ionic liquid. Some of these new, air- and moisture-stable ionic liquids 9 have been used as solvents for rhodium and nickel catalysed hydrogenation reactions,10,11 electrochemical studies using the ionic liquid as both solvent and electrolyte 12,13 and in solar cell applications.14 The Belfast team 15 is currently investigating the wide range of properties exhibited by diVerent ionic liquids in order to develop a database of materials which can be interrogated to obtain a desirable set of properties for a particular application or reaction.As part of this programme, we have studied the range of rheological and chemical properties exhibited by ionic liquids as a function of alkyl chain length, anion type, etc., in order to develop more logical approaches to modifying their physical properties. We have demonstrated, in recent publications, that long-chain 1-alkyl-3-methylimidazolium chloride, tetrahalogenometalate 16 and hexafluorophosphate 17 salts display thermotropic smectic liquid crystalline mesomorphism over a relatively wide temperature range, increasing with increasing chain length.Related symmetrically substituted long-chain 1,3-dialkylimidazolium hexafluorophosphate salts have also been recently reported.18 The incorporation of liquid crystalline properties into these neoteric solvents 19 promises to extend even further their range of useful solvent and catalytic properties for commercially important processes.We report here the preparation and physical properties of 1-alkyl-3-methylimidazolium tetrafluoroborate salts with alkyl chains (CnH2n 1 1) varying from n = 0 to 18 (designated as [Cnmim][ BF4] except for the commonly used [emim]1 and [bmim]1 for n = 2 and 4 respectively, and [H-mim]1 for n = 0).The salts with short alkyl chain lengths (n = 2–10) are liquids at room temperature and form glasses on cooling to 280 8C, whereas the longer chain salts (n = 12–18) are low melting solids which display enantiotropic mesomorphism with an extensive thermotropic mesophase range. The two imidazolium salts N-substituted with a proton or methyl group (n = 0 and 1 respectively) are low melting crystalline solids. Experimental Microanalyses were performed by A.S.E.P. at the Queen’s University of Belfast.All chemicals were used as received unless otherwise stated. Infrared spectra were recorded as liquid or glassy films between KBr plates using a Perkin-Elmer 1600 or Bio-Rad FTS-185 spectrometer, 1H NMR spectra in propanone-d6 solution using a Bruker WM500 spectrometer; proton chemical shifts are recorded relative to an internal TMS standard. All the tetrafluoroborate salts with n > 3 were prepared by metathesis in aqueous solution from the respective chloride or bromide salts with either HBF4 or NaBF4; the salts with n = 1–3 were prepared using the literature method for [emim][BF4]5 by metathesis with silver tetrafluoroborate, and [H-mim][BF4] 20 was obtained by direct reaction of 1-methylimidazole with tetrafluoroboric acid. Representative syntheses for each case are described.Analytical data are shown in Tables 1 and 2. The melting points, clearing points and glass transition temperatures were determined by diVerential scanning calorimetry (Perkin-Elmer Pyris 1 DSC equipped with dinitrogen cryostatic cooling, 5–10 mg samples, 5 8C min21 heating and cooling rates) and heated-stage polarising optical microscopy where appropriate (transition temperatures > 0 8C) using an Olympus BX50 microscope equipped with a Linkam TH600 hot stage and TP92 temperature controller. The X-ray powder diVraction pattern of [C14-mim][BF4] was determined using a Siemens D5000 powder diVractometer with Cu-Ka X-rays, l = 1.542 Å.Data were recorded from between 2 and 208 in steps of 0.058. N N CnH2 n + 1 [C n-mim]+, n = 0-18 +2134 J. Chem. Soc., Dalton Trans., 1999, 2133–2139 Preparations 1-Methylimidazolium tetrafluoroborate, [H-mim][BF4].20 Tetrafluoroboric acid (11.8 cm3, 0.092 mol, 48% solution in water) was added dropwise to stirred 1-methylimidazole (7.5 g, 0.092 mol) cooled to 0 8C in an ice-bath over 30 min. The reaction mixture was stirred for 2 h and the aqueous solvents were removed in vacuo to give the product as a colourless oil, which solidified on cooling. Yield 15.55 g (100%). 1-Ethyl-3-methylimidazolium tetrafluoroborate, [emim]- [BF4].5,21 Tetrafluoroboric acid (15.2 cm3, 0.116 mol, 48% solution in water) was added slowly to a rapidly stirred slurry of silver(I) oxide (13.49 g, 0.058 mol) in water (50 cm3) over 15 min. The reaction mixture was covered with aluminium foil to prevent photodegradation and stirred for a further hour until all the silver(I) oxide had completely reacted to give a colourless solution. A solution of 1-ethyl-3-methylimidazolium bromide (22.24 g, 0.116 mol) in water (200 cm3) was added to the reaction mixture and stirred at room temperature for 2 h.The resulting yellow precipitate of silver bromide was removed by filtration and the solvent removed from the supernatant liquor by heating at 70 8C, initially under reduced pressure and finally in vacuo, to yield the tetrafluoroborate salt as a pale yellow liquid.Yield 21.36 g, 93%. IR (liquid film); n& /cm21 3167, 3125 (s, aromatic C–H stretch), 2990 (m, aliphatic C-H stretch), 1576 (s, sym. ring stretch) 1457 (s, sym. ring stretch), 1173 (s, sym. ring stretch) and 1052 (br s, BF4 stretch). 1-Decyl-3-methylimidazolium tetrafluoroborate, [C10-mim]- [BF4]. Tetrafluoroboric acid (10.2 cm3, 0.078 mol, 48% solution in water) was added dropwise to a cooled, rapidly stirring solution of 1-decyl-3-methylimidazolium chloride (20.2 g, 0.078 mol) in water (200 cm3) over 10 min.The lower ionic liquid layer was collected via a separating funnel, dissolved in dichloromethane (300 cm3) and washed with water (2 × 150 cm3). The organic layer was collected, dried over anhydrous MgSO4, filtered and the solvent removed in vacuo to yield the tetrafluoroborate salt as a colourless liquid. Yield 22.55 g, 93%. 1-Methyl-3-tetradecylimidazolium tetrafluoroborate, [C14- mim][BF4].A solution of NaBF4 (6.60 g, 0.060 mol) in water (50 cm3) was added slowly to a cooled, rapidly stirring solution of 1-methyl-3-tetradecylimidazolium chloride (19.84 g, 0.060 mol) in water (200 cm3). The product precipitated as a waxy solid and was collected by filtration, dissolved in dichloromethane (300 cm3) and washed with water (2 × 100 cm3). The organic layer was collected, dried over MgSO4, filtered and the solvent removed in vacuo to yield the tetrafluoroborate salt which was recrystallised from methanol as a colourless powder (14.74 g, 88% yield).The salts [C1-mim][BF4] and [C3-mim][BF4] were prepared using the method described for [emim][BF4]; [bmim][BF4] and [C5-mim][BF4] were prepared using the method for [C14-mim]- [BF4]. (The minimum volume of water was used to dissolve the chloride salts. The tetrafluoroborate salt did not initially separate from the aqueous phase, but was preferentially extracted by dichloromethane).The salts [C7-mim][BF4], [C8- mim][BF4] and [C9-mim][BF4] were prepared using the method for [C10-mim][BF4] by metathesis with HBF4; [C6-mim][BF4], [C11-mim][BF4], [C12-mim][BF4], [C13-mim][BF4], [C15-mim]- [BF4], [C16-mim][BF4] and [C18-mim][BF4] were prepared using the method for [C14-mim][BF4] by metathesis with Na[BF4]. Results and discussion Synthesis of the salts The tetrafluoroborate salts are air and water stable. In addition, it was assumed from the previously published examples, [emim][BF4] (n = 2) 5,21,22 and [bmim][BF4] (n = 4),11 that their melting points should be generally lower than those of their chloride or hexafluorophosphate analogues. It was also expected that the tetrafluoroborate salts would have a greater miscibility with water than the corresponding hexafluorophosphate salts, but that they would be less hygroscopic than the chloride and bromide salts, thus placing their properties in an intermediate position between those of the [PF6]2 and Cl2 salts.The tetrafluoroborate salts were prepared by simple metathesis reactions from the corresponding chloride or bromides. The short alkyl chain salts, [emim][BF4] and [bmim][BF4], had been prepared previously from the corresponding chloride salts using aqueous silver(I) tetrafluoroborate 21 or from propanone using sodium tetrafluoroborate.11 Unfortunately, the former route can be prohibitively expensive at moderate or large scale (even when recycling of silver halide by-products is utilised) whereas, in practice, the latter method leaves appreciable amounts of chloride contamination (although the solubility of sodium chloride in propanone is only 5.5 × 1026 mol l21, it increases significantly with only small concentrations of water and with additional dissolved salts 23) in the ionic liquids.Recently a silver-free preparation of [emim][BF4] using ammonium tetrafluoroborate in propanone has been described.24 We found that for alkyl chains longer than n > 3 the salts can be prepared by metathesis from the chloride or bromide salts in water followed by extraction into dichloromethane, and have developed an improved laboratory-scale metathesis and extraction route which allows a cost-eVective high purity preparation of the tetrafluoroborate salts using either HBF4 or Na[BF4] in aqueous solution. For most of the series, the desired tetrafluoroborate salt either separates from the aqueous reaction mixture as a dense liquid (n = 6–10) or precipitates as a solid (n > 10) which can be isolated by decantation or filtration as appropriate.The water-soluble short chain salts (n = 4 or 5) can be extracted and purified from aqueous solution by partitioning into an organic solvent (dichloromethane is excellent on a laboratory scale). Unfortunately, for the shortest chain salts (n < 4), the water :CH2Cl2 partition coeYcient is too low to permit eYcient extraction of the salts from the aqueous phase and the published route using aqueous Ag[BF4] was used to prepare n = 1–3.The 1-methylimidazolium salt ([H-mim][BF4]) was synthesized by direct combination of methylimidazole and HBF4 using a published procedure.20 The particular method employed in each case is shown in Table 1. In all cases, except n = 4 and 5, the tetrafluoroborate salts were isolated with unoptimised yields of better than 85% after work-up; variation in yields represents mechanical losses during the extraction and work-up process.For n = 4 and 5 where the tetrafluoroborate salts have a higher solubility in aqueous media, the water : CH2Cl2 partition is less eYcient and variable yields between 50 and 70% were obtained. For the water soluble salts (n < 6), testing for the presence of residual halide using silver nitrate solution gave a negative result (i.e. no AgCl precipitation). 1-Alkyl-3-methylimidazolium tetrafluoroborate salts ([Cnmim][ BF4]) were prepared for n = 0 to 18 and characterised by a combination of 1H NMR and infrared spectroscopy, microanalysis and DSC.The CHN microanalyses and 1H NMR data are summarised in Tables 1 and 2, respectively. The infrared spectra for all the dialkylated salts are similar, except for an increase in aliphatic C–H stretch and bend intensities with n, and show characteristic stretching bands for the imidazolium cation and a broad, unresolved band due to the [BF4]2 anion.The shorter chain salts are hygroscopic, although to a much lesser extent than their chloride analogues. Both miscibility with water and the hygroscopic nature of the salts decreases markedly with increasing chain length. The salts can be dried eVectively by heating at 70–90 8C in vacuo for 6–8 h, then stored under dinitrogen in a glove-box. Karl–Fisher measurements 25 show typical water contents of dried samples to be <200 ppm. In addition to the Karl–Fischer titration measurements, theJ.Chem. Soc., Dalton Trans., 1999, 2133–2139 2135 absence (or presence) of OH stretching absorbances in the 3500–3800 cm21 region of the infrared spectra was used to monitor for the presence of water in samples, the presence of a band indicating that further drying in vacuo was required. In addition, for thoroughly dried samples, no broad absorption bands in the region 3000–3100 cm21 which could have been ascribed to CH ? ? ? F hydrogen bonding between the [BF4]2 anion and the organic cation were observed.This appears to indicate that there is no significant hydrogen bonding present, in contrast to imidazolium based ionic liquids with chloride anions.26 The 1H NMR chemical shift of the C(2)–H protons of the tetrafluoroborate salts both in solution in propanone-d6 and neat [Cn-mim][BF4] (n = 2–8) are all remarkably similar and do not show the large downfield shift related to hydrogen bonding observed for the chloride salts.26 However, there is a disparity between these observations and the confused and contradictory reports in the literature about which observations can be used to demonstrate the presence or absence of hydrogen bonding between the anions and cations in the [emim][BF4] and [bmim] Table 1 CHN Microanalytical data for the anhydrous tetrafluoroborate salts prepared. Note, for n = 0 and 1, known compounds, satisfactory analyses could not be obtained due to the hygroscopic nature of the salts Analysis, found (calc.) (%) n Synthesis C H N 23456789 10 11 12 13 14 15 16 18 AgBF4/H2O AgBF4/H2O NaBF4/H2O NaBF4/H2O NaBF4/H2O HBF4/H2O HBF4/H2O HBF4/H2O HBF4/H2O NaBF4/H2O NaBF4/H2O NaBF4/H2O NaBF4/H2O NaBF4/H2O NaBF4/H2O NaBF4/H2O 35.91 (36.40) 38.91 (39.66) 42.13 (42.51) 44.71 (45.03) 45.87 (47.27) 49.10 (49.28) 50.96 (51.09) 52.63 (52.72) 54.95 (54.21) 55.24 (55.57) 56.53 (56.82) 58.19 (57.96) 58.59 (59.02) 60.13 (60.05) 60.78 (60.92) 63.01 (62.56) 5.67 (5.60) 6.06 (6.18) 6.86 (6.69) 7.08 (7.14) 7.53 (7.54) 7.80 (7.89) 8.14 (8.22) 8.70 (8.51) 9.05 (8.77) 9.28 (9.02) 9.32 (9.24) 9.34 (9.44) 9.87 (9.63) 10.09 (9.81) 9.77 (9.97) 10.12 (10.26) 13.76 (14.15) 12.92 (13.21) 12.24 (12.39) 11.44 (11.67) 10.85 (11.03) 10.85 (10.45) 9.58 (9.93) 9.44 (9.46) 8.98 (9.03) 8.51 (8.64) 8.47 (8.28) 7.87 (7.95) 7.37 (7.65) 7.23 (7.37) 7.01 (7.10) 6.83 (6.63) [BF4] ionic liquids.Suarez et al. have repeatedly reported 11,13 that [bmim][BF4] exhibits hydrogen bonding in the liquid phase as observed from bands between 3000 and 3200 cm21 in the IR spectrum, yet these IR bands are not described in the experimental section of their paper.In addition, the original communication describing, amongst other salts, [emim][BF4] by Wilkes and Zaworotko5 is cited as evidence for similar hydrogen bonding in related salts. However, the only reports of hydrogen bonding in this paper are in the crystal structures of the nitrate, nitrite and sulfate salts with well established, strong C–H ? ? ?O hydrogen bonding in the solid state which cannot readily be compared to potential C–H ? ? ? F interactions in liquid tetra- fluoroborate or hexafluorophosphate 5 examples.Although it appears very likely that interactions occur (as implied by the diVerences in water solubility of the tetrafluoroborate and hexafluorophosphate salts), these are too weak to be seen unambiguously in the IR spectra. The tetrafluoroborate salts were obtained as white, lowmelting solids (n = 0, 1, or 11–18) or colourless to pale yellow liquids (n = 2–11).All the salts prepared are water- and airstable under ambient conditions, and may be handled under normal laboratory conditions. The shorter chain salts (n = 0–5) are somewhat hygroscopic, but can be dried by extended heating at 70 8C under high vacuum, until no water signals are observed in the infrared spectrum around 3500 cm21. The salts were then stored under dinitrogen in a dry-box.Interestingly, in contrast to the analogous hexafluorophosphate salts, the salts with n < 6 are soluble in water, although the aqueous solubility drops rapidly through the series. Thus, while [emim][BF4] is both hygroscopic and totally soluble in water, [C6-mim][BF4] is essentially insoluble in water, though water is soluble in the ionic liquid at up to 30 000 ppm25 and the intermediate [bmim][BF4] is completely miscible in water at room temperature, but on cooling to close to 0 8C phase separation occurs as the solubility of [bmim][BF4] in water drops.22 All the tetrafluoroborate salts are soluble in propanone and in dichloromethane, but exhibit liquid clathrate formation in trichloromethane; based on observation and 1H NMR measurements the clathrates have a 1 : 3 [Cn-mim][BF4] :CHCl3 ratio.We have also observed similar clathrate formation between CHCl3 and hexafluorophosphate ionic liquids. It is possible that this is induced by C–H ? ? ? F hydrogen bonding between the relatively acidic trichloromethane proton and the [BF4]2 or [PF6]2 anions.Table 2 1H NMR chemical shifts d (relative to TMS internal standard) and coupling constants J/Hz of [Cn-mim][BF4] salts in propanone-d6 solution (integrals and coupling constants are given in parentheses as appropriate) n H2 H4 H5 NCH2 NCH3 NCH2CH2 Alkyl Terminal CH3 0 123 4 567 89 10 11 12 13 14 15 16 18 8.92 (s) 8.90 (s) 8.88 (s) 9.00 (s) 8.83 (s) 9.05 (s) 9.05 (s) 9.02 (s) 9.05 (s) 9.28 (s) 9.04 (s) 9.35 (s) 9.04 (s) 9.08 (s) 9.09 (s) 9.16 (s) 9.20 (s) 9.09 (s) 7.71 (t, J = 1.7) 7.70 (d, J = 1.8) 7.72 (d, J = 1.8) 7.78 (d, J = 1.8) 7.69 (d, J = 1.8) 7.80 (d, J = 1.8) 7.78 (d, J = 1.8) 7.78 (1 H, d, J = 1.8) 7.81 (d, J = 1.8) 7.84 (d, J = 1.8) 7.78 (s, J = 1.8) 7.86 (s, J = 1.8) 7.78 (s, J = 1.8) 7.80 (s, J = 1.8) 7.79 (s, J = 1.8) 7.81 (s, J = 1.8) 7.82 (s, J = 1.8) 7.82 (s, J = 1.8) 7.69 (t, J = 1.7) 7.65 (d, J = 1.8) 7.74 (d, J = 1.8) 7.63 (d, J = 1.8) 7.73 (d, J = 1.8) 7.72 (d, J = 1.8) 7.72 (d, J = 1.8) 7.75 (d, J = 1.8) 7.77 (d, J = 1.8) 7.72 (s, J = 1.8) 7.79 (s, J = 1.8) 7.72 (s, J = 1.8) 7.74 (s, J = 1.8) 7.74 (s, J = 1.8) 7.78 (s, J = 1.8) 7.76 (s, J = 1.8) 7.76 (s, J = 1.8) 12.68 (1 H, br, NH) 4.35 (q, J = 7.3) 4.34 (t, J = 7.3) 4.34 (t, J = 7.3) 4.37 (t, J = 7.3) 4.36 (t, J = 7.4) 4.38 (t, J = 7.3) 4.40 (t, J = 7.3) 4.41 (t, J = 7.4) 4.37 (t, J = 7.4) 4.41 (t, J = 7.33) 4.37 (t, J = 7.4) 4.39 (t, J = 7.4) 4.38 (t, J = 7.4) 4.40 (t, J = 7.4) 4.39 (t, J = 7.4) 4.40 (t, J = 7.3) 4.09 (s) 4.05 (s) 4.00 (s) 4.07 (s) 4.04 (s) 4.06 (s) 4.05 (s) 4.07 (s) 4.08 (s) 4.04 (s) 4.05 (s) 4.09 (s) 4.06 (s) 4.07 (s) 4.07 (s) 4.09 (s) 4.09 (s) 4.09 (s) 1.99 (m, 1J = 7.3, 2J = 7.3) 1.95 (m, 1J = 2J = 7.4) 1.97 (m) 1.96 (m) 1.98 (m) 1.99 (m) 1.99 (m) 1.98 (m) 1.99 (m) 1.98 (m) 1.99 (m) 1.99 (m) 2.00 (m) 1.98 (m) 2.00 (m) 1.41 (2 H, m, 1J = 2J = 7.4) 1.38 (4 H, m) 1.35 (6 H, m) 1.32 (8 H, m) 1.31 (10 H, m) 1.32 (12 H, m) 1.32 (14 H, m) 1.32 (16 H, m) 1.32 (18 H, m) 1.32 (20 H, m) 1.33 (22 H, m) 1.34 (24 H, m) 1.33 (26 H, m) 1.33 (30 H, m) 1.54 (t, J = 7.3) 0.99 (t, J = 7.3) 0.98 (t, J = 7.4) 0.92 (t, J = 7.0) 0.90 (t, J = 7.4) 0.91 (t, J = 7.1) 0.91 (t, J = 7.3) 0.91 (t, J = 7.1) 0.91 (t, J = 7.0) 0.91 (t, J = 7.0) 0.92 (t, J = 7.0) 0.92 (t, J = 6.6) 0.92 (t, J = 6.8) 0.93 (t, J = 7) 0.92 (t, J = 7) 0.92 (t, J = 7)2136 J.Chem.Soc., Dalton Trans., 1999, 2133–2139 Table 3 Thermal data from DSC and optical microscopy observations. Transition temperatures (8C) measured from the peak positions for first order transitions (i.e. melting and clearing points) and transition midpoints for glass transitions from DSC; enthalpy (kJ mol21) given in parentheses Phase n Treatment Cr SA Iso 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 18 Heat Coolb Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool Heat Cool ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 52.4 a (2.9) 103.4 (9.8) 73.6 (29.6) 5.8 c (10.1) d 213.9 d 271.0 d,e 287.5 d 288.0 d 282.4 d 280.4 d 281.9 d 278.5 d 280.5 d 277.2 d 280.0 d 24.2 (7.3) 224.7 (212.8) 21.4 (24.7) 22.5 (223) 26.4 (29.6) 7.4 (223.7) 49.1 (32.2) 17.3 (224.1) 42.4 (20.7) 29.3 (216.8) 55.2 (33.4) 35.0 (223.2) 49.6 (31.5) 45.1 (224.7) 66.8 (23.8) 64.5 (227.6) — ——— — — — —————————— ? ? ? ? ? ? ? ? ? ? ? ? 38.5 (0.3) 37.0 (20.2) 92.7 (0.5) 91.7 (20.6) 129.5 (0.6) 130.0 (20.6) 148.1 (0.9) 147.5 (20.9) 182.0 (1.7) 172.2 (20.8) 214.8 (1.0) 213.4 (21.1) ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? a 55 8C,20 b Did not crystallise on cooling.c 15 8C,5 12 8C.21 d Glass transition. e 281 8C.11 The preparations of chloride-free [bmim][BF4] and higher homologues from aqueous solution† using this low cost metathesis route with Na[BF4] or HBF4 permit relatively large quantities (up to kilo scale in the laboratory) of these interesting ionic liquid solvents to be prepared, which will increase the scope for the investigation of synthetic reactions using the ionic liquids as solvents.In contrast, the existing metathetic routes to [emim][BF4] using silver salts fail under these criteria and, as has been noted by Chauvin et al.,10 chloride impurities can have a major disadvantageous eVect on some transition metal catalysed reactions in ionic liquid solvents. Clearly, this is the reason for the recent increase in publications of syntheses using [bmim][BF4] in preference to [emim][BF4] as an ionic liquid solvent. The preparation of the higher homologues described in this paper opens up new opportunities for the use of a wider range of ionic liquid solvents.Characterisation The melting points and mesophase clearing points (as appropriate) of the tetrafluoroborate salts are shown in Table 3 and Fig. 1. In all cases, the salts exhibited a marked tendency to supercool before crystallising or to form glasses; transition temperatures are given from DSC measurements for both heating and cooling cycles (except for the short chain salts, n = 0–4, † Chloride ion concentrations determined in aqueous solution using a Cl2 ion selective electrode for the [Cn-mim][BF4] ionic liquids (n = 2–10), prepared from the respective chloride salts, were typically below 0.2 ppm.where it proved impossible to obtain reproducible results on cooling cycles) and data are only presented for the melting transition (as indicated in Table 1). A typical DSC trace for the mesomorphic [C14-mim][BF4] salt is shown in Fig. 2 revealing the characteristically large enthalpy for the crystal (Cr)–liquid crystal transition and small enthalpy for the mesophase– isotropic (Iso) transition. Melting points and transition temperatures were also determined from heated-stage optical Fig. 1 Phase diagram for 1-(CnH2n 1 1)-3-methylimidazolium tetra- fluoroborate ionic liquids and liquid crystals showing the melting (closed square), glass (open square) and clearing (circle) transitions measured by DSC.J. Chem. Soc., Dalton Trans., 1999, 2133–2139 2137 microscopy observations as appropriate. The transition temperatures measured on heating were, in all cases, within ±1 8C of the values obtained by DSC. Thermal investigations.Melting points. The phase diagram for the salts as a function of chain length is illustrated in Fig. 1. Only [H-mim][BF4], [C1-mim][BF4], and the salts with n > 9 crystallise on cooling. The ionic liquids [Cn-mim][BF4] (n = 2–9) show a strong tendency to supercool, forming progressively more viscous liquids, and finally glasses, with no indication of crystallisation even on slow cooling (1 8C min21) to 2170 8C. On increasing the chain length over the range n = 0–4 the glass transition temperatures fall progressively until a lower limit around 280 8C is reached.It is notable that for the salts with n = 4–9 the glass transition temperatures are all between 270 and 290 8C at a cooling rate of 5 8C min21. This Tg is very similar to the lowest glass transition points (296 8C) observed for the [emim]Cl–AlCl3 ionic liquid system at 66% AlCl3 composition.27 Above n = 9, the true melting points rise quite dramatically and liquid crystalline mesomorphism is observed from n = 12.The salts (from n = 10) all supercool before crystallising; the degree of supercooling decreases with increasing chain length. The eVect on the transition behaviour from altering the alkyl chain length is continuous below n = 10; there is no overall diVerence in behaviour between the odd and even chain lengths (in contrast to the odd–even eVect observed in the melting points of the mesomorphic salts described below). Hence there appears to be no advantage to be gained (in terms of solvents for potential reactions) in preparing the more expensive oddchain length compounds in order to alter the properties.Mesomorphism. None of the salts with n < 12 displayed thermotropic mesomorphism, however for n = 12–18 a single enantiotropic mesophase was observed on both cooling from the isotropic liquid and heating from the crystalline solid. In general, for all the mesomorphic salts, on cooling from the isotropic liquid, small bâtonnets first form, but rapidly disperse into a homeotropic texture (i.e.the optical axis is perpendicular to the slide) with birefringence only observed around airbubbles and at the edges of the liquid. When the mesophase film is deformed, regions of focal conic and oily-streak texture are observed indicative of a layered smectic A (SA) mesophase.28 In addition, free standing droplets display a typical spherulitic texture. When the salts are heated between two glass microscope slides (as opposed to a more typical microscope slide and cover slip) spontaneous formation of a single homeotropic Fig. 2 The DSC trace of [C14-mim][BF4] for the second heating (upper) and cooling (lower) cycles showing the characteristically sharp Cr-SA (smectic) and SA-Iso transitions. monodomain is frustrated and a fan-like texture was observed (see Fig. 3). These mesophases are miscible with those formed by the corresponding mesomorphic 1-alkyl-3-methylimidazolium and N-alkylpyridinium hexafluorophosphate and tetrachloronickelate salts, which had been assigned as smectic A (SA) on the basis of the textures observed by polarised optical microscopy.16,17 However, considering both the amphiphilic structure of the mesogen and the optical texture, this mesophase could be equally well described as an anhydrous La lamellar phase by analogy with the lamellar phases 29,30 formed by anhydrous soaps.31,32 On the basis of the current observations, it is not possible to diVerentiate between these two descriptions of the mesophase; further small angle X-ray diffraction studies are being considered to investigate the structure of these mesophases.It is important to note that although the tetrafluoroborate salts (n > 6) have only low solubility in water and lyomesomorphism was not observed with any of the salts at temperatures below their respective melting points using the Lawrence penetration experiment,33 the thermotropic mesophase displayed by all the salts with n > 11 was observed to swell and become more extensive on contact with water.This clearly indicates a degree of water solubility within the ionic liquid crystal mesophase and is in contrast to the hexafluorophosphate salts which show no miscibility with water and the more water soluble chloride and bromide salts which do exhibit general surfactant properties and form lyotropic mesophases in water 34 with behaviour similar to that of the analogous N-alkylpyridinium salts.The melting points increase only relatively slowly with increasing (n = 12–18) chain length (i.e. from 25 to 56.4 8C) whereas the clearing temperatures show a marked dependency on the chain length (from 39.4 to 214 8C) leading to an increased thermal range of the mesophase from a modest 14.5 8C for [C12-mim][BF4] to 149 8C for [C18-mim][BF4], the longest chain compound studied. The transition temperatures and energies are shown in Table 3 and Fig. 1. A small, but significant odd–even eVect was observed in the melting points; the odd chain length compounds have melting points which are increased by ca. 10 8C compared to those of their immediate neighbours with even length chains. This is due to the deviation of the even length chains from the linear structure of the favoured all-trans configuration found for odd chain lengths. It is notable that an enantiotropic mesophase is observed for the compounds with n = 12 and 13, in contrast to the [Cn-mim][PF6] ionic liquid crystals 17 where mesomorphism is only observed from n = 14.The phase behaviour follows the same pattern observed for the previously studied 1-alkyl-3-methylimidazolium chloride, 16 tetrachlorometalate 16 and hexafluorophosphate 17 ionic liquid crystals. Compared to the hexafluorophosphate salts, the Fig. 3 The fan-like texture of the mesophase of [C14-mim][BF4] under crossed polarisers at 126 8C.2138 J. Chem. Soc., Dalton Trans., 1999, 2133–2139 tetrafluoroborate salts display a wide thermal range of the mesophase though not as extensive as those for the chloride and tetrachlorometalate salts. 16 The enthalpy changes on melting (see Table 3) are also similar to those published for N-octadecylpyridinium chloride 16 and for 1-alkyl-3-methylimidazolium and 1-alkylpyridinium hexafluorophosphates,17 and are relatively large (20–30 kJ mol21) which suggests a significant structural change on melting.In contrast, the clearing enthalpy is small and probably is solely due to weak van der Waals interactions between alkyl chains. On cooling, all the examples show extensive supercooling before crystallisation, which is typical for these imidazolium systems. Powder X-ray diVraction studies. Although it was not possible to use variable temperature X-ray powder diVraction to study the mesophase structure of the liquid crystalline salts, the crystal powder pattern of the representative [C14-mim][BF4] salt was determined at room temperature and the unit cell parameters were calculated.The salt was determined to have a triclinic space group with cell parameters a = 7.69, b = 10.92, c = 30.23 Å, a = 90.9, b = 102.0, g = 88.18; the powder diVraction pattern is shown in Fig. 4. The unit cell parameters are consistent with an interdigitated bilayer structure, comparable to that typically observed for other long-chain imidazolium and N-alkylpyridinium salts.The c axis is long in comparison with the corresponding hexafluorophosphate salt which has a monoclinic unit cell (c = 24.2 Å), which is probably due to diVerences in the interactions between [BF4]2 or [PF6]2 with the cationic imidazolium headgroups. However, in the absence of a single crystal structure of any related tetrafluoroborate salt few conclusions can be safely drawn. Thermogravimetric analysis. The thermal stability of representative [bmim]1, [C15-mim]1 and [C18-mim][BF4] salts was determined by thermogravimetric analysis (TGA) over the temperature range 30–500 8C under N2, heating at 10 8C min21.In each case, the samples showed no weight loss below 250 8C. The salts [bmim][BF4] and [C18-mim][BF4] show a small weight loss of 3.5 (7.9) and 7.9 wt% (24.5 g mol21) respectively between 280 and 320 8C. No further degradation was observed until the temperature reached 360 8C, from which point progressive decomposition and mass loss occurred in all three samples between 360 and 450 8C.Thus, the tetrafluoroborate ionic liquids are more thermally stable than the halide and tetrachloroaluminate( III) analogues. Conclusion We have described the preparation and properties of a series of tetrafluoroborate salts which are liquid or liquid crystalline at, Fig. 4 Powder X-ray diVraction pattern of [C14-mim][BF4] at 25 8C. or close to, ambient temperature. The salts described can be separated into three distinct groups on the basis of their thermal behaviour: short chain, crystalline solids with strong interactions in the solid state and relatively high melting points; intermediate ionic liquids, which have a wide liquid range (up to 400 8C) and form glasses on cooling due to the reduction of lattice energies through disruption of packing eYciencies; and mesomorphic salts which have an amphiphilic structure, long alkyl chains (n > 11) and properties that are governed by the microphase separation of the hydrophilic and hydrophobic components of the amphiphilic cations which induces mesophase formation and crystallisation in layered structures. As the chain length increases the glass transition temperatures decrease rapidly due to an elongation of the molecular length, disruption of crystal packing with increased flexibility of the alkyl chains and a reduction in the lattice energy, coupled (on increasing alkyl chain) with a distinct tendency towards glass formation rather than crystallisation on cooling.The glass transition temperatures all tend towards 290 8C, which also represents the lower limit of glass formation for the [emim]Cl– AlCl3 system 27 and for [Cn-mim][PF6] salts.17 On increasing alkyl-chain length, the attractive van der Waals interactions increase and from n = 10 the salts start to exhibit amphiphilic character. This produces an increase in the melting point and also stabilises microphase separation of the hydrophilic ionic headgroups and hydrophobic alkyl chains which leads to increasing orientational order and the layered structure observed in the solid state.Crystalline rather than glassy solids are formed with a discontinuity in the melting points, although relative to other long-chain salts the melting points, at between 25 and 60 8C, are low. On melting the long-chain salts display a single ordered thermotropic mesophase, the clearing points increasing with chain length. For the mesomorphic salts the variation in melting points is very small compared with the dramatic increase in clearing point with chain length.This is consistent with the behaviour observed for other alkylimidazolium salts containing e.g. [NiCl4]22, [CoCl4]22, [PF6]2 and Cl2, although data for a complete series of alkyl chain substituents has not previously been correlated. In the case of the tetrafluoroborate salts the melting points are lower than in the previous examples. These observations follow the general, empirically observed trends for alkylimidazolium salts that the melting points decrease on changing the anion in the order Cl2 > [PF6]2 > [BF4]2.The liquid range of the short chain salts is exceptionally wide; the preparation of the short chain salts [emim][BF4] and [bmim][BF4] and their use as solvents for catalytic reactions and electrochemistry have previously been reported. However, dif- ficulties in preparing pure [emim][BF4] (especially removing halide impurities) means that future work is most likely to use the longer chain homologues, especially with [bmim]1 and [C6- mim]1 cations.There are interesting changes in the solvent properties of the salts with increasing chain length; for example, the transition from water miscibility to immiscibility in a progressive, controlled manner. The low melting points of the longer chain, mesomorphic salts enables access to liquid crystalline solvents with large mesophase ranges from close to ambient temperature and the potential to use these materials as ordered reaction media.Acknowledgements The authors would like to thank Susanne Johnston and Mark Nieuwenhuyzen for assistance with synthesis, DSC and powder X-ray measurements and Maria Torres and Anne Stark for water and halide determinations. We also thank the ERDF Technology Development Programme and the QUESTOR Centre (J. D. H.) for financial support, and the EPSRC and Royal Academy of Engineering for the award of a Clean Technology Fellowship (to K.R. S.). The referees are thankedJ. Chem. Soc., Dalton Trans., 1999, 2133–2139 2139 for invaluable comments, especially regarding the nature of the mesophases formed. References 1 C. L. Hussey, Adv. Molten Salt Chem., 1983, 5, 185. 2 C. L. Hussey, Pure Appl. Chem., 1988, 60, 1763. 3 K. R. Seddon, Molten Salt Forum, 1998, 5, 53. 4 M. Freemantle, Chem. Eng. News, 1998, 76, 32. 5 J. S. Wilkes and M. J. Zaworotko, J.Chem. Soc., Chem. Commun., 1992, 965. 6 E. I. Cooper and E. J. M. O’Sullivan, in Molten Salts, eds. R. J. Gale, G. Blomgren and H. Kojima, The Electrochemical Society Proceedings Series, Pennington, NJ, 1992, vol. 16, p. 386. 7 P. Bonhôte, A. P. Diaz, N. Papageorgiou, K. Kalyanasundaram and M. Grätzel, Inorg. Chem., 1996, 35, 1168. 8 M. Fields, F. V. Hutson, K. R. Seddon and C. M. Gordon, World Pat., WO 98/06106, 1998. 9 Y. Chauvin, Actual. Chem., 1996, 7, 44. 10 Y.Chauvin, L. Mussmann and H. Olivier, Angew. Chem., Int. Ed. Engl., 1995, 34, 2698. 11 P. A. Z. Suarez, J. E. L. Dullius, S. Einloft, R. F. de Souza and J. Dupont, Polyhedron, 1996, 1217; A. L. Monteiro, F. K. Zinn, R. F. de Souza and J. Dupont, Tetrahedron-Asymmetry, 1997, 8, 177; P. A. Z. Suarez, J. E. L. Dullius, S. Einloft, R. F. de Souza and J. Dupont, Inorg. Chim. Acta, 1997, 255, 207. 12 C. Nanjundiah, S. F. McDevitt and V. R. Koch, J. Electrochem. Soc., 1997, 144, 3392. 13 P. A. Z. Suarez, V. M. Selbach, J. E. L. Dullius, S. Einloft, C. M. S. Piatnicki, D. S. Azambuja, R. F. de Souza and J. Dupont, Electrochim. Acta, 1997, 42, 2533; P. A. Z. Suarez, S. Einloft, J. E. L. Dullius, R. F. de Souza and J. Dupont, J. Chim. Phys., 1998, 95, 1626. 14 P. Bonhôte, A. P. Dias, N. Papageorgiou, K. Kalyanasundaram and M. Grätzel, Inorg. Chem., 1996, 35, 1168. 15 See for example, http://www.ch.qub.ac.uk/staV/personal/krs/recent– papers.html 16 C. J. Bowlas, D. W. Bruce and K. R. Seddon, Chem. Commun., 1996, 1625. 17 C. M. Gordon, J. D. Holbrey, A. R. Kennedy and K. R. Seddon, J. Mater. Chem., 1998, 8, 2627. 18 K. M. Lee, C. K. Lee and I. J. B. Lin, Chem. Commun., 1997, 899. 19 K. R. Seddon, J. Chem. Technol. Biotechnol., 1997, 68, 351. 20 S. Christie, S. Subramanian, L. Wang and M. J. Zaworotko, Inorg. Chem., 1993, 32, 5415. 21 J. Fuller, R. T. Carlin, H. C. De Long and D. Haworth, J. Chem. Soc., Chem. Commun., 1994, 299. 22 J. E. L. Dullius, P. A. Z. Suarez, S. Einloft, R. F. de Souza, J. Dupont, J. Fischer and A. De Cian, Organometallics, 1998, 17, 815. 23 A. Seidell, Solubilities of Inorganic and Metal-organic Compounds, American Chemical Society, Washington DC, 1995, vol. II. 24 J. Fuller, R. T. Carlin and R. A. Osteryoung, J. Electrochem. Soc., 1997, 144, 3881. 25 K. R. Seddon, M. J. Torres and A. Stark, unpublished results. 26 A. Elaiwi, P. B. Hitchcock, K. R. Seddon, N. Srinivasan, Y.-M. Tan, T. Welton and J. A. Zora, J. Chem. Soc., Dalton Trans., 1995, 3467. 27 A. A. Fannin, L. A. King, J. S. Landers, B. J. Piersma, D. J. Stech, R. L. Vaughan and J. S. Wilkes, J. Phys. Chem., 1986, 88, 2614. 28 G. W. Gray and J. W. Goodby, Smectic Liquid Crystals, Textures and Structures, Leonard Hill, Glasgow, 1984. 29 P. A. Winsor, in Liquid Crystals and Plastic Crystals, eds. G. W. Gray and P. A. Winsor, Ellis Horwood, Chichester, 1974, vol. 1, pp. 199– 287. 30 G. J. T. Tiddy, Phys. Rep., 1980, 57, 1. 31 A. Skoulis and V. Luzzati, Acta Crystallogr., 1961, 14, 278. 32 V. Busico, P. Cernicchiaro, P. Corradini and M. Vacatello, J. Phys. Chem., 1983, 87, 1631. 33 A. C. S. Lawrence, in Liquid Crystals 2, ed. G. H. Brown, Gordon and Breach, London, 1969, part 1, p. 1. 34 J. D. Holbrey and K. R. Seddon, unpublished observations. Paper 9/02818H

ISSN:1477-9226    

DOI:10.1039/a902818h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 2139 Spectroscopic and structural studies on adducts of silver(I) cyanide with ER3 ligands (E 5 P, As or Sb) Graham A. Bowmaker,a EVendy,b,c Jason C. Reid,a Clifton E. F. Rickard,a Brian W. Skelton b and Allan H. White b a Department of Chemistry, University of Auckland, Private Bag 92019, Auckland, New Zealand b Department of Chemistry, The University of Western Australia, Nedlands, W.A. 6907, Australia c Jurusan Pendidikan Kimia, FPMIPA IKIP, Malang, Jalan Surabaya 6, Malang, 65145, Indonesia The adducts AgCN:EPh3 (E = P, As or Sb) (1 :2) and AgCN:PPh3 (1 : 1) have been characterized by room temperature single crystal structure determination and low frequency vibrational spectroscopy.The compounds AgCN:EPh3 (E = P or As) (1 : 2) [as pyridine (py) monosolvates] are isomorphous and of the form [(Ph3E)3- Ag(CN)Ag(EPh3)3][Ag(CN)2], an aqua–2-methylpyridine solvate and an acetonitrile solvate also being structurally characterized for the E = P adduct.For E = Sb the complex is a simple single stranded polymer ? ? ? Ag(SbPh3)2(CN)Ag(SbPh3)2(CN) ? ? ? as is AgCN :PPh3 (1 : 1), ? ? ? Ag(PPh3)2(CN)Ag(CN)Ag(PPh3)2 ? ? ?. An adduct of the form AgCN:PR3 (1 : 1) is also obtained for tris(2,4,6-trimethoxyphenyl)phosphine (tmpp) as a pyridine hemisolvate, now ionic but with the pyridine loosely co-ordinated to the anion, [Ag(tmmp)2][Ag(CN)2- (py)]. By contrast, in the AgCN :P(o-tol)3 (tol = o-tolyl) :py(1 : 0.5 : 0.5) adduct, Ag{P(o-tol)3}(py) moieties are linked in a single-stranded polymer by NCAgCN units, ? ? ? Ag[P(o-tol)3](py)(NCAgCN) ? ? ?.The structural characterization of [Ag(PPh3)3(CN)] has also been made. The far-infrared spectra of AgCN?2AgNO3, [Ag(SbPh3)2- (CN)] and [Ag(SbPh3)3(CN)] show n(AgX) bands (X = CN) at 435, 358 and 310 cm21. The relationship between the n(AgX) wavenumbers and the Ag]X bond lengths r(AgX) for the case of terminally bound CN groups has been established, and this has the same form as that found previously for a range of AgX complexes (X = Cl, Br or I) with phosphine and arsine ligands.The n(AgX) wavenumbers for the CN-bridged compounds AgCN?2AgNO3 and [Ag(SbPh3)2(CN)] are about 50 cm21 higher than those for compounds with terminal Ag]CN bonds of similar bond length. Similar behaviour is observed for the bridging CN groups in the cations of the ionic 1 : 2 complexes [{Ag(EPh3)3}2CN]1[Ag(CN)2]2?3py (E = P or As), where the n(AgX) wavenumbers (301, 354 cm21 respectively) are about 80 cm21 higher than those predicted for terminal bonding. Structurally defined adducts of silver(I) cyanide with Group V unidentate donors are very sparse, being restricted to a recent report of the 1 : 3 triphenylstibine adduct [Ag(SbPh3)3(CN)],1 presumably a consequence of a tendency of such species to polymerize by virtue of the ambidentate nature of the cyanide ligand.In recent work2,3 we have extended the array to 1 : 1 and 1 : 2 adducts with tricyclohexylarsine, the 1 : 1 adduct obtained in various forms being one-stranded infinitely polymeric.We have extended these studies to encompass further ligand types, and record here syntheses, room temperature single crystal determinations, and low frequency vibrational spectroscopy for adducts of silver cyanide with bases of the form EPh3 for E = P, As or Sb of 1 : 2 AgCN:EPh3 stoichiometry, and of 1 : 1 stoichiometry for E = P, defining novel species incorporating bridging cyanide moieties, most being solvated by non-co-ordinating pyridine base.Adducts of 1 : 1 and 1 : 0.5 stoichiometry, whose characterization rests on X-ray studies, have also been obtained with with tmpp [= tris(2,4,6-trimethyoxyphenyl)phosphine] and P(o-tol)3 (o-tol = o-tolyl) solvated with pyridine, but with the pyridine more or less co-ordinated, yielding novel arrays in these instances. We record the results of these studies herein, together with the structure determination of [Ag(PPh3)3(CN)].Experimental Syntheses Colourless crystalline materials were obtained in all cases by dissolving appropriate millimolar stoichiometries of silver(I) cyanide and pnicogen ER3 base in pyridine base (5–10 cm3) with warming and allowing to cool slowly, in some cases crystalline material depositing only after volume reduction by evaporation. The materials are generally solvated, the solvent in most cases being readily lost, so that useful analyses were obtained for a limited number of species only, AgCN:PPh3:py (1:2:1) (Found: C, 67.6; H, 5.0; N, 4.0.C42H35AgN2P2 requires C, 68.40; H, 4.78; N, 3.80%) and AgCN:AsPh3:py (1:2:1) (Found: C, 60.6; H, 4.5; N, 3.6. C42H35AgAs2N2 requires C, 61.11; H, 4.27; N, 3.39%). Crystals for the X-ray work were mounted in capillaries. In addition to the general method involving reactions carried out in pyridine base as described above, the following products, obtained from acetonitrile solvent, require explicit comment. [(Ph3P)3Ag(CN)Ag(PPh3)3][Ag(CN)2]?2MeCN and [{(Ph3P)2- Ag(CN)Ag(CN)}n]. Silver cyanide (0.265 g, 2.0 mmol) and PPh3 (1.558 g, 6.0 mmol) were dissolved, with stirring, in boiling MeCN (10 cm3).The resultant solution was filtered, and on cooling a mixture of white ‘needle-like’ crystals and white ‘cube-shaped’ crystals formed in the filtrate. A portion of the product was filtered oV and air-dried.During the drying process the cube-shaped crystals rapidly became opaque. Some of the needle-like white crystals were physically separated from the bulk of the dried sample and subjected to an X-ray study, showing them to be a 1 : 1 AgCN:PPh3 complex. A ‘cube-shaped’ specimen was collected from the mother-liquor for an X-ray study, and determined to be the 1: 2 AgCN:PPh3 adduct, [{Ag(PPh3)3}2CN][Ag(CN)2]?2MeCN. [Ag(SbPh3)2CN]. This was prepared by dissolution of AgCN (0.094 g, 0.70 mmol) and SbPh3 (0.773 g, 2.2 mmol) in boiling MeCN (25 cm3).The resultant solution was filtered, and on2140 J. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 cooling white crystals of the complex formed in the filtrate. M.p. 147–151 8C (Found: C, 53.0; H, 3.4; N, 1.6. C37H30AgNSb2 requires C, 52.90; H, 3.60; N, 1.67%). This sample was shown by X-ray diVraction measurements to be identical to that obtained from pyridine solution. AgCN?2AgNO3.This complex was prepared for IR studies by the following reaction: AgNO3 (3.42 g, 20 mmol) was dissolved in water (10 cm3). To the resulting solution was added AgCN (1.34 g, 10 mmol). The mixture was stirred, with heating for 4 h. After this time some of the AgCN remained undissolved, and the solution was filtered. Upon cooling, fine white crystals of the complex formed in the filtrate. These were filtered oV. Yield 0.335 g (Found: C, 2.6; H, 0.0; N, 8.6. CAg3N3O6 requires C, 2.54; H, 0.00; N, 8.87%).Structure determinations General procedures and caveats concerning cyanide modelling are given in ref. 3, specific details as follows; as a general particular caveat here we reiterate a common inability to distinguish cyanide C, N atoms crystallographically in bridging arrays, which were modelled as composites. (a) AgCN:EPh3:py (1:2:1) (×3) º [(Ph3E)3Ag(CN)Ag- (EPh3)3][Ag(CN)2]?3solv. º C111H90Ag3E6N3?3solv. (i) solv = py, E = P or As. Triclinic, space group P1� (Ci 1, no. 2), Z = 1. a (E = P). C126H105Ag3P6N6, M = 2212.8, a = 15.335(6), b = 14.338(5), c = 13.45(1) Å, a = 69.30(5), b = 85.56(6), g = 85.69(3)8, U = 2755 Å3, Dc = 1.33 g cm23, F(000) = 1134, mMo = 6.0 cm21, specimen 0.26 × 0.40 × 0.65 mm, A*min,max = 1.16, 1.24, 2qmax = 508, N = 9653, No = 5809, R = 0.046, R9 = 0.048. b (E = As). C126H105Ag3As6N6, M = 2476.5, a = 15.482(4), b = 14.414(3), c = 13.650(6) Å, a = 68.80(3), b = 86.06(3), g = 85.83(2)8, U = 2830 Å3, Dc = 1.45 g cm23, F(000) = 1242, mMo = 22.0 cm21, specimen 0.22 × 0.22 × 0.41 mm, A*min,max = 1.50, 1.63, 2qmax = 458, N = 7280, No = 3699, R = 0.050, R9 = 0.050.Variata. One of the pyridine solvent molecules is disposed about a crystallographic inversion centre, modelled with C/N composites at the putative nitrogen sites (both E = P/As). (ii) 3S = 2(2-Methylpyridine =7H7N)?H2O. C123H106Ag3P6- N5O, M = 2179.7, triclinic, space group P1� , a = 25.847(6), b = 14.743(5), c = 14.545(3) Å, a = 85.48(3), b = 84.96(2), g = 79.79(3)8, U = 5422 Å3, Dc (Z = 2) = 1.34 g cm23, F(000) = 2236, mMo = 6.8 cm21, specimen 0.55 × 0.20 × 0.26 mm, A*min,max = 1.13, 1.19, 2qmax = 47.58, N = 15 688, No = 8489, R = 0.049, R9 = 0.050.Variata. Solvent molecules were modelled as shown, the 2- methylpyridine groups being modelled as constrained geometry rigid bodies, two being disordered about crystallographic inversion centres, and site occupancies set at unity after trial refinement.The bridging cyanide was modelled as disordered about a crystallographic inversion centre. (b) AgCN:PPh3 :MeCN (1:2:2/3) (×3) º [{Ag(PPh3)3}2- CN][Ag(CN)2]?2Me3CN. C115H96Ag3N5P6, M = 2057.4, triclinic, space group P1� , a = 13.635(3), b = 13.934(3), c = 16.747(3) Å, a = 98.92(3), b = 110.04(3), g = 112.95(3)8, U = 2595 Å3, Dc (Z = 2) = 1.317 g cm23, F(000) = 1052, mMo = 7.0 cm21, specimen 0.55 × 0.46 × 0.38 mm, 2qmax = 588, N = 11 515, No = 10 417, R = 0.037, R9 = 0.130.Variata. Data were collected on a Siemens SMART diVractometer with a CCD area detector at a sample temperature of 203(2) K. A full sphere of data was collected by a series of w scans and a semiempirical absorption correction applied by comparing equivalent reflections. The bridging cyanide ion lies on a centre of symmetry and was refined using composite scattering forms. The silver atom of the [Ag(CN)2]2 group also lies on a centre of symmetry and the cyanide group was modelled as C-bound.(c) AgCN:SbPh3 (1 : 2). C37H30AgNSb2, M = 840.0, monoclinic, space group P21/c (C5 2h, no. 14), a = 14.249(5), b = 9.314(5), c = 28.81(1) Å, b = 117.55(2)8, U = 3390 Å3, Dc (Z = 4) = 1.65 g cm23, F(000) = 1428, mMo = 20.1 cm21, specimen 0.29 × 0.96 × 0.40 mm, A*min,max = 1.52, 2.51, 2qmax = 508, N = 5943, No = 4495, R = 0.048, R9 = 0.057. Variata. Phenyl ring 22 was modelled as disordered over two sets of sites, occupancies set at 0.5 each after trial refinement, isotropic thermal parameter forms being used for the disordered atoms.Refinement behaviour suggests localization of C, N of the cyanide bridge in this case as shown, consistent with diVerent bond lengths to the silver. (d) AgCN:PPh3 (1: 1) (×2). C38H30Ag2N2P2, M = 792.4, monoclinic, space group P21/c, a = 9.708(2), b = 20.159(8), c = 19.152(4) Å, b = 114.96(2)8, U = 3398 Å3, Dc (Z = 4) = 1.55 g cm23, F(000) = 1584, mMo = 12.8 cm21, specimen 0.28 × 0.32 × 0.09, A*min,max = 1.14, 1.22, 2qmax = 608, N = 7988, No = 4437, R = 0.046, R9 = 0.047.Variata. Cyanide C, N were indistinguishable on the basis of refinement, but were modelled as C-bound within the NCAgCN component. (e) AgCN:P(o-tol)3 : py (1: 0.5 : 0.5) (×2). C28H26Ag2N3P, M = 651.2, monoclinic, space group P21/c, a = 10.335(8), b = 19.823(4), c = 15.340(4) Å, b = 90.16(4)8, U = 3143 Å3, Dc (Z = 4) = 1.54 g cm23, F(000) = 1296, mMo = 13.2 cm21, specimen 0.42 × 0.16 × 0.28 mm, A*min,max = 1.21, 1.42, 2qmax = 508, N = 5520, No = 3518, R = 0.036, R9 = 0.036.Variata. Atoms C, N were assigned as localized, consistent with an NCAgCN component model, on the basis of refinement behaviour. ( f ) AgCN:tmpp:py (1:1:0.5) (×2) 5 [Ag(tmpp)2][Ag- (CN)2(py)]. C61H71Ag2N3O18P2, M = 1412, monoclinic, space group C2/c (C6 2h, no. 15), a = 19.229(6), b = 22.672(3), c = 16.08(1) Å, b = 115.47(4)8, U = 6327 Å3, Dc = 1.48 g cm23, F(000) = 2904, mMo = 7.4 cm21, specimen 0.25 × 0.08 × 0.41, A*min,max = 1.06, 1.24, 2qmax = 508, N = 4943, No = 2974, R = 0.047, R9 = 0.044.Variata. In space group C2/c the anion is modelled with the pyridine, disposed on a crystallographic 2 axis, approaching the (NC)Ag(CN) anion normal to its quasi-axis. The silver atom is slightly displaced from the 2 axis, Ag ? ? ? Ag being 0.28(2) Å; although thermal motion of associated atoms is large, no associated disorder was resolvable. Cyanide C, N were assigned as localized on the basis of refinement behaviour within the NCAgCN component.( g) AgCN:PPh3 (1 : 3). C55H45AgNP3, M = 920.8, monoclinic, space group P21/n (C5 2h, no. 14, variant), a = 18.843(7), b = 13.748(6), c = 17.630(3) Å, b = 95.69(2)8, U = 4545 Å3, Dc (Z = 4) = 1.35 g cm23, F(000) = 1896, mMo = 5.2 cm21, specimen 0.10 × 0.12 × 0.40 mm, A*min,max = 1.05, 1.07, 2qmax = 508, N = 7983, No = 3313, R = 0.048, R9 = 0.042. Variata. As described elsewhere 1 this complex is isomorphous with a considerable family of [Ag(EPh3)3X] arrays, and was refined in the setting described therein; C, N were assigned as localized on the basis of refinement behaviour, the anion being modelled as C-bound.CCDC reference number 186/985. See http://www.rsc.org/suppdata/dt/1998/2139/ for crystallographic files in .cif format. Spectroscopy Infrared spectra were recorded at 4 cm21 resolution as NujolJ. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 2141 mulls between KBr plates on a Digilab FTS-60 Fourier transform spectrometer employing an uncooled deuteriotriglycine sulfate detector.Far-infrared spectra were recorded at 4 cm21 resolution at room temperature as polythene discs on a Digilab FTS-60 Fourier transform infrared spectrometer employing an FTS-60V vacuum optical bench with a 6.25 mm mylar film beam splitter or a 5 lines mm21 wire mesh beam splitter, a mercury lamp source and a pyroelectric triglycine sulfate detector. Results and discussion Crystal structures Adducts of silver(I) cyanide with Group V tertiary unidentate bases ER3, E � N, have been defined for a number of stoichiometries 1 : n; in a number of cases the compounds have been obtained only with diYculty, as single crystals in isolation among a bulk matrix possibly composed of mixtures of various poly- or oligo-meric species rather than specimens representative of bulk samples.In cases where the latter have been obtained, usually from a pyridine base solution, solvation is frequent with the solvent often being lost rather readily under ambient conditions.For n = 3, only the SbPh3 adduct is recorded in the literature as a structurally characterized example;1 in the course of the present work the PPh3 analogue was isolated in small quantities from pyridine solution, its structure determination being recorded herewith. The structural parameters for the PPh3 and SbPh3 complexes are compared in Table 1. Despite the limited data available for the PPh3 complex (only the structure), the result is of considerable interest.The complex crystallizes in a form widespread throughout [Ag(EPh3)3X] arrays (no copper analogues in this form are known), being monoclinic, P21/n, Z = 4, one independent molecule comprising the asymmetric unit of the structure, refinement indicating not entirely unambiguously (thermal motion on the peripheral atom being rather high) that the cyanide moiety is linearly C-(rather than N-) bound.This P21/n array is the most common form thus far defined for unsolvated [M(EPh3)3X], and is found (for M = Ag) for E = P and X = CN (here), I, NCO, ONO2 or FBF3, for E = As and X = Br, I, NCO, SCN or ONO2 and for E = Sb and X = I, CN, ONO2 or SCN, as tabulated in ref. 1 (interestingly, the example E = As, X = CN as yet is not recorded). This structural type is remarkable among those generally found for [M(EPh3)3X] arrays, by virtue of the substantial perturbations from 3m symmetry found in the E3MX environment and con- firmed at low temperature for the E = P iodide recently.4 While, seemingly, the general eVect must be ascribed to lattice forces, it is of interest that here, in the context of what is probably the most tightly bound anionic moiety, the excursions are at their most extreme [e.g., cf.the E = P iodide 5 where Ag]P are 2.780(3), 2.573(2), 2.544(2) Å at room temperature] with Table 1 Selected geometries (distances in Å, angles in 8) for [Ag- (PPh3)3(CN)]. Values for the isomorphous E = Sb analogue (second entry, from ref. 1) are given for comparison Ag]C C]N Ag]P(1) N]C]Ag C]Ag]P(1) C]Ag]P(2) C]Ag]P148(8), 2.09(3) 1.14(1), 1.14(4) 3.087(2), 2.843(3) 175.7(8), 177(2) 100.4(2), 106.8(5) 117.0(2), 116.8(5) 114.3(2), 115.1(6) Ag]P(2) Ag]P(3) P(1)]Ag]P(2) P(1)]Ag]P(3) P(2)]Ag]P(3) 2.539(2), 2.737(2) 2.529(2), 2.734(2) 104.89(7), 105.62(7) 100.97(6), 103.38(6) 115.87(9), 107.83(7) Angles Ag]P(n)]C(n11, n21, n31) are: n = 1, 109.8(2), 108.2(3), 130.0(3); n = 2, 112.3(2), 119.3(2), 114.2(2); n = 3, 114.7(2), 114.8(2), 115.4(2)8.Ag]P(1) well on the way to dissociation; for the CAgP(1,2) array here, the angle sum is 347.28. By contrast, in the E = Sb cyanide (see Table 1), Ag]E are more similar with angles about the phosphorus rather more tetrahedral; residual strain is still evident, however, inasmuch as Ag]Sb(1)]C(131) is 133.3(5)8, cf. 130.0(3)8 in the present compound.For n = 2 the present work defines the totality of the array for E = P, As or Sb. Unlike the silver halide 1 : 2 counterparts,6,7 the present are not discrete mono- or bi-nuclear entities, but, rather more complex ionic (E = P or As) or polymeric aggregates. As crystallized from pyridine, solutions of 1 : 2 AgCN:EPh3 (E = P or As) yield isomorphous adducts of 1:2:1 AgCN:EPh3:py stoichiometry, the array being ionic m-cyano-bis{tris[triphenylphosphine( arsine)]silver(I)} dicyanoargentate tris(pyridine) solvate, [(Ph3E)3Ag(CN)Ag(EPh3)3][(NC)Ag(CN)]?3py, one half of this array modelling the asymmetric unit of the structure, entailing the location of crystallographic inversion centres at the centres of cation, anion and one of the pyridine moieties, with concomitant disorder of the nitrogen atoms of the bridging cyanides of the cations with their associated carbon atoms, and that of the centrosymmetric pyridine, with its inversionrelated carbon; the central silver atom of the centrosymmetric anion is also located on a crystallographic inversion centre.The structure of the [(Ph3P)3Ag(CN)Ag(PPh3)3]1 cation in the E = P pyridine solvate is shown in Fig. 1. The corresponding complex crystallized from 2-methylpyridine is not isomorphous, but is nevertheless similarly formulated with three solvent molecules, modelled as two 2-methylpyridine and one adventitious water molecule; in this structure the cation and solvent molecules do not conform to any crystallographic symmetry, but there are two independent anions, one with the silver atom located at a crystallographic centre of symmetry and the other located near a crystallographic centre of symmetry as modelled in the present space group, the two components of the silver atom being separated by 1.47 Å.By contrast, the E = P adduct crystallized from acetonitrile, although not isomorphous with the pyridine solvates, shows the same centrosymmetric character as the latter, but with only two solvate (MeCN) molecules in the unit cell.Geometries about the cationic silver atoms for these species are proposed in Table 2; given the deficiency in Table 1 in respect of data for any comparable arsenic complex, the present array would seem to oVer the possibility of an alternative approach to its access. However, the comparability of the phosphine complex data between Tables 1 and 2 is poor, in consequence of the previously noted perturbations; the data of Table 2, showing greater evenness in the parameters of the various metal atom environments, appear to oVer a more useful approach to Fig. 1 Projection of the [(Ph3P)3Ag(CN)Ag(PPh3)3]1 cation normal to the Ag ? ? ? Ag axis, as in the [Ag(CN)2]2 tris(pyridine) solvate array.Cations of the arsenic and the bis(2-methylpyridine) aqua solvate analogues are similar2142 J. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 E3AgCN metal environment parameters, and, in this context, data for E = Sb are lacking.Of interest, despite the putative disorder in the CN bridge, is the suggestion that ·Ag]EÒ increases on passing from E = P to E = As adducts, accompanied by a diminution in Ag]C/N. Although the cations in these complexes provide the first examples of the unusual linearly bridged silver complexes [{Ag(EPh3)3}2CN]1, it may be noted that an analogous copper(I) species, with the tetracyanoethylene radical anion as the counter ion, has recently been reported.8 As noted above, the 1 : 2 E = Sb adduct does not conform to the E = P or As ionic structural type, but, instead, is an infinite one dimensional polymer ]Ag(SbPh3)2]CN]Ag(SbPh3)2]CN], one Ag(SbPh3)2(CN) unit comprising the asymmetric unit, with the polymer generated by the (short) 21 screw axis, b, of space group P21/c (Fig. 2). Here the refinement behaviour suggests the C, N components of the cyanide bridge to be localized as shown, with a significant disparity in Ag]C, N (Table 3), and C]Ag]Sb(x) significantly larger than N9]Ag]Sb(x).In view of the expected relative strength of the silver–cyanide interaction, the angle Sb]Ag]Sb, near the tetrahedral value, may be considered larger than expected. For the 1 : 1 AgCN:ER3 complexes the E = P, R = Ph array (obtained from acetonitrile solution) is the only structurally Fig. 2 Projection of the AgCN :SbPh3 (1 : 2) polymer, normal to the polymer axis Table 2 Selected geometries (distances in Å, angles in 8) for [(Ph3E)3Ag(CN)Ag(EPh3)3]1.Values are given in the following order: silver atoms 1, 2 of the E = P 2-methylpyridine solvate; the silver atom of the E = P acetonitrile solvate; the silver atom of the E = P, As pyridine solvates Ag]C/N Ag]E(1) Ag]E(2) Ag]E(3) C/N]Ag]E(1) C/N]Ag]E(2) C/N]Ag]E(3) E(1)]Ag]E(2) E(1)]Ag]E(3) E(2)]Ag]E(3) 2.27(1), 2.26(1); 2.279(3); 2.27(1), 2.19(1) 2.572(2), 2.554(2); 2.550(2); 2.572(2), 2.612(2) 2.560(2), 2.567(2); 2.5635(9); 2.559(2), 2.608(2) 2.550(2), 2.564(2); 2.5869(10); 2.573(2), 2.619(2) 107.2(3), 107.6(3); 108.47(8); 108.6(2), 110.6(3) 105.0(3), 106.0(3); 110.72(7); 106.7(2), 109.4(3) 105.5(4), 104.7(3); 101.80(7); 104.8(2), 105.6(3) 112.68(7), 112.95(7); 111.62(4); 113.24(6), 111.59(6) 112.83(7), 112.28(7); 112.42(4); 111.57(6), 109.84(6) 112.88(7), 112.60(7); 111.36(4); 111.34(5), 109.63(5) Table 3 Selected geometries (distances in Å, angles in 8) of AgCN: SbPh3 (1 : 2).Primed atoms belong to the adjoining moiety Ag]C Ag]N9 Ag]Sb(1) Ag]C]N C]N]Ag9 C]Ag]N9 C]Ag]Sb(1) 2.133(7) 2.263(7) 2.7208(9) 174.4(5) 168.7(6) 115.7(5) 120.6(2) Ag]Sb(2) C]N C]Ag]Sb(2) N9]Ag]Sb(1) N9]Ag]Sb(2) Sb(1)]Ag]Sb(2) 2.794(1) 1.122(9) 112.9(2) 104.6(2) 91.7(2) 107.24(4) defined example. The PPh3 complex, found in unsolvated form, is augmented by the example with tris(2,4,6-trimethoxyphenyl)- phosphine, the latter being a solvated ionic array, with the pyridine solvent molecule having an incipient co-ordinating role (see below).The AgCN:PPh3 (1 : 1) compound is, like AgCN: SbPh3 (1:2), a one dimensional polymer ]Ag(PPh3)2] NC]Ag]CN], also in space group P21/c, with the generator now the unit a translation, with Ag(EPh3)2 units spaced this time not by simple CN anionic units, but by NCAgCN anionic moieties (Fig. 3). In the latter, the assignment of C, N atoms is not clear-cut, despite the disparity of Ag(1)]C/N distances, and they are modelled as disordered.Of considerable interest here is the closing of the C/N]Ag]C/N angle about Ag(1), vis-à-vis the enlargement of the opposite P]Ag]P angle (Table 4). A related structure is observed for the 1 : 0.5 : 0.5 adduct of AgCN with P(o-tol)3 and pyridine. Adducts of AgX :PPh3:py base (1:1:1) stoichiometry are well known and generally of binuclear [(Ph3P)(py base)Ag(m-X)2Ag(py base)(PPh3)] form with four-co-ordinate silver(I).9 In the present case a linear polymer of the form ]Ag(PR3)(py)]NC]Ag]CN] is found, also with four-co-ordinate silver(I), similar to that of the AgCN: SbPh3 (1 : 2) adduct, but with the stoichiometry reduced to AgX:ER3 (1 : 0.5) by subsuming the additional AgCN unit required into a bridging NCAgCN unit, modelled with C]Ag bonds on the basis of the refinement behaviour (Fig. 4). Within the anionic NCAgCN unit the geometries are largely as expected: Ag(2)]C(1,2) 2.065(6), 2.063(6), C]N (anions 1,2) Fig. 3 Projection of the AgCN :PPh3 (1 : 1) (× 2) polymer, normal to the polymer axis Fig. 4 Projection of the AgCN :P(o-tol)3:py (2:1:1) polymer, normal to its axis Table 4 Selected geometries (distances in Å, angles in 8) of AgCN:PPh3 (1 : 1) [Ag(1) only] Ag(1)]N(1) Ag(1)]N(2) N(1)]Ag]N(2) N(1)]Ag]P(1) N(1)]Ag]P(2) 2.274(7) 2.380(5) 99.6(2) 116.8(2) 103.5(1) Ag(1)]P(1) Ag(1)]P(2) N]Ag]P(1) N(2)]Ag]P(2) P(1)]Ag]P(2) 2.466(2) 2.471(2) 101.4(2) 109.8(2) 123.45(6)J. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 2143 1.132(8), 1.125(8) Å, C(1)]Ag(2)]C(2) 173.6(2), angles at C(1,2) 175.6(5), 178.1(6)8; about the four-co-ordinate silver(I), Ag(1)]N(1,2) are somewhat disparate and shorter than Ag(1)]N(11) (py), with an irregular geometry with the largest angle, perhaps surprisingly, N(11)(py)]Ag(1)]P (Table 5).A single unit of the polymer comprises the asymmetric unit of the structure, the polymer being generated by the 21 screw. For the two polymeric structures that involve bridging NC]Ag]CN (Figs. 3 and 4) the ]Ag]NC]Ag]CN] chains are non-linear, mainly because of a ‘kinking’ at the silver atom bearing the organic ligands. This kinking may be a consequence of the fact that the bond angles about this silver atom must be less than 1808 because of its higher co-ordination number (four). Other structures (e.g. cyclic oligomers or zigzag chains) which could accommodate such a non-linear angle while maintaining all linear Ag]C]N or Ag]N]C angles are conceivable.However, the present structures, which involve kinks in an otherwise linear ]Ag]NC]Ag]CN] chain (with concomitant bending of the Ag]N]C bonds involving the four-co-ordinate silver), are apparently preferred. The 1: 1 adduct of AgCN: tmpp is a pyridine hemisolvate and, like many tmpp complexes of silver(I) salts, is ionic, ligand disproportionation yielding the familiar [Ag(tmpp)2]1 cation, with an [Ag(CN)2]2 anion solvated by pyridine, [Ag(CN)2- (py)]2 : bis[tris(2,4,6-trimethoxyphenyl)phosphine]silver(I) dicyano( pyridine)argentate(I) (Fig. 5). The silver atom in the cation is located on a crystallographic 2 axis which relates the two ligands; P]Ag]P is 177.38(7)8 and Ag]P 2.417(2) Å, one of the more precisely determined of such examples. The linear [NCAgCN]2 anion, refined as C-bound but non-definitively so, lies with the silver atom close to, but not precisely coincident with, a crystallographic 2 axis, lying 0.28(2) Å from its rotation image and 1.90(2), 2.17(2) Å from the pair of associated carbon atoms, whose locations must be considered as poorly determined in the circumstances, with their large thermal envelopes, as well as those of the associated nitrogen atoms encompassing probable disorder. The angle C]Ag]C is 168.4(4)8, presumably bent from linearity in response to the approach by the solvent pyridine, which lies about the two-fold axis that passes through the nitrogen and opposite carbon.The distance Ag ? ? ? N is long [3.16(2) Å], indicative that the interaction may simply be a consequence of a fortuitous packing, an adventitious void lying close by.Infrared spectroscopy Previous studies of copper(I) and silver(I) halide complexes have shown that IR spectroscopy provides a valuable tool with which to characterize them. In particular, the n(MX) vibrational modes in the far-IR region give characteristic patterns of bands for the various types of structures that can occur for these compounds, and the wavenumbers of these modes are sensitive to the strength of the M]X bonds.10–16 Thus, the n(MX) wavenumbers have been empirically correlated with the M]X bond length r via relationship (1) where b and m are conn/ cm21 = b(r/Å)2m (1) stants.12,16 Somewhat surprisingly, this relationship appears to be valid for vibrations involving both terminal and bridging Table 5 Selected geometries (distances in Å, angles in 8) of AgCN:P(o-tol)3:py (2:1:1) Ag(1)]N(1) Ag(1)]N(29) N(1)]Ag(1)]N(29) N(1)]Ag(1)]N(11) N(1)]Ag(1)]P 2.288(5) 2.332(5) 111.1(2) 96.3(2) 118.6(1) Ag(1)]N(11) Ag(1)]P N(29)]Ag(1)]N(11) N(29)]Ag(1)]P N(11)]Ag(1)]P 2.375(5) 2.451(2) 95.1(2) 106.8(1) 126.5(1) halogen atoms X.13,15,16 The most likely reason for this is that the M]X]M angles in the bridged complexes are close to 908, so that the vibrations of the two M]X bonds involved in the bridge are essentially uncoupled, and so are independent of each other.In contrast to the situation for the halide complexes, Fig. 5 (a) The [Ag(tmpp)2]1 cation projected normal to its axis, (b) the associated [Ag(CN)2(py)]2 array, projected normal to the 2 axis and (c) the unit cell contents, projected down c2144 J.Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 Table 6 The Ag]X bond distances and n(AgX), n(CN) IR band positions for some AgCN adducts containing (a) terminal and (b) bridging cyanide groups Species r(Ag]X)/Å Ref. n(AgX)/cm21 n(CN)/cm21 Ref.(a) Terminal CN AgCN (monomer) [Ag(CN)2]2 [Ag{P(C6H11)3}2(CN)] [Ag{As(C6H11)3}2(CN)] [Ag(SbPh3)3(CN)] 2.087 2.05 2.153 2.14 2.090 17 a 14 21 364 396 (IR) 360 (Raman) 288 311 310 2094 2140 (IR) 2146 (Raman) 2107 2111 2115 17 18 14 b 2 b b (b) Bridging CN AgCN?2AgNO3 [Ag(SbPh3)2(CN)] [{Ag(PPh3)3}2(CN)]1 [{Ag(AsPh3)3}2(CN)]1 2.04, 2.06 2.133, 2.263 2.27 2.19 19 bb ,c b,c 435 358 301 354 2119 2133 2108 2124 bbb ,c b,c a Average value for compounds containing this ion (see text).b This work. c For the compounds [{Ag(EPh3)3}2(CN)][Ag(CN)2]?3py (E = P or As); the IR spectra also showed n& (AgX) = 394 and n& (CN) = 2135, 2133 cm21 respectively due to the [Ag(CN)2]2 ion. relatively little is known about the M]X vibrational modes of Group 11 metal cyanides (X = CN). While the situation is expected to be similar in many respects to that of the halide complexes, the data that have been available to date have not permitted an examination of this point. Also, the cyano complexes diVer from the halide complexes in the important respect that, where cyanide bridging occurs, this normally involves a linear M]CN]M arrangement, so that the vibrations of the two ‘M]X’ bonds (M]C and M]N) will be strongly coupled, in contrast to the situation for terminal M]CN bonding where the n(M]X) frequency depends only on the properties of the M]C bond.Since the present study, together with other recently published or concurrent studies, have yielded examples of AgCN complexes with a variety of structures, we have attempted to assign the n(M]X) bands in the IR spectra of a number of complexes with a view to establishing the relationship between the n(M]X) frequencies and the structures of the complexes concerned.As a starting point we summarize the results for the relatively few AgCN complexes whose vibrational spectra have been previously assigned. These data are given in Table 6. Considering first the cases involving terminal Ag]CN bonding, the simplest example of this is the isolated triatomic Ag]CN molecule.No experimental data for this have been reported to date, but a theoretical study has yielded n& (AgX) = 364 cm21 for this species.17 The simplest species involving terminal Ag]CN bonding that has been well characterized experimentally is the linear [Ag(CN)2]2 ion. Here the situation is complicated by the presence of coupling between the two collinear Ag]C bonds, which gives rise to two n(AgX) modes {396 cm21, IR and 360 cm21, Raman for solid K[Ag(CN)2]}.18 The average n& (Ag]X) = 378 cm21 can be taken to be representative of this bonding situation, and it can be noted that this is quite close to the value calculated for isolated Ag]CN (Table 6).Two further species that involve well defined terminal Ag]CN bonding are [Ag{E(C6H11)3}2(CN)] (E = P or As).2,14 These compounds contain monomeric units with trigonal planar E2AgC bonding, and the IR spectra show n& (Ag]X) = 288 and 311 cm21 respectively for the Ag]C bonds.The Ag]C bond length data are available for all of the above compounds, although the value of 2.13 Å that was originally reported for K[Ag(CN)2] is considered to be rather unreliable,20 and results for a range of other compounds containing [Ag(CN)2]2 ions, including the ones in the present study, yield smaller values of about 2.05 Å.21 This, and the other experimental data for K[Ag(CN)2] and [Ag{E(C6H11)3}2(CN)], yield a good fit using equation (1) with b = 15 560, m = 5.17.This can be compared with the values previously determined for Ag]Cl complexes: b = 22 340, m = 5.09.16 The best-fit curve is compared in Fig. 6 to those for the AgX (X = Cl, Br or I) complexes. The experimental data points for the AgCl complexes are the same as those reported in the previous study, with the exception that a point for [AgCl2]2 has been included. This species shows two n(AgCl) modes (333 cm21, IR and 268 cm21, Raman in tri-n-butyl phosphate solution).22 The average n& (AgCl) = 301 cm21 and the Ag]Cl bond length 2.329(2) Å 23 yield a point that lies exactly on the curve for the other AgCl complexes (Fig. 6), and this provides some justification for the procedure that has been used to incorporate the [Ag(CN)2]2 data into the correlation for the terminal Ag]CN species (see discussion above). The simplest system that involves the linear Ag]CN]Ag bridge bonding mode is solid AgCN.This consists of infinite linear chains of Ag atoms connected by linear CN bridging groups, and a single n(AgX) mode has been assigned at 480 cm21.24 The main reason for the considerable increase in frequency relative to the terminal Ag]CN case is the coupling of the Ag]C and Ag]N coordinates that occurs in the linear bridge bonding case, as discussed above. Another example of this bonding situation occurs in the compound AgCN? 2AgNO3.19 This contains infinite ]Ag]CN]Ag]CN] chains, similar to solid AgCN itself, but the chains are separated by Ag1 and NO3 2 ions.Since no vibrational studies have been reported to date for this complex we have prepared it and recorded its far-IR spectrum (Fig. 7). The spectrum is similar to that of solid AgCN,24 and by analogy with this the weak band at 435 cm21 is assigned to the n(AgX) mode of the chain. Fig. 6 Plots of the wavenumber of the n(AgX) band against Ag]X bond length.Data are for X = Cl (d), Br (j), I (m) (from ref. 16) and for terminal (r) and bridging (.) CN. The curves are the best fit using equation (1) for the first four casesJ. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 2145 Although the Ag]X distances for this compound are the same as those of [Ag(CN)2]2, the n(AgX) frequency is considerably higher (Table 6, Fig. 6). This is due to the fact that this frequency depends on the sum of the Ag]C and Ag]N force constants in the Ag]CN]Ag bridge, due to coupling of these coordinates as discussed above.Thus, it is not possible to represent the frequencies of the bridging n(AgX) modes by an equation of the type (1), since these frequencies depend on the strengths of the two bonds involved in the bridge. Nevertheless, the above result suggests that the frequencies of bridging CN groups will be greater than those of terminal CN groups for similar Ag]C bond lengths. This latter point is well illustrated by the results for [Ag{Sb- (PPh3)3}n(CN)] (n = 2 or 3).The far-IR spectra of these compounds are shown in Fig. 7. These show strong bands at about 450 and 270 cm21 due to the co-ordinated Ph3Sb.25,26 The bands at 358 and 310 cm21 for the n = 2 and 3 compounds respectively are assigned as the n(AgX) modes of the CN groups. Initially it was rather surprising to find that the wavenumber for the n = 2 compound is greater than that for n = 3, as the Ag]C bond length in the latter compound is shorter than that in the former (Table 6), thus indicating a weaker bond.This apparent anomaly is attributed to the eVects of linear bridging, as discussed above. The n = 2 complex has an infinite chain structure with bridging CN groups, whereas the n = 3 complex contains terminally bound CN. While the data for the n = 3 complex lie close to the best fit line for terminally bound CN [the wavenumber predicted from equation (1) is 344 cm21], those for the n = 2 complex lie considerably above this line [the wavenumber predicted from equation (1) using r(AgC) = 2.133 Å is only 310 cm21].Fig. 6 shows that the frequencies for both of the infinite chain compounds that contain bridging CN groups are about 50 cm21 higher than those for analogous compounds that involve terminal CN bonding. This provides a clear contrast to the situation for the n(AgX) modes of halide complexes, where angular rather than linear bridging occurs, and the data for both types of complex lie on the same curve, as shown in Fig. 6.The far-IR spectra of the ionic 1 : 2 adducts [{Ag(EPh3)3}2- CN]1[Ag(CN)2]2?3py (E = P or As) are shown in Fig. 8. These spectra show bands due to co-ordinated EPh3, which can be assigned with reference to the spectra of uncomplexed EPh3.25,26 Fig. 7 Far-infrared spectra of (a) AgCN?2AgNO3, (b) [Ag(Sb- Ph3)2(CN)] and (c) [Ag(SbPh3)3(CN)]. Bands assigned to the n(AgX) modes are labelled with their wavenumbers They also show a band at 394 cm21 that can be assigned to the antisymmetric n(AgC) mode of [Ag(CN)2]2 {cf. 396 cm21 for K[Ag(CN)2]}.18 We have assigned the bands at 301 and 354 cm21 in the E = P and As compounds respectively to the n(AgX) modes of the linear bridging CN groups in the cations. The increase in wavenumber from E = P to As is in agreement with the trend in Ag]C/N bond lengths, Table 6. These data are included in the plot of n(AgX) vs. r(AgX) in Fig. 6. This shows that, as in the case of the compounds with infinite chain structures, the frequencies for these compounds are significantly higher (in this case by about 80 cm21) than those for analogous compounds that involve terminal CN bonding.Again, this can be attributed to the eVects of linear bridging, as discussed above. Thus, while the behaviour of the n(AgX) modes in cyano-complexes (X = CN) is clearly more complicated than that of halogeno complexes (X = Cl, Br or I), the data obtained for the present series of complexes provide a good basis for understanding this unusual behaviour. The n(CN) wavenumbers for several of the compounds prepared in this study were measured from their IR spectra, and the values are compared with those of some related compounds in Table 6.The relationship between n(CN) and r(AgX) for the AgCN/ER3 complexes is shown in Fig. 9. From this it is clear that there is some overlap between the range of n(CN) values for the terminal and bridging bonding modes of the CN group, although the values for the bridging groups tend to be higher.There is a clear trend of increasing n(CN) with decreasing r(AgX), and the rate of this increase is greater for the bridging than for the terminal bonding mode. These results can readily be explained in terms of the Ag/CN bonding interactions; co-ordination of the CN group by s donation to a silver atom results in an increase in n(CN). For a given Ag]X bond length the eVect on n(CN) is greater in the bridging mode, since two silver atoms are acting as electron acceptors.Hence the correlation line for bridging should lie above that for terminal bonding, and the rate of increase should be greater for the bridging case, as observed. One as yet unexplained anomaly concerning the results in Table 6 is the relatively low n(CN) for AgCN?2AgNO3. In this compound AgCN forms linear chains with bridging CN groups, as has also been proposed for solid Fig. 8 Far-infrared spectra of [{Ag(EPh3)3}2CN]1[Ag(CN)2]2?3py: (a) E = P, (b) E = As. Bands assigned to the n(AgX) modes are labelled with their wavenumbers2146 J. Chem. Soc., Dalton Trans., 1998, Pages 2139–2146 AgCN itself.19,24 Extrapolation of the correlation line for bridging bonding (Fig. 9) to the r(AgX) value reported for AgCN?2AgNO3 yields a n(CN) value that is considerably higher than the observed value of 2119 cm21 (Table 6). However, a similar extrapolation for solid AgCN itself yields a value close to the observed n& (CN) = 2168 cm21 for this compound,24 suggesting that the anomaly in the AgCN?2AgNO3 adduct is peculiar to this compound.Acknowledgements We acknowledge support of this work by grants from the Australian Research Grants Scheme and the University of Auckland Research Committee. We thank Ms. Catherine Hobbis for recording some of the far-IR spectra, and Dr. L.-J. Baker for assistance with the X-ray crystallography. References 1 EVendy, J.D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 587. 2 G. A. Bowmaker, EVendy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1998, 2123. Fig. 9 Plots of the wavenumber of the n(CN) band against Ag]X bond length for AgCN/ER3 complexes containing terminal (r) and bridging (.) CN groups 3 G. A. Bowmaker, EVendy, P. C. Junk and A. H. White, J. Chem. Soc., Dalton Trans., preceding paper. 4 D. E. Hibbs, M. B. Hursthouse, K. M. Malik, M. A. Beckett and P.W. Jones, Acta Crystallogr., Sect. C, 1996, 52, 884. 5 L. M. Engelhardt, P. C. Healy, V. A. Patrick and A. H. White, Aust. J. Chem., 1987, 40, 1873. 6 G. A. Bowmaker, EVendy, J. V. Hanna, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1993, 1387. 7 G. A. Bowmaker, EVendy, E. N. de Silva and A. H. White, Aust. J. Chem., 1997, 50, 627, 641. 8 M. M. Olmstead, G. Speier and L. Szabó, J. Chem. Soc., Chem. Commun., 1994, 541. 9 S. Gotsis, L. M. Engelhardt, P. C. Healy, J. D. Kildea and A. H. White, Aust. J. Chem., 1989, 42, 923; corrigendum, EVendy, L. M. Engelhardt, P. C. Healy, B. W. Skelton and A. H. White, Aust. J. Chem., 1991, 44, 1585. 10 B. K. Teo and D. M. Barnes, Inorg. Nucl. Chem. Lett., 1976, 12, 681. 11 G. A. Bowmaker and P. C. Healy, Spectrochim. Acta, Part A, 1988, 44, 115. 12 G. A. Bowmaker, P. C. Healy, J. D. Kildea and A. H. White, Spectrochim. Acta, Part A, 1988, 44, 1219. 13 G. A. Bowmaker, R. D. Hart, B. E. Jones, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1995, 3063. 14 G. A. Bowmaker, EVendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2449. 15 G. A. Bowmaker, EVendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2459. 16 G. A. Bowmaker, EVendy, J. D. Kildea and A. H. White, Aust. J. Chem., 1997, 50, 577. 17 F. C. Veldkamp and G. Frenking, Organometallics, 1993, 12, 4613. 18 T. M. Loehr and T. V. Long, J. Chem. Phys., 1970, 53, 4182. 19 D. Britton and J. D. Dunitz, Acta Crystallogr., 1965, 19, 815. 20 A. G. Sharpe, The Chemistry of Cyano Complexes of the Transition Metals, Academic Press, London, 1976, p. 274. 21 C. Kappenstein, A. Quali, M. Guerin, J. Cernak and J. Chomic, Inorg. Chim. Acta, 1988, 147, 189; M. Carcelli, C. Ferrari, C. Pelizzi, G. Pelizzi, G. Predieri and C. Sollinas, J. Chem. Soc., Dalton Trans., 1992, 2127; J. Cernak, M. Kanuchova, J. Chomic, I. Potocnak, J. Kamenicek and Z. Zak, Acta Crystallogr., Sect. C, 1994, 50, 1563. 22 D. N. Waters and B. Basak, J. Chem. Soc. A, 1971, 2733. 23 G. Helgesson and S. Jagner, Inorg. Chem., 1991, 30, 2574. 24 G. A. Bowmaker, B. J. Kennedy and J. C. Reid, Inorg. Chem., 1998, submitted. 25 K. Shobatake, C. Postmus, J. R. Ferraro and K. Nakamoto, Appl. Spectrosc., 1969, 23, 12. 26 F. W. Parrett, Spectrochim. Acta, Part A, 1970, 26, 1271. Received 6th February 1998; Paper 8/01085D

ISSN:1477-9226    

DOI:10.1039/a801085d

出版商:RSC    

年代:1998

数据来源: RSC